Scientific Sessions

...

4 downloads 2266 Views 2MB Size

B Scientiﬁc Sessions (SS)

Friday ....................... 139 Saturday................... 195 Sunday ..................... 243 Monday .................... 271 Tuesday.................... 303 137

Scientific Sessions

138

A

B

C

D

E

F

G

Scientific Sessions

Friday

Friday, March 7

A

B

C

D

E

F

G

139

Scientific Sessions room A 2nd level

room B 2nd level

room C 2nd level

room E1 entr. level

room E2 entr. level

room F1 entr. level

room F2 entr. level

room G/H lower level

room I lower level

07:00

07:00

07:30

07:30

08:00

08:00

08:30 CC 117 Multidetector CT Made Easy MDCT implementation: Making the best of the technology race (p. 8)

CC 116 MRI MR techniques: From basic to advanced (p. 8)

09:00

09:30

NH 1 New Horizons Session MR diffusion and perfusion: Can they replace PET? (p. 9)

RC 110 Musculoskeletal Foot and ankle (p. 10)

E³ 120a Foundation Course: Interventional Radiology Arteries (p. 10)

E³ 120b Interactive Teaching Session The imaging management of breast cancer (p. 11)

SF 1 Special Focus Session Uterine ﬁbroid embolisation – the case for and against: Gynecologist, radiologist, patient (p. 11)

08:30

RC 111 Neuro Stroke (p. 12)

RC 115 Vascular Non-invasive vascular imaging (p. 13)

09:30

10:00

10:00

10:30

10:30 SS 210 Musculoskeletal Peripheral nerves and muscle disorders (p. 142)

11:00

11:30

SS 201b GI Tract Advances in abdominal imaging (p. 144)

SS 206b Molecular Imaging Preclinical studies (p. 146)

E³ 220 Foundation Course: Interventional Radiology Veins (p. 18)

SS 207 Genitourinary The uterus (p. 148)

SS 202 Breast Advances in digital mammography (p. 150)

SS 208 Head and Neck Lesion detection and characterisation (p. 152)

SS 215 Vascular Carotid imaging (p. 154)

12:00 Opening Ceremony/ Presentation of Gold Medal Award and Honorary Members

12:30

13:00

12:30

13:00

13.30

13:30

14:00

14:00 SS 310 Musculoskeletal Spine and shoulder (p. 170)

14:30

15:00

SS 301a Abdominal Viscera Challenges in hepatobiliary diseases (p. 172)

SY 1 Bayer Schering Pharma Satellite Symposium (p. 536)

E³ 320 Foundation Course: Interventional Radiology Gastrointestinal tract (p. 18)

SS 307 Genitourinary Imaging of the pelvis (p. 174)

SS 302 Breast Innovations in breast diagnosis (p. 176)

SS 308 Head and Neck Tumors: Metastases and imaging in therapy (p. 178)

SS 315 Vascular CT angiography (p. 180)

15:30

16:00 CC 417 Multidetector CT Made Easy The heart: A newly found land (p. 19)

16:30

registration

EPOS™ - scientiﬁc exhibition

17:30

technical exhibition

17:00

A

B

SA 4 State-of-the-Art Symposium The radiologist as the therapist in liver tumors (p. 19)

CC 418 Breast: From Basics to Advanced Imaging From pathology to breast imaging: Radiologicpathologic correlations (p. 20)

E³ 420a Foundation Course: Interventional Radiology Genitourinary (p. 21)

E³ 420b Interactive Teaching Course: Useful Signs in Imaging Signs in neuroradiology and in head and neck imaging (p. 21)

16:00 RC 402 Breast Digital mammography (p. 22)

SF 4 Special Focus Session Dual source CT/ Dual energy CT: Doubly informative? (p. 23)

RC 415 Vascular Imaging in carotid disease (p. 23)

16:30

17:00

17:30 Lifetime Achievement Award

18:00

Inaugural Lecture (p. 28)

18:30

ECR Opening Concert

19:00

140

14:30

15:00

15:30

18:30

11:00

11:30

12:00

18:00

09:00

C

D

19:00

E

F

G

Scientific Sessions room K lower level

room L/M 1st level

room N/O 1st level

room R1 1st level

room R2 1st level

room R3 1st level

room U 2nd level

room X 2nd level

room XI, Z & Y 2nd level

room Z 2nd level 07:00

07:30

07:30

08:00

08:00

08:30

09:00

09:30

08:30 RC 105 Computer Applications What can PACS do today? (p. 13)

MC 125 Women’s Imaging: Female Pelvis Assessment of anatomy and imaging in benign lesions (p. 14)

MC 119 Molecular Imaging Molecular imaging (p. 15)

RC 104 Chest Volumetric HRCT of diffuse lung diseases (p. 16)

RC 101 Abdominal and Gastrointestinal Imaging of malignant liver lesions (p. 16)

RC 108 Head and Neck Imaging of the orbit (p. 17)

09:00

09:30

10:00

10:00

10:30

10:30

11:00

11:30

SS 205 Computer Applications CAD and automated image analysis (p. 157)

SS 211 Neuro CT angiography and perfusion (p. 159)

SS 206a Contrast Media Side effects (p. 161)

SS 204 Chest Advances in tumor imaging (p. 163)

SS 203 Cardiac Advances in cardiac CT (p. 165)

SS 209 Interventional Radiology Vascular interventions (1) (p. 167)

WS 222 Experience Vascular Procedures Using Simulators

WS 224 Test-Driving your Workstation

11:00

11:30

12:00

12:00

12:30

12:30

13:00

13:00

13:30

13:30

14:00

14:30

15:00

14:00 SS 305 Computer Applications Image viewing and PACS (p. 182)

SS 311a Neuro Advanced spinal MRI (p. 184)

SS 311b Neuro Brain tumors (p. 186)

SS 304 Chest Screening/ Pulmonary nodule (p. 188)

SS 303 Cardiac Advances in cardiac MR (p. 190)

SS 309 Interventional Radiology Thoracic interventions (p. 192)

WS 24A1 Test-Driving your Workstation

WS 322 Experience Vascular Procedures Using Simulators

15:00 WS 24A2 Test-Driving your Workstation

15:30

16:00

16:30

17:00

RC 405 Computer Applications PACS in evolution (p. 24)

MC 425 Women’s Imaging: Female Pelvis Malignant tumors (p. 25)

RC 408 Head and Neck Head and neck imaging: A practical approach (p. 25)

RC 404 Chest A practical approach for interpreting the chest radiograph (p. 26)

RC 401 Abdominal and Gastrointestinal Diseases of the mesentery and peritoneum (p. 27)

14:30

15:30

16:00

RC 409 Interventional Radiology Endovascular treatment of aortoiliac occlusive disease: How far can interventional radiology go? (p. 27)

16:30

17:00

17:30

17:30

18:00

18:00

18:30

18:30

19:00

19:00

A

B

C

D

E

F

G

141

Friday

07:00

Scientific Sessions 10:30 - 12:00

Room B

SS 210 Peripheral nerves and muscle disorders Moderators: R. Arkun; Izmir/TR M.S. Baramia; Tbilisi/GE

10:30

Phosphocreatine kinetics during incremental calf exercise in patient with diffuse atherosclerosis: A 31P MRS study M.F.H. Schocke, R. Esterhammer, H. Messner, W.R. Jaschke, G. Fraedrich, A. Greiner; Innsbruck/AT ([email protected]) Purpose: To investigate the HEP metabolism in the calf muscle of patients with diffuse PAD during incremental plantar ﬂexion exercise with the help of time-resolved phosphorus-31 magnetic resonance spectroscopy (31P MRS). Methods and Materials: Using a 1.5 Tesla whole body MR scanner, 20 patients with uni- or bilateral diffuse PAD (63.4p7.8 years) and 24 healthy controls (65.3p8.6 years) underwent serial 31P MRS during exercise increments at 2, 3, 4 and 5 W. The patients had at least one painful leg induced by exercise and received an MR angiography by using FLASH 3D sequences with moving bed technique. For 31P MRS, we used serial free induction decay sequences with a time resolution of 10 s and NOE enhancement. The phosphocreatine (PCr) time constants were determined for each exercise increment and recovery. Results: Patients with diffuse PAD showed normal PCr time constants during the increment at 2, 3 and 4 W and during recovery compared to the healthy controls. The patient group exhibited signiﬁcantly increased PCr time constants during the last exercise increment at 5 W. Conclusion: Our ﬁndings indicate that diffuse PAD is associated with a slowed PCr metabolism at high work intensities within the calf muscle, whereas the PCr recovery time constants are normal. This ﬁnding indicates a normal mitochondrial function and might be probably caused by changes in muscle ﬁber distribution due to PAD. Those alterations could be treated by exercise training.

B-002

10:39

Different patterns of involvement in myotilinopathy and desminopathy detected by a novel neuromuscular whole-body MRI protocol N. Schramm, C. Born, S. Weckbach, P. Reilich, M.C. Walter, M.F. Reiser; Munich/DE ([email protected]) Purpose: Whole-body MRI (WB-MRI) has been successfully applied for oncologic and cardiovascular purposes. In contrast, myopathies were mostly examined with dedicated MRI limited to particular body parts. The aim of this study was to evaluate a novel comprehensive neuromuscular WB-MRI protocol. Methods and Materials: Whole-body imaging was performed on a 1.5 T scanner using parallel imaging. Examination time was 41 min 26 s. 18 patients were included (8 female, 10 male; mean age 56 years). Coronal and axial T1w as well as coronal STIR sequences of the whole body were acquired. Images were analyzed by two experienced radiologists. Results: In contrast to dedicated examinations whole-body imaging was able to demonstrate affectation of the muscles of the upper arm (16.7%), the shoulder (27.9%) and the trunk (erector spinae: 61.1%). Concerning desminopathy and myotilinopathy, which are distal myopathies, different patterns of involvement could be detected: Patients with myotilinopathy showed frequent involvement of the rhomboid muscles (4/5), the erector spinae (5/5), the biceps femoris and the semimembranosus (5/5), while the semitendinosus was relatively spared (2/5). In contrast, in patients with desminopathy the ilipsoas (3/4), the sartorius, (3/4), the gracilis (3/4) and the semitendinosus (3/4) were typically involved, while the semimembranosus was spared (1/4). Conclusion: As shown for rare distal myopathies WB-MRI is an appropriate modality to analyze the complete muscle status with only one examination. It could be useful to narrow the differential diagnosis of myopathies and to determine the adequate site for muscle biopsy.

142

A

B

10:48

Diffusion tensor imaging (DTI) and ﬁber tractography of skeletal muscle: Optimization of b-value for imaging at 1.5 Tesla N. Saupe, L.M. White, J. Stainsby, G. Tomlinson, M.S. Sussman; Toronto, ON/CA ([email protected])

Musculoskeletal

B-001

B-003

C

D

E

F

G

Purpose: To determine the optimal b-value for diffusion tensor (DT) MRI and ﬁber tractography of in-vivo human skeletal muscle at 1.5 T. Methods and Materials: Five healthy volunteers (mean age 35 years) were included in this prospective study. DT-MRI of the proximal calf was performed using 15 noncollinear directions of diffusion sensitization. Parallel imaging with an acceleration factor of 2 was applied. Each single shot spin-echo EPI sequence was performed using eight different b-values ranging from b=125 s/mm2 to b=1000 s/mm2. Fiber tractography of two different calf muscles (anterior tibialis (AT) and lateral gastrocnemius (LG) muscles) was performed. Qualitative analysis of tractography images was performed by two readers independently. Quantitative fractional anisotropy (FA), signal-to-noise ratio (SNR) and ﬁber density index (FDi) were assessed. Statistics for all acquisitions and the best-through-worst within-subject b-values were computed and 95% conﬁdence interval calculated. Results: Fiber tractography using a b-value of 625 s/mm2 was qualitatively ranked ﬁrst (best image quality) or second (second best image quality) in 100% of tractography images assessed. FA was not signiﬁcantly different between both muscles (p=0.43) and different b-values (p=0.21-0.82). SNR was highest for both muscles using a b-value of 125 s/mm2 (AT=66.2, LG=60.1). Highest FDi results were found for both muscles using a b-value of 625 s/mm2 (AT=7.82, LG=5.75). Conclusion: The optimal b-value for in vivo DT imaging and tractography assessment of human skeletal muscle in the calf at 1.5 T MR imaging was found to be 625 s/mm2.

B-004

10:57

Intracranial internal carotid artery calciﬁcations: Association with vascular risk factors and cerebrovascular symptomatology T.T. de Weert, H. Cakir, S. Rozie, S. Cretier, E. Meijering, D.W.J. Dippel, A. van der Lugt; Rotterdam/NL ([email protected]) Purpose: Atherosclerotic calciﬁcations in the intracranial internal carotid arteries are very frequent. However, their association with cardiovascular risk factors and cerebrovascular events has not been studied extensively. Therefore, we assessed the association between intracranial internal carotid artery calciﬁcations and cardiovascular risk factors in patients with cerebrovascular disease, and the relation between calciﬁcations and the presence of cerebrovascular symptomatology. Methods and Materials: CTA of the carotid arteries was performed in 406 patients (age 62p14 years; 242 men). A semi-automatic custom-made system to quantify calciﬁcations was developed. Calciﬁcations were deﬁned as pixels within the arterial wall with an attenuation above 500 Hounsﬁeld Units. The relation between calciﬁcations and age was analyzed with correlation coefﬁcients. The association between cardiovascular risk factors (smoking, hypertension, diabetes, hypercholesterolemia, previous cardiac disease and previous cerebrovascular disease) and the volume of calciﬁcations, and the relation between calciﬁcations and cerebrovascular symptomatology were assessed with logistic regression. Results: Men had a signiﬁcantly (p 0.003) higher calciﬁcation volume (66 mm³) than women (33 mm³). Calciﬁcation volume was positively associated with age in both men and women (correlation coefﬁcient 0.548 and 0.501, respectively). Previous cardiac disease and previous cerebrovascular disease were independently related to the volume of calciﬁcation (OR was 2.1 and 1.8, respectively). No relation was found between calciﬁcations and the presence or type of symptomatology. Conclusion: Intracranial carotid calciﬁcations can reproducibly be quantiﬁed, and the volume of calciﬁcations is associated with age, gender and cardiovascular risk factors. Calciﬁcations are not related to cerebrovascular symptomatology.

B-005

11:06

Chronic compartment syndrome of the forearm in motocross racers: Findings on MRI J.L.M.A. Gielen1, B. Peersman1, G. Peersman2, P. Van Dyck1, F.M.H.M. Vanhoenacker1; 1Edegem/BE, 2Antwerp/BE ([email protected]) Purpose: To demonstrate the ﬁndings of MRI in CCS of the forearm. Methods and Materials: Five male professional motocrossers with proven CCS and a control group of two motocrossers without CCS, and three normal male individuals not involved in strenuous activities at the forearm were included. MRI with SI measurements in the muscles was performed before and immediately after activity. Results: MRI after exercise shows a diffuse and marked increase in SI on T2-WI in bilateral ﬂexor digitorum superﬁcialis and profundus muscles in all the patients.

Scientific Sessions

B-006

11:15

Anatomy of vastus medialis obliquus: An MRI study P.K.R. Mereddy, V. Roberts, S. Hakkalamani, P. Evans, N.J. Donnachie; Upton, Wirral/UK ([email protected]) Purpose: To clarify the anatomy of VMO and its insertion on patella. This study also determines the inﬂuence of variables such as age, gender and diagnosis of concurrent knee osteoarthritis on VMO insertion. Methods and Materials: We analysed 198 knee MR scans performed in 2006. Patients over 16 years were included and the scans with neoplasm or without complete set of ﬁlms were excluded. There were 106 female and 92 male patients with a mean age of 43 years (range 16-86). Patellar height and VMO length were measured on the axial scans. Based on the VMO insertion on the patella, patients were divided into proximal and distal groups. Knee radiographs were assessed for presence of osteoarthritis. Results: Overall there were 119 (60%) patients in the proximal group and 79 (40%) in the distal group. In patients above 50 years, 20% had distal VMO insertion. In patients below 50 years, 50% had distal VMO insertion. Statistically signiﬁcant difference was noted between VMO length and age (p 0.0005). Statistically signiﬁcant difference was also noted between the patellar height and VMO length in relation to gender (p 0.0001 and p 0.0005). Statistically signiﬁcant inverse relationship was noted between age and VMO insertion (p 0.0001). Statistically signiﬁcant relationship was also noted between age and osteoarthritis (p=0.001). Conclusion: The contributing factors for the anatomical changes in VMO are age and the presence of osteoarthritis. Older patients are more likely to have a VMO insertion at or more proximal to the mid-point of patella.

B-007

11:24

Meralgia paresthetica: Ultrasound ﬁndings and ultrasound guided therapy G. Bodner, M. Bernathova, M. Felfernig, T. Boerger; Gibraltar/UK ([email protected]) Purpose: Neuropathy of the lateral femoral cutaneous nerve (LFCN) is a painful compression neuropathy, also known as meralgia paresthetica. The aim of the study is to describe ultrasound ﬁndings in 18 patients and to report results from ultrasound-guided therapeutic injections. Methods and Materials: Eighteen patients (mean age 54 years, range 38-77 years, 10 males, eight females), suffering from pain and sensory loss or burning sensation over the lateral and anterolateral thigh (10 left, eight right), underwent ultrasound examinations. Ultrasound examination was performed with a high frequency probe working with 17 MHz (Philips, IU 22). Ultrasound guided injection of local anesthetics (one ml) adjacent to the affected nerve was performed in order to conﬁrm the ultrasound ﬁndings. If patient was pain free 20 mg of cortisone was injected. Pain level was registered before and after the injection (Visual analogue scale). Patients were monitored for 6-12 months. Results: Ultrasound showed a fusiform swelling of the affected LCFN at the area where the nerve pierces the inguinal ligament. Comparison was made with the unaffected side. In all patients injection of local anaesthetics was immediately successful; cortisone injection released pain in 12 cases. In four cases additional cortisone injections were necessary. Conclusion: Ultrasound allows visualizing compression of the FCLN in patients with meralgia paresthetica. Ultrasound guided injection of local anaesthetics and cortisone may be a valid procedure for pain management.

B-008

11:33

Value of an additional cross sectional area measurement of the median nerve in carpal tunnel syndrome patients compared to healthy volunteers: Assessment by ultrasound T. De Zordo, D. Junker, G. Feuchtner, A. Mallouhi, M. Starzinger, A. Klauser; Innsbruck/AT ([email protected]) Purpose: The aim of the study is to compare two different cross-sectional area (CSA) measurements of the median nerve at the level of carpal tunnel and more proximal to assess individual nerve thickness in order to resolve the problem of different cut off values discussed in the literature. Methods and Materials: Ninety wrists in 75 consecutive CTS patients and 40 wrists of 20 healthy volunteers underwent US examination (Siemens Acuson Sequoia, 15L8w). Difference between maximal CSA value obtained at the carpal tunnel (CSAc) minus CSA value obtained at the level of the pronator quadratus muscle at the distal forearm (CSAp) was calculated. Results: Mean CSAc in patients was 13.83 mm2 (p3.58, range 9.0-34.1) compared to healthy volunteers with 8.90 mm2 (p1.12, range 7.3-10.8) (p 0.01). 18% of patients presented CSAc values in the range of healthy volunteers, but showed increased thickness of the median nerve at the carpal tunnel when compared to CSAp. Calculated difference of CSAc minus CSAp measurements in patients compared to healthy volunteers was statistically signiﬁcant (p 0.001). Conclusion: Based on our preliminary results, an additional CSA measurement at the distal forearm level (pronator quadratus) is helpful and quick to perform to diagnose nerve thickening at the carpal tunnel in the assessment of CTS in respect to individual patients’ hand physiognomy.

B-009

11:42

Nerve ultrasound in acromegaly A. Tagliafico, E. Resmini, R. Nizzo, F. Bianchi, D. Ferone, C. Martinoli; Genoa/IT (alberto.tagliaﬁ[email protected]) Purpose: To examine nerves of acromegalic patients with US and to determine whether nerve abnormalities correlate with clinical parameters and nerve conduction studies (NCS). Methods and Materials: We prospectively examined the median and ulnar nerves in 34 non-diabetic, acromegalic patients and 34 age-matched controls (C) with 175 MHz US. The median nerve was examined at the carpal tunnel (MN-Ct) and at the mid-forearm (MN-f) level; the ulnar nerve at the mid-forearm (UN-f) and at distal arm (UN-a). A total of 272 nerve cross-sectional areas (CSA) were recorded from both patients and C. 22 patients underwent NCS of median and ulnar nerves. Results: Nerves of acromegalic patients had signiﬁcantly (p 0.0001) greater CSA compared to C. NCS displayed at least one abnormality in 13/22 (59%) patients. Acromegalic patients were grouped according to disease activity (13 controlled; 8 partially controlled; 13 uncontrolled) and medical therapy with somatostatin analogs (18 treated; 16 untreated). All these groups had signiﬁcantly (p 0.0001) greater CSA compared with C. Nerve CSA were signiﬁcantly greater in uncontrolled patients compared to controlled at MN-Ct (15.46p4.17 mm2 vs 18.41p4.14 mm2; p 0.01) and at UN-f (12.34p2.77 mm2 vs 14.95p3.98 mm2; p 0.01) level. A positive correlation between IGF-1 values and CSA (r=0.34) was present, no correlation with GH levels was present. A positive correlation was found between disease duration and both nerve CSA and NCS (r=0.33 and r=0.31). Conclusion: US identiﬁed a signiﬁcantly increased volume of nerves in acromegalic patients. Nerve enlargement appears as an intrinsic feature of the disease related to clinical control, disease duration and IGF-I.

B-010

11:51

Magnetic resonance imaging as a new method for monitoring nerve regeneration after median or ulnar nerve repair A.R. Viddeleer1, P.E. Sijens1, P.M.A. van Ooijen1, S.E.R. Hovius2, M. Oudkerk1; 1 Groningen/NL, 2Rotterdam/NL ([email protected]) Purpose: After operative repair of peripheral nerve lesions, it is important to closely monitor the nerve regeneration process, as reintervention may be indicated if regeneration fails. Reintervention has a better chance of success if attempted early. Currently, electromyography (EMG) is the standard for monitoring nerve regeneration after nerve repair. However, EMG has some disadvantages: EMG can be painful, results are clinician- and patient-dependent and difﬁcult to interpret. Therefore, new tests are needed which monitor nerve regeneration in a noninvasive, standardized and objective manner. It has been shown that Short Tau Inversion Recovery (STIR) MRI sequences have a high sensitivity and speciﬁcity for detecting denervation, as differences in image intensity between healthy and denervated muscles occur. Therefore, STIR-MRI could be a viable candidate to monitor the

A

B

C

D

E

F

G

143

Friday

Also, a minor SI increase in the extensor carpi radialis longus was noted. In the non-motocross controls, a comparable markedly increased SI of the ﬂexor digitorum superﬁcialis and profundus and the extensor carpi radialis longus muscles was noted on both sides in all three individuals. In the motocross controls, only a minor increased SI of the ﬂexor digitorum superﬁcialis and profundus and extensor carpi radialis longus muscles was noted. Conclusion: The muscular edema in our non trained control individuals is well known in submaximal repetitive eccentric post exercise circumstances. This edema is not demonstrated in our non symptomatic motocrossers. In contrast, in motocrossers with CCS, there is a dissociation in the ﬁndings between the ﬂexor compartments and lateral compartment as muscle edema exists predominantly at the symptomatic ﬂexor compartments. Thus it is important to control exercise circumstances to exclude non pathologic muscular edema on MRI in particular in non trained individuals/compartments that may mimic muscle edema that is found in CCS.

Scientific Sessions process of peripheral nerve regeneration. Methods and Materials: A standardized and optimized image acquisition and postprocessing protocol was used to follow 20 patients with complete transection of the median or ulnar nerve in the forearm. In the year following nerve repair, MRI-scans of the hand were obtained at 1, 3, 6, 9 and 12 months. Results: All denervated muscles showed abnormal image intensity at the ﬁrst scan. Image intensity was maximal at 3 months (intensity ratio 1.43p0.08, 95% CI). During nerve regeneration, image intensities gradually normalized before muscle function returned. Non-reinnervated muscles kept their abnormal appearance in all patients. Conclusion: These results strongly suggest that MRI can be used for monitoring nerve regeneration by early differentiation of de- and (re)innervated muscles.

10:30 - 12:00

Room C

GI Tract

SS 201b Advances in abdominal imaging Moderators: E. Biscaldi; Genoa/IT Y. Imai; Kanagawa/JP

B-011

10:30

Signiﬁcance of incidental focal FDG-uptake in the gastrointestinal tract: Correlation of PET/CT with endoscopic and histological ﬁndings P. Aschoff, T. Kratt, M.P. Lichy, M.Ö. Öksüz, C.D. Claussen, C. Pfannenberg; Tübingen/DE ([email protected]) Purpose: The signiﬁcance of unexpected focal FDG-uptake in the gastrointestinal tract (GIT) detected in FDG-PET examinations often remains unclear. In combined PET/CT examinations these ﬁndings can be localized more precisely in the GIT but the differentiation of benign from malignant lesions remains difﬁcult. We correlated these foci with endoscopic and histological ﬁndings to evaluate their signiﬁcance. Methods and Materials: Retrospective evaluation of 1250 whole body FDGPET/CT examinations of patients with oncological and non-oncological diseases. Correlation of focal FDG-uptake in the GIT not explained by known malignancies of the GIT with endoscopic and histological ﬁndings. Results: Unexpected focal FDG-uptake in the GIT was detected in 4.5% (56/1250) of the FDG-PET/CT examinations. In 61% (34/56) of these patients further endoscopic examinations with biopsies of suspicious lesions were performed. In 47% (16/34) of these patients malignant lesions and in 38% (13/34) benign lesions were found. In 5/34 patients (15%) no correlating lesion was found. In the subgroup of 14 patients with FDG-foci in the upper GIT, malignant lesions were found in 64% (9/14, 6 carcinomas, 2 metastases, 1 NHL), inﬂammatory lesions in 14% (2/14) and no corresponding lesion in 21% (3/14). In the subgroup of 20 patients with FDGfoci in the lower GIT, malignant lesions were found in 35% (7/20, 4 carcinomas, 3 metastases), adenomas with dysplasia in 30% (6/20), inﬂammatory lesions in 25% (5/20) and no corresponding lesion in 10% (2/20). Conclusion: Unexpected FDG-foci in the GIT detected by PET/CT require a further endoscopic examination because they frequently represent malignancies.

B-012

10:39

FDG uptake of GIST with the relatio to tumor size and mitosis T. Nakagawa, M. Yamada, T. Uno, Y. Suzuki, A. Yokota, K. Yoshida, I. Umehara, A. Toriihara; Chiba/JP ([email protected]) Purpose: To evaluate the F (18)-ﬂuoro-deoxyglucose (FDG) uptake of gastrointestinal stromal tumor (GIST) with the relation of tumor size and mitosis by using MIB1 index. Methods and Materials: Twelve patients with surgically resected GISTs undertook PET/CT prior to the operation. The maximal diameter and standardized uptake value (SUVmax) were measured. Mitotic activity was counted per 50 high power ﬁelds (HPFs) from the area showing the most numerous mitoses. Corresponding section was stained by proliferative biomarker Ki-67 (clone; MIB-1). The fraction of Ki-67 positive cells to the total cells on a HPF was deﬁned as MIB1 index (%). The maximum MIB1 index (MIB-1max) was assessed from the area of highest positivity. And MIB-1 standard deviation (MIB-1SD) from ﬁve randomly selected areas was calculated. Results: All the patients with SUV 6.5 were categorized into high risk (n=8). The MIB-1max and MIB-1SD were related to mitotic activity (r=0.940, p 0.0001; r=0.966, p 0.0001 respectively). Among the high risk group, highly mitotic GISTs

144

A

B

C

D

E

F

G

( 30/50hpf) showed high MIB-1max ( 9.5%) and MIB-1SD ( 2.7%). The rest GISTs with mitotic activity less than 10/50hpf showed low MIB-1max ( 7.5%) and MIB-1SD ( 1.7%). In the high MIB-1max GISTs (n=5), SUVmax was signiﬁcantly correlated to MIB-1max (r=0.927, p=0.0235) and not to diameter (p=0.360). In the low MIB-1max GISTs (n=7), SUVmax was signiﬁcantly correlated to the diameter (r=0.938, p=0.0017) and not to MIB-1max (p=0.0735). Conclusion: Highly proliferative GISTs with intratumoral heterogeneity showed FDG uptake depending on highly proliferative area. In contrast, homogeneous GISTs showed FDG uptake depending on tumor size.

B-013

10:48

Is further investigation justiﬁed for incidental unusual bowel uptake on PET-CT? H. Gupta, S. Olson; Aberdeen/UK ([email protected]) Purpose: Different patterns of uptake are noted in bowel on PET-CT, most being non-speciﬁc and physiological. However, patients having examination for nongastrointestinal tract primary tumour are sometimes found to have an area of uptake suspicious of underlying bowel abnormality and further investigations are recommended. Purpose of the study was to evaluate the outcome and justiﬁcation of further investigation. Methods and Materials: Reports of PET-CT of patients from all over Scotland from April 2006 to Mid-February 2007 were reviewed retrospectively. Patients in whom unusual bowel uptake was reported and further investigation suggested were the subjects of this study. Patients with known colorectal malignancy or history of inﬂammatory bowel disease were excluded. A questionnaire was formulated and sent to all the referring physicians. Results: 39 patients were selected (Non-Hodgkin’s-4, Lung Cancer-33, other paraneoplastic-2). Questionnaire was returned in 74% patients. Further investigations performed in 72% patients and not done in 28% (eight patients: one died, three poor prognosis). 52% patients had signiﬁcant abnormalities on gastrointestinal investigations, including 14% having colonic malignancy and remainder included polyps, gastrointestinal stromal tumour and ulcerative colitis (requiring hemicolectomy). Conclusion: Results of this study indicate that further investigations of unexpected focal or unusual 18F-FDG uptake in the bowel was justiﬁed with 52% of our study patients having malignant or premalignant lesions.

B-014

10:57

Perfusion CT in the follow-up of patients with gastrointestinal stromal tumors under speciﬁc molecular pharmacotherapy F. Berger, M. Schlemmer, T. Saam, K. Nikolaou, C. Becker, M.F. Reiser; Munich/DE ([email protected]) Purpose: Selected tyrosine kinase inhibitors with activity against gastrointestinal stromal tumors (GIST) have anti-angiogenic effects. Aim of this study is to optimize the CT perfusion protocol and evaluation of perfusion patterns of metastatic GIST lesions under speciﬁc pharmacotherapy. Methods and Materials: Twenty-ﬁve patients with metastatic GIST under tyrosine kinase inhibition were evaluated using a 64 slice CT scanner. Perfusion CT of metastatic lesions was carried out with the optimized parameters: Delay 6 sec/10 sec (hepatic/extrahepatic abdominal lesions), image frequency: 4 images/second for 40 seconds, (4x40 images), voltage 100 KV, tube current 80 mAs, contrast media bolus/ﬂow: 50 ml, 5 ml/sec. The perfusion scan was followed by a venous phase CT. Flow, blood volume and hepatic perfusion index measurements of lesions were carried out utilizing the Siemens syngo Body Perfusion software. Results: Twenty-one patients with liver lesions and four patients with intraperitoneal lesions were evaluated. Three patients showed clinical signs of progressive disease; they demonstrated liver metastases with global hyperperfusion (Hepatic perfusion index 78.8p21.2%). One patient showed a solid nodule with hyperperfusion (blood ﬂow 109p18.8 ml/100 ml/min) on the border of a pseudocystic hypoperfused hepatic lesion. In the other 21 patients, all with no clinical signs of relapse, the hepatic or intraperitoneal lesions showed hypoperfusion with low blood ﬂow values (4.3p9.7 ml/100 ml/min). Conclusion: Quantitative CT perfusion measurements of GIST-lesions are feasible. Our preliminary data indicate that this technique may contribute to an optimized monitoring of molecular therapies in GIST.

Scientific Sessions B-015

11:06

The role of DCE-MRI in the evaluation of response to angiogenesis inhibitors S. Bogaert1, P. Smeets1, J. Vercruysse1, P. Duyck1, P. Van Laer2; 1Gent/BE, 2 Brussels/BE ([email protected])

B-016

11:15

Diagnostic accuracy of 64-slice CT with isotropic multiplanar reconstructions (MPRs) in the pre-operative diagnosis of peritoneal carcinomatosis: Correlation with intraoperative and histopathologic ﬁndings M. Baski, D. Marin, C. Catalano, M. Di Martino, A. Guerrisi, G. De Filippis, S. Sibio, A. Di Giorgio, R. Passariello; Rome/IT ([email protected]) Purpose: To investigate the diagnostic accuracy of 64-slice CT with isotropic multi-planar reconstructions (MPRs) in the preoperative detection of peritoneal carcinomatosis. Methods and Materials: This study was approved by the institutional review board with a waiver of informed consent. Thirty-ﬁve patients (median age, 57.5 years) with peritoneal carcinomatosis underwent preoperative 64-slice CT (Sensation 64, Siemens) and surgical exploration within 40 days. CT parameters included 0.6x64 mm collimation, 3.0-mm section thickness, 250 mAs, 120 kVp. All patients received nonionic contrast medium (400 mgI/mL; Iomeron 400, Bracco) at a rate of 4 mL/sec. A dual-phase protocol was started 18 and 60 s after reaching a trigger threshold of 150 HU in the abdominal aorta. Two additional image sets were reformatted in the coronal and sagittal planes (3.0-mm section thickness). Images were evaluated independently by three radiologists on a four-point conﬁdence scale. CT ﬁndings were compared with intraoperative and histopathologic results. Diagnostic accuracy was evaluated using the AFROC method. Sensitivity and speciﬁcity with 95% conﬁdence intervals were determined. Results: The diagnosis of peritoneal carcinomatosis was conﬁrmed by exploratory laparotomy in all patients. Overall, CT sensitivity and speciﬁcity were 83 (158/191) and 97% (129/133), respectively. For lesions smaller than ﬁve mm, sensitivity and speciﬁcity were 48 (27/56) and 97% (57/59). The mean area under the AFROC curve of CT was 0.80. Conclusion: Although 64-slice CT with isotropic MPRs is accurate for the detection of peritoneal implants r5 mm, its performance is markedly decreased for lesions less than ﬁve mm.

B-017

11:24

Graded compression ultrasonography and computed tomography in acute colonic diverticulitis: A meta-analysis of test performance W. Laméris, A. van Randen, S. Bipat, P.M.M. Bossuyt, M.A. Boermeester, J. Stoker; Amsterdam/NL ([email protected]) Purpose: Computed tomography (CT) and graded compression ultrasonography (US) are the initial radiological investigations in acute diverticulitis. This study aims to determine their accuracy more precisely and compare the diagnostic performance of both modalities. Methods and Materials: We performed a meta-analysis of the performance of CT and US in diagnosing acute diverticulitis. Summary sensitivities and speciﬁcities were calculated using a bivariate random effects model. Post-test probabilities after

B-018

11:33

MSCT in arterial and portalvenous contrast phases for detection of intestinal bleeding M. Dobritz, A. Schneider, H.P. Engels, H. Wieder, E.J. Rummeny, J. Stollfuss; Munich/DE ([email protected]) Purpose: To evaluate Multislice-Computed-Tomography (MSCT) in arterial and portalvenous contrast phases in the detection of intestinal bleeding using an experimental bowel-bleeding model. Methods and Materials: A high pressure injector tube with a single perforation (1 mm) was placed in 10 m small bowel specimen of a pig. The bowel was ﬁlled with water/contrast solution of 30-40 HE and was incorporated in a phantom model containing vegetable oil to simulate mesenteric fat. Intestinal bleeding in different locations and bleeding velocities varying from 0 to 1 ml/min (0.05 ml/min increments, constant bleeding duration of 20 sec) was simulated with an injection of water/contrast solution (300 HE). 19 complete datasets in arterial and portalvenous phases using increasing bleeding velocities and seven negative controls were measured using a 64-row MSCT (3 mm slice thickness, 1.5 mm reconstruction increment). Three radiologists blinded to the experimental settings evaluated the datasets in a random order. The likelihood for intestinal bleeding was assessed using a 5-point scale with subsequent ROC analysis. Results: A bleeding velocity of 0.1-0.5 and 0.5-1.0 ml/min was detected with a mean sensitivity of 0.50 and 0.98, respectively. The speciﬁcity was 1.00. The area under curve was measured to be 0.88, 0.89 and 0.73 for readers 1-3, respectively. Conclusion: MSCT provided high sensitivity and speciﬁcity in the detection of intestinal bleeding with bleeding velocities of 0.5-1.0 ml/min. Therefore, MSCT should be considered as a primary diagnostic technique in the management of patients with suspected intestinal bleeding.

B-019

11:42

Minimal-preparation abdomino-pelvic CT in frail and elderly patients: Prognostic value of colonic and extracolonic ﬁndings C.S. Ng1, W. Wei1, T.C. Doyle2, H.M. Courtney3, A.K. Dixon4, A.H. Freeman4; 1 Houston, TX/US, 2Dunedin/NZ, 3Bracknell/UK, 4Cambridge/UK ([email protected]) Purpose: Minimal preparation abdomino-pelvic CT (MPCT) has been used to evaluate colorectal cancer (CRC) in frail and elderly patients. Our objectives were to examine the overall survival of patients who had been referred for MPCT, and to assess the prognostic value of the colonic and extracolonic ﬁndings detected. Methods and Materials: The survival of a cohort of 1029 elderly and frail patients, with clinical symptoms and signs suspicious for CRC, who had undergone MPCT between 1995 and 1998, was analyzed. Univariate and multivariate survival analyses were undertaken according to the presence of CRC and extracolonic abnormalities (ECA). Results: The overall median survival following MPCT was 5.4 years (median age 79.4 years), and 6.6 years if no abnormality was detected. On multivariate analysis, age, sex, CRC status, and number of ECAs were signiﬁcant factors in overall survival. Median survival for those with conﬁrmed CRC (n=91 (prevalence, 8.8%)) was 1.1 years, compared with 5.9 years without CRC (p 0.0001), and 2.4 years for those with one or more ECA (n=245 (prevalence, 23.8%)), compared to 6.1 years without ECA (p 0.0001). Survival was progressively shorter for increasing numbers of ECAs, and shorter for previously unknown non-CRC malignancies (n=24) compared to CRC (p 0.0001). Conclusion: MPCT appears to have prognostic potential in this patient population, with signiﬁcant reductions in survival if a CRC or ECA is detected. The detection of extracolonic abnormalities would appear to have at least as important an impact on the utility of the examination as does the detection of colorectal carcinoma.

A

B

C

D

E

F

G

145

Friday

Purpose: To investigate if DCE-MRI would be an appropriate technique to evaluate, in a quick and reliable way, the response to angiogenesis inhibitors used for colorectal cancer and liver metastases. The techniques currently used to evaluate tumour response (as biopsy, in vivo measurement of the oxygen level in the tumour, tumour markers in the blood, visible shrinkage etc.) happen to be invasive, unreliable and/or liable to sampling bias. Moreover, these techniques are only useful long after therapy onset. Could DCE-MRI overcome the disadvantages of these techniques? Methods and Materials: Post contrast T1 dynamic MR series were acquired from 15 patients, before and after therapy onset. For every acquisition, a time intensity curve was made and compared before and after therapy. Results: A successful therapy (12 patients) resulted in a quick fall of the contrast uptake-phase, a reduction of the amplitude of the plateau-phase and a ﬂatter washout-phase. The TI curves of three patients showed no difference before and after therapy. Conclusion: DCE-MRI is less invasive and liable to sampling bias than biopsy. It appears more reliable than quantiﬁcation of tumour markers in the blood and is for sure able to evaluate the response to therapy much earlier than appraisal of tumour shrinkage. With DCE-MRI it is possible to appraise tumour response to angiogenesis inhibitors 48 hours after therapy start, which makes it possible to change therapy rapidly if necessary. This evolution could save lives, time and money.

CT and US were calculated for the lower and upper prevalence of diverticulitis in the included studies. Results: Two head-to-head comparative studies and 10 studies evaluating US or CT separately were identiﬁed: Six US studies evaluating 630 patients and eight CT studies evaluating 684 patients. Mean prevalence of diverticulitis was 55% (range 36-68%) in US studies and 53% (range 36-68%) in CT studies. Summary sensitivity estimates were 92% (95% CI: 80-97%) for US versus 94% (95% CI: 8797%) for CT (p=0.65). Summary speciﬁcity estimates were 90% (95% CI: 82-95%) for US versus 99% (95% CI: 90-100%) for CT (p=0.07). In the reported range of prevalences, positive post-test probabilities ranged between 84% and 95% for US and between 98% and 99% for CT; negative post-test probabilities ranged between 5% and 16% for US and between 3% and 11% for CT. Conclusion: The diagnostic performance of US and CT in diagnosing diverticulitis is not signiﬁcantly different. Both modalities can be used as an initial diagnostic tool in the assessment of patients suspected of diverticulitis.

Scientific Sessions B-020

B-022

11:51

Early detection of acute mesenteric ischemia with diffusion weighted MRI (DWI) using parallel imaging C.A. Schwartz1, P. Haage2, C. Ocklenburg1, G. Mühlenbruch1, G.A. Krombach1, T. Niendorf1, R.W. Günther1, C. Hohl1; 1Aachen/DE, 2 Wuppertal/DE ([email protected]) Purpose: To evaluate the early diagnosis of mesenteric ischemia with diffusion weighted (DWI) imaging in a swine model. Methods and Materials: Early detection of mesenteric ischemia was studied in 10 female pigs with artiﬁcially induced mesenteric ischemia. Under control of conventional ﬂuoroscopy a branch of the superior mesenteric artery was embolized with tissue glue. DWI was performed at a 1.5 T MR-scanner (Achieva 1.5, Philips, Best, Netherlands) 30 and 60 minutes after embolization using a 32-channel receiver coil. The scan protocol included SE-EPI DWI (TE=88 ms, TR=5513 ms, FA=90°, b=0, 250, 500 s/mm2) in axial planes employing sensitivity encoding based parallel imaging (SENSE) with an acceleration factor R=2. Additionally, ADC-maps were calculated. The scans were performed in breathhold mode with i.v.-administration of butylscopolamine to avoid motion artifacts. The ﬁndings were correlated to the angiographic and autoptic ﬁndings. Image quality of DWI was assessed by means of SNR-measurements and the diagnostic impact by CNR-measurements. Variance analysis using the restricted maximum likelihood method was performed to test the factors inﬂuencing the ﬁndings. Results: After embolization, DWI yielded a reduction in the water diffusion coefﬁcient in the ADC map in all animals. Variance analysis showed a good correlation between CNR-measurements and infarction areas. Conclusion: DWI using parallel imaging techniques is a promising diagnostic tool for an early detection of acute mesenteric ischemia.

10:30 - 12:00

Room E1

Preclinical studies Moderators: I. Carrió; Barcelona/ES F.M.A. Kiessling; Heidelberg/DE

10:30

Targeted in vivo MRI of atherosclerosis with P947 and co-localization of ﬂuorescent Europium-P947 with matrix metalloproteinases in atherosclerosis S. Amirbekian1, V. Amirbekian2, F. Hyaﬁl3, E. Vucic3, J.S. Aguinaldo3, E. Lancelot4, C. Corot4, Z.A. Fayad3; 1Atlanta, GA/US, 2Boston, MA/US, 3New York, NY/US, 4 Paris/FR ([email protected]) Purpose: MMPs are implicated in atherosclerosis and in plaque rupture. P947 (Guerbet) is a peptide ligand for MMPs with a DOTA-Gadolinium chelate attached, making it into an MRI contrast-agent. Here we investigated the in vivo efﬁcacy of P947 in detecting atherosclerotic plaques and the localization of ﬂuorescent Europium-P947. Methods and Materials: Atherosclerotic ApoE-/- mice (n=15) underwent in vivo MRI of the abdominal aorta using 9.4T-MRI. Pre-contrast-enhanced (CE) and post-CE MRI were performed at 1, 2, 3 and 22 hours post-injection. As one control (n=4), we used the peptide contrast-agent P1135, which is P947 with the peptide sequence scrambled (speciﬁcity for MMPs is lost). As a second control, ApoE-/- mice (n=5) were injected with the nontargeted Gd-DOTA. Fluorescent EuP947 was made using Europium instead of Gadolinium. Eu-P947 was injected in ApoE-/-mice to track where P947 goes using laser-scanning confocal-microscopy. Fluorescence co-staining for MMP-2, 3 and 9 was peformed to investigate possible co-localization with Eu-P947. Results: In ApoE-/-mice using P947 the change in MRI signal measured by the NER was 2.25p0.19 (125% enhancement) at one hr, 1.74p0.14 (74%) at two hr, 1.31p0.13 (31%) at three hr, and 1.18p0.06 (18%) at 24 hr. Using P1135 (P947 scrambled) we saw enhancements of only 1.31p0.14 (31%) in the ApoE-/- mice at one hr, 1.19p0.11 (19%) at two hr, 1.09p0.1 (9%) at three hr, and 1.07p0.05 (7%) at 24 hr. MMP zymography demonstrated signiﬁcant MMP activity in imaged sections. Laser-scanning confocal-microscopy revealed co-localization of ﬂuorescent Eu-P947 with speciﬁc MMPs in atherosclerotic plaques of ApoE mice. Conclusion: Targeting MMPs with P947 showed highly signiﬁcant MRI enhancement of aortic atherosclerotic plaque in Apo E-/-mice. Eu-P947 speciﬁcally colocalized with MMPs in plaques.

146

A

B

C

D

10:48

Diagnosis of granulation tissue after lumbar spine surgery with an ICGenhanced integrated OI-/X-ray imaging system R. Meier1, S. Boddington1, C. Krug2, F.L. Acosta1, T.D. Henning1, S. Tavri1, E.J. Sutton3, E.J. Rummeny2, J.C. Lotz1, H.E. Daldrup-Link1; 1San Francisco, CA/US, 2 Munich/DE, 3Boston, MA/US ([email protected])

SS 206b

B-021

Purpose: Myeloid Related Proteins (MRP) 8 and 14, cytoplasmic proteins of phagocytes, have been demonstrated to be released during phagocyte-activation. An increased expression of MRP8/MRP14 is found in numerous inﬂammatory diseases, such as sepsis, rheumatoid arthritis or cancer at sites of inﬂammation amplifying the inﬂammatory responses of phagocytes. The purpose of this study was to establish a ﬂuorescence-labelled antibody targeted to MRP14 for in vivo monitoring of inﬂammation particularly with regard to phagocyte activity. Methods and Materials: An Anti-MRP14 antibody was coupled to Cy5.5-NHS-ester. To induce allergic eczema, the right ears of female BALB/C mice were exposed to Croton Tiglium D4, while left ears remained untreated as controls. 24 hours after induction of inﬂammation, mice were iv. injected with the ﬂuorochrome (2nM AntiMRP14-Cy5.5, n=7). To discriminate unspeciﬁc label distribution IGG-Cy5.5 (n=5) or free Cy5.5 (n=5) were injected instead of Anti-MRP14-Cy5.5. MRP14-knock-outmice (n=3) served as additional controls (n=3). Fluorescence Reﬂectance Imaging (FRI) was performed up to 96 hours p.i. and contrast-to-noise-ratios (CNR) were calculated. For statistical analysis, unpaired t-tests were performed. Results: In BALB/C-mice, injection of Anti-MRP14-Cy5.5 resulted in CNR which was more than ten-fold higher compared to those of knock-out-mice (115.07 vs 10.52; p 0.01) and still signiﬁcantly higher than CNR after injection of IGG-Cy5.5 (25.03; p 0.01) or free-Cy5.5 (54.74; p 0.015), conﬁrming that the measurable signal was indeed due to probe-to-target-binding. With decreasing inﬂammatory response 48 h after dye-injection, ﬂuorescence-intensity started to fade. Conclusion: Anti-MRP-Cy5.5 combined with FRI allows sensitive and speciﬁc detection of inﬂammatory activity represented by MRP14 expression in vivo.

B-023

Molecular Imaging

10:39

MRP-targeting in epidermal inﬂammation: A mouse-model M. Eisenblaetter, T.J. Vogl, A. Wall, W.L. Heindel, J. Roth, C. Bremer; Münster/DE ([email protected])

E

F

G

Purpose: To establish a sensitive, quick and inexpensive imaging tool for the diagnosis of postoperative granulation tissue. Methods and Materials: 12 male Sprague Dawley rats underwent intervertebral disk surgery of the lumbar spine. The spine of these rats was investigated with a new, integrated optical imaging (OI)/X-ray imaging system (Kodak In-Vivo Imaging FX) at 7 and 14 d after surgery. Each scan was obtained before and continuously up to 25 minutes after intravenous injection (p.i). of 1 mg/kg (n=6) or 10 mg/kg (n=6) indocyanine green (ICG). The ﬂuorescence signal of paravertebral soft tissues was compared between different time points after surgery and different ICG dose groups using the Wilcoxon test. Integrated OI/X-ray data were compared with histopathology. Results: In both dose groups, ICG provided a signiﬁcant enhancement of soft tissue in the area of surgery (p 0.05), which correlated to granulation tissue on histopathology. The dose of 1 mg/kg provided an early and short enhancement; the dose of 10 mg/kg provided a delayed and prolonged enhancement. The peak enhancement and the time interval of signiﬁcant enhancement of this paravertebral tissue was signiﬁcantly higher in the 10 mg/kg group than in the 1 mg/kg group (p 0.05). The enhanced paravertebral soft tissue correlated to granulation tissue on histopathology. Integrated OI/X-ray images allowed an accurate depiction of the anatomical location and extent of this granulation tissue, which correlated with the corresponding histopathology. Conclusion: ICG-enhanced OI/X-ray images provide a sensitive, fast and inexpensive tool for the diagnosis of granulation tissue after lumbar spine surgery.

B-024

10:57

Utility of a herpes oncolytic virus for the detection and imaging of neural invasion by cancer P. Brader, Z. Gil, K. Kelly, J.P. Shah, Y. Fong, H. Hricak, R.G. Blasberg, R.J. Wong; New York, NY/US ([email protected]) Purpose: Prostate, pancreatic, and head and neck carcinomas have a high propensity to invade nerves at an early stage of disease. The development of an imaging method that is able to detect nerves invaded by cancer (NIC) may potentially guide surgical resection. Methods and Materials: The infection of pancreatic (MiaPaCa2), prostate (PC3 and

Scientific Sessions

B-025

11:06

Spatio-temporal control of the intra renal expression of a reporter gene, in vivo via a cellular vector: A feasibility study in rats O. Eker, F. Couillaud, B. Quesson, N. Frulio, J. Arsaut, C. Deminière, C.T. Moonen, N. Grenier; Bordeaux/FR ([email protected]) Purpose: To test the feasibility of controlling spatio-temporally the intra-renal expression of a reporter gene by using a heat-inducible promoter, via a cellular vector, in rats. Methods and Materials: Luciferase expression was characterised both in vitro using luminometry and in vivo using bioluminescence imaging (BLI), within three genetically modiﬁed cell lines (Glial C6 or mesenchymal stem cells) expressing the ﬁreﬂy luciferase gene under control of either a constitutive or a heat-inducible promoter (Hsp-70B). Eight rats underwent a kidney heating, after surgical superﬁcialization, using focused ultrasounds (FUS) controlled by real time MR thermometry. After intra-arterial injection, engineered cells were followed-up within kidney by luciferase activity. Results: Kidneys from eight rats were heated at 43 °C, for 2 min. The standard deviation of MR thermometry was p1 °C. Histological analysis revealed no lesion (n=5), focal glomerular and tubulo-interstitial necrosis (n=2) and ﬁbrinoid necrosis (n=1). Among the cell lines tested in vitro, a C6 Hsp-luc cell line was found to exhibit a high rate of luciferase expression after heating. Injected in the kidney, this cell line allowed in vivo photonic signal detection with BLI after heating by FUS. Conclusion: MRI guided FUS technique allows a focal and precisely controlled, low level hyperthermia of rat kidney in vivo, with limited histological consequences. Using a cell line expressing luciferase under control of a heat-inducible promoter, light is detectable in vivo within the superﬁcialized kidney cortex. This opens possibility for in vivo spatio-temporally controlled transgene expression in the kidney and follow-up with BLI in rats.

B-026

11:15

Detection of breast cancer with MR imaging using the folate-receptor targeted contrast agent P1133 R. Meier1, T.D. Henning1, S. Tavri1, S. Arora1, E.J. Rummeny2, C. Corot3, H.E. Daldrup-Link1; 1San Francisco, CA/US, 2Munich/DE, 3Paris/FR ([email protected]) Purpose: Due to its over-expression on cancer cells, the folate receptor (FR) presents a promising target for tumor-speciﬁc contrast agents. The purpose of this study was to assess the uptake of a new FR-targeted USPIO in breast cancers. Methods and Materials: The FR-targeted USPIO P1133 (Guerbet, France) was incubated for 24 h with six different FR-positive human breast cancer cells, with and without free folic acid (FFA) as competitor. Labeling efﬁciencies were evaluated by MRI and ICP-mass-spectrometry. Subsequently, 12 athymic rats with implanted MDA-MB-231 breast cancers underwent MRI at 3 T before, up to 1 h and 24 h p.i. of P1133 (n=6), P1133+FFA (n=3) or the non-FR-targeted USPIO P904B (n=3). Tumor signal-to-noise-ratios (SNR) were compared before and after USPIO-injection and between different animal groups using generalized estimating equations. MR data was correlated with histopathology. Results: MR and spectrometry data showed different levels of P1133 uptake in different tumor cells. The P1133 uptake was highest in MDA-MB-231 cells and was inhibited in vitro by adding FFA. In vivo studies demonstrated a progressive enhancement of central tumor areas with both P1133 and P904B. Corresponding SNR data were signiﬁcantly higher for P1133 compared to P904B, indicating at least

a component of FR-speciﬁc enhancement with P1133. The P1133 tumor uptake was not signiﬁcantly inhibited by FFA, most likely due to its rapid metabolism in the liver and the shorter half-life of FFA compared to USPIO. Conclusion: The FR-targeted USPIO P1133 provides a signiﬁcant enhancement in MDA-MB-231 breast cancers due to a combined non-speciﬁc and speciﬁc accumulation in the tumor tissue.

B-027

11:24

1.5 T and 7T in vivo magnetic resonance imaging of SPIO-labeled pancreatic islets transplanted into streptozocin-induced mouse model of type-1-diabetes A. Esposito, F. De Cobelli, M.L. Malosio, A. Poletti, C. Losio, T. Canu, P. Mafﬁ, A. Secchi, A. Del Maschio; Milan/IT ([email protected]) Purpose: Pancreatic islets transplantation is a promising and effective therapy for type-1-diabetes. However, the success is not always guaranteed and monitoring graft loss should be a very important issue. Aim of the study was to label islets with clinically approved super-paramagnetic-iron-oxide agent (SPIO) and to visualize them in-vivo, after transplantation, with both 7.0T small-bore and 1.5 T clinical magnets. Methods and Materials: Murine and human islets were labeled by 16-hours incubation with SPIO (Endorem,Guerbet) after optimization of different parameters to achieve the best possible uptake preserving viability, insulin secretion, gene expression proﬁle and proinﬂammatory status. The labeled islets (SPIO-islets) were transplanted by portal vein infusion in four mice and under the kidney capsule in the other four. All mice had streptozocin-induced diabetes. In-vivo MRI was performed once a week for six months, at 1.5 T (Philips) and 7T (Bruker), including T2*sequences. Two mice were sacriﬁced for histological analysis. Results: SPIO-islets did not show altered viability, reduced insulin secretion or higher activation of pro-apoptotic pathways, in-vitro. Diabetic mice transplanted with 450 SPIO-islets became normoglycemic. SPIO-islets were detectable in-vivo by MRI both at 1.5 T and 7T, up to six months after transplantation. Histological analysis conﬁrmed the presence of SPIO-islets in the sites indicated by MRI. Conclusion: Our results showed that islets can be labeled with SPIO and imaged at 7.0T and 1.5 T. This approach could be very helpful to improve the success rate of graft. The use of a commercial agent and a clinical magnet makes our protocol easy to be transferred to humans.

B-028

11:33

MRI assays for the evaluation of dimethyl sulfoxide (DMSO) effects on experimental human breast cancer vasculature C.C. Cyran, B. Sennino, B. Chaopathomkul, Y. Fu, V.S. Rogut, M.F. Wendland, D.M. Shames, D.M. McDonald, R.C. Brasch; San Francisco, CA/US ([email protected]) Purpose: To evaluate the potential of dynamic contrast-enhanced MRI assays of endothelial permeability and fractional plasma volume to investigate and quantify effects of dimethyl sulfoxide (DMSO) on experimental human breast cancers and correlate MRI results with microscopy. Methods and Materials: 14 athymic rats implanted with human breast cancer xenografts (MDA-MB-435) were imaged at baseline and again after a one-week, three-injection treatment course of DMSO by dynamic MRI at 2.0T following enhancement with the macromolecular contrast agent albumin-Gd-(DTPA)27-(biotin)11. Quantitative MRI estimates of cancer microvessel permeability (KPS; μl/min/100 cm3) and tumor vascular richness (fPV; %), based on a two-compartment kinetic model, were compared to microscopic measurements of vascular density and contrast media extravasation. Results: DMSO suppressed endothelial permeability signiﬁcantly. KPS values in the DMSO-treated tumors (n=7) yielded a signiﬁcant decrease in endothelial permeability from baseline to day 7 (KPSbaseline=185p89 to KPS day7=51p92; p 0.05). KPS values in the saline-treated tumors (n=8) did not change signiﬁcantly. The amount of extravasated albumin-Gd-(DTPA)27-(biotin)11, as assayed by a streptavidin-stain for biotin, was signiﬁcantly lower in DMSO-treated cancers than in saline-treated cancers (58p6 vs 34p5, p 0.05). Conclusion: Reductions in cancer microvascular leakiness induced by a one-week course of DMSO could be detected and quantiﬁed by dynamic MMCM-enhanced MRI and were conﬁrmed by microscopic measurements of the leaked macromolecular agents in the same cancers. Results support the robustness of an MMCMenhanced MRI approach to the characterization of cancers as well as providing ﬁrst evidence for an in vivo effect of DMSO on cancer blood vessels.

A

B

C

D

E

F

G

147

Friday

DU145), and adenoid cystic carcinoma (ACC3) cell lines with replication-competent oncolytic herpes simplex virus 1 (HSV1) that expresses eGFP (NV1066) was assessed in vitro using eGFP expression. A murine model of NIC was established by implanting tumors into the sciatic nerves of nude mice and then nerves were injected with the NV1066 virus. Positron emission tomography (PET) scanning with [18F]-FEAU and intraoperative ﬂuorescent stereoscopic imaging were performed 48 hours after infection. Results: Infection of all three cell lines with NV1066 induced high expression of eGFP that was effectively imaged with ﬂuorescence microscopy. PET scanning using [18F]-FEAU showed signiﬁcantly higher uptake in NV1066-treated NIC animals than in controls. Also intraoperative ﬂuorescent stereoscopic imaging revealed strong eGFP signal only in NV1066-treated NIC animals. NV1066 virus was conﬁrmed by polymerase chain reaction. Conclusion: We developed an imaging method for detection of NIC using a herpes simplex vector (HSV) NV1066, expressing tyrosine kinase (HSV1-TK) and enhanced green ﬂuorescent protein (eGFP). This method may improve the diagnosis and treatment of patients with neurotrophic cancers by facilitating identiﬁcation of inﬁltrated nerves and requiring resection, thus reducing injury to normal nerves.

Scientific Sessions B-029

10:30 - 12:00

11:42

Monitoring E. coli Nissle 1917 colonization of tumors with PET and optical imaging P. Brader1, J. Stritzker2, C.C. Riedl1, J. Grimm1, A.A. Szalay2, Y. Fong1, R.G. Blasberg1, H. Hricak1; 1New York, NY/US, 2San Diego, CA/US ([email protected]) Purpose: Bacteria-based tumor-targeted therapy is a modality of growing interest in anti-cancer strategies. Imaging bacteria speciﬁcally targeting and replicating within tumors using radiotracers and optical imaging would provide conﬁrmation of successful colonization of malignant tissue in humans. Methods and Materials: The uptake of radiolabeled pyrimidine nucleoside analogs and [18F]-FDG by E.coli Nissle 1917 (EcN) was assessed in vitro and in vivo. Targeting of EcN to 4T1 breast tumors in nude mice was monitored by PET and optical imaging. The correlation of radiotracer accumulation in the tumors with the number of bacteria was investigated. Optical imaging based on bioluminescence was performed using EcN bacteria that encode luciferase genes under the control of an L-arabinose inducible promoter system. Results: We showed that EcN can be detected using radiolabeled pyrimidine nucleoside analogs, [18F]-FDG and PET. Importantly, this imaging paradigm does not require transformation of the bacterium with a reporter gene. Imaging with 18 F-FDG provided lower contrast than 18F-FEAU due to high FDG accumulation in control (non-treated) tumors and surrounding tissues. A linear correlation was demonstrated between and the number of viable bacteria in tumors and the accumulation of [18F]-FEAU, but not [18F]-FDG. The presence of EcN was also conﬁrmed by bioluminescence imaging. Conclusion: EcN can be imaged by PET, using the expression of endogenous E.coli thymidine kinase. This imaging paradigm could be translated to patient studies for the detection of solid tumors. Optical imaging provides a low cost alternative to PET imaging in small animals.

B-030

11:51

Molecular proﬁling of tumor neovascularisation by multi-target molecular ultrasound imaging M. Palmowski1, J. Huppert1, G. Ladewig2, P. Hauff3, M. Reinhardt3, M.M. Mueller1, M. Maurer2, W. Semmler1, F. Kiessling1; 1Heidelberg/DE, 2Erlangen/DE, 3Berlin/DE ([email protected]) Purpose: Molecular ultrasound is capable of elucidating the expression of angiogenic markers in vivo. However, it is still uncertain whether it can assess changes in the molecular proﬁle of tumors during therapy. Particularly, the investigation of more than one target within one examination has not been reported. Methods and Materials: In this study, the expression of the vascular endothelial growth factor receptor 2 (VEGFR2) and of AvB3-Integrins in squamous cell carcinoma xenografts of the HaCat-ras RT3 cell line in mice treated with the MMP-inhibitor AG3340 was investigated by molecular ultrasound. Therefore, cyanoacrylate microbubbles linked to speciﬁc ligands were generated and their accumulation in tumor vessels was determined by the quantitative ultrasound method SPAQ (sensitive particle acoustic quantiﬁcation). Results: Due to the rapid blood clearance and destruction of stationary microbubbles by ultrasound, consecutive assessment of both angiogenic markers was successful. Molecular ultrasound indicated a signiﬁcant increase of VEGFR2 and AvB3-Integrin expression during tumor growth and a considerable decrease in both marker densities after MMP-inhibitor treatment. Histological data suggests decreased vessel density after treatment as potential cause for decreased marker densities. Conclusion: Targeted ultrasound quantiﬁed by SPAQ is suited for molecular proﬁling of tumor angiogenesis and for the sensitive assessment of therapy effects.

148

A

B

C

D

E

F

G

Room F1

Genitourinary

SS 207 The uterus Moderators: E.J. Adam; London/UK R. Girometti; Udine/IT

B-031

10:30

Dynamic MR-hysterosalpingography (dMR-HSG) replacing conventional X-ray HSG: First experience W. Pegios1, C. De Geyter1, S. Steimann1, J. Fröhlich2, W. Steinbrich1; 1Basle/CH, 2 Zurich/CH ([email protected]) Purpose: To optimize the new method for imaging of the uterine cavity and Fallopian tube patency by dynamic MR-HSG without the procedure of conventional HSG. Methods and Materials: 37 consecutive infertile women underwent MR imaging in a 1.5 T MR unit (Avanto, Siemens, Erlangen). For imaging of the true pelvis standard axial and sagittal T2-TSE-FS sequences were acquired. A balloon catheter (Ch. 5, Rüsch, Kernen, Germany) was placed into the uterine cavity and blocked at its position. Dynamic MR-HSG consisted of injection of 20 ml of gadolinium-polyvidone solution (18.4 mM Dotarem 1:20 with Polyvidon) into the uterine cavity while acquiring ﬁve consecutive FLASH-3D T1-weighted MR-sequences. Results: Dynamic MR-HSG allowed good 3D visualization of the uterine cavity and conﬁrmation of bilateral Fallopian tube patency (n=27). In patients who showed a one-sided occlusion of the Fallopian tubes (n=3) catheterisation was performed. In case of double-sided occlusion (n=3) correlation with laparoscopy ﬁndings was performed. However, in two of these patients (n=4) the examination was still partially diagnostic, causing dislocation of the catheter. Conclusion: More information about all organs of the true pelvis and especially of avoidance of exposure to ionizing radiation. In case of succeeding in, at least in a part of patients, achieving the Fallopian tube patency with selective catheterisation and with the passage of contrast medium in MR imaging, not only will it be possible to lower costs, but also to avoid unnecessary operations.

B-032

10:39

Hysterosalpingogram, does the instrument used make a difference? W.R. Davis, S. Punwani, C. Allen; London/UK Purpose: Hysterosalpingograms (HSG) at our institution are performed using three different instruments, the Leech Wilkinson cannula, the Margolin cannula and the HSG balloon catheters. A retrospective analysis on 214 examinations was conducted to determine whether the choice of instrument affected study performance and outcome. Method: Every HSG examination performed between November 2005 and November 2006 was reviewed by a consultant uro-radiologist (CA). The time taken to perform the procedure and the radiation dose used for each study were recorded. Studies were further evaluated for image quality (diagnostic or sub diagnostic quality) and report accuracy. Results: The Margolin cannula was used most often (154 procedures, Leech Wilkinson 34, HSG catheter 23) and was associated with a signiﬁcantly (p 0.001) lower radiation dose (31.6 cGy cm2) in comparison to the other devices. In addition procedures performed with the Margolin cannula took signiﬁcantly (p 0.05) less time (mean 129 seconds) than the other devices. However, use of the Leech Wilkinson catheter produced the highest proportion of studies of diagnostic quality (85%) compared to the Margolin (55%) and the HSG balloon catheters (68%). In addition 85% of Leech Wilkinson were accurate in comparison to 72% for the Margolin and 76% for HSG catheters. Conclusion: Use of the Leech Wilkinson cannula provided the best chance of obtaining diagnostic quality images and an accurate report. Choice of catheter was associated with statistically signiﬁcant differences in performance and interpretation characteristics of HSG studies.

Scientific Sessions B-033

10:48

Purpose: To describe CEUS ﬁndings before and after SUFE and to understand if CEUS can play a role in the assessment of outcomes and in SUFE follow-up. Methods and Materials: We assessed 51 ﬁbroids in 26 patients by CEUS immediately before and after embolization directly in the angiographic room. Due to ﬁbroids size (range 3.4-8.9 cm, mean size 5.1 cm), all ultrasound examinations were performed with a transabdominal approach. We assessed only the three largest ﬁbroids in case of multiple lesions. In all patients, a single bolus of contrast agent (4.8 ml of SonoVue®, Bracco, Italy) was injected. Follow-up protocol included CEUS evaluation after one month and CEUS and dynamic magnetic resonance (MR) evaluation after six months. Results: In 48/51 cases, CEUS showed total devascularization of the ﬁbroids immediately after embolization, with non signiﬁcant intralesional signals delayed in comparison to myometrial enhancement. In 3/51 cases of multiple lesions, one among the embolized ﬁbroids mantained its vascularization. These ﬁndings are consistent with angiographic data and with the purpose of performing a complete devascularization. No recurrence was observed during follow-up. One patient reported the reappearance of symptoms after 18 months and CEUS showed the persistence of intralesional vascularization. Conclusion: CEUS is effective to demonstrate the degree of vascular occlusion at the end of SUFE procedures. CEUS ﬁndings correlate with clinical results at one and six months. CEUS is reliable and cost-effective when compared to MR.

B-034

10:57

Novel MRI techniques in advanced cervical cancer show early prediction of treatment outcome V.N. Harry, S.I.K. Semple, D.E. Parkin, F.J. Gilbert; Aberdeen/UK ([email protected]) Purpose: The emergence of functional imaging techniques that can integrate pathophysiological and morphological changes offers great promise in providing early markers of tumour response and the development of customised patient-speciﬁc regimes. Diffusion weighted MRI (DWI) can non-invasively characterise biological tissues based on their water diffusion properties and allows for the calculation of the apparent diffusion coefﬁcient (ADC). Similarly, dynamic contrast-enhanced MRI (DCE-MRI) can exploit the information obtained from enhancement characteristics to assess vascularity changes. We investigated the use of DWI and DCE-MRI as early predictors of treatment outcome in women with advanced cervical cancer. Methods and Materials: Fifteen women with advanced cervical cancer who received concurrent chemo/radiotherapy were included. MR imaging (1.5 T) was carried out prior to treatment and repeated after two weeks of therapy. This included a diffusion weighted epi acquisition and dynamic T1 images using Gd-DTPA contrast. Change in ADC and change in peak and mean enhancement was tested for correlation with tumour response as determined by clinical assessment. Results: Percentage tumour response ranged from 10 to 80% (median 60%, SD 21.7%). The change in ADC (R=0.586, p=0.022) as well as change in peak (R=0.690, p=0.004) and mean (R=0.677, p=0.006) tumour enhancement after two weeks of treatment all showed a signiﬁcant correlation with clinical response. Conclusion: Diffusion and dynamic MRI parameters predicted treatment response as early as two weeks following the start of treatment. These techniques may ultimately prove to be useful biomarkers for predicting response and offer improved patient care by individualising treatment.

B-035

11:06

Diagnosis and evaluation of uterine cervical cancer using diffusionweighted imaging Y. Ren, Q.Y. Guo, Z.M. Lu; Shenyang/CN ([email protected]) Purpose: To evaluate the feasibility of diffusion-weighted imaging (DWI) in diagnosis and staging of cervical cancer of uterus. Methods and Materials: 17 consecutive female patients with cervical cancer and 15 female volunteers were included. Preoperative MRI examinations were performed. In addition to routine T1 weighted and T2 weighted imaging (T1WI and T2WI), DWI with b factors of 0 and 800 s/mm2 was also performed. The ADC values of tumor and normal cervical tissue were measured and were compared. The diagnostic accuracy acquired with routine T1WI and T2WI combined with DWI and ADC value was compared with that obtained from T1WI and T2WI only. The extent of cervical cancer was also evaluated in different sequences.

Results: On DWI, all the cervical cancers appeared with high signal intensity without clear layers of the cervix as compared with normal cervical tissue, and the ADC value of tumor was statistically lower (0.861p0.08x10-3 mm2/s) than that of normal tissue (cervical mucosa 1.83p0.37; conjunctive band 1.23p0.24; muscular layer 1.87p0.24x10-3 mm2/s). The diagnostic accuracy was 83.4% for T1WI and T2WI only, and increased to 88.2% when DWI and ADC were also taken into consideration. The extent and border of the cancerous foci were more obvious on DWI than on other sequences. Conclusion: DWI and ADC measurement has a potential ability to differentiate between cancer and normal tissues of the uterine cervix and is helpful for preoperative staging.

B-036

11:15

Early stage endometrial carcinoma: Value of diffusion-weighted MR imaging in the assessment of myometrial tumor invasion F. Invernizzi, G. Rechicchi, A. Lissoni, P. Perego, D. Ippolito, C. Mangioni, S. Sironi; Monza/IT ([email protected]) Purpose: To determine the value of diffusion-weighted MR imaging (DWI) in the assessment of myometrial invasion in patients with early stage endometrial carcinoma. Methods: This prospective study included 37 patients with clinical stage I endometrial carcinoma who underwent MR imaging and then radical hysterectomy. MR conventional protocol included multiplanar FSE T2-weighted, and dynamic gadolinium-enhanced SE T1-weighted images. Diffusion-weighted images were also acquired with b value of 0 and 800 s/mm2, using single-shot echo-planar sequences (SE EPI SS). To establish the additional value of DWI, myometrial tumor invasion was assessed with the use of conventional MR images alone, and with the combined use of conventional MR images and DWI. Histopathologic ﬁndings served as the reference standard. Fisher exact test was used for statistical analysis. Results: At histopathology, no myometrial tumor invasion was found in eight cases; invasion of the inner half of myometrium was found in 23 cases, and invasion of the outer half in the remaining six. The sensitivity, speciﬁcity, positive predictive value (PPV), negative predictive value (NPV), and accuracy of conventional MR images in evaluating the depth of myometrial tumor invasion were 84.5, 93.7, 89.3, 95.5 and 87.3%, respectively. The corresponding values for combined conventional MR images and DWI were 93.5, 94.6, 97.2, 96, and 91.6%. A signiﬁcant difference (p 0.005) was found between the sensitivity and PPV values obtained with the two evaluation modalities. Conclusion: DWI can play an important role in the assessment of myometrial invasion by endometrial carcinoma.

B-037

11:24

Differentiation between early endometrial carcinoma and endometrial hyperplasia using MR imaging: Emphasis on dynamic enhancement patterns and perfusion weighted images L. Xiang-Sheng, S. Yun-Long, Z. Wan-Shi, W. Dong; Beijing/CN ([email protected]) Purpose: To evaluate the usefulness of MRI including dynamic enhancement and perfusion weighted images in differentiating early endometrial carcinoma from endometrial hyperplasia. Methods and Materials: This study included 40 patients with early endometrial carcinoma (earlier than FIGO stage II) and 35 patients with endometrial hyperplasia, who received MR examination. The enhancement patterns on dynamic contrast-enhanced study and various MR perfusion parameters were compared between two pathologies. Results: On dynamic contrast-enhanced imaging, 67% (27/40) of endometrial cancers showed early peak enhancement to be reached within 1 min after administration of contrast medium and the enhancement gradually decreased. Eighty-six percent (30/35) of endometrial hyperplasia showed late peak enhancement to be reached in 2-3 min following intravenous administration of contrast material. There was statistically signiﬁcant difference in the enhancement patterns between early endometrial carcinoma and endometrial hyperplasia (C2=7.35, p 0.05), and the perfusion parameters including positive enhancement integral (PEI), time to peak (TP) and maximal slope of increase (MSI) of the group of endometrial carcinoma were lower than that of the group of endometrial hyperplasia, respectively (t=6.546, 7.019, 6.893, p 0.05). Conclusion: Different enhancement patterns on dynamic contrast-enhanced MR images and different perfusion parameters can provide useful information for the differentiation between endometrial carcinoma and endometrial hyperplasia.

A

B

C

D

E

F

G

149

Friday

Contrast-enhanced ultrasound (CEUS) versus dynamic magnetic resonance (MR) in the assessment of outcomes and in the follow-up of uterine ﬁbroids after superselective uterine ﬁbroids embolization (SUFE) P. Gazzo1, F. Lacelli1, L.M. Sconfienza2, M. Gravano1, N.M. Gandolfo1, N. Perrone2, G. Seraﬁni1; 1Pietra Ligure/IT, 2Genoa/IT (io@lucasconﬁenza.it)

Scientific Sessions B-038

11:33

Number of nodes and myometrial invasion better than tumor volume in predicting disease free survival in cervical cancer patients undergoing radiation therapy M.A. Haider, E. Maheshwari, A. Fyles, G. Lockwood, M. Milosevic; Toronto, ON/CA Purpose: To assess the value of node status and myometrial invasion as independent predictors of disease free survival (DFS) in cervical cancer patients undergoing chemo and radiation therapy. Methods and Materials: 91 cervical cancer patients underwent MRI prior to therapy as part of this prospective trial from 1999 to 2006. All nodes 5 mm in short axis were measured and classiﬁed as below (BB) and at or above the common iliac bifurcation (AAB). Maximum node size, number of nodes 8 mm, number of nodes 5 mm, presence of nodes 8 mm AAB, age, clinical FIGO stage, myometrial invasion ( 2 cm), and tumor volume were assessed. A Cox regression model was used to determine independent predictors of DFS. Subgroup testing was preformed for 39 patients with no nodes 8 mm. Results: The number of nodes 8 mm (p=0.002) was the single best predictor of DFS. The only other non-nodal parameter that was an independent predictor was myometrial invasion (p=0.02). Although the number of nodes 5 mm was an independent predictor of DFS, the model ﬁt was better for an 8 mm size threshold. FIGO stage, node location and tumor volume were not independent predictors. When considering patients without lymphadenopathy (no nodes 8 mm), myometrial invasion was the most signiﬁcant independent predictor of DFS (p=0.04) while age, tumor volume and FIGO stage were not. Conclusion: The number of nodes 8 mm and myometrial invasion 2 cm are independent factors predicting DFS in cervical cancer patients undergoing radiation therapy. In patients without enlarged nodes, myometrial invasion may be the best independent predictor of DFS.

B-039

11:42

Lymph node involvement in uterine cervical cancer: A comparison between PET-CT and MRI with diffusion sequences P.-Y. Roudaut, M. Pouquet, J. Monteil, Y. Aubard, A.J.M. Maubon; Limoges/FR ([email protected]) Purpose: Compare sensitivity and speciﬁcity for the detection of malignant lymph node involvement in a series of uterine cervical cancers. Methods and Materials: Consecutive series of female patients with biopsy proven cervical cancer. All patients, prior to treatment, had a PET-CT (Siemens biograph 6) and pelvic MRI at 1.5 T (Philips Achieva) with Diffusion Weighted Imaging sequences (DWI) (b=0 and 1000 sec m-2; acquisition time 52 sec). Comparison of scores of nodal detection on a patient basis, with histologic analysis after laparoscopic lymph node ablation. Results: Sixteen patients were included (mean age 58.3 years, range 27-78 years). FIGO stages of cervical cancer: stage Ib n=4, stage IIa n=8, stage IV n=4. Sensitivity and speciﬁcity were 94 and 86% versus 100 and 60% for PET-CT and MRI with DWI, respectively. Conclusion: PET-CT and DWI during MRI are complementary in the evaluation of lymph node involvement. MRI evaluation of cervical cancer should include a DWI sequence.

B-040

11:51

Fusion of high b-value diffusion-weighted and T2-weighted MR images improves identiﬁcation of lymph nodes in the pelvis N. Mir, A. Sohaib, D. Collins, D.-M. Koh; London/UK ([email protected]) Purpose: Accurate identiﬁcation of lymph nodes facilitates nodal assessment by size, morphological and MR lymphographic criteria. We compared the MR detection of lymph nodes in patients with pelvic cancers using T2-weighted images and fusion of diffusion-weighted (DWI) and T2-weighted images. Methods and Materials: 20 patients with pelvic tumours underwent 5 mm axial T2weighted and DWI (b-values 0-750 s/mm2) MR imaging on a 1.5 T system. Fusion images of b=750 s/mm2 DWI and T2-weighted images were created. Two radiologists evaluated in consensus the T2-weighted images and fusion images independently. For each image set, the location and diameter of pelvic nodes were recorded; and nodal visibility scored using a four-point scale (0-3). Nodal visualisation was compared using RIDIT analysis. The mean RIDIT score describes the probability that a randomly selected node will be better visualised relative to the other image set. Results: 114 pelvic nodes (mean 5.9 mm; 2-10 mm) were identiﬁed on T2-weighted images and 161 nodes (mean 4.3 mm; 2-10 mm) on fusion images. Using fusion images 47 additional nodes were detected compared with T2-weighted images (eight external iliac, 24 inguinal, 12 obturator, two peri-rectal, one presacral). Nodes detected only on fusion images were 2-9 mm (mean 3.7 mm). Nodal visualisation was better using fusion images

150

A

B

C

D

E

F

G

compared with T2-weighted images (mean RIDIT score 0.689 versus 0.302). Conclusion: Fusion of DWI with T2-weighted images improves identiﬁcation of pelvic lymph nodes compared with T2-weighted images alone. This may aid future nodal characterisation in patients with pelvic cancers.

10:30 - 12:00

Room F2

Breast

SS 202 Advances in digital mammography Moderators: L. Khoo; London/UK R.L. Lavayssiere; Sarcelles/FR

B-041

10:30

Digital breast tomosynthesis versus mammography: Multireader study - ﬁrst results G. Gennaro, E. Baldan, E. Bezzon, M. La Grassa, L. Pescarini, C. di Maggio; Padova/IT ([email protected]) Purpose: To provide an initial assessment of clinical performance of digital breast tomosynthesis over standard mammography. Methods and Materials: A clinical study including symptomatic women was approved by the competent authorities to compare clinical performance of digital breast tomosynthesis (DBT) and standard mammography in a diagnostic population. Patients with informed consent underwent digital mammography (CC and MLO views) and tomosynthesis (MLO only) on both breasts. The DBT equipment was a clinical prototype based on a standard digital mammography unit (Senographe DS, GE Healthcare), modiﬁed to acquire 15 projections over a 40° arc and using dose levels equivalent to those of standard mammography examinations. A ﬁrst group of 45 women was evaluated independently by three experienced radiologists. Sets of DBT reconstructed slices, sampled at 1 mm intervals and standard digital mammograms of each breast were reviewed in separate sessions on a dedicated workstation and without additional patient information. Findings were localized and classiﬁed in terms of probability of cancer using a 7-step scale. Image assessments from each radiologist were compared with the truth and standard ROC analysis applied to a dataset of 89 breasts. Results: The areas under the ROC curve (AUC) for the three radiologists were 0.92, 0.85, 0.85 and 0.73, 0.76, 0.80 for tomosynthesis and mammography, respectively. The AUC difference was signiﬁcant in favour of DBT for one radiologist (p-values 0.003, 0.229, 0.522). Conclusion: Despite the limited number of cases and short clinical experience with tomosynthesis, results showed a potential beneﬁt of DBT over mammography.

B-042

10:39

Performance of breast tomosynthesis as an adjunct imaging modality to digital mammography A. Smith, L. Niklason, Z. Jing; Bedford, MA/US ([email protected]) Purpose: To test the efﬁcacy of the use of tomosynthesis as an adjunctive imaging modality to digital mammography. Methods and Materials: Forty patients were imaged using both digital mammography (2D) and digital breast tomosynthesis (3D), and ground truth was collected for these cases. The cases included normals, recalls, benign biopsies, and cancers. 14 breast radiologists with zero-minimal experience with tomosynthesis were trained to read tomosynthesis images in a four-hour learning session. They then read the blinded cases, ﬁrst evaluating the 2D images, and then the same case using both the 2D and 3D images. Cases were evaluated using BIRADS, and Probability of Malignancy (POM) metrics. Performance of the radiologists was evaluated to generate Receiver Operating Characteristics (ROC) curves. Results: ROC curves were generated from both the BIRADS and POM data. The radiologists’ performance improved when reading 2D+3D, compared to 2D alone, as measured using the area under the ROC curves. Cancers were given higher mean POM, and non-cancers lower mean POM when using 3D, indicating improved conﬁdence. The 2D+3D performance was seen after minimal training, and is indicative of the facility with which breast radiologists can learn to evaluate tomosynthesis images. Images illustrative of how 3D improves 2D evaluation will be shown. Conclusion: The inclusion of tomosynthesis images when reading digital mammograms improved radiologists’ performance as measured by ROC curve analysis. The use of tomosynthesis as an adjunctive imaging modality may speed the introduction of the technology into common practice, and improve clinical accuracy.

Scientific Sessions B-043

10:48

Purpose: We present preliminary data from the ﬁrst pilot study. This study compares detection and diagnosis of breast cancer with traditional mammography and Photon Counting Tomosynthesis. 3D mammography has the potential to improve detection and diagnosis, at a dose similar to conventional two-view mammography. Methods and Materials: Data were collected, over two days, using the ﬁrst prototype system on 13 patients recalled for suspicious ﬁndings at screening mammography. The detector is made of crystalline silicon which directly converts the X-rays to electronic charge, an intelligent electronic circuit processes each X-ray. Photon counting rejects all electronic noise and almost all scattered radiation, a particular advantage in 3D imaging when many images, each at a very low exposure, are required. The tomosynthesis exposure time was ﬁxed at three seconds. Three experienced radiologists read the tomosynthesis cases and the ﬁlm screen cases, three months apart, scoring each case for cancer probability on a scale from one (no cancer) to nine (deﬁnite cancer). Results: Currently, we have radiology opinion for 10 complete cases. When combined, the diagnostic performance was superior with tomosynthesis in six cases. Performance was superior with ﬁlm in two cases and equivalent in two cases. Radiation dose: mean average glandular dose for the 10 cases was 1.1 mGy for tomosynthesis compared to 1.05 mGy for ﬁlm. Conclusion: The new technique is promising; the clinical protocol worked well and only minor modiﬁcations are required for the full clinical trial to proceed within the HIGHREX project at ﬁve European sites.

B-044

10:57

Digital mammographic tomosynthesis system: First clinical results R. Schulz-Wendtland, E. Wenkel, L. Michael, B. Catrin, B. Werner, M. Thomas; Erlangen/DE ([email protected]) Purpose: To compare the sensitivity of digital two-dimensional (2-D) mammographic projection imaging with digital mammographic tomosynthesis with respect to the detectability and extension of breast lesions. Methods and Materials: Since Aug 2006, we examined 25 patients with microcalciﬁcations and masses ﬁrst with digital mammography in cc and oblique projection (Siemens Mammomat NovationDR). After diagnosis (BI-RADS 4/5) and before histological assessment, we used a digital breast tomosynthesis system based on the commercial FFDM system (Siemens Mammomat NovationDR), but modiﬁed for a wide angle tube motion and equipped with a fast read-out amorphous selenium detector. We acquired tomosynthesis series of projection views. As typical radiographic techniques, the anode ﬁlter combinations Mo/Mo and W/Rh were used. Slice images through the breast parallel to the detector were reconstructed with 1 mm distance employing a ﬁltered backprojection algorithm. The image data sets were read by ﬁve radiologists and evaluated with respect to the detectability and extension of the lesions. All lesions were histologically proved. Results: With tomosynthesis there was a higher, but not signiﬁcant, rate in the diagnostic detectability. The accuracy of description in the extension of the lesions compared with the histology was 25% more exact in contrast to digital 2-D mammographic projection. Conclusion: Digital mammographic tomosynthesis has the potential to increase the sensitivity in diagnostic detectability and in the description of the extension of microcalciﬁcations and masses in contrast to 2-D digital mammography, but further studies are necessary for conﬁrmation of the results.

B-045

11:06

Challenges introduced by using telematics for soft copy 2nd reading of digital mammography images acquired in independent institutions A. Van Steen, C. Van Ongeval, T. Deprez, E. Bellon, G. Marchal; Leuven/BE ([email protected]) Purpose: Our radiologists use workstations for second reading for breast cancer screening on images sent over the Internet by sites that operate widely different acquisition devices and information systems. From our experience we draw attention to technical-organizational challenges in such a decentralized setup. Methods and Materials: Hospitals and private radiologists acquire mammograms using CR/DR devices from different manufacturers, and connect to the communication infrastructure using different PACS/RIS solutions. This freedom of choice is fundamental as our central service for second reading must be able to support

many radiological centers. The images arrive in our central PACS to be reported using soft copy reading. Results: Challenges included providing efﬁcient reporting work lists (particularly in this transition phase in which hard and soft copy are mixed), up-front detection of transmission failures (which can be automated only if additional information about ﬁrst reading is transmitted electronically) and providing follow-up in a situation in which different institutions do not deploy a common ID for all patients. We were surprised to ﬁnd out that a problem of suboptimal image presentation quality in our central screening center was caused by a reconﬁguration error in an intermediate PACS at the distant site. Conclusion: Most problems are not directly related to technology but to difﬁculties in aligning organizational aspects. Remedies require organizational re-engineering. Although Internet transmission as such does not inﬂuence the images, seemingly unrelated parts of the complex overall transmission chain may ultimately affect image quality. Remedies for this may require fundamentally reconsidering image quality assurance programs.

B-046

11:15

Overlooked ﬁndings of computer-aided detection (CAD) applied to screening full-ﬁeld digital mammograms previously obtained in women with breast cancer N. Cho, W.K. Moon; Seoul/KR ([email protected]) Purpose: To evaluate the lesions, marked by computer-aided detection (CAD), judged as non-actionable by radiologists, and later developed breast cancers in screening full-ﬁeld digital mammograms. Methods and Materials: During a 23-month period, commercially available CAD system (ImageChecker M1000-DM, version 3.1; R2 Technology, Sunnyvale, CA), embedded in full-ﬁeld digital mammography (Senographe 2000D, GE Medical Systems, Buc, France) was prospectively applied to the over 30,000 screening mammograms. Each mammogram was interpreted by one of three radiologists after review of CAD marks. All images with CAD marks were saved in a picturearchiving communication system. Thirty-eight mammogram pairs - an initial negative screening mammogram and a subsequently obtained mammogram showing cancer - were collected. The median interval between examinations was 12 months (range 5-20). Two radiologists retrospectively analyzed the lesions marked by the CAD, but judged as non-actionable by radiologists initially. Results: On the initially negative screening mammograms, of the 38 areas where cancer later developed, the CAD system marked 18 lesions (47%; 11 masses, 7 calciﬁcations). Of the 18 cancers, 17 were invasive and one was ductal carcinoma in situ. Eight (44%) of the 18 lesions were reassessed as actionable and 10 (56%) lesions were reassessed as non-actionable. Among the eight overlooked lesions, the reason of overlooking was distracting lesion (2, 25%), misinterpretation (5, 63%), and slow growing lesion (1, 13%). Conclusion: The CAD system marked correctly 47% (18/38) of the initially negative screening mammograms previously obtained in women with breast cancer. Fortyfour (8/18) % of the correct CAD markings were overlooked by radiologists.

B-047

11:24

Effect of introduction of digital mammography with CAD in a population based screening program N. Karssemeijer1, D. Beijerinck2, R. Visser1, J. Deurenberg2, M. Beekman2, A. Bartels-Kortland2; 1Nijmegen/NL, 2Utrecht/NL ([email protected]) Purpose: To compare digital and ﬁlm-based screening three years after introduction of full ﬁeld digital mammography (FFDM) and CAD. Methods and Materials: In a breast cancer screening center, one of six conventional mammography units was replaced by FFDM (Selenia, Hologic). The center is part of a nationwide organisation in which women aged 50-75 are screened every two years. Radiologists used a dedicated mammography workstation and CAD. Six radiologists were involved in reading. All mammograms were double read. We compared recall rates and detection in the ﬁrst three years of the project. Results: In total 212,194 ﬁlm-based and 23,767 digital screening exams were carried out, of which 26,098 were initial screens with ﬁlm and 5,792 were initial with FFDM. With ﬁlm 1049 cancers were found while FFDM detected 136 cancers. Recall rates for initial and subsequent screens were higher with FFDM (4.4 and 1.8%) than with ﬁlm (2.3 and 1.1%). Recall rates for initial and subsequent FFDM screenings gradually decreased to 3.8 and 1.6% during the project. The detection rate for initial screenings was higher with digital mammography, 0.76% compared to 0.56% with ﬁlm, though the difference was not signiﬁcant (p=0.053). For subsequent screenings the detection rate was similar for digital and ﬁlm, 0.51 vs 0.48%. The fraction of DCIS found with FFDM was signiﬁcantly larger than with ﬁlm, 27.7 vs 17.5% (p 0.001).

A

B

C

D

E

F

G

151

Friday

HIGHREX, photon counting tomosynthesis: A new concept for X-ray imaging of the breast M.G. Wallis1, C. Wright2, U. Bick3, M. Dalielsson4, F. Diekmann3, W.L. Heindel5, K. Leiﬂand4, B. Seradour6, M. Aslund4; 1Cambridge/UK, 2Birmingham/UK, 3Berlin/DE, 4 Stockholm/SE, 5Munster/DE, 6Marseille/FR ([email protected])

Scientific Sessions Conclusion: Recall rate increased after introduction of FFDM with CAD and detection increased for initial screens. Signiﬁcantly more DCIS was found with digital screening.

B-048

11:33

Electronic digital zooming vs geometric digital magniﬁcation of microcalciﬁcations: A comparative study E. Vianello, M. Lorenzon, V. Londero, A. Linda, B. Falcone, C. Zuiani, M. Bazzocchi; Udine/IT ([email protected]) Purpose: To compare Electronic Digital Magniﬁcation Mammography (EDMM) and Geometric Digital Magniﬁcation Mammography (GDMM) for detection and characterization of microcalciﬁcations. Methods and Materials: 68 patients with microcalciﬁcations undergoing stereotactic vacuum-assisted biopsy were enrolled. For each patient, a 1.8X magniﬁcation of microcalciﬁcations was obtained with both EDMM and GDMM. Image quality (1-5), number of microcalciﬁcations, Le Gal and BIRADS classiﬁcation were rated by two experienced readers (A-B) and by seven mid-experienced radiologists, independently. BIRADS 4-5 resulting in invasive or in-situ carcinoma at histologic examination were considered as true positive. Results: Comparing EDMM to GDMM, sensitivity, speciﬁcity and overall accuracy were respectively 79-89%, 63-61%, 68-69% for A, and 95-89%, 59-61%, 69-69% for B. No signiﬁcant differences (chi-square test, p 0.05) between intra- and interobserver values were found. Cohen's Kappa (K) applied over diagnostic accuracy data was 0.714-0.735 (A-B), while agreement in BIRADS evaluation, tested with Weighted Cohen's Kappa (wK), was 0.698-0.586. Microcalciﬁcation classiﬁcation (Le Gal) was different in 12/68 cases for A (K=0.703) and in 17/68 cases for B (K=0.545). Agreement on number of microcalciﬁcation evaluations resulted in a wK of 0.297-0.441 (A-B). Image quality average value was 3.16-3.20 (wK=0.297) for A, and 4.06-4.07 (wK=0.255) for B. Data related to mid-experienced radiologists are heterogenous, but mostly similar to those of experienced radiologists. Conclusion: Compared to EDMM, GDMM reduces the number of both exposures and recalls, without a signiﬁcant difference in diagnostic accuracy, BIRADS and Le Gal evaluation (good agreement), and in number of microcalciﬁcation evaluations and image quality (fair agreement).

B-050

11:51

Monitor reading of lossy compressed digital mammograms R.W.R. Loose1, R. Simmler1, K. Detmar1, M. Wucherer1, R. Adamus1, A. Linke1, R. Schulz-Wendtland2; 1Nuremberg/DE, 2Erlangen/DE ([email protected]) Purpose: Evaluate the range of compression factors without loss of diagnostic image quality in digital mammography with monitor reading. Methods and Materials: Mammograms with normal ﬁndings, with solid masses and with micro calciﬁcations were acquired with a digital FD and CR system (Mammomat 3000 Nova, Siemens, 24x29 cm, matrix 3328x2560, 12 bit, 70 μm and CR 75.0, Agfa, 8x10”, matrix 4640x3560, 12 bit, 50 μm). All images were processed on a Mammo Workstation (ImageDiagnost) using JPEG2000/wavelet algorithms. Eight lossy compression factors from 20 to 160 were used. Eight experienced radiologists and two medical physics experts had to compare compressed and uncompressed images simultaneously on two monitors (RadiForce G51, 5MP, Eizo) with randomized order and side of the compressed images. Images were displayed in “full screen” mode (FS) and magnifed by 1:1 pixel replication. For 540 image pairs observers had to identify which monitor displays the original image, to rank image quality on a 5-point scale and to identify images with loss of diagnostic quality. Results: The evaluation of FS reading showed no signiﬁcant degradation in image quality up to 1:80 compression rate. 1:1 reading showed an earlier inﬂuence of compression at 1:60. A siniﬁcant loss of diagnostic quality was not observed for any compression. Conclusion: JPEG2000/wavelet compression of digital mammograms does not reduce the diagnostic quality of monitor reading up to 1:80 compression rates. Storage space in PACS can be reduced and network transmission is no longer a bottleneck.

10:30 - 12:00

Room G/H

Head and Neck

SS 208 Lesion detection and characterisation

B-049

11:42

A new generation algorithm for digital mammography designed to reproduce the performance of expert radiologists in detecting actionable clusters P. Bamberger1, I. Leichter1, N. Merlet1, G. Fung2, R. Lederman1; 1Jerusalem/IL, 2 Malvern, PA/US ([email protected]) Purpose: To assess the performance of a new generation mammography algorithm, designed to detect clusters, deemed actionable by expert radiologists. Methods and Materials: 53 cases with clusters were culled from 212 biopsy proven cases collected consecutively from three digital mammography screening facilities. 45 of the 53 cases (25 malignant) were considered actionable by at least two of three independent expert radiologists who interpreted the cases retrospectively. These 45 cases together with 208 normal cases collected consecutively from the above facilities were run on two versions of a prototype detection algorithm (Siemens), designed to detect all clusters deemed actionable by expert radiologists. In the advanced algorithm, multi-step ﬁltration is replaced by global classiﬁcation of candidate micro-calciﬁcations. Moreover, the advanced algorithm considers interdependence between various stages of the parametric clusterization process and implements automatic performance optimization. The performance of the algorithms was compared. Results: The advanced algorithm improved the sensitivity from 80 to 98%. For malignant lesions, the sensitivity improved from 92 to 100% while for benign lesions deemed actionable by experts it increased from 67 to 94%. The ﬁrst algorithm yielded a sensitivity of 80% in dense breasts and 81% in non-dense breasts while the sensitivity of the advanced algorithm was 100 and 96%, respectively. The calciﬁcation false mark rate per view was reduced from 0.27 to 0.14. Conclusion: The new generation algorithm achieved the goal of reproducing the performance of expert radiologists with 98% sensitivity and very few false marks. The algorithm performed equally well in dense and non-dense breasts.

152

A

B

C

D

E

F

G

Moderators: A. Borges; Lisbon/PT V. Vandecaveye; Leuven/BE

B-051

10:30

Dynamic contrast-enhanced CT of head and neck tumors: Comparison of ﬁrst-pass and permeability perfusion measurements using two different commercially available tracer kinetics models S. Bisdas1, M. Baghi1, J. Wagenblast1, T. Koh2, T.J. Vogl1, M. Mack1; 1 Frankfurt a. Main/DE, 2Singapore/SG ([email protected]) Purpose: To evaluate the interchangeability of perfusion parameters derived from two software packages for post-processing of dynamic contrast-enhanced (DCE) CT images of head and neck tumors. Methods and Materials: DCE-CT images of 75 patients with head and neck malignancies were post-processed using a software package based on the maximum-slope approach and Patlak analysis, and a software package with deconvolution-based analysis which incorporates an adiabatic approximation of the tissue homogeneity (ATH) model. The evaluated perfusion parameters included blood ﬂow (F), blood volume (v), and permeability surface area product (PS). Region-of-interest (ROI) analysis of the tumors and the associated lymph nodes was performed. The perfusion parameters were correlated with Spearman’s aparametric criterion and were compared using Wilcoxon matched pairs test and Bland-Altman plots. Results: 152 ROIs of tumors and nodes were outlined and analyzed. Moderate to good correlations were demonstrated between the various perfusion values (r:0.560.72, p 0.0001). The Wilcoxon test revealed signiﬁcant difference between the two methods (p 0.001) with F, v, and PS values obtained by the maximum-slope approach and Patlak analysis being higher than those obtained by deconvolutionbased analysis with assumptions of the ATH model. The Bland-Altman plots for F and v values revealed a proportionality trend with outliers, which were strongly associated with the magnitude of the parameters. Analysis of the PS values did not show any systematic bias. Conclusion: F, v, and PS values showed signiﬁcant disagreement between the two software packages and thus, these parameters derived from the software packages are not directly interchangeable.

Scientific Sessions B-052

10:39

Diffusion-weighted echo-planar MRI: A valuable tool for differentiating primary parotid gland tumors? C.R. Habermann, P. Gossrau, C. Arndt, J. Graessner, F. Reitmeier, M. Jaehne, G. Adam; Hamburg/DE ([email protected])

B-053

10:48

Tumour perfusion measured by dynamic multi slice detector computer tomography in head and neck cancer during and after combined radio- and chemotherapy F.M. Hinkmann1, M. Lell1, S. Lettmaier1, E. Klotz2, M. Geiger1, R. Sauer1, W.A. Bautz1; 1Erlangen/DE, 2Forchheim/DE ([email protected]) Purpose: To quantify changes in tumour perfusion with MDCT of primary head and neck squamous cell carcinomas and lymph node metastases during and after combined radiochemotherapy (RCT) as a prognostic factor of treatment response. Methods and Materials: 64-MDCT perfusion scans were performed in patients with advanced squamous cell carcinomas of the pharynx. All patients were examined before, during (after application of 50 Gy) and after RCT (three and six months). The scan protocol included a non-enhanced scan of the tumour region to deﬁne the area for the perfusion measurements. 2.6 cm were covered by the perfusion scan. A biphasic contrast injection protocol was employed; image acquisition started 5 sec after the initiation of the contrast injection. Blood volume, perfusion and permeability values were calculated with dedicated perfusion software. Results: Hyperperfused areas were detectable on the pre-treatment scans in each tumour. After the ﬁrst part of radiochemotherapy we measured a signiﬁcant (p 0.05) decrease of perfusion, blood volume and permeability in primary tumours and lymph node metastases. This perfusion pattern was stable in the control examinations after three and six months. Conclusion: Quantiﬁcation of tumour perfusion by MDCT proved to be a useful tool to evaluate tumour response to RCT.

B-054

10:57

Merkel cell carcinoma: Imaging ﬁndings of PET, sonography, CT, and MRI with clinicopathologic correlation in 16 patients P. Peloschek1, C. Novotny1, C. Mueller-Mang1, M. Weber1, J. Sailer1, C. Czerny1, R. Dudczak1, K. Kletter1, A. Becherer2; 1Vienna/AT, 2Feldkirch/AT ([email protected]) Purpose: To review the morphological and functional imaging ﬁndings and to assess their potential impact on clinical management in 16 patients with biopsyproved MCC. To our knowledge, no equivalent series of MCC has been described in imaging literature. Methods and Materials: In this IRB-approved retrospective study 16 patients with biopsy-proven Merkel cell carcinoma were included. Patients ranged in age from 62 to 88 years, with a mean of 75 years. Twenty-nine whole body PET scans in 16 patients were retrospectively reviewed in consensus by radiologists and nuclear medicine specialists with regard to tracer uptake in six anatomical sites per patient. A combined standard of reference was applied, which consisted of histopathology

B-055

11:06

Ultrafast MR-sialography: Comparison of two coil systems including a high resolution surface coil T.F. Weber1, M.C. Cramer2, D. Aldefeld2, F. Weiss2, G. Adam2, C.R. Habermann2; 1 Heidelberg/DE, 2Hamburg/DE ([email protected]) Purpose: To compare the visualisation of the salivary ducts by ultrafast magnetic resonance sialography (MRS) using a high resolution surface coil and a conventional head-neck coil. Methods and Materials: In 11 healthy volunteers ultrafast MRS (acquisition time: 2.8 ms) was performed before and after oral application of a sialogogue. Each subject received examinations with both a high resolution carotid coil (A, PACCSS15, Machnet) and a conventional head-neck coil (B, Neck Array Coil, Siemens Medical Solutions). The depiction of parotideal and submandibular duct systems was graded from 1 (very good) to 5 (poor) by three independent radiologists. ANOVA served for statistical analysis of duct rankings, and interobserver variability was determined by intraclass correlation. Results: With a high interobserver correlation of 0.96, both coil systems offered a symmetric visualisation of the salivary ducts, and image quality increased after oral application of a sialogogue (p 0.001). Overall rating was worse for A than for B (2.13p1.24 vs 1.45p0.65, p 0.001). While B depicted extra- and intraglandular components in an equivalent quality, A resulted in decreased visualisation of extraglandular portions (p 0.001). Conclusion: Irrespective of the coil system, ultrafast MRS offers a detailed delineation of salivary duct components. However, most notably due to reduced visualisation of the extraglandular ducts, the high resolution coil used in this study is inferior in image quality. MRS is at present preferentially performed using a conventional head-neck coil system.

B-056

11:15

MRI of parotid lesions: Typical lesion characteristics in MR improve the discrimination between benign and malignant disease C. Waldherr, A. Christe, P. Zbaeren, H.C. Thoeny; Berne/CH ([email protected]) Purpose: To determine whether malignant and benign parotid lesions can be discriminated by MR imaging. Methods and Materials: Fifty-nine consecutive patients (m: 31, f: 28; median age: 57 y; range: 15-86 y) with parotid gland tumours who underwent MRI before undergoing surgery were retrospectively analyzed. Axial T1-, T2-weighted and TIRM sequences and axial, coronal and sagittal contrast-enhanced fat-suppressed T1weighted images were acquired with a 1.5-T MR unit. Histology was available for all tumours. Following MR parameters were analyzed: signal intensity (in the respective sequence), margins (well vs ill-deﬁned margin), localization (deep/superﬁcial lobe), extension into neighbouring structures, growth patterns (circumscribed, diffuse), and contrast enhancement (marked/poor; homogeneity). Results: Histology revealed 37 benign lesions (adenoma: n=22, warthin tumor: n=11, epidermoid cyst: n=1, oncocytoma: n=2, papilloma: n=1), and 22 malignant (adenoid cystic carcinoma (ca): n=4, lymphoma: n=1, spinalioma: n=1, mucoepidermoid ca: n=3, salivary duct ca: n=1, squamous-cell-ca-metastases: n=6, malignant myoepithelioma: n=1, acinic cell ca: n=2, malignant lymphoepithelial tumor: n=1, sarcoma: n=1, undifferentiated ca: n=1). Substantial discriminative factors for malignancy were: T2 intermediate signal intensity (p=0.008), blurred margins pre-contrast (p=0.001), and post-contrast (p=0.06), deep lobe localization (p=0.06), extension into both lobes (p=0.0003), diffuse growth pattern (p=0), and inﬁltration of neighbouring structures (p=0). Contrast enhancement is not a discriminative factor (p=0.7). Conclusion: A combination of different MR parameters (T2 intermediate signal intensity, blurred pre- and post-contrast margins, extension into both lobes, diffuse growth pattern, and inﬁltration into neighbouring structures) is highly suggestive of malignancy.

A

B

C

D

E

F

G

153

Friday

Purpose: The purpose of the study was to determine the value of diffusion-weighted (DW) echo-planar (EP) magnetic resonance (MR) imaging in differentiating various entities of parotid gland tumors. Methods and Materials: 109 consecutive patients with a clinically suspected tumor of the parotid gland were examined with a DW EP sequence using a 1.5 T MR unit. Image analysis was performed by two radiologists independently. Histological diagnosis was obtained in every patient. For comparison of apparent diffusion coefﬁcients (ADC), paired two-tailed Student’s t test with Bonferroni correction was used. An intraclass correlation coefﬁcient was computed to evaluate interobserver variability. Results: In 96 patients a primary parotid gland tumor was conﬁrmed by histology. Among the two observers a high correlation was calculated (0.98). ADC values of pleomorphic adenomas were signiﬁcantly higher than all other entities, except for myoepithelial adenomas (p=0.54). ADC values of Warthin tumors were signiﬁcantly different from myoepithelial adenomas, lipomas, and salivary duct carcinomas (p 0.001, 0.013, and 0.037, respectively), whereas mucoepidermoid carcinomas, acinic cell carcinomas, and basal cell adenocarcinomas were not differentiable from Warthin tumors (p=0.094, 0.396, and 0.604, respectively). Conclusion: DW-EP MR imaging has a potential for differentiation of various entities of parotid gland tumors. Due to an overlap not only within the group of benign and malignant lesions, but also an overlap between these groups, diagnoses should not be addressed based solely on ADC values.

(lymphadenectomy or biopsy) or clinical or radiological follow-up for at least 12 months. Sensitivity and speciﬁcity of PET and morphological imaging methods (sonography, contrast-enhanced computed tomography and magnetic resonance imaging) were calculated and compared with 95% conﬁdence intervals for the difference between binomial proportions based on paired data and a McNemar test. Results: The region-based assessment of diagnostic value, in consideration of the standard of reference, resulted in a sensitivity of 85.7% and a speciﬁcity of 96.2% of FDG-PET (n=127) and in a combined sensitivity of 95.5% and a speciﬁcity of 89.1% for morphological imaging methods (n=68). In one patient FDG-PET led to a true up-staging. Conclusion: FDG-PET is a highly useful whole body staging method of comparable value compared to conventional imaging methods with restricted ﬁeld of view. PETCT will combine the advantages of functional and morphological imaging.

Scientific Sessions B-057

11:24

Functional imaging of salivary glands using diffusion-weighted echoplanar MRI: Evaluation of different stimulation substances C.R. Habermann, T. Ries, J. Graessner, K. Nelting, J. Scheel, G. Adam; Hamburg/DE ([email protected]) Purpose: To evaluate the inﬂuence of different oral stimulation substances of measuring different functional conditions of parotid and submandibular glands using diffusion-weighted (DW) echo-planar imaging (EPI). Methods and Materials: 12 volunteers were examined with a DW EPI sequence using a 1.5 T system once prior to and 14 times after oral stimulation. Volunteers were prospectively randomized into two groups and examined three times. Group A was stimulated for ﬁrst and third examinations with 5 cc of a commercially available lemon juice, and secondly with a 500 mg tablet of ascorbic acid. Group B was stimulated at ﬁrst and third examinations with ascorbic acid, and secondly with lemon juice. Measurements were performed by two independent observers. For statistical evaluation an intraclass correlation, Student’s t test, and an ANOVA with post hoc comparison was used. Results: The observers showed a high correlation (0.974). The mean values for both groups prior to stimulation showed no signiﬁcant differences (p=0.527-0.726). In all volunteers, ﬁrst measurement after stimulation showed signiﬁcant increase of ADC (p 0.0001). The ADC after stimulation with ascorbic acid proved to be signiﬁcantly higher compared to lemon juice for both glands (A=0.001-0.005). Over 14 measurements, ADC decreased stepwise to the initial ADC value for both substances. Both groups showed no statistical signiﬁcance between ﬁrst and third examinations (p=0.53-0.99). Conclusion: DW EPI imaging displays functional changes in salivary glands with high reproducibility. The stimulation substance has an inﬂuence on the extent of ADC changes, but not on the course.

B-058

11:33

MR-sialography at 1.5 and 3 T: Quantiﬁcation of salivary duct visualisation using an ultrafast imaging sequence T.F. Weber1, M.C. Cramer2, D. Aldefeld2, F. Weiss2, G. Adam2, C.R. Habermann2; 1 Heidelberg/DE, 2Hamburg/DE ([email protected]) Purpose: To evaluate the visualisation of salivary duct components by ultrafast magnetic resonance sialography (MRS) at 3 T compared to 1.5 T. Methods and Materials: In 27 healthy volunteers ultrafast MRS was performed before and after oral application of a sialogogue. On consecutive days each subject was examined at both 1.5 T (head-neck-coil, Magnetom Symphony, Siemens Medical Solutions) and 3 T (head-coil, Intera, Philips Medical Systems). The depiction of parotideal and submandibular duct systems was graded from 1 (very good) to 5 (poor) by four independent radiologists. ANOVA served for statistical analysis of duct rankings, and interobserver variability was determined by intraclass correlation. Results: With a high interobserver correlation of 0.94, overall duct visualisation did not differ between 1.5 T and 3 T (2.26 vs 2.22). However, the parotideal ducts were superiorly depicted at 3 T (p 0.001), and the submandibular ducts were inferiorly depicted at 3 T (p 0.001). Image quality increased after oral application of a sialogogue (p 0.001). Conclusion: Ultrafast MRS at 3 T produces discordant image qualities of the parotideal and submandibular duct components. As this is most notably due to a suboptimal conﬁguration of the available coil system, it is to be expected that 3 T signiﬁcantly improves overall visualisation of the salivary ducts if adequate coil coverage of the submandibular glands could be achieved.

B-059

Purpose: To determine the most accurate parameter for prediction of malignancy in parotid tumors using diffusion weighted MR & dynamic contrast enhanced T2*weighted perfusion MR imaging. Methods and Materials: Diffusion & dynamic contrast enhanced T2*-weighted ﬁrst pass perfusion weighted MR imaging was conducted on 45 patients (26M, 19 F; mean age is 55ys) with parotid tumors using a single shot EPI. Diffusion MR images were acquired with a b-factor of 0.500 & 1000 sec/mm2. Apparent diffusion coefﬁcient (ADC) value of the tumor was calculated. Dynamic contrast enhanced MR was performed after bolus injection of Gadolinium-DTPA (0.3 ml mol/kg BW). The maximum signal intensity loss of the tumor was calculated. Multivariate logistic regression analysis was done.

A

B

B-060

C

D

E

F

G

11:51

First-pass dynamic contrast-enhanced MR imaging of the squamous cell carcinoma in the upper aerodigestive tract: Initial results and comparison with ﬁrst-pass dynamic contrast-enhanced CT imaging S. Bisdas, L. Medov, M. Baghi, T. Koh, T.J. Vogl, M. Mack; Frankfurt a. Main/DE ([email protected]) Purpose: To examine the feasibility of dynamic-contrast-enhanced (DCE) MR imaging of the squamous cell carcinoma (SCC) in the upper aerodigestive tract and to compare the DCE-MR parameters with the corresponding DCE-CT parameters in the same patient population. Methods and Materials: 17 patients with SCC underwent DCE-MR and CT studies. An echo-planar-imaging (EPI) MR-sequence was applied for ﬁrst-pass DCE-T2*-weighted imaging. A deconvolution-based analysis was applied to derive tumor blood ﬂow (BF), blood volume (BV), and mean transit time (MTT) values from DCE-MR and CT data. Results: The BF and BV values from ROIs through the normal tissue were signiﬁcantly lower in both modalities when compared with the mean values of the SCC. The BF, BV, and MTT values of the primary SCC in DCE-MR were 84.68 ml/ min/100 g, 9.99 ml/100 g, and 6.82 s, respectively. The BF, BV, and MTT values of the primary SCC in DCE-CT were 75.30 ml/min/100 g, 5.74 ml/100 g, and 8.49 s, respectively. The BF and BV values were signiﬁcantly correlated between the two modalities, r=0.78 (p 0.0001) and r=0.49 (p=0.04), respectively. Bland-Altman graphical analysis showed that the mean difference of the BF values was -8.514 (CI -24.269 to 7.240) and the BV value was -3.778 (CI -6.823 to -0.733), respectively. Logistic regression analysis showed: BFMR=0.89*BFCT+17.17 (p=0.000, R2=0.60) and BVMR=1.15*BVCT+2.83 (p=0.04, R2=0.22). Conclusion: T2*-weighted DCE-MR of the SCC is feasible and the BF and BV values correlate signiﬁcantly with the corresponding values from the DCE-CT imaging. The BF values may be also used interchangeably between the two modalities.

10:30 - 12:00

Room I

Vascular

SS 215 Carotid imaging Moderators: R.A. Bucek; Vienna/AT M.W. de Haan; Maastricht/NL

B-061

11:42

Prediction of malignancy in parotid tumor using diffusion weighted MR and dynamic contrast enhanced T2*-weighted perfusion MR imaging A.A.A. Abdel Razek, A. Kandeel, A. Motamed, N. Nada; Mansoura/EG ([email protected])

154

Results: Signiﬁcant differences in the ADC value & maximum signal intensity loss were noted between benign & malignant parotid tumors (P 0.001 & 0.003, respectively).The threshold value of 1.30X10-3 mm2/sec for ADC value to predict malignancy provided accuracy of 86%, sensitivity of 84%, speciﬁcity of 90%, PPV of 94% and NPV of 76%. Selection of 24% as a threshold parameter for maximum signal intensity loss has 88% accuracy, 85% sensitivity, 88% speciﬁcity, 92% PPV & 72% NPV. The combination of ADC value & maximum signal intensity loss resulted in accuracy, sensitivity, speciﬁcity, PPV & NPV of 90%, 93%, 91%, 89% & 83%, respectively. Conclusion: We concluded that combination of apparent diffusion coefﬁcient values and maximum signal intensity loss is the most accurate parameter for prediction of malignancy in parotid tumors.

10:30

Application of 64 multidetector-row CT angiography in the evaluation of carotid stenosis Y.G. Gao, K.L. Li, X.D. Du, Y.S. Shen; Beijing/CN ([email protected]) Purpose: To evaluate the diagnostic value of 64 MDCT angiography for ICA stenosis. Methods and Materials: All TIA patients with interpretable CTA and DSA of the cervical carotid arteries were selected in 2007. This yielded a total of 80 vessels. The CTA curved planar reformations and DSA images referenced to the distal ICA were graded by two senior neuradiologists blindly, according to the North American Symptomatic Carotid Endarterectomy Trial guidelines. The DSA was the gold standard, the 50 and 70% stenoses were used as the cut-off values, the sensitivity, speciﬁcity, the negative and positive predictive values of CTA were calculated. In order to avoid the underestimation of the stenosis grade in using the NASCET criteria, the maximal diameter of the ICA distal to the carotid bulb could be converted to the mean distal-ICA diameter of symptomatic arteries for calculating the stenosis percentage. The paired-t test was used to verify the statistically

Scientific Sessions

B-062

10:39

Comparison between three different criteria for quantiﬁcation of carotid stenosis by means of 64-slice multidetector CT: The role of the absolute area with respect to the percentage of vessel stenoses G. Pontone, D. Andreini, A. Annoni, A. Formenti, E. Bertella, G. Ballerini, S. Galli, P. Ravagnani, M. Pepi, P. Montorsi; Milan/IT ([email protected]) Purpose: The AHA/ACC guidelines indicate three methods of quantiﬁcation of carotid stenosis: the CC, the ECST and the NASCET methods. The MDCT has made it possible to obtain a direct measurement of the area of the carotid artery at the level of the stenosis (A). The aim of this study was to ascertain which of the various quantiﬁcation analysis criteria best correlates with A. Methods and Materials: We studied 112 patients (pts) with known carotid bulb disease. Each patient underwent MDCT (64-slice GE VCT). We measured the diameters and the areas of the common carotid artery, the carotid bulb, the internal carotid artery at the level of the carotid plaque and the internal carotid artery downstream of the bulb. We calculated the percentage of stenosis by means of the CC method based on diameter (CC-d) and on area (CC-A), the ECST method on diameter (ECST-d) and area (ECST-A), the NASCET method on diameter (NASCET-d) and area (NASCET-A) and correlations with A using Pearson’s correlation coefﬁcient. Results: The quantitative analysis showed an A=14.4p11.6 mm2. The percentages of stenosis assessed by means of the three criteria were: CC-d: 58p17; CC-A: 71p19; ECST-d: 55p16; ECST-A: 62p23; NASCET-d: 39p18; NASCET-A: 66p25. The best correlation (r) was between A and CC-A (r: 0.87), and the worst between A and NASCET-d (r: 0.56). Conclusion: Our results show that the quantiﬁcation criterion that best correlates with both absolute areas of the vessel at the level of the stenosis is the CC-A method.

B-063

10:48

ECG-gated dual source CT angiography (ECG-DSCTA) provides 4D-data to evaluate the distensibility of carotid arteries S. Rozie, P.J. Homburg, M.L. Dijkshoorn, D.W.J. Dippel, A. van der Lugt; Rotterdam/NL ([email protected]) Purpose: Atherosclerosis may lead to arterial stiffness and measurement of arterial distensibility may be important for risk prediction. We hypothesize that arterial distensibility can be assessed with 4D-reconstructed images of an ECG-DSCTA scan. Methods and Materials: Fourteen patients (eight male, mean age 64p10 years, range 51-77) with cerebrovascular symptoms underwent multidetector CTA of the carotid arteries. Scanning was performed on a 64-slice MDSCT scanner (Siemens, SOMATOM deﬁnition, Forchheim, Germany). A 4D dataset at every 8% of the cardiac RR interval was reconstructed (13 images of each plane) providing motion visibility and minimal and maximal lumen areas of the carotid arteries. Images were analyzed on a dedicated workstation using 3D-analysis software (Siemens, CT workplace/MMWP, InSpaceAVA). On axial MPR images, perpendicular to the central lumen line, we measured lumen area at every 8% of the cardiac cycle in the common carotid artery (CCA) and the internal carotid artery (ICA). Using the end-diastolic diameter (D), the absolute change in diameter ($D) during systole and the pulse pressure ($P) we calculated the distensibility coefﬁcient (DC) as 2 ($D/D)/$P (10-3 kPa). Results: In all 30 carotid arteries distensibility was detected. The DC was 5.6p2.6 (10-3 kPa) (range 3.1-11.2). Lumen area in the CCA in systole was 28p4% (range 23-33%) larger than in diastole. Lumen area in the ICA in systole was 29p22% (range 14-63%) larger than in diastole. Conclusion: Arterial distensibility of the carotid artery can be evaluated with 4Dreconstructed data of an ECG-gated DSCTA scan. A signiﬁcant increase in lumen area occurs during the systole.

B-064

10:57

Utility of direct dual-energy hard plaque removal image using dual-source CT in quantiﬁcation of calciﬁed carotid artery stenosis: Correlation with digital subtraction angiography K. Uotani1, Y. Watanabe1, T. Nakazawa1, M. Higashi1, Y. Hori1, T. Fukuda1, S. Kanzaki1, N. Yamada1, K. Otani2, T. Itoh2, K. Iihara1, H. Naito1; 1Suita/JP, 2 Tokyo/JP ([email protected]) Purpose: Dual-source CT (DSCT) allows separation of bone and calciﬁcations from iodine using dual-energy acquisition, and makes it possible to quantify carotid stenosis with dense calciﬁcations. The purpose of this study was to compare direct dual-energy hard plaque removal (DE-Hard Plaque Removal) with conventional digital subtraction angiography (DSA) regarding quantiﬁcation of calciﬁed carotid artery stenosis. Methods and Materials: Ten vessels of eight patients with densely calciﬁed carotid stenosis were enrolled. CT angiography was performed using a DSCT scanner. Dual-energy data were acquired with one tube being operated at 140 kV while the other was operated at 80 kV. Subtraction images were obtained from two datasets acquired at different kV using commercial software. Carotid artery stenosis was quantiﬁed according to NASCET criteria on MIP images and on DSA images in the same plane. Correlation between CTA and DSA was determined by means of cross tabulation, and accuracy for detection and grading of stenosis were calculated. Results: Evaluation of stenosis was possible for all vessels with DE-Hard Plaque Removal, although conventional CTA did not allow evaluating stenosis in eight out of 10 vessels. Good correlation for quantiﬁed carotid stenosis between DE-Hard Plaque Removal images and DSA images were observed (r2=0.9738). Sensitivity and speciﬁcity to detect hemodynamically relevant (50-99%) stenosis were 88 and 100%, respectively. Severe stenosis (over 85% at DSA) tended to be overestimated and displayed as short occlusion in DE-Hard Plaque Removal images. Conclusion: DE-Hard Plaque Removal is useful for the evaluation of carotid stenosis with severe calciﬁcation.

B-065

11:06

Symptomatic carotid plaques have the same plaque volume and composition as the contra lateral asymptomatic carotid plaque: An analysis with MDCTA S. Rozie, T.T. de Weert, P.J. Homburg, H.L.J. Tanghe, D.W.J. Dippel, A. van der Lugt; Rotterdam/NL ([email protected]) Purpose: Cerebrovascular events are related to atherosclerotic disease in the carotid artery. Atherosclerotic plaque rupture leading to thromboembolisation occurs in a speciﬁc type of atherosclerotic plaque (vulnerable plaque). We therefore hypothesize that symptomatic and asymptomatic plaques have a different plaque volume (PV) and plaque composition. Methods and Materials: We studied 130 consecutive patients (74 male; mean age 61p15.3 years, range 19-88 years) with TIA or minor ischemic stroke who underwent MDCT angiography. Scanning was performed on a 16-slice MDCT scanner (Siemens, Sensation 16, Erlangen, Germany). We measured plaque volume (PV) manually by drawing the outer contour of the carotid artery with a custom-made software tool. Luminal boundary was assessed automatically based on a HounsﬁeldUnit (HU) threshold. Plaque composition was measured with ranges of HU-values (calciﬁcation 130 HU, ﬁbrous tissue 60-130 HU, lipid core 60 HU). Results: Eighty patients had atherosclerotic plaque in the carotid artery. Sixty-one patients had bilateral plaques. The PV of the symptomatic artery was 836p738 mm3 and the PV of the asymptomatic artery was 614p530 mm3. In patients with bilateral plaques, the PV of the symptomatic plaque was not signiﬁcantly larger than the PV of the asymptomatic plaque (p-value 0.10). No signiﬁcant difference was found in the proportion of lipid, ﬁbrous tissue and calciﬁcations. Conclusion: Atherosclerosis is a bilaterally symmetric disease. Risk prediction cannot be based on plaque volume and plaque composition.

B-066

11:15

Carotid plaque ulceration assessed with multidetector computed tomography angiography T.T. de Weert, S. Cretier, S. Rozie, H. Groen, J. Wentzel, D.W.J. Dippel, A. van der Lugt; Rotterdam/NL ([email protected]) Purpose: Atherosclerotic plaque rupture in the carotid artery is considered an underlying mechanism of ischemic stroke. We analysed the frequency and location of plaque rupture in patients with cerebrovascular symptoms, and analysed the relation with severity of stenosis and cardiovascular risk-factors. Methods and Materials: MDCTA images from 404 symptomatic patients were evaluated. Atherosclerotic plaque surface was classiﬁed as smooth, irregular or

A

B

C

D

E

F

G

155

Friday

signiﬁcant difference between these two methods. Results: When the 70% stenosis was used as the cut-off value, the sensitivity, speciﬁcity, negative predictive value and the positive predictive value of CTA were 97, 95, 95 and 98%, respectively. There was no signiﬁcant difference in calculating the stenosis percentage from two methods (p=0.14). Conclusion: From the comparative study between CTA and DSA examination of head and neck, it was proved that CTA had a high sensitivity, high speciﬁcity and high negative and positive predictive values for carotid disease. CTA appears to be an excellent screening test for internal carotid stenosis.

Scientific Sessions ulcerated. The ulcer location was deﬁned as upstream or downstream from the point of maximum stenosis. The severity of stenosis was measured, and patient characteristics on cardiovascular risk-factors (smoking, hypertension, diabetes, hypercholesterolemia, previous cardiac disease and previous cerebrovascular disease) were assessed. The association between ulceration and stenosis degree and between ulceration and risk-factors was determined with logistic regression. Results: Carotid arteries with atherosclerotic plaque were classiﬁed as: smooth: 276 (62%), irregular: 99 (22%) and ulcerated: 73 (16%). In the 73 ulcerated plaques 90 ulcerations were present. Sixty two (69%) of the ulcerations were located upstream. Ulcerated plaques were more common in carotid arteries with stenosis 30% (p 0.001). The odds ratio for the association between plaque ulceration and a 10% increase in severity of stenosis is 1.5. There is a trend towards an association between smoking and plaque ulceration and/or irregular surface (OR 1.9, p=0.08). Symptomatic carotid arteries had more ulcerated plaques than asymptomatic carotid arteries (11 vs7%; p=0.04). Conclusion: MDCTA is an excellent technique to visualize plaque surface morphology. In the carotid bifurcation, plaque rupture is very frequent, commonly located upstream and more often present in the more severe stenotic symptomatic artery.

B-067

11:24

Increased carotid plaque echolucency predicts adverse cardiovascular events in high risk patients M. Reiter, I. Effenberger, R. Bucek, S. Puchner, M. Schillinger; Vienna/AT ([email protected]) Purpose: Ultrasonic measurement of carotid plaque echolucency has been suggested as potential marker of plaque instability. We hypothesized that low carotid gray scale median (GSM) levels determined by repeat ultrasound examinations identify patients at high risk for future clinical events of atherosclerosis. Methods and Materials: We studied 558 consecutive patients of a prospective registry with carotid stenosis 30%, who were initially asymptomatic with respect to carotid disease. Carotid ultrasound investigations at baseline and after median 7.5 months (range 6-9) were performed to determine the GSM level. Patients were then followed clinically for median 3.2 years for the occurrence of major adverse cardiovascular events (composite MACE: myocardial infarction, percutaneous coronary or peripheral interventions, coronary or vascular surgery, amputation, stroke, and all-cause mortality). Results: GSM levels at baseline and control visit were not related to an increased risk of MACE during the follow-up period (p 0.05). However, patients with a decrease of GSM levels between baseline and control visit had a signiﬁcantly increased risk for cardiovascular events compared to the patients with stable or increased GSM levels (p 0.001). Conclusion: Increased carotid plaque echolucency detected by decreased GSM levels within a 6-9 month interval seems to identify unstable plaques and predicts midterm clinical adverse events of atherosclerosis in high risk patients affecting the coronary, cerebrovascular and peripheral circulation.

B-068

11:33

High resolution carotid black blood 3 T MRI with parallel imaging: Initial experience T. Saam1, J.G. Raya1, K. Bochmann1, G. Meimarakis1, S. Sourbron1, U. Frey1, A. Werf2, H. Gerhards3, M.F. Reiser1, K. Nikolaou1; 1Munich/DE, 2TD Elde/AN, 3 Kerpen/DE ([email protected]) Purpose: To evaluate a multi-sequence carotid black blood protocol using parallel imaging and a dedicated 4-channel surface coil (Machnet, The Netherlands) in combination with neck and spine coils. Methods and Materials: Ten healthy volunteers and 10 patients with 50% stenosis by duplex ultrasound had their bilateral carotid arteries imaged at 3.0 T (Siemens Trio, Germany) using a multi-sequence protocol (TOF, pre-contrast T1, PD and T2 in the volunteers and post-contrast T1 and dynamic contrast-enhanced sequences [DCES] in patients; best-in-plane resolution 0.5x0.5 mm2). Imaging time for TOF, T1, PD, T2 and DCES images were 4:11, 4:38, 2:08, 2:08 and 5:00 minutes, respectively. To assess intra-scan reproducibility, the volunteers were scanned twice within two weeks. Results: All MRI exams were of diagnostic image quality. Intra-scan reproducibility for quantitative measurements of lumen, wall and outer wall areas was excellent with intra-class correlation coefﬁcients 0.9 and measurement errors 10%. One 57-year old healthy volunteer with no stenosis in TOF images had a large necrotic core. In the patients with 50% stenosis, characteristics of advanced carotid atherosclerosis were found, such as American Heart Assocation Lesion Type VI (hemorrhage, ﬁbrous cap ruptures, calciﬁed nodules and thrombus). Plaque compositional analysis was feasible and necrotic core, loose ﬁbrous matrix, calciﬁcation

156

A

B

C

D

E

F

G

and hemorrhage areas could be identiﬁed. Conclusion: First experience indicates that high resolution carotid black blood 3 T MRI with parallel imaging is a fast, reproducible and robust method to assess carotid atherosclerotic plaque in vivo and this method is ready to be used in clinical practice.

B-069

11:42

Enhancement of the atherosclerotic plaque and the vessel wall of the carotid artery after injection of a blood pool agent S. Rozie, P.A. Wielopolski, M. Ouhlous, W.J. Niessen, T. van Walsum, A. van der Lugt; Rotterdam/NL ([email protected]) Purpose: We hypothesized that a blood-pool agent enhances the atherosclerotic carotid plaque in patients with severe and symptomatic stenosis of the carotid artery. Methods and Materials: We studied 26 atherosclerotic plaques of 15 consecutive patients (14 male; mean age 65.3p10.1 years, range 47-80 years) with severe carotid stenosis ( 50%) and cerebrovascular symptoms. These patients underwent multiple sequences on a 3.0 T MRI (Signa Excite, GE Healthcare, USA) with a 4-channel bilateral coil to identify the various plaque components. We injected 0.03 mmol/kg of Vasovist®, (Gadofosveset, Schering, Germany) and an axial 3DT1-weighted scan was performed before injection and 20 minutes. We used a custom-made 3D pointbased registration tool to match the images before and after contrast injection in order to correct for patients movements. In the axial images of the 3DT1-weighted scans we measured the signal intensity (SI) of the sternocleidomastoideus muscle (SCM), the non-calciﬁed part of the plaque and the vessel wall (outer border) of the carotid artery. The percentage contrast enhancement was calculated as: [(SI (enhanced)-SI (unenhanced))/SI (unenhanced)]x100. Results: The mean enhancement on the 20 minute delay scan was 24.6p17.5% in the SCM, 40.9p32.5% (range 1-120%) in the non-calciﬁed part of the plaque and 60.0p21.7% in the vessel wall of the symptomatic carotid arteries. Conclusion: The atherosclerotic plaque and the vessel wall of the carotid artery enhance after injection of Vasovist®. The enhancement of the non-calciﬁed part of the plaque has a large range which may reﬂect differences in plaque composition.

B-070

11:51

Carotid plaque imaging: High-resolution MR imaging with 3-inch surface coil M. Li, Y. Zhu, F. Feng, H. You, L. Hu, Z. Jin; Beijing/CN ([email protected]) Purpose: To study the value of 3-inch surface coil in MR imaging of carotid plaques. Methods and Materials: 33 patients with atherosclerosis underwent MR imaging with a 1.5-T system (Signa Excite HD) and two 3-inch coil (GE Medical System Pvt. Ltd, India). A standardized protocol was performed to obtain four different contrast-weighted images (3D TOF, T1, and PD/T2-weighted) of the carotid arteries. An image-quality rating (5-point scale: 1=poor, 5=excellent) for each sequence was evaluated by two radiologists. Also four components of carotid plaque were analyzed based on four sequences. Results: Totally 264 sequences (bilateral carotid arteries) were reviewed. Images with image quality b2 were found in only 27 sequences (10.2%), and images with quality r4 were found in 147 sequences (55.7%). Images with quality r3 were included for plaque analysis, including 50 carotid arteries. Among them 32 plaques were found in 20 patients. Components analysis demonstrated 16 plaques with unambiguous lipid-rich/necrotic core (LR/NC), 10 plaques with hemorrhage, 15 plaques with calciﬁcation and ﬁve plaques with ruptured ﬁbrous cap. The inter-reader agreement (Kappa, 95%CI) was excellent for evaluation of LR/NC (0.81; 0.61-1.01), hemorrhage (0.86; 0.66-1.04), calciﬁcation (0.87; 0.70-1.00), and moderate for evaluation of ruptured ﬁbrous cap (0.45; 0.05-0.85). Conclusion: High-resolution MR imaging with 3-inch surface coil is capable of producing satisfactory image quality in most of the carotid arteries. The evaluation of imaging characteristics of different plaque components is accurate with an excellent to moderate inter-reader agreement.

Scientific Sessions 10:30 - 12:00

Room K

Computer Applications

SS 205 Moderators: D. Loeckx; Leuven/BE R. Passariello; Rome/IT

B-071

10:30

Automated segmentation and volumetric, RECIST and WHO measurements of abdominal and pelvic lymph nodes in multi-slice CT V. Yaghmai1, M. Rezvani1, H. Soud2, R. Salem1, S. Berggruen1; 1Chicago, IL/US, 2 Malvern, PA/US ([email protected]) Purpose: To evaluate the feasibility of automated volumetric, RECIST and WHO measurements of lymph nodes on multi-slice CT. Methods and Materials: Seventy-ﬁve lymph nodes on contrast-enhanced multislice CT scans of 23 patients were evaluated (para-aortic N=28; mesenteric N=26; iliac chain N=11; inguinal N=10). Prototype software program (Siemens Medical Solutions, Forchheim, GER) was utilized to measure lymph nodes automatically using RECIST and WHO criteria. Manual measurements were also performed independent of the automatic measurements by two board-certiﬁed radiologists. Correlation coefﬁcients (r) for automated and manual measurements were calculated. The software automatically calculated the volume of each lymph node. The quality of automated segmentation for each lymph node was evaluated on a ﬁve-point scale (1:poor-5:excellent). Results: The software successfully segmented all the lymph nodes. Segmentation scores were as follows: 5=43 lymph nodes; 4=23 lymph nodes; 3=8 lymph nodes; 2=2 lymph nodes. The average quality of segmentation was 4.4 (SD=0.79). The average segmentation score based on the location was: para-aortic=4.2; mesenteric=4.6; iliac chain=4; inguinal=4.8.There was no correlation between segmentation score and the size of lymph nodes (r=0.04). The correlation coefﬁcients between manual and automated measurements using RECIST and WHO criteria were 0.94 (p 0.0001) and 0.95 (p 0.0001), respectively. Conclusion: Our results suggest that automated segmentation and volumetric, RECIST and WHO measurements of the abdominal and pelvic lymph nodes may be performed accurately and may replace manual measurement. Its role in the management and treatment follow-up of oncology patients remains to be studied.

B-072

10:39

CT, PET and software-based-image-fusion: A comparison in patients with malignant lymphoma B. Henninger, E. Gunsilius, D. Kendler, C. Uprimny, I. Virgolini, W.R. Jaschke, R. Bale; Innsbruck/AT ([email protected]) Purpose: To compare the accuracy of 2-deoxy-2-[ﬂuorine-18]ﬂuoro-D-glucose (FDG) positron emission tomography (PET) with computed tomography (CT) in the imaging of non-Hodgkin’s lymphoma (NHL) and Hodgkin’s disease (HD) using software-based-image-fusion and to assess the clinical impact of this method. Methods and Materials: Seventy-seven patients, aged 20-77 years with NHL (n=58) or HD (n=19) underwent an FDG-PET scan, a contrast enhanced CT and a subsequent image-fusion. 109 examinations of each modality were evaluated and compared with each other. Sensitivity and speciﬁcity were determined by deﬁning positive ﬁndings of the fusion as “true positive”. Negative ﬁndings at the fusion, i.e. evaluated negative after discussing all ﬁndings in an interdisciplinary image-fusion conference, were deﬁned as true negative. Finally, the results of the image-fusion were compared with the outcome, i.e. a complete follow-up regarding all diagnostic possibilities for malignant lymphoma. Results: In a patient based analysis, sensitivity and speciﬁcity for detecting malignant lymphoma were 90 and 78% for CT, 94 and 91% for PET, respectively. A lymph node region based statistic, that should show how many pre-deﬁned lymph node regions (14) were truly affected, revealed 82% sensitivity and 98% speciﬁcity for CT, 96 and 99% for FDG-PET, respectively. Only three of 109 image fusion ﬁndings were incorrect. Conclusion: The combination of the datasets by using image fusion improves the accuracy of staging, restaging and therapy monitoring in patients with malignant lymphoma and may reduce the need for other (invasive) diagnostic procedures.

10:48

Software for analysis of tracheal lumen on cine-CT M. Owsijewitsch, D. Simon, A. Mehndiratta, J. Ley-Zaporozhan, F.L. Giesel, H.-U. Kauczor, S. Ley; Heidelberg/DE ([email protected]) Purpose: Cine-CT during different breathing maneuvers is recommended for the assessment of tracheal collapse. The purpose of our project was to develop a minimal-interactive tool for analysis of tracheal area on cine-CT. Methods and Materials: The application was developed using the MeVisLab development environment. The trachea is segmented on time resolved (cine-) CT data sets using a region-growing-algorithm. A seed point is set manually on an arbitrary slice. The center of gravity of the trachea on the previous slice is used as a dynamic seed point for the following slice. The tracheal cross sectional area (CSA) is calculated for each slice. The maximal change of the tracheal CSA during a forced expiratory maneuver was compared with results of area calculation based on visual selection of slices showing maximal and minimal tracheal CSA. The time needed for the evaluation was determined for both techniques. Results: The developed tool ran stable in all 21 patient data sets selected for testing the tool. The mean difference of the decrease of tracheal CSA between the software tool and analysis based on visual slice selection was 4.6p3.1%. The software tool needed on average 96p4.1 seconds (visual method 259p18.7 seconds) for processing each dataset containing 101 slices. Conclusion: An in-house developed minimal-interactive tool allows for analysis of the tracheal area changes on cine-CT with the same precision but within much shorter time than analysis based on visual slice selection.

B-074

10:57

Variability in repeated automatic measurements of left ventricular function on dual source computed tomography datasets using three different software packages G.J. de Jonge, P.M.A. van Ooijen, J.H. Kasemier, P.A. van der Vleuten, G.H. de Bock, M. Oudkerk; Groningen/NL ([email protected]) Purpose: To repeatedly assess left ventricular (LV) function as automatically as possible using three different workstations on the same Dual Source Computed Tomography (DSCT) datasets and to compare the results between the three different workstations and the three different measurements on each workstation. Methods and Materials: Twenty-four patients, undergoing cardiac DSCT (Somatom Deﬁnition, Siemens Medical Solutions, Germany), were included (18 men, mean age 59p12 yrs). Reconstructions were made at every 10% of the RR-interval. LV function analysis was performed on three different workstations: Syngo (Siemens Medical Systems, Germany), Vitrea ®2 (Vital Images, Minnetonka, MN) and AquariusWorkstation (TeraRecon, San Mateo, CA). In all software packages, three repeated measurements were made by the same observer. ANOVA test was used for statistical analysis. Results: Mean ejection fraction (EF) was 38p15% and 59p14% for Siemens (ﬁxed and manually deﬁned end-diastolic (ED) and end-systolic (ES) phases, respectively). Mean EF was 63p11% for Vitrea and 60p10% for TeraRecon. ANOVA tests showed no signiﬁcant difference between the repeated measurements on each workstation (p=0.97, 0.99, 1.00 and 0.11, respectively). ANOVA showed a signiﬁcant difference between the three workstations when Siemens with ﬁxed ED and ES phase setting was included (p 0.01) but no signiﬁcance was shown when Siemens with manually determined ED and ES phases was used (p=0.59). Conclusion: No signiﬁcant difference existed between the three repeated measurements for all workstations; this indicates that each software package is consistent. No signiﬁcant differences were found between the three workstations.

B-075

11:06

Accuracy of individually adapted, interactive oblique multiplanar reconstructions vs. semiautomated coronary segmentation and rotated curved planar reformations for stenosis detection in coronary CT angiography K. Anders, S. Achenbach, A. Kuettner, U. Ropers, M. Wechsel, W.A. Bautz; Erlangen/DE ([email protected]) Purpose: To evaluate the diagnostic accuracy of semiautomated centerline deﬁnition and rotated curved planar reformations (CPR) compared to freely adapted, interactive oblique multiplanar reconstructions (ioMPR) for the detection of stenoses in coronary CT angiography (CTA). Methods and Materials: Four readers (1, 1, 3 and 4 years of experience) blinded to clinical results evaluated a total of 50 CTA datasets each. Data sets were randomly assigned to one of the following evaluation techniques: (1) ioMPR excluding CPR, and (2) axial slices and CPR along an automatically generated centerline provided

A

B

C

D

E

F

G

157

Friday

CAD and automated image analysis

B-073

Scientific Sessions by a commercially available software (syngoCirculation, Siemens, Forchheim, Germany). Results were compared to invasive angiography. Readers’ conﬁdence was determined on a 3-point scale (1=low; 3=high). Results: 796 coronary arteries were evaluated (398 per each technique, no left main in one patient). Using CPR, 66 of 84 stenoses were correctly identiﬁed, whereas ioMPR correctly detected 74/84. Stenosis was correctly ruled out in 282 and 283 of 314 cases (sensitivity 79 vs 88% [n.s.], speciﬁcity 90 vs 90% [n.s.], NPV 94 vs 97% [n.s.]). On a per-patient basis, sensitivity for detection of patients with at least one stenosis was 95% for ioMPR and 85% for CPR (p 0.001). Speciﬁcity was 65% each. Mean readers' conﬁdence was 2.07 for syngoCirculation vs 2.22 for ioMPR reading (p=0.0028). Mean conﬁdence in case of misdiagnosis was 1.67. Conclusion: Dedicated software with automatic vessel extraction and freely rotatable CPR does not improve detection of coronary artery stenoses.

B-076

11:15

Automated assessment of non-calciﬁed atherosclerotic plaque burden at coronary CT angiography K.N. Blackmon, B. Ruzsics, J. Streck, C. Thilo, S.A. Nguyen, P. Costello, U.J. Schoepf; Charleston, SC/US ([email protected]) Purpose: To evaluate a post-processing software tool for automated volumetric quantiﬁcation of non-calciﬁed coronary artery plaque burden at non-invasive coronary CT angiography (cCTA). Methods and Materials: Thirty-seven patients underwent contrast enhanced (Bracco) 64-slice (Siemens) cCTA. Two experienced observers in consensus evaluated each scan and identiﬁed 40 non-calciﬁed coronary artery plaques. Also in consensus, they performed manual, three-dimensional planimetric measurements on each plaque as the reference standard. The same two observers then performed volumetric measurements utilizing a threshold-based software tool (TeraRecon) for automated assessment of non-calciﬁed plaque burden. Two different, less experienced observers also performed both manual and automated measurements of each plaque, ﬁrst independently and then in consensus. Spearman rank correlation was used to determine association between variables. Results: Automated volumetric assessment was successfully performed on each plaque. There was excellent (R=0.920) correlation between the expert manual measurements (average plaque burden=33.58 mm3, 18.16 STD) and measurements using the automated volumetry tool (35.64 mm3, 16.42 STD). For the less experienced observers, there was better correlation with expert manual measurements using the automated tool (R=0.885) than with manual measurements (R=0.854). Interreader correlation for volume measurements by the two less experienced observers without and with use of the automated tool signiﬁcantly (p 0.001) increased from R=0.781 to R=0.919. Conclusion: Compared with expert manual planimetric measurements, automated postprocessing tools enable accurate volumetry of non-calciﬁed plaque burden. Reproducibility of plaque measurement between different observers is improved. Use of such tools should facilitate fast, objective and reproducible assessment of non-calciﬁed plaque burden for risk stratiﬁcation and therapeutic monitoring.

B-077

11:24

Computed tomographic quantiﬁcation of coronary artery stenosis: Accuracy of different dedicated cardiac software packages R. Dikkers1, H.A. Marquering2, H. Adams1, G.J. de Jonge1, P.M.A. van Ooijen1, T.P. Willems1, M. Oudkerk1; 1Groningen/NL, 2Leiden/NL ([email protected]) Purpose: To investigate the non-invasive detection and quantiﬁcation of coronary artery stenosis using different commercially available dedicated cardiac software packages (SP). Methods and Materials: 26 CT datasets of vessel phantoms and CT datasets of 15 patients were evaluated by two observers using four different SP (S, V, T, and B). The area and diameter stenosis and the time to perform the measurements were recorded. Results: The variation between the true measure and the measurements of the SP were, for the vessel phantoms, signiﬁcantly lower for the manual measurements compared to the (semi)automatic measurements (p 0.001), except for V. For all SP, the duration to perform the manual measurements in the vessel phantoms was signiﬁcantly longer than for the (semi)automatic measurements (p 0.001), although the accuracy was signiﬁcantly lower for the (semi)automatic measurements. In patient examination soft plaques results in poor to moderate inter-software agreement for manual measurements (kappa value 0.36-0.64). Calciﬁed and mixed plaques had poor inter-software agreement for manual measurements (kappa value 0.02-0.38). Conclusion: In dedicated SP, manual adjustments of automatic determined vessel contours are essential for accurate stenosis quantiﬁcation. For manually determined

158

A

B

C

D

E

F

G

stenosis quantiﬁcation in vessel phantoms, no signiﬁcant difference was found between the different SP. In patient examination the presence of calciﬁed and mixed plaques strongly decreases inter-software agreement.

B-078

11:33

Automatic detection of atherosclerotic carotid plaque from combined magnetic resonance angiography and vessel wall images I. Isgum1, R. van ‘t Klooster1, P.J.H. de Koning1, F. Jabi2, K. DeMarco2, J.H.C. Reiber1, R.J. van der Geest1; 1Leiden/NL, 2East Lansing, MI/US ([email protected]) Purpose: To develop a computer-aided diagnosis system for automatic detection of atherosclerotic plaques in carotid arteries from combined magnetic resonance angiography (MRA) and multi-sequence vessel wall images (VWI). Methods and Materials: For 22 subjects both carotids were imaged (GE 3.0 T, Signa Excite, carotid 4-element coil). Scan parameters for T1, T1-PC and T2 were: FOV 16x13 cm, scan matrix 256x256 reconstructed to a 512x512 image, 18 slices, 2 mm slice thickness. 3D-IRFSPGR, TOF and MRA were reformatted to identical geometrical parameters. A trained observer set the reference standard by manually delineating carotid inner and outer walls, hemorrhage, lipid and calcium in VWI. VWI sequences were registered to MRA. Carotid wall was automatically segmented in all sequences. A statistical pattern recognition system was designed to detect hemorrhage, lipid and calcium. Each wall pixel was described with 239 features computed from texture analysis and vessel segmentation. A three-stage classiﬁcation employing linear discriminant and nearest neighbor classiﬁer with feature selection was performed. The system was evaluated in leave-one-subject-out fashion. Results: For hemorrhage, lipid and calcium lesion-wise sensitivity was 82.8, 58.5 and 61.3%, with an average 0.3, 1.2 and 1.1 false positive lesions per carotid, respectively. On average in terms of area hemorrhage lesions were underestimated by 12.6%, lipid overestimated by 26.2%, and calcium underestimated by 31.2%, with average false positive lesion area of 3.2, 8.7 and 6.8 mm2 per carotid, respectively. Conclusion: The system detected hemorrhage with high accuracy. However, small individual lipid and calcium lesions caused relatively large disagreement with the reference standard.

B-079

11:42

Semi-automatic measurement of left ventricular function on dual source computed tomography using four different software tools in comparison with magnetic resonance imaging G.J. de Jonge, P.A. van der Vleuten, D.D. Lubbers, H.J. van der Zaag-Loonen, P.M.A. van Ooijen, M. Oudkerk; Groningen/NL ([email protected]) Purpose: To assess left ventricular (LV) function using four different workstations on the same Dual Source Computed Tomography (DSCT) datasets and to compare the results with those of MRI. Methods and Materials: Twenty-six patients, undergoing cardiac computed tomography (DSCT, Somatom Deﬁnition, Siemens Medical Solutions, Germany) were included (20 men, mean age 59p12 yrs). Multiphase reconstructions were made at every 10% of the RR-interval. LV function analysis was performed on four different, commercially available workstations: Syngo Siemens (Siemens Medical Systems, Germany), Vitrea®2 (Vital Images, Minnetonka, MN), AquariusWorkstation (TeraRecon, San Mateo, CA) and QmassCT (Medis, Leiden, The Netherlands). In all software packages, semi-automatic LV function measurements were made, if necessary with manual corrections. Within 0-22 days (mean 6 days), patients were scanned on a 1.5 T MRI-system (Magnetom Sonata, Siemens Medical Systems, Germany) using a steady-state free precession sequence. Epi- and endocardial contours were drawn manually (Qmass v6.2.1, Medis, Leiden, The Netherlands). Bland-Altman analysis was performed to calculate limits of agreement between each workstation for CT and MRI. Results: For all four workstations, mean ejection fraction (Siemens 63%p10, Vitrea 64%p10, TeraRecon 61%p10 and QMassCT 58%p8) correlated well with measurements on MRI (58%p11) (r=0.88, 0.90, 0.80 and 0.83, respectively). Bland-Altman analysis in the comparison of CT and MR showed a mean difference of 5%p5, 7%p5, 3%p7 and 2%p5, respectively, for ejection fraction. Conclusion: Semi-automatic LV function analysis on DSCT correlates well with MRI measurements for four different CT software tools. A slight overestimation in ejection fraction for CT results compared with MRI was observed.

Scientific Sessions B-080

B-082

11:51

10:39

Comparison of two approaches for bone removal in intracranial CT angiography (CTA): Dual energy CT (DE-CT) and subtraction of a bone mask with “NeuroDSA” K. Papke, M. Pörner, D. Meila, F. Brassel; Duisburg/DE ([email protected])

Purpose: To provide a novel algorithm for the classﬁcation of lung-tissue depicted by multidetector-CT based on the topology of CT-attenuation-values and to compare the results with clinically established densitometric methods. Methods and Materials: The study comprises seven subjects without pulmonary pathology, seven with emphysema, and seven patients with ﬁbrosis. From each dataset, 7-21 cubic volumes-of-interest (VOI) (edge-length 40 pixels) are obtained, 275 probes in total. For each of these datasets, which are visually consensusclassiﬁed by two experienced radiologists, we calculate the Minkowski-Functionals (MF) as a function of gray-level-threshold. For comparison, we obtain the mean and median CT-density, and the lowest-ﬁfth-percentile of the density-histogram. Using receiver-operator-characteristic and Fisher´s Discriminant-Analysis, we compare the two groups of gray-level-topology- and density-based parameters with respect to correct classiﬁcation of the datasets. Results: For the densitometric parameters, normal and abnormal VOIs can be distinguished with an area-under-the-curve (AUC) ranging from 0.78 to 0.85. By quantitative topological texture-analysis, correct classiﬁcation is achieved with an AUC=0.87-0.96. Correlation between the two groups of parameters is non-signiﬁcant (pr0.36). By discriminant-analysis applied to a combination of densitometric and topological quantities, correct diagnosis can be increased to 92%. Conclusion: The topological method performs well for identiﬁcation of pathologic lung tissue. By combining the complementary information of densitometry and topology, the number of correctly classiﬁed datasets is increased further. Topological analysis of pulmonary tissue by MF may well be suited to enhance density-based analysis and thus improve monitoring texture-changes in the course of progression of pulmonary disease.

Purpose: Bone removal in intracranial CTA is helpful for the display of vascular structures close to bone (e.g., arteries at the skull base, bridging veins close to the calvarium) and facilitates rotated maximum intensity projection of large volumes. This study compares the image quality of two different approaches for bone removal. Methods and Materials: Fifteen patients obtained an unenhanced head CT and intracranial CTA in a Dual Energy CT scanner (Siemens Somatom Deﬁnition) with standard clinical protocols. Bone removal was performed with (a) use of two different tube voltages for discrimination between bone and opaciﬁed vessels (DE-CT) and (b) subtraction of bone with the tool “NeuroDSA” using the unenhanced CT as a mask. Overall image quality and the presence of residual bone after removal were rated using maximum intensity projections of the whole dataset. Artiﬁcial narrowing or extinction of arteries within the skull base and of bridging veins adjacent to the calvarium were detected by direct visual comparison of axial source images before and after bone removal. The quality for each criterion was rated on a scale from 1 (=excellent) to 4 (=poor). Results: Bone removal with “NeuroDSA” was superior to DE-CT in all criteria; mean quality was 1.8 for “NeuroDSA” and 2.7 for DE-CT. All differences were statistically signiﬁcant. Conclusion: In intracranial CTA, bone subtraction with “NeuroDSA” should be preferred over bone removal with DE-CT.

10:30 - 12:00

Room L/M

Neuro

SS 211 CT angiography and perfusion Moderators: B. Lubicz; Brussels/BE J. Walecki; Warsaw/PL

B-081

10:30

Dual energy direct bone removal CT angiography for evaluation of intracranial aneurysm or stenosis: Comparison with conventional digital subtraction angiography Y. Watanabe1, K. Uotani1, T. Nakazawa1, M. Higashi1, Y. Hori1, S. Kanzaki1, T. Fukuda1, N. Yamada1, T. Itoh2, K. Otani2, H. Naito1; 1Suita/JP, 2Tokyo/JP ([email protected]) Purpose: Dual energy CT can differentiate iodine from bone and calciﬁcations and can be applied to direct bone-removal CT angiography (CTA). The aim of this study was to compare the results of dual energy direct bone removal CTA (DEBoneRemoval-CTA) to that of digital subtraction angiography (DSA). Methods and Materials: CTA and DSA were obtained in three patients with intracranial aneurysms and three patients with intracranial carotid stenosis. These patients were scanned on a dual-source CT system in dual energy mode at tube voltages of 140 and 80 kVp. The software registers the two-energy data sets and selectively removes bone structures. 3D-volume-rendered images and maximum-intensity projection (MIP) images with and without bone removal were reviewed and compared to DSA. Results: Bone removal was successful in all patients. For four patients, bone removal was good with only tiny bone remnants and MIP images were easy to read. For two patients, bone removal was moderate with some bone remnants but this did not disturb the 3D visualization. Two aneurysms adjacent to the skull base were only partially visible in conventional CTA but were fully visible in DE-BoneRemovalCTA. In three patients with intracranial carotid stenosis, the stenotic vessel was not visible in conventional CTA. With DE-BoneRemoval-CTA, all intracranial stenotic lesions were visible on the MIP images. Compared to DSA, DE-BoneRemoval-CTA leads to an overestimation of severe stenosis. Conclusion: DE-BoneRemoval-CTA is able to eliminate bone structure using only a post-contrast scan and is useful to evaluate intracranial aneurysms and stenosis.

B-083

10:48

Deﬁning the shape of intracranial aneurysms with 16-row multidetector CT angiography (MDCTA): Comparison with 3D rotational angiography (3DRA) in a phantom model K. Papke, F. Brassel; Duisburg/DE ([email protected]) Purpose: Precise shape deﬁnition of intracranial aneurysms (in particular of the neck region, including the relation between the diameters of aneurysm neck and dome) is important for treatment decisions in intracranial aneurysms (e.g. neurointerventional coiling vs neurosurgical clipping, coiling plus stenting). This study compared measurement results of MDCTA and 3D-RA in a phantom model of intracranial aneurysms. Methods and Materials: Twelve saccular aneurysms with similar dome diameters (6.5-7.5 mm) but different neck sizes (1.5-5.6 mm) were manufactured from polyethylene tubes of three mm lumen diameter and ﬁlled with diluted iodine contrast agent diluted to a density of 200 HU. Aneurysms were placed in a water tank of 10.5 cm diameter and imaged with 16-row MDCT (Siemens Somatom Sensation 16, 120 kV, 130 mAs, TF=12 mm/rot., collimation: 0.75 mm, reconstruction thickness/ increment: 0.75/0.5 mm). For gold standard, aneurysms were ﬁlled with undiluted iodine contrast agent (240 mg/ml) and imaged using 3D-RA (Siemens Axiom Artis BA). Measurements of dome and neck diameter were performed in orthogonal thin slice reformations parallel and orthogonal to the phantom vessel axis. Results: There was a high correlation between MDCTA and 3D rotational angiography measurements (r=0.75 for mean aneurysm diameter, r=0.9 for mean neck diameter). Conclusion: MDCTA is highly accurate in the deﬁnition of shape parameters relevant for treatment decisions in intracranial aneurysms. However, results need conﬁrmation in vivo to establish the comparative value of MDCTA versus 3D-RA in a clinical setting.

B-084

10:57

The application of angiographic computed tomography in the periinterventional cerebral diagnostic imaging M.-N. Psychogios, J.-H. Buhk, C. Sohns, A. Mohr, M. Knauth; Goettingen/DE ([email protected]) Purpose: Evaluation of angiographic computed tomography (ACT) as a fast available diagnostic tool after neuroendovascular procedures. Comparison with periinterventional multislice computed tomography (MSCT) and magnetic resonance imaging (MRI). Methods and Materials: 52 peri-interventional ACT-acquisitions were obtained and evaluated, in 40 cases after coil embolisation of cerebral artery aneurysms and in 12 patients after intracranial angioplasty with stent placement. Digital subtraction angiography was performed on an AXIOM Artis biplane angiography

A

B

C

D

E

F

G

159

Friday

Differentiation between normal and pathologic lung tissue in multidetector-CT of patients with emphysema and lung-ﬁbrosis using a novel texture metric based on the topology of CT-attenuation-values H.F. Boehm1, C. Fink2, U.I. Attenberger1, C. Becker1, M. Reiser1; 1Munich/DE, 2 Mannheim/DE ([email protected])

Scientific Sessions system equipped with ﬂat-panel detectors (Siemens Medical Solutions, Erlangen, Germany). Postprocessing was performed on a Leonardo workstation (Siemens Medical Solutions) and multiplanar reconstruction slices were generated. The reference studies were performed either directly after the intervention or in a 48-hour period after the endovascular procedure. The MRI examinations were performed in a 3 Tesla scanner (Magnetom Trio, Siemens Medical Solutions) and the conventional CT studies in a 16 MSCT scanner (Aquilion 16, Toshiba Medical Systems, Japan). All studies were independently evaluated by two blinded neuroradiologists. Results: ACT images were of sufﬁcient diagnostic quality in all cases, concerning intracranial spatial relations (midline shift, supra-/infratentorial ventricular system). While diagnostic evaluation of the grey matter (basal ganglia, insular cortex, opercula) was possible in 77% of the cases, ACT proved to be a reliable tool for the detection of peri-interventional subarachnoidal or intracerebral hemorrhage in 100% of the cases. Conclusion: After neuroendovascular procedures and within the angiography suite, ACT enables an immediate detection/exclusion of peri-interventional hemorrhage or hydrocephalus. However, for the detection of early papenchymal signs of cerebral infarction, ACT is not yet reliable enough.

B-085

11:06

Can dynamic CTA substitute for helical CTA of 64-row volume CT in diagnosing the stenosis of MCA in TIA patient? Y.G. Gao, L. Kuncheng, X.D. Du, Y.S. Shen; Beijing/CN ([email protected]) Purpose: To evaluate the image quality and accuracy of dynamic CTA compared with helical CTA in diagnosing middle cerebral artery (MCA) stenosis using 64-row Volume CT (VCT). Methods and Materials: Eighty-ﬁve transient ischaemic attack patients were examined by GE 64-row VCT with the following protocol: non-contrast axial CT of the brain, followed by dynamic CT perfusion (5 mm thick; 8 interesting slices; 40 mm range), and helical cerebral CTA.Using the same raw data of volume CTP, the dynamic CTA of Willis circle in 0.625.1.25, 2.5 and 5 mm thickness of 40 mm range were reconstructed respectively. Helical cerebral CTA 0.625 mm thick was reconstructed in 40 mm range including Willis circle and in the same WW and WL of dynamic CTA. Cerebral arteries were all evaluated using maximum intensity projection (MIP) images. The analysis of variance was used to verify the statistically signiﬁcant difference between helical CTA and all kinds of thickness of dynamic CTA in measuring the CT value of MCA. Results: There was no statistically signiﬁcant difference in measuring the CT value of MCA between helical CTA and dynamic CTA of 0.625 mm and 1.25 mm thickness (p 0.05), respectively. There was statistically signiﬁcant difference in measuring the CT value of MCA between helical CTA and dynamic CTA of 2.5 mm (p 0.05) and 5 mm thickness (p 0.01). Conclusion: As a result of high spatial resolution and greater scan volume, dynamic CTA of Willis circle in 0.625 and 1.25 mm thickness can replace the helical cerebral CTA of Willis circle to diagnose the stenosis of MCA and to reduce total radiation dose.

B-086

11:15

Relation between vasospasm and cerebral perfusion after subarachnoid hemorrhage M. Rijsdijk; Utrecht/NL ([email protected]) Purpose: The presence of vasospasm after subarachnoid hemorrhage (SAH) does not necessarily lead to delayed cerebral ischemia (DCI). We investigated the effects of vasospasm on cerebral perfusion. Methods and Materials: We obtained CT angiography (CTA) and CT perfusion (CTP) scans on admission (within 72 hours after SAH) in 37 patients. All patients underwent a follow-up CTP and CTA scan at time of the clinical deterioration or, when patients remained clinically stable, about one week after admission. The presence of vasospasm (none ( 25% decrease of luminal diameter), moderate (25-50%), severe ( 50%)) and cerebral perfusion were scored for all patients. The relation between vasospasm and cerebral perfusion values in the corresponding ﬂow territories was assessed. Additionally, for patients with or without vasospasm, we compared whether cerebral perfusion differed in patients with or without DCI. Results: Twenty-three patients (62%) had vasospasm of whom 10 patients (43%) developed DCI. With increasing degree of vasospasm, CBF values decreased (none 59.51 ml/100 g/min, moderate 52.74, severe 38.19; p-value=0.1), CBV values decreased (none 4.28 ml/100 g, moderate 3.96, severe 3.51; p-value 0.01), MTT values increased (none 4.81 sec, moderate 4.98, severe 8.10; p-value 0.01) and TTP values increased (none 22.75 sec, moderate 23.46, severe 26.99; p-value=0.1). Perfusion was lower in patients later developing DCI, both for patients with and without vasospasm. Perfusion was lowest in patients with vasospasm who developed DCI.

160

A

B

C

D

E

F

G

Conclusion: Our ﬁndings suggest that cerebral perfusion is affected by vasospasm; mainly by severe vasospasm. A decrease in cerebral perfusion does not necessarily lead to DCI.

B-087

11:24

Global and focal cerebral perfusion after aneurysmal subarachnoid haemorrhage in relation with delayed cerebral ischaemia M. Rijsdijk; Utrecht/NL ([email protected]) Purpose: The pathogenesis of delayed cerebral ischaemia (DCI) after subarachnoid haemorrhage (SAH) is unclear. We assessed whether DCI relates to focal or global cerebral perfusion on admission and on follow-up imaging. Methods and Materials: Twenty-seven SAH patients underwent CT Perfusion (CTP) on admission and at clinical deterioration, or one week after admission in clinically stable patients. We compared global and focal (least perfused territory) perfusion in patients with DCI (n=12), clinically stable patients (n=7) and patients with non-DCI related deterioration (n=8). Results: Global cerebral blood ﬂow (CBF) increased on follow-up: 29% (95% conﬁdence interval (CI) 15% to 43%) in patients with DCI, 12% (95%CI -1% to 25%) in stable patients and 20% (95%CI 4% to 36%) in patients with non-DCI related deterioration. Focal CBF decreased in patients with DCI, (-23%; 95%CI -58% to 12%) but increased in patients with non-DCI related deterioration (23%; 95%CI -26% to 55%) and stable patients (7%; 95%CI -30% to 45%). On follow-up global CBF was lower in patients with DCI (70.0 ml/100 g/min) than in clinically stable patients (81.6; difference 11.6; 95%CI 0.8 to 22.5 ml/100 g/min), but comparable to patients with non-DCI related deterioration (67.6; difference -2.4; 95%CI -11.9 to 7.2 ml/100 g/min). Focal CBF was lower in patients with DCI (30.7) than in clinically stable patients (53.6; difference 22.9; 95%CI 5.1 to 40.6 ml/100 g/min) and patients with non-DCI related deterioration (46.6; difference 15.9; 95%CI -2.6 to 28.4 ml/100 g/min). Conclusion: Our results suggest that DCI is more likely a focal or multifocal than a global process.

B-088

11:33

Comparative study of perfusion CT parameters to differentiate penumbra from infarct core in the acute stroke setting F.L.G. Borny, P.C. Seynaeve, H.H. Pottel, J.S. Meersschaert, E.E. Pattyn, S.S. Smet, P.M. Vanderdonckt, D.G. Peeters; Kortrijk/BE ([email protected]) Purpose: PerfusionCT (PCT) is gaining interest in the evaluation of stroke patients. The absence of validated parameters is a major disadvantage. In several studies the mean transit time has been proposed as the most sensitive parameter in the acute stroke setting. As it lacks speciﬁcity the aim of this study is to compare all previously published perfusion CT parameters. Methods and Materials: A total of 68 acute stroke patients were included. CT perfusion was performed shortly upon admission using a 64 slice CT (GE, Milwaukee, USA). A native CT, a cine CT covering a 40-mm slab upwards from the base of the lateral ventricles and a post contrast CT were performed. The deconvolution method was used to obtain different PCT parameters. Sensitivity, speciﬁcity and accuracy results were calculated to differentiate penumbra from infarct core. Results were compared with T2 FLAIR and diffusion MRI results. Results: Using an absolute cerebral blood ﬂow (aCBF) cut off value of 10 ml/100 g/ min in a region of relative prolonged MTT to evaluate the presence of infarction resulted in a sensitivity (S) of 65%, a speciﬁcity (Sp) of 100% and an S+Sp of 165%. Using absolute mean transit time values, absolute cerebral blood volume values, relative CBF values, relative cerebral blood volume (CBV) values, CBFxCBV values or different aCBF values only resulted in lower values of S+Sp. Conclusion: The study demonstrates that the use of aCBF cut off value of 10 ml/ 100 g/min is the best parameter in differentiating penumbra from infarct core.

B-089

11:42

Changes in perfusion parameters of double CT perfusion studies as predictor of early mortality in ischemic stroke S. Langner, S. Stumpe, M. Kirsch, A.V. Khaw, N. Hosten; Greifswald/DE ([email protected]) Purpose: The knowledge of predictors of mortality in ischemic stroke, which will help to identify patients at a higher risk of fatal outcome and to optimize therapeutic strategies in stroke units, is most important. The aim of the study was to evaluate the predictive value of changes in repeated perfusion CT parameters within 48 hours for early neurologic outcome. Methods and Materials: 44 consecutive patients with clinical signs of acute ischemic stroke were evaluated prospectively. All examinations were performed in a

Scientific Sessions

B-090

11:51

Cost-effectiveness of CT scanning in minor head injury: Selective strategies versus scanning all patients M. Smits1, D.W.J. Dippel1, D.A. van Rijssel1, H.M. Dekker2, P.E. Vos2, D.R. Kool2, P.J. Nederkoorn3, P.A.M. Hofman4, A. Twijnstra4, H.L.J. Tanghe1, M.G.M. Hunink1; 1 Rotterdam/NL, 2Nijmegen/NL, 3Amsterdam/NL, 4Maastricht/NL ([email protected]) Purpose: CT scanning is routinely used for timely diagnosis of neurocranial complications after minor head injury (MHI). Prediction rules can be used to select patients for head CT. We assessed the cost-effectiveness of selective CT scanning strategies, compared with routine scanning. Methods and Materials: We evaluated ﬁve strategies: scanning all MHI patients; selective scanning according to the New Orleans Criteria (NEJM; 343:100-5), Canadian CT Head Rule (CCHR; Lancet; 357:1391-6) or CT in Head Injury Patients rule (AnnIntMed; 146:397-402); no scanning. Model parameters were primarily based on our own study of 3,181 MHI patients (JAMA; 294:1519-25). We used a decision tree for short-term, and a Markov model for long-term costs and effectiveness. Outcome measures were ﬁrst year and lifetime costs, quality adjusted life years (QALYs), and incremental cost-effectiveness ratio (ICER). The model’s robustness was tested against varying the model parameters across their 95% conﬁdence intervals. Results: First year costs were highest for patients requiring neurosurgery not having initially undergone CT (35,437 euro). The scanning strategies showed little difference in effectiveness; the CCHR was the least costly (6,941 euro). The model was robust against varying most parameters, except for the prediction rules’ sensitivities to identify patients requiring neurosurgery. At sensitivities below 97% scanning all patients was cost-effective (ICER 60,000 euro/QALY) compared with the CCHR. At sensitivities below 89%, scanning all patients became less costly and more effective. Conclusion: The CCHR was most cost-effective. However, if the prediction rules’ sensitivities for identifying patients requiring neurosurgery were below 97%, scanning all MHI patients was cost-effective compared with selective scanning.

10:30 - 12:00

Room N/O

Contrast Media

SS 206a Side effects Moderators: J. Grimm; New York, NY/US S.K. Morcos; Sheffield/UK

B-091

10:30

Iodinated contrast media inhibit oxygen consumption in freshly isolated proximal tubular cells from humans and diabetic rats F. Palm, A. Fasching, P. Hansell, P. Liss; Uppsala/SE ([email protected]) Purpose: Injection of iodinated contrast media (CM) is associated with increased risk for acute kidney failure in patients with pre-existing renal dysfunction. The exact mechanism mediating the nephropathy is not known, but earlier studies have proposed alterations in the renal oxygen metabolism as a potential mediator. In this study, we compared the effect of iopromide and iodixanol on the oxygen consumption (QO2) in freshly isolated proximal tubular cells (PTC) from nephrectomized patients and identiﬁed a potential mechanism for the observed effects on PTC

from normoglycemic and diabetic rats. Methods and Materials: PTC were isolated by a collagenase technique from nephrectomized normoglycemic patients and from adult healthy and streptozotocin-induced diabetic Sprague-Dawley rats. In vitro oxygen consumption was measured in a gas-tight chamber on PTC with and without incubation with iopromide or iodixanol (20 mg I/ml media). Ouabain was used to inhibit Na+/K+-ATPase and L-NAME to inhibit nitric oxide (NO) synthase. Results: Both CM decreased QO2 in the normoglycemic human PTC (-35%). However, PTC from normoglycemic rats were unaffected by iopromide, whereas iopromide decreased QO2 in PTC from diabetic rats (-28%). The effect on diabetic PTC was completely blocked by L-NAME. Conclusion: These observations suggest that iodinated CM reduce in vitro QO2 in human PTC, similar to what was observed in PTC from diabetic rats. Furthermore, the mechanism for the effects on the diabetic PTC involves NO release since the effect was blocked by L-NAME. These novel ﬁndings might suggest an involvement of mitochondrial ATP production for the development of CM-induced nephropathy.

B-092

10:39

Proapoptotic and pronecrotic effects of iodinated contrast media on renal tubular cells following repeated and single exposure J.-M. Idée, K. Pallier, S. Ballet, P. Prigent, E. Lancelot, C. Corot; Roissy/FR ([email protected]) Purpose: To investigate the proapoptotic and pronecrotic effects of iodixanol following single and repeated exposure to proximal tubular cells and to compare those of several contrast media (CM). Methods and Materials: All CM were incubated with with LLC-PK1 cells. The proapoptotic/pronecrotic effects of iodixanol (120 mg I/mL) were measured following two 6-h incubations at 24 and 48 hours interval. Diatrizoate, ioxaglate, iodixanol, ioxilan and iohexol were also compared following single 6-h incubation (120 mg I/mL). Apoptosis and necrosis were measured by ﬂow cytometry. Actinomycin-D (5 μM) and 5% glucose were used as positive and negative controls, respectively. Statistical analysis by Mann-Wittney’s test. Results: Re-exposure of iodixanol led to a signiﬁcant increase in apoptosis at 48 hours (ﬁrst exposure: 33.3+4.7% of cells; 2nd exposure at 24 h: 41.9+4.7%, NS; 2nd exposure at 48 h: 62.1+3.6%, p 0.05 vs 1st exposure) whereas re-exposure to 5% glucose had no effect. Such a delayed exposure-dependent effect was not found for necrosis. All CM induced signiﬁcant apoptosis vs 5% glucose. Iodixanol (33.8+2.8%), iohexol (34.6+3.9%) and ioxilan (27.4+3.3%) were more proapoptotic than ioxaglate (15.3+3.0%) and diatrizoate (10.7+2.2%) (p 0.05). The greatest pronecrotic effect was observed with diatrizoate (21.0+1.7% of cells, p 0.05 vs other CM). Conclusion: Re-exposure of tubular cells to iodixanol led to an increase of the proapoptotic but not the pronecrotic effect. Unlike the low-osmolar CM tested, diatrizoate-induced cell death was mostly pronecrotic.

B-093

10:48

A randomized, double-blind comparison of contrast-induced nephropathy in high-risk patients: The PREDICT study M.J. Kuhn1, N. Chen2, D.V. Sahani3, D. Reimer4, E.J.R. van Beek5, J.P. Heiken6, G.J. So7; 1Springﬁeld, IL/US, 2Shanghai/CN, 3Boston, MA/US, 4Mobile, AL/US, 5 Iowa City, IA/US, 6St. Louis, MO/US, 7Torrance, CA/US ([email protected]) Purpose: To compare the incidence of CIN following intravenous (IV) administration of low-osmolar or iso-osmolar contrast media (CM) in patients with diabetes mellitus (DM) and chronic kidney disease (CKD) undergoing computed tomography (CT) using a multicenter, randomized, double-blind, parallel-group study design. Methods and Materials: In the PREDICT (Patients with REnal impairment and DIabetes undergoing Computed Tomography) study, 248 patients with moderate-tosevere CKD (eGFR 20-59 mL/min) and DM were randomized to receive a minimum of 65 mL of iopamidol-370 or iodixanol-320 administered IV for their CT procedure. Serum creatinine (SCr) was measured at screening, baseline and 48-72 h postcontrast. CIN was deﬁned as a post-contrast increase in SCrr25% from baseline. The incidence of CIN was compared using Fisher's exact test. Results: A total of 125 patients received iopamidol-370 and 123 received iodixanol-320. The two groups were comparable in age, gender, body weight, IV hydration, CM dose, concomitant administration of nephrotoxic drugs, and baseline SCr (iopamidol-370: 1.46 mg/dL; iodixanol-320: 1.42 mg/dL; p=0.35). Baseline renal function was slightly worse in the group receiving iopamidol-370 (46.3 vs 48.6 mL/min; p=0.05). Increases in SCrr25% occurred in seven (5.6%) patients receiving iopamidol-370 and six (4.9%) subjects receiving iodixanol-320 (95% CI iopamidol-iodixanol=[-4.8%, 6.3%], p=1.0). Mean SCr changes from baseline were

A

B

C

D

E

F

G

161

Friday

16 row MSCT. CT perfusion was performed on admission and repeated within 48 hours (two slices, 12 mm slice thickness, 1 scan/s for 40 s, 80 kV, 209 mAs, 40 ml contrast agent, ﬂow 5.9 mL/s). Perfusion data was analyzed using the vendor’s Neuroperfusion software. Perfusion maps for rCBV, rCBF and TTP were calculated for each hemisphere. Relative changes in perfusion parameters were calculated. Neurological assessment was done on admission and daily until discharge using NIHSS. Spearman correlation and ROC curves were used for statistical analysis. Results: Perfusion CT was diagnostic in all patients. Improvement or stable NIHSS was observed in 32/44 patients (74.2%) and worsening of symptoms in 12/44. In 17 patients an increase in rCBF correlated with a stable or improved NIHSS on discharge (p=0.037) with a cut-off value of 3.76 in percent change. Increased rCBV correlated with improved NIHSS in 13/31 patientens (p=0.049). There was no correlation for TTP. Conclusion: Changes in rCBF of repeated CT perfusion within 48 hours have a high predictive value for early neurologic outcome.

Scientific Sessions 0.04 mg/dL for both groups (ANCOVA p=0.80). No patient required dialysis or died from acute renal failure. Conclusion: The incidence of CIN in patients with DM and CKD receiving IV CM is not signiﬁcantly different after iopamidol-370 or iodixanol-320.

B-094

10:57

Preclinical studies to investigate the development of NSF: Experiments in renally impaired rats H. Pietsch, J. Walter, P. Lengsfeld, T. Frenzel, J. Huetter, M.A. Sieber; Berlin/DE ([email protected]) Purpose: Nephrogenic Systemic Fibrosis (NSF) is only observed in patients with severe renal dysfunction and a role for Gadolinium-based contrast agents (GBCAs) as a possible trigger has been suggested. The objective of this study was to evaluate the impact of prolonged circulation time of GBCAs on the onset of NSF-like signs in rats. Methods and Materials: The Gadolinium (Gd) concentration in the skin of 5/6nephrectomized rats (simulating renal impairment), and of healthy Han-Wistar rats, was determined after application of Gadodiamide (Omniscan®), Gadoversetamide (OptiMARK®), Gd-DTPA (Magnevist®), and Gadobutrol (Gadovist®). The animals received single injections of 2.5 mmol Gd/kg into the tail vein on ﬁve consecutive days and the Gd-concentration in skin samples was determined by ICP-MS at several time points post-injection. Animals were examined daily for clinical ﬁndings. Results: Prolonged circulation times for all GBCAs were observed in the 5/6 nephrectomized rats as compared to healthy rats. For non-ionic linear compounds, the prolonged circulation time led to a higher Gd-concentration in the skin over time, while for linear ionic compounds, the increase in Gd-concentration was less pronounced. For the macrocyclic compounds, no long-term Gd retention in the skin could be observed for any of the animals. Conclusion: In the study a prolonged circulation time for linear contrast agents due to reduced renal elimination correlated with an increased Gd retention in skin, in particular after administration of non-ionic linear GBCAs. For macro-cyclic compounds, no long-term retention of Gd in the skin could be observed.

B-095

11:06

Stability of gadolinium-based contrast agents in human serum T. Frenzel; Berlin/DE ([email protected]) Purpose: To study the stability of gadolinium-based contrast agents (GBCAs) in human serum at 37 °C, which may play a role in the context of nephrogenic systemic ﬁbrosis. Methods and Materials: The following GBCAs were investigated: non-ionic linear complexes: Omniscan® and OptiMARK®. As both products contain large excess of free ligand, the respective drug substances gadodiamide and gadoversetamide were also investigated. Ionic linear complexes: Magnevist®, MultiHance®, Primovist®, and Vasovist®; macrocyclic complexes: Dotarem®, Gadovist®, and ProHance®. Each compound was incubated in human serum from healthy volunteers at a concentration of 1 mmol/l at 37 °C. Samples were repeatedly analyzed up to 25 days by ion-exchange HPLC-ICP-MS to quantify intact Gd-complex and released Gd and to estimate the half-life of Gd-complex dissociation in serum. Results: For the non-ionic drug substances, dissociation half-lives were in the range of a few weeks. Excess free ligand in the formulation delayed the Gd-release from Omniscan and OptiMARK by 2-3 days but did not alter the dissociation rate thereafter signiﬁcantly. The dissociation half-lives for the ionic GBCAs were in the range of several months (Magnevist and MultiHance) to a few years (Primovist and Vasovist). For the macrocyclic GBCAs half-lives could not be calculated as no Gd-release above the detection limit of ~0.1% was found after 25 days, but extrapolation yielded lower limits for the half-lives of many decades. Conclusion: The study results demonstrate differences among the GBCAs regarding Gd-release in serum of healthy volunteers at 37 °C. Non-ionic linear complexes exhibit the lowest stability, while macrocyclic complexes are the most stable.

B-096

11:15

Long-term retention of gadolinium in the skin of rodents following the administration of gadolinium based contrast agents M.A. Sieber, T. Frenzel, P. Lengsfeld, J. Hütter, H. Pietsch; Berlin/DE ([email protected]) Purpose: Several recent publications suggest a role for Gadolinium based contrast agents (GBCAs) as a possible trigger for Nephrogenic Systemic Fibrosis (NSF). The aim of the study was to evaluate the possible long-term retention of Gadolinium (Gd) in the skin of rodents following administration of different GBCAs. Methods and Materials: Gd-concentration in the skin was measured after applica-

162

A

B

C

D

E

F

G

tion of linear non-ionic (Omniscan® and OptiMARK®), linear ionic (Magnevist®), and macrocyclic GBCAs (Gadovist®, ProHance® and Dotarem®) in Han-Wistar rats. The GBCAs were injected i.v. once daily at a dose of 2.5 mmol Gd/kg for ﬁve consecutive days. The Gd-concentration in skin biopsies was determined at various time points (up to 250 days p.i). by ICP-MS. Results: Regarding the Gd-concentration in the skin, we observed statistically signiﬁcant differences between the different GBCA classes. For linear non-ionic compounds, accumulation during the injection period (ﬁve days) and high Gdconcentration was maintained over time in the skin (up to 250 days). For the linear ionic compounds, signiﬁcantly lower Gd retention was observed over time in the skin. Beginning 50 days after the last injection, the Gd values in the skin observed after application of all macrocyclic compounds were in the same range as observed in saline treated and in untreated animals. Conclusion: We observed a correlation between the complex stability of GBCAs and the amount of residual Gd in the skin up to several months after application of GBCAs. No long-term retention of Gd in the skin could be detected after application of macrocyclic GBCAs.

B-097

11:24

A center case review of nephrogenic systemic ﬁbrosis (NSF) T.C. Lauenstein, K.N. Salman, R. Morreira, S. Tata, D.R. Martin; Atlanta, GA/US ([email protected]) Purpose: To retrospectively analyze nephrogenic systemic ﬁbrosis (NSF) cases at our Center and to evaluate predisposition factors in dialysis patients for the development of NSF. Methods and Materials: Four databases (pathology, MRI, dialysis and medical records) were cross-referenced for identiﬁcation and evaluation of NSF patients. Medical history of NSF patients was assessed for previous deep venous thrombosis (DVT), surgery or infections. Laboratory data (creatinine, anion gap, calcium, phosphorus and albumin) and concurrent medication were evaluated. Findings were compared to those of a control group of non-NSF dialysis patients. Results: Between 10/2003 and 03/2007, a total number of 9 NSF cases were identiﬁed. All patients had undergone contrast-enhanced MRI prior to the diagnosis of NSF. Only one gadolinium compound had been used at our Center (Omniscan®, gadodiamide, GE Healthcare). Eight of 9 patients were receiving dialysis at the time of the MRI scan. The remaining patient had a transplant kidney with impaired function. During the same time, 312 dialysis patients received gadodiamide. Thus, the relative risk within dialysis patients to develop NSF amounted to 2.6%. There were no differences between the control group and the NSF group for previous history of DVT, surgery, infections, laboratory blood tests or medication. Conclusion: Our ﬁndings are consistent with early reports showing a strong correlation between gadodiamide exposure in patients with severely impaired renal function and development of NSF. NSF evolved in approximately 3% of dialysis patients after gadodiamide exposure. Although other cofactors are suspected, none were found in our cohort.

B-098

11:33

Prevalence of nephrogenic systemic ﬁbrosis in renal insufﬁciency patients: Results of the FINEST study S. Karie1, N. Janus1, V. Launay-Vacher1, G. Choukroun2, C. Frances1, G. Deray1, O. Clément1; 1Paris/FR, 2Amiens/FR ([email protected]) Purpose: NSF is an emerging cutaneous and systemic disorder characterized by widespread tissue ﬁbrosis, ﬁrst described in 2000. Recently, it has been suggested that gadolinium chelates may be responsible for NSF in patient with renal insufﬁciency (RI). The FINEST study (FIbrose Néphrogénique SysTémique) investigated the prevalence of NSF after a Magnetic Resonance Imaging (MRI) examination with or without gadolinium chelate administration in RI patients. Methods and Materials: All patients with RI and who had undergone MRI examination, over a period of one year, with or without gadolinium, were included. Data were retrospectively collected from nine Nephrology Departments in France, and included sex, age, renal function, type of gadolinium administered and cutaneous disorders (during a follow-up period of 4 months after MRI procedure). Patients were classiﬁed with the K/DOQI-K/DIGO deﬁnition according to their renal function. Results: 308 patients were included: mean age 59.9 years, 58.8% men, 53.6% had stage 5 RI (table), and among them 60% were dialyzed. 76.3% of the patients received gadolinium, most of them (57%) received Gadoterate and only a minority (2.3%) received Gadodiamide. No patients presented cutaneous disorders. Conclusion: Those results conﬁrm that NSF is a very rare disease. According to data available in the literature on the incidence/prevalence of NSF with Gadodiamide in patients with RI (~3.5%), cases should have been observed in our study. Most patients received a macrocyclic gadolinium chelate for which no case of NSF has been observed worldwide (Gadoterate), thus explaining why we did not observe any case of NSF.

Scientific Sessions B-099

11:42

Prevalence of NSF following administration of gadolinium-based contrast agents in patients with chronic renal failure G. Heinz-Peer, A. Neruda, B. Watschinger, A. Vychytil, M. Haumer, A. Geusau, M. Weber; Vienna/AT ([email protected])

B-100

11:51

Survival of patients on renal replacement therapy diagnosed with nephrogenic systemic ﬁbrosis compared to those without G.H. Roditi, P. Thomson, T. Collidge, P. Mark; Glasgow/UK ([email protected]) Purpose: Accurate survival data for patients diagnosed with nephrogenic systemic ﬁbrosis (NSF) are not known; previous crude mortality analysis of our cohort suggested no signiﬁcant difference as compared to other patients on renal replacement therapy (RRT)(1). The purpose of this investigation was to better determine accurate survival in an accepted manner for patients on RRT. Methods and Materials: The survival of all patients on RRT in the West of Scotland renal registry between 01/01/2000 to 01/07/2006 was analysed using data abstracted from the electronic patient record system. For all patients, survival from date of commencement of RRT to date of death was ascertained. Survival was analysed by Kapan-Meier/Log rank analysis. Results: 1824 patients in total, 14 diagnosed with NSF. Crude mortality of those without NSF 62.0% (688/1122) and those with 57.1% (6/8), overall 62.0%. Mean and median survival times for patients diagnosed with NSF were not statistically signiﬁcantly different - Log Rank (Mantel-Cox) Chi Square 0.778 with one degree freedom and p=0.378. Conclusion: Our data do not demonstrate a survival difference in patients with NSF compared to those without in the RRT cohort.

Room R1

Chest

SS 204 Friday

Purpose: To evaluate the prevalence of nephrogenic systemic ﬁbrosis (NSF) in patients with chronic renal failure who intravenously received gadolinium based contrast agents during MR-examinations. Methods and Materials: After institutional review board approval the radiological records of 552 patients with endstage renal disease being on hemodialysis or CAPD between 1997 and 2007 were retrospectively reviewed to identify whether the patients underwent MRI with intravenous administration of Gd-contrast media. The number of contrast injections, the type and the cumulative doses of Gd-based contrast agents, and possible cofactors regarding pathogenesis of NSF were recorded. Diagnoses of NSF were conﬁrmed or excluded either by clinical evaluation including deep skin biopsy in case of suspicious ﬁndings or by medical and pathological records. Results: 195/552 proved to have MR examinations with intravenous administration of Gd-based contrast agents. Seven different contrast agents were involved. Six of 195 patients had a diagnosis of NSF. One patient died of NSF. No signiﬁcant difference regarding cumulative dose and kind of contrast agent was found between NSF-affected and non-affected patients, although there was a slight tendency towards gadodiamide. History of thrombosis and recent surgery proved to be signiﬁcant risk cofactors for development of NSF. History of hyperparathyroidismus did not prove to be a risk factor. Conclusion: The data conﬁrm previously published ﬁgures on the prevalence of NSF among chronic renal failure patients undergoing MR examinations with intravenous administration of Gd-based contrast agents; they do not support a positive association between cumulative dose and speciﬁc type of contrast agent used.

10:30 - 12:00

Advances in tumor imaging Moderators: M. Boijsen; Gothenburg/SE A. Oikonomou; Alexandropoulis/GR

B-101

10:30

Compared efﬁcacy for M-stage assessment between whole-body MR imaging and whole-body FDG-PET-CT in non-small cell lung cancer patients Y. Ohno, H. Koyama, M. Nogami, D. Takenaka, Y. Onishi, K. Sugimura; Kobe/JP ([email protected]) Purpose: To directly compare the efﬁcacy of whole-body MR imaging (MRI) and FDG-PET-CT for M-stage assessment in non-small cell lung cancer (NSCLC) patients. Methods and Materials: 227 consecutive NSCLC patients prospectively underwent whole-body MRI, FDG-PET/CT, pre-therapeutic standard radiological examinations for diagnosis of M-stage and more than one-year follow-up examinations. Final diagnosis of M-stage in each patient was determined according to the results of all radiological and follow-up examinations. As whole-body MRI, STIR turbo spin-echo images, dual-phase T1-weighted gradient-echo images with and without contrast-media and diffusion-weighted images were obtained by using a moving-table system on a 1.5 T scanner. Two chest radiologists having more than three years of experience with PET independently evaluated probabilities of metastases on both methods by using 5-point scales on a per patient basis. Then, ﬁnal diagnosis in each patient was made by consensus of two readers. To compare diagnostic capabilities between two methods on a per patient basis, ROC analyses were performed. Then, sensitivity, speciﬁcity and accuracy were also compared by using McNemar’s test. Results: Area under the curve (Az) of whole-body MRI (Az=0.87) had no signiﬁcant difference with that of FDG-PET/CT (Az=0.89, p 0.05). Sensitivity, speciﬁcity and accuracy of whole-body MRI were 70.0 (28/40), 92.0 (150/163) and 85.7% (178/203). Sensitivity (62.5% [25/40]), speciﬁcity (94.5% [154/163]) and accuracy (88.2% [179/203]) of FDG-PET/CT had no signiﬁcant difference with those of whole-body MRI (p 0.05). Conclusion: Whole-body MRI is as effective as FDG-PET/CT for M-stage assessment in non-small cell lung cancer patients.

B-102

10:39

Assessment of mediastinal lymphadenopathy with diffusion weighted MR imaging A.A.A. Abdel Razek, A. Elmorsi, A. Elshafeai, T. Elhadedy, O. Hamza, N. Nada; Mansoura/EG ([email protected]) Purpose: To assess the role of diffusion-weighted MR imaging in patients with mediastinal lymphadenopathy. Methods and Materials: Prospective study was conducted on 47 patients with untreated mediastinal lymphadenopathy (32M & 15 F aged 5-70 ys). They underwent single shot echo planar diffusion-weighted MR imaging of the mediastinum with b-factor of 0, 300 & 600 sec/mm2. The scanning parameters were: TR=10000 ms, TE=108 ms, bandwidth =125 kHz, slice thickness=4 mm. Apparent diffusion coefﬁcient (ADC) maps were reconstructed. The ADC value of the mediastinal lymph nodes was calculated. The ﬁnal diagnosis was performed by biopsy. The ADC value was correlated with the biopsy results and statistical analysis was done. Results: The mean ADC value of metastatic LN (0.96p0.32X10-3 mm2/sec), non-Hodgkin lymphoma (0.97p0.24X10-3 mm2/sec) and Hodgkin disease (1.07p0.11X10-3 mm2/sec) was signiﬁcantly lower (p 0.04) than that of sarcoidosis (1.94p0.51X10-3 mm2/sec) and tuberculous (3.01p0.58X10-3 mm2/sec) mediastinal lymph nodes. There is an insigniﬁcant difference in ADC value within malignant nodes (p 0.006). When apparent diffusion coefﬁcient value of 1.85X103 mm2/sec was used as a threshold value for differentiating malignant mediastinal nodes from benign nodes, the best results were obtained with an accuracy of 83.9%, sensitivity of 96.4%, speciﬁcity of 71.4%, positive predictive value of 77.1% and negative predictive value of 95.2%. Conclusion: Diffusion-weighted MR imaging is a new non-invasive promising imaging modality that can be used for assessment of mediastinal lymphadenopathy and differentiation of malignant from benign mediastinal lymph nodes.

A

B

C

D

E

F

G

163

Scientific Sessions B-103

10:48

Characterization of mediastinal mass with diffusion weighted single shot echo planar MR imaging A.A.A. Abdel Razek, A. Elmorsi, A. Elshafeai, T. Elhadedy, O. Hamza, N. Nada; Mansoura/EG ([email protected]) Purpose: To assess the clinical usefulness of diffusion weighted single shot echo planar MR imaging in characterization of mediastinal mass. Methods and Materials: Prospective study was conducted on 67 consecutive patients (42M, 15 F aged 5-67 ys; mean 34 ys) with mediastinal mass. They underwent diffusion weighted single shot echo planar MR imaging of the mediastinum with b-factor of 0, 300 and 600 sec/mm2. The scanning parameters were: TR=10000 ms, TE=108 ms, bandwidth=125 kHz, slice thickness=4 mm. Apparent diffusion coefﬁcient (ADC) maps were reconstructed. The ADC value of the mediastinal mass was calculated and correlated with the histopathological ﬁndings. Results: The mean ADC value of malignant mediastinal tumors was 1.07p0.29X103 mm2/sec, benign solid masses was 2.53p0.62X10-3 mm2/sec and cystic lesions were 3.99p0.11X10-3 mm2/sec. There was signiﬁcant difference in the mean ADC values between malignant and benign mediastinal masses (p 0.001) and within benign masses (p 0.029) and insigniﬁcant difference within mediastinal malignancy (p 0.65). When an ADC value of 2.05X10-3 mm2/sec was used as a threshold value for differentiating malignant from benign mediastinal mass, the best results were obtained with an accuracy of 90.9%, sensitivity of 97.2%, speciﬁcity of 84.6%, positive predictive value of 86.4% and negative predictive value of 96.7%. Conclusion: Apparent diffusion coefﬁcient map is a new imaging modality used for characterization of mediastinal mass as it can differentiate benign mediastinal mass from malignant tumor.

B-104

10:57

Peripheral lung carcinoma: Correlation of angiogenesis and ﬁrst-pass perfusion parameters of 64-detector row CT Y. Li, Z.-G. Yang, T.-W. Chen, H.-J. Chen, J.-Y. Sun; Chengdu/CN ([email protected]) Purpose: To investigate the feasibility of 64-detector row CT perfusion imaging to assess the hemodynamics of the tumor and evaluate the correlation between ﬁrstpass perfusion parameters and angiogenesis of peripheral lung carcinomas. Methods and Materials: Forty-six surgically peripheral lung carcinomas were examined with 64-detector row CT. First-pass CT perfusion study comprised 12 repeated spiral acquisitions over 60 sec following a 50-ml intravenous bolus of contrast medium at 6-7 ml/sec. Perfusion, peak enhancement intensity (PEI), time to peak (TTP), blood volume (BV) and microvessel density (MVD) of the tumor were compared by means of one-way ANOVA among histological types, sizes and metastases. Pearson correlation coefﬁcients were calculated to represent relationships between perfusion parameters and MVD of the tumor. Results: Mean values for perfusion, PEI, TTP, and BV of the 46 tumors were 70.3p39.4 ml/min/ml, 67.0p37.6 Hu, 36.9p11.2 sec and 34.9p17.9 ml/100 g, respectively. No statistically signiﬁcant differences were found in perfusion parameters among histological types (p 0.05). Considerable differences with higher perfusion, PEI and BV were noted in tumor b30 mm than in tumor 30 mm (p 0.05). No signiﬁcant differences were found in perfusion parameters between nodule metastasis positive and negative groups (p 0.05). Perfusion, PEI and BV were positively correlated with extent of MVD (r=0.715, 0.681, 0.762, respectively, all p 0.001), whereas no signiﬁcant correlation was found between TTP and MVD (r=-0.154, p 0.05). Conclusion: 64-detector row CT ﬁrst-pass perfusion technique is a valuable noninvasive method in assessing tumor angiogenesis of lung carcinoma. The perfusion parameters could reﬂect the underlying extent of tumor angiogenesis.

B-105

11:06

BALI (BLAST Accelerated Lung Imaging) - Kt blast dynamic contrast enhanced multiphasic imaging of lung masses: A study of 22 cases R.M.S.V. Vadapalli, J. Akkaraju, T. Mandapal; Hyderabad/IN ([email protected]) Purpose: To evaluate the role of DCE MRI of lung tumors for characterization using BALI (kt BLAST accelerated lung imaging). Methods and Materials: 21 patients with CECT and biopsy proven cases of lung cancer with age range 39-54 years and M:F ratio 2:1 were included in the study after informed consent. All patients were imaged on a 3 Tesla Achieva XR system (Phillips Medical Systems) with a 12 channel body coil. 2D-balanced TFE axial, coronal and sagittal breathhold sequences followed by contrast-enhanced 3D GRE imaging with kt BLAST (k-space time Broad-use Linear Acquisition Speed-up) technique with acceleration factor of 8, number of phases 18 covering entire lungs, with three breathholds generating six

164

A

B

C

D

E

F

G

phases per breathhold with a temporal resolution 2 sec per phase. The multiphasic data were then post processed on a functool platform and wash in and out kinetic curves, signal enhancement ratios, and time-to-peak indices with color-coded images were generated and analysed. Results: All the lung cancers showed enhancement with contrast. Bronchoalveolar carcinoma (n=7) showed type 2B pattern with rapid initial enhancement with plateau with washout. Squamous cell cancers (n=10) showed type 3 curve with rapid washout. Adenocarcinoma (n=3) showed type 3 curve with rapid washout when ROI was placed in periphery and type 2b pattern when ROI was placed in matrix necrosis (area corresponding to bubble lucencies). Microvessel density and permeability indices were calculated for all the tumors and results were computed. Conclusion: kt BLAST-based dynamic contrast-enhanced multphasic lung imaging has an emerging role in lung tumor perfusion and characterization.

B-106

11:15

Quantitative prediction of post-surgical recurrence in non-small cell lung cancer patients with integrated whole-body FDG-PET-CT D. Takenaka, Y. Ohno, Y. Onishi, H. Koyama, M. Nogami, S. Matsumoto, M. Fujii, K. Sugimura; Kobe/JP ([email protected]) Purpose: To determine the capability of whole-body FDG-PET-CT for quantitative prediction of post-surgical recurrence in non-small cell lung cancer (NSCLC) patients. Methods and Materials: 53 consecutive NSCLC patients (32 men, 21 women) prospectively underwent integrated whole-body PET-CT, radiological examination for lung cancer staging, tumor marker measurements, surgical resection, and pathological and follow-up examinations. According to the results of biopsy and follow-up examination, all patients were divided into two groups: non-recurrence (n=46) and recurrence (n=7). Then, maximum of standard uptake value (SUVmax) within primary tumor, tumor markers and tumor size were compared between the two groups by using Student’s t-test. Radiological subtypes on CT, histological subtypes, and clinical and pathological TNM stages were also compared by using chi-square test. To determine the capability for prediction of post-surgical recurrence, SUVmax and other parameters having signiﬁcant difference between both groups were assessed by using logistic regression analysis. Finally, ROC-based positive test was performed to determine the feasible threshold value for distinguishing recurrence from non-recurrence group, and the results were tested on a per patient basis. Results: SUVmax, CEA and pathological N-stage had signiﬁcant differences between the two groups (SUVmax: p 0.0001, CEA: p=0.04 and pathological N stage: p=0.01). SUVmax was determined as the signiﬁcant predictor (p=0.01, odds ratio=1.5). When 4.0 was adopted as threshold value for SUVmax, recurrence group was completely distinguished from non-recurrence group (sensitivity: 100.0 [7/7], speciﬁcity: 63.0 [29/46] and accuracy: 67.9% [36/53]). Conclusion: Whole-body PET-CT has a potential in predicting post-surgical recurrence in NSCLC patients quantitatively.

B-107

11:24

Lung tumoral angiogenesis at dual-source CT: Technical approach with pathologic-CT correlations N. Tacelli1, M. Rémy-Jardin1, A. Scherpereel1, E. Klotz2, M.-C. Copin1, J. Rémy1; 1 Lille/FR, 2Forchheim/DE ([email protected]) Purpose: To evaluate tumoral perfusion with dual-source CT in patients with nonsmall cell lung carcinoma. Methods and Materials: Nine consecutive patients with histologically proved nonsmall cell lung cancer (adenocarcinoma: n=7; squamous cell carcinoma: n=2) were prospectively enrolled in this study aimed at evaluating the technical conditions for CT analysis of tumoral perfusion. Dual-source 64-slice MDCT studies were obtained prior to antiangiogenic chemotherapy (n=6) and lung surgery (n=3), the latter group enabling correlations between CT and histologic ﬁndings. Institutional review board approval and informed consent were obtained for this study. Results: Each perfusion study consisted of a noncontrast scan followed by nine successive contrast-enhanced studies over the whole tumor, obtained every 6 seconds (80 kV; 150 mAs), with administration of 108 mL of a 30% contrast agent at a decreasing rate over 90 seconds. Whole tumor coverage (mean height: 5.5 cm) was possible in all patients with generation of colored parametric maps of permeability and relative blood volume using Patlak analysis. Areas with high proportion of tumoral cells (tumor/stroma ratio: 90/10) showed the highest relative blood volume (mean value: 18.62 mL/100 mL) with a mean capillary permeability of 11.16 mL/100 mL/min. The existence of intratumoral necrosis was depicted as areas of lower blood volume (mean value: 13.2 mL/100 mL) and lower permeability (mean value: 2.87 mL/100 mL/min) compared to the cellular zones. The mean DLP value of perfusion studies was 628 mGy cm. Conclusion: Whole tumor perfusion analysis is technically accessible to dualsource CT which could be used in monitoring of antiangiogenesis drugs.

Scientific Sessions B-108

11:33

Quantitative assessment of perfusion of esophageal squamous cell carcinoma using 64-slice MDCT: Initial observation Z.G. Yang, T.-W. Chen, Y. Li, Z.-l. Li, J. Yao, J.-Y. Sun; Chengdu, Sichuan/CN ([email protected])

B-109

11:42

Quantitative CT perfusion measurements of solitary pulmonary nodules: Initial experience M.A. Mazzei1, M. Scialpi2, M. Carcano3, M. Vincenti1, A. Mannini1, S.F. Carbone1, L. Volterrani1; 1Siena/IT, 2Perugia/IT, 3Milan/IT ([email protected]) Purpose: To investigate the clinical application value of CT perfusion in order to differentiate between benign and malignant solitary pulmonary nodules (SPNs). Methods and Materials: Thirty-two patients with SPNs, diameter b3 cm, 17 malignant (nine peripheral lung cancer and eight metastases) and 15 benign, including six inﬂammatory nodules, underwent CT perfusion with 64-detector scanner (VCT, GE Healthcare) (contrast material 100 ml, 4 ml/sec). All 32 perfusion studies were analyzed by using commercially available software (CT Perfusion 3 - Body tumour protocol; GE Healthcare Technologies). The parameters of CT perfusion including blood volume (BV), blood ﬂow (BF), mean transit time (MTT), permeability surface (PS) and PH-ratio (peak height enhanced ratio of SPN to aorta) were measured. Results: There were signiﬁcant differences in PS (p 0.01) between peripheral lung cancer and metastasis, benign and inﬂammatory nodules. The PH-ratio of the malignant nodules were signiﬁcantly higher than that of benign nodules (p 0.01). For the diagnosis of peripheral lung cancer, when PSr10 ml/100 g/min was set as the diagnostic threshold, the sensitivity, speciﬁcity, positive (PPV) and negative predict values (NPV) were 66.6, 93.3, 85.7, and 82.3%, respectively; when PH-ratio r15% was set as the diagnostic threshold the sensitivity, speciﬁcity, PPV and NPV were 77.7, 95.6, 87.5, and 91.6%, respectively; when PSr10 ml/100 g/min and PH-ratio r15% were associated the sensitivity, speciﬁcity, PPV and NPV values were 77.7, 100, 100, and 92%, respectively. Conclusion: CT perfusion can represent a method of quantitative information about blood ﬂow patterns for differentiating between benign and malignant SPNs.

B-110

11:51

First-pass CT perfusion imaging of solitary pulmonary nodules: Comparison of perfusion parameters of malignant and benign peripheral lung lesions Y. Li, Z.-G. Yang, T.-W. Chen, J.-Q. Yu, Y.-K. Guo, Z.-L. Lin; Chengdu/CN ([email protected]) Purpose: To determine whether ﬁrst-pass perfusion CT imaging could reveal signiﬁcant differences between malignant and benign solitary pulmonary nodules (SPNs). Methods and Materials: 77 consecutive patients with SPNs including malignant (n=46), benign (n=22), and active infection groups (n=9) were examined with 64detector row CT. First-pass perfusion study comprised twelve series of scans over

10:30 - 12:00

Room R2

Cardiac

SS 203 Advances in cardiac CT Moderators: G. Hadjidekov; Sofia/BG C.S. White; Baltimore, MD/US

B-111

10:30

What are the realistic radiation risks to patients undergoing cardiac CT? W. Huda, U.J. Schoepf, J.A. Abro, J.G. Ravenel, P. Costello, D.G. Frey; Charleston, SC/US ([email protected]) Purpose: Previous studies on radiation risk from cardiac CT are based on presumptive scan protocols and not on actual patient cohorts. We aimed to estimate the realistic risk of radiation-induced cancer based on an actual, clinical patient population undergoing cardiac CT. Methods and Materials: We studied 104 consecutive patients undergoing 64-slice cardiac CT (Siemens Sensation Cardiac). Data were obtained on patient demographics (sex, age, weight) and CT technique (kV, mAs, scan length, DLP). Organ dose estimates were obtained using the ImPACT dosimetry spreadsheet. Organ doses were converted into patient risk of developing radiation-induced fatal and non-fatal cancer using the BEIR VII approach. Patient cancer risks were adjusted taking into account patient sex, age and weight. Results: The majority of patients (62%) were males. Mean age of the overall patient population was 58p12 years, with a mean weight of 92p21 kg. Mean DLP was 1124p200 mGy cm. In normal sized patients, the organs receiving the highest radiation doses were the lung (75 mGy) and the female breasts (89 mGy). The average patient radiation risk for a radiation-induced cancer was 0.11%, of which 90% would be fatal. Approximately 75% of the radiation risk arises from the irradiation of the lung. Conclusion: In an actual clinical population undergoing cardiac CT, patients are predominantly male and in their ﬁfties-seventies. In such a patient population, the realistic risk of radiation induced cancer is 0.11%, which is substantially lower than previously reported for general populations.

B-112

10:39

Usefulness of dual source CT in preoperative evaluation of patients with aortic regurgitation H.-J. Hwang, J.-W. Kang, J. Seo, E. Chae, Y. Jang, T. Lim; Seoul/KR ([email protected]) Purpose: To evaluate the clinical value of dual source computed tomography (DSCT) for preoperative evaluation of coronary artery stenosis and severity of aortic valve regurgitation (AR). Methods and Materials: A total of 39 consecutive surgical candidates (28 men, 11 women: mean age, 60p12.77) for AR were studied by DSCT (Somatom Deﬁnition, Siemens) without heart rate control. This study was approved by our institutional review board (IRB). The diagnostic accuracy of DSCT coronary angiography in detection of signiﬁcant coronary artery stenosis ( 50%) was evaluated in seg-

A

B

C

D

E

F

G

165

Friday

Purpose: To investigate tumor perfusion of esophageal squamous cell carcinoma with ﬁrst-pass perfusion technique at 64-slice MDCT and to determine the threshold value of perfusion values for the discrimination of microcirculation of the tumor from that of normal esophagus. Methods and Materials: Forty-one consecutive patients with pathologically conﬁrmed esophageal squamous cell carcinomas were enrolled in our study as test group. Forty patients with normal esophagus served as controls. All patients underwent volume-based perfusion scan with 64-slice MDCT. Perfusion parameters including perfusion (PF), peak enhanced (PE), blood volume (BV) and time to peak (TTP) were measured using perfusion software. The differences in perfusion parameters between test and control groups, and between subgroups were compared with Student’s t test. Cutoff values were obtained statistically to discriminate the microcirculation of esophageal squamous cell carcinomas from that of normal esophagus. Results: BV and TTP were signiﬁcantly higher in patients with esophageal squamous cell carcinoma than in control group (p 0.05), whereas no statistical differences in PF and PE were found between test and control groups (p 0.05). There were no statistical differences in all parameters between subgroups of different pathological grades, and between subgroups with and without lymph-node metastases (p 0.05). For the discrimination of microcirculation of esophageal squamous cell carcinoma from that of normal esophagus, a threshold value of 6.65 ml/100 g was determined and achieved a sensitivity of 95.1% and speciﬁcity of 90.0%. Conclusion: First pass perfusion technique at 64-slice MDCT could provide valuable information to assess tumor microcirculation of esophageal squamous cell carcinoma.

60 seconds following a 50-ml intravenous bolus of contrast medium at 6-7 ml/sec. Perfusion, peak enhancement intensity (PEI), time to peak (TTP), and blood volume (BV) were compared by means of one-way ANOVA among different groups. Receiver operating characteristic (ROC) analysis was performed to evaluate the usefulness of perfusion parameters as markers for differentiating malignant from benign nodules. Results: Mean perfusion, PEI and BV for the benign SPNs were signiﬁcantly lower than those for the malignant SPNs (p 0.05) and active infections (p 0.05), but the differences were not statistically signiﬁcant between malignant tumors and active infections (p 0.05). ROC curve analysis showed that SPNs with a perfusion greater than 30.7 ml/min/ml, PEI higher than 24.3 HU or BV larger than 9.2 ml/100 g were more likely to be malignant. Sensitivity, speciﬁcity, positive predictive value, negative predictive value and accuracy in distinguishing malignant and benign modules were 87.0, 86.4, 93.9, 76.0 and 86.8% for perfusion, respectively, 87.0, 95.5, 97.6, 80.8 and 91.2% for PEI, respectively, and 95.7, 86.4, 93.7, 95.0 and 92.6% for BV, respectively. Conclusion: First-pass perfusion parameters could permit prediction of potential malignancy with high sensitivities, speciﬁcities, positive predictive values, negative predictive values and accuracies.

Scientific Sessions ment-by-segment comparison with conventional coronary angiography, which was performed in all patients. In addition, the correlation between central valvular leakage area (aortic regurgitation area) during diastole planimetrically measured with DSCT (ARACT) and the grade of AR severity as measured by transthoracic echocardiography (TTE) were evaluated by Spearman's rank coefﬁcient. Results: The average heart rate during scanning was 69 beats per minute (bpm)p17.35 (range, 47-133 bpm). Overall, 496 (99%) of 507 coronary artery segments were classiﬁed as assessable. DSCT yielded a sensitivity of 93.8%, speciﬁcity of 99.6%, positive predictive value of 88.2%, and negative predictive value of 99.8% in detecting signiﬁcant coronary artery stenosis. In subanalysis, both heart-rate subgroups ( 70 bpm and r70 bpm) yielded similar diagnostic accuracy. The ARACT showed signiﬁcant correlation with the grade of AR as measured by TTE (r=0.738, p=0.0001). Conclusion: With improved temporal resolution, DSCT may be a useful tool for comprehensive evaluation in patients with aortic valvular disease without heart rate control.

B-113

10:48

Coronary CT angiography: Quantitative assessment of myocardial perfusion using test bolus data - Initial experience A. Groves1, V.J. Goh2, S. Rajasekharan1, I. Kayani1, R. Endozo1, J. Dickson1, L. Menezes1, M. Shastri1, S. Habib3, P. Ell1, B. Hutton1; 1London/UK, 2Northwood/UK, 3 Nottingham/UK Purpose: Current CT technology has evolved such that combined assessment of coronary artery anatomy, myocardial perfusion, and left ventricular function is potentially achievable in a single examination in clinical practice. Our aim was to assess if quantiﬁcation of myocardial perfusion is feasible from the test bolus performed during routine coronary CT angiography. Methods and Materials: Following institutional review board approval and informed consent, 14 patients (mean age 66.5 years; eight male; six female) with suspected coronary artery disease underwent a retrospectively EKG-gated dynamic test bolus acquisition at the mid left ventricular level (120 kV; 35 mAs; 5 mm coverage) prior to combined coronary CT angiography and cardiac 82Rubidium perfusion PET. Mean myocardial perfusion at rest was assessed for two cardiac phases: 40-50% (endsystole) and 75-100% (end-diastole) using CT software based on the slope method: maximum-slopem/peaka by two observers to allow assessment of observer agreement. CT coronary angiography and 82Rubidium perfusion PET were interpreted by two experienced radiologists, and nuclear medicine physicians. Results: Mean resting myocardial perfusion of 0.89 ml/min/g (0.62-1.47p0.27) and 0.93 ml/min/g (0.61-1.36p0.21) at end-systole and end-diastole, respectively, were within the expected range for normal myocardium. Inter-observer agreement was good with a coefﬁcient of variation of 11.6-13.9%. Normal resting perfusion at the level of the perfusion study was conﬁrmed by 82Rubidium perfusion PET. Nine of the 14 patients studied had signiﬁcant segmental stenoses on coronary CT angiography conﬁrmed by 82Rubidium perfusion PET. Conclusion: Quantiﬁcation of myocardial perfusion from a routine coronary CT angiography test bolus is possible.

B-114

10:57

Coronary MSCT angiography: Effect of experience and training on reader performance F. Pugliese, K. Gruszczynska, M.G.M. Hunink, F. Alberghina, R. Malago’, N. Van Pelt, N.R. Mollet, P.J. de Feyter, G.P. Krestin; Rotterdam/NL ([email protected]) Purpose: To prospectively evaluate the effects of radiologist experience and increasing exposure to coronary MSCT angiography on reader performance. Methods and Materials: Four radiologists independently analyzed 50 coronary MSCT data sets and their diagnostic performances were compared over a period of 1 year after their appointment within a dedicated cardiac MSCT unit. Conventional angiography was referred as reference standard technique. Readers had no prior experience of coronary MSCT angiography and before starting the study they received feedback and training by three independent experienced radiologists in the interpretation of 15 mentored cases. They also received lectures on cardiac anatomy and MSCT angiography technique. For the duration of the study, the radiologists were involved in the acquisition and interpretation of coronary MSCT studies (12-15 per week). Results: At the study outset, average sensitivity was 53% (CI: 48-57), speciﬁcity 85% (CI: 84-86) and NPV 92% (CI: 91-93). After six months, average sensitivity was 61% (CI: 56-66), speciﬁcity 85% (CI: 84-87) and NPV 94% (CI: 92-94). After one year, average sensitivity was 70% (CI: 66-74), speciﬁcity 89% (CI: 88-90) and NPV 95% (CI: 94-96). Sensitivity, speciﬁcity and NPV after one year were signiﬁcantly

166

A

B

C

D

E

F

G

better than those after six months (P-value 0.05). The diagnostic odds ratios were 6.4 at the study outset, 9.1 after six months and 18.7 after one year. Conclusion: The interpretative learning curve in cardiac CT is not rapid. Competence cannot be assumed after 6-month training.

B-115

11:06

Non-invasive visualization of the coronary venous system using 64-slice spiral computed tomography and a comparison with conventional invasive angiography Y. Hou1, Q.Y. Guo1, Y.X. Sun1, Y. Yue1, J.H. Chen2; 1Shenyang/CN, 2BeiJing/CN ([email protected]) Purpose: To compare 64-slice CT and conventional invasive angiography in depicting the cardiac venous system especially on the coronary sinus (CS), its tributaries and the arterio-venous relationship. Methods and Materials: Both 64-slice CT (SOMATOM Sensation 64, Siemens, Germany) and invasive coronary venography (CVG) were performed on 37 patients who were referred for ablation of posteroseptal and left posterior accessory pathways (CSAPs) and nine patients who were referred for percutaneous coronary intervention (PCI). Three-dimensional (3D) post-processing was done to evaluate morphology, course, diameter of the coronary veins and the arterio-venous relationship. CVG was taken as gold standard. The diameter of coronary sinus ostium (CSO), middle cardiac vein ostium (MCVO) and posterior vein of left ventricle ostium (PVLVO) were measured on CT and CVG. The correlation between the data acquired on MSCT and CVG were evaluated. Results: From the 3D images of 64-slice CT, all of the CS and its major branches could be seen, small cardiac veins (SCV) were found in 12 patients, coronary sinus diverticulum was shown in one patient and persistent left superior venae cavae were shown in three patients. The course of vessels and arterio-venous relationship in MSCT were in accordance with CVG. The diameter of CSO, MCVO, PVLVO with CT are 16.0p4.6, 5.8p1.6, and 3.5p1.7 mm, respectively. The linear correlations between these two methods are 0.89, 0.83, 0.95, respectively (p 0.001). Conclusion: 64-slice CT can accurately depict the anatomy of the cardiac venous system as needed for mapping, pacing and radiofrequency ablation.

B-116

11:15

Computerized patient-based dosing of contrast media delivery at dualsource CT coronary angiography H. Lee1, P. Suranyi1, J.F. Kalafut2, C.A. Kemper2, J.A. Abro1, P.G. Kogos1, P. Costello1, U.J. Schoepf1; 1Charleston, SC/US, 2Pittsburgh, PA/US ([email protected]) Purpose: To determine the effectiveness of a computer-algorithm-based, patientspeciﬁc contrast media injection protocol integrated with the injector for cardiac dual-source CT (DSCT). Methods and Materials: In 40 consecutive patients undergoing cardiac DSCT, a computerized algorithm was prospectively evaluated for determining individualized contrast (370 mg I/ml) volume and injection parameters of a triphasic (contrast, contrast/saline and saline phase) injection protocol. The algorithm is based on scan duration, patient-speciﬁc variables and the time-attenuation response in the aorta following a test bolus. A control group of another 40 consecutive patients who were injected with our routine contrast protocol were retrospectively analyzed to compare the level and homogeneity of enhancement within the aorta, coronary arteries, myocardium and ventricles. The quality of diagnostic display of cardiac anatomic landmarks was rated. Results: Mean coronary attenuation ranged from 359.5p17.7 to 450.6p19.9 HU, and from 297.3p11.1 to 429.8p10.9 HU in the study and control groups, respectively. Mean attenuation in the distal LAD, mid and distal LCx, and the right ventricle were signiﬁcantly (p 0.05) higher in the study group than in the control group. Intra-subject variability of vascular attenuation was signiﬁcantly (p 0.002) lower in the study group (SD=49p19.5) than in the control group (SD=63.2p22.1). The diagnostic display of right heart structures (papillary muscles, pulmonary valve, ventricular myocardium), left main, LAD and proximal right coronary arteries were rated signiﬁcantly (p 0.05) higher in the study group than in the control group. Conclusion: This patient-speciﬁc contrast protocol enables higher and more uniform coronary enhancement and improved diagnostic display of the heart at cardiac DSCT.

Scientific Sessions B-117

11:24

Comparison of X-ray radiation exposure during digital subtraction angiography (DSA) and multislice CT angiography (CTA) B. Liu; Beijing/CN ([email protected])

B-118

11:33

Vessel wall hemodynamics and coronary artery plaque formation: Morphological analysis with coronary CT angiography B. Enrico1, P. Suranyi1, C. Thilo1, L. Bonomo2, P. Costello1, U.J. Schoepf1; 1 Charleston, SC/US, 2Rome/IT ([email protected]) Purpose: To investigate the spatial distribution and composition of atherosclerotic plaques using coronary CT angiography (cCTA) to better understand the mechanism of plaque formation and its relationship to vascular hemodynamics. Methods and Materials: Seventy-three patients underwent 64-slice cCTA. Data were analyzed using transverse sections and curved multiplanar reformats. Two radiologists reviewed all cCTAs for plaques and classiﬁed all lesions according to their location, conﬁguration, and composition. Bifurcation lesions were characterized using the Medina classiﬁcation. Eccentric plaques were classiﬁed regarding their location in the vessel wall. Results: A total of 382 plaques were analyzed. 8% were in the LM, LAD: 48%, LCx: 18%, RCA: 26%. 25% of plaques were completely calciﬁed, 72% mixed, and 3% purely non-calciﬁed. 51% of plaques were classiﬁed as bifurcation lesions. The most frequent (40%) Medina pattern was 1.1.0 (lesion starts before and extends beyond bifurcation, sparing sidebranch). 80% of plaques were eccentric. A signiﬁcant (p 0.01) majority (55%) were in the myocardial side of the vessel wall, while 17% were lateral, and 28% epicardial. Of all non-calciﬁed and mixed plaques, 45% (p 0.01) were myocardial, whereas only 14% were lateral, 21% epicardial, and 20% concentric. 45% of eccentric LM lesions were lateral, while 48% of LAD plaques and 49% of RCA plaques were myocardial. Conclusion: Coronary segments with turbulent ﬂow (bifurcations) and low hemodynamic shear stress (myocardial side) are areas of predilection for plaque deposition, especially those with non-calciﬁed plaque components. Analysis of atherosclerotic disease patterns based on non-invasive cCTA may enhance our understanding of coronary plaque pathogenesis.

B-119

11:42

Cardiac-output based individualized contrast media injection protocol to minimize contrast volume at coronary dual-source CT angiography P. Suranyi1, J.M. Kerl1, H. Lee1, J.F. Kalafut2, C.A. Kemper2, J.A. Abro1, G.E. Gilbert1, R.E. Carter1, P. Costello1, U.J. Schoepf1; 1Charleston, SC/US, 2 Pittsburgh, PA/US ([email protected]) Purpose: To evaluate a cardiac-output based, individualized contrast media injection algorithm to achieve a predetermined level of coronary enhancement and minimize contrast volume at coronary dual source CT (DSCT) angiography. Methods and Materials: 70 patients underwent coronary DSCT angiography. A computerized algorithm was prospectively used to individualize contrast (300 mg I/ml) volume and injection parameters. The algorithm is based on patient-specifc variables and the time-attenuation response in the pulmonary artery and the aorta

B-120

11:51

Multi-detector CT coronary angiography: Sublingual nitroglycerine improves image quality signiﬁcantly due to peripheral coronary vasodilatation O. Klass, S. Mutlu, K. Hohl, S. Feuerlein, M. Jeltsch, H.J. Brambs, M.H.K. Hoffmann; Ulm/DE ([email protected]) Purpose: In this study we investigated the inﬂuence of sublingual glycerol trinitrate on the peripheral diameter, contrast enhancement and image quality of coronary arteries during CTCA. Methods and Materials: Thirty consecutive patients with nitrate application were matched to 30 patients without nitrate. The diameters of the left anterior descending coronary artery (LAD), the left circumﬂex coronary artery (LCX) and the right coronary artery (RCA) were measured at 1, 4, and 8 cm lengths of each vessel and the contrast agent attenuation along the LAD. Vessel diameters and contrast attenuation at 4 and 8 cm were referenced against the values at 1 cm and processed as percentage reduction. Image quality of the posterior descending artery (PDA) was assessed subjectively by two independent observers using a scale from 1 to 5. Results: The percentage of peripheral vessel diameter reduction as well as the peripheral attenuation of contrast agent for all measured coronary arteries was signiﬁcantly smaller in the group with nitrate administration (p 0.001). The image quality of the posterior descending artery (PDA) was signiﬁcantly higher in the group with nitrate. Conclusion: Sublingual administration of nitrate prior to CTCA results in improved diagnostic image quality due to a signiﬁcant dilatation and improved contrast agent density of the peripheral vessels.

10:30 - 12:00

Room R3

Interventional Radiology

SS 209 Vascular interventions (1) Moderators: B. Nemes; Budapest/HU S. Soimakallio; Tampere/FI

B-121

10:30

10-years follow-up of endovascular aneurysm treatment with talent stentgrafts M.B. Pitton, T. Scheschkowski, S.C.A. Herber, R. Lachmann, A. Neufang, W. Schmiedt, C. Düber; Mainz/DE ([email protected]) Purpose: To evaluate clinical results, complications, and secondary interventions during long-term follow-up. Methods and Materials: From 1997 to March 2007, 128 patients (12 female,116 male, age 72.6p7.2 years) with abdominal aortic aneurysms were included. In 54 patients, all tributaries of the aneurysm sac were occluded prior to stentgrafting (group I). Group II included cases with incomplete or without coil embolisation (at least one patent vessel in the aneurysm, n=74). Follow-up included contrast enhanced MDCT, MRI and radiographs at three, six, 12 months, and annually. Results: All implantations were successful without primary conversion surgery. Follow-up is 30.4p28.6 months (range 3-108). 26 patients died not related to

A

B

C

D

E

F

G

167

Friday

Purpose: To compare radiation risk during digital subtraction angiography (DSA) and CT angiography (CTA). Methods and Materials: Eighty patients randomly divided into two groups of 40 patients each were evaluated for coronary atherosclerotic heart disease with 64-slice spiral CT angiography and DSA, respectively. For CTA group we divided patients into two subgroups: regular mode and ECG modulation mode. Matrix of thermoluminescent dosimeters (TLDs) were placed on patients’ skin surface during each examination to access peak skin dose (PSD). A male ART phantom was equipped with TLDs in six different positions to assess achievable organ dose. Effective dose was calculated by commercially available computer programs. Results: For DSA, PSD was 64 mGy. Phantom organ doses for lung, heart, stomach, liver, kidney and brain were 11.07, 5.87, 1.42, 2.11, 6.76 and 0.001 mGy, respectively, and effective dose was 4 mSv. For regular mode CTA, PSD was 74.5 mGy, and CTDIvol was 37.9 mGy. Phantom organs dose for lung, heart, stomach, liver, kidney and brain were 58, 64.2, 6.6, 9.9, 2.7, 0.072 mGy, respectively, and effective dose was 16 mSv. For ECG modulation mode, PSD was 50.9 mGy, and CTDIvol was 26.3 mGy. Conclusion: According to measured PSD value, deterministic radiation risk for regular mode CTA was slightly higher than DSA. However, for stochastic radiation risk which can be indicated by organ dose and effective dose, regular mode CTA was much higher than DSA. Comparing with regular mode, ECG modulation mode can effectively reduce both stochastic and deterministic radiation risks.

following a test bolus injection as an approximation of cardiac-output. A targeted attenuation level of 250 HU was chosen. The eventual contrast attenuation in the aorta and the coronaries (proximal, mid and distal) was measured. The actually injected contrast volume was compared to the volume that would have been injected using the routine scan-duration-based injection protocol. Results: All studies were diagnostic. Mean (pSD) contrast attenuation ranged from 237p56 (distal LAD) to 277p48 HU (proximal RCA). Attenuation in the LM (272p52), proximal LAD (270p49) and LCx (267p42), and throughout the RCA (276p48, 268p56, 274p56) were signiﬁcantly (p 0.05) higher than 250 HU. Attenuation in the distal LAD (237p56) and LCx (241p36) were not signiﬁcantly lower than the targeted 250 HU. Attenuation in the aorta, mid LAD, and mid LCx were not signiﬁcantly higher than 250 HU. Mean overall contrast savings was 19.7p24.1 mL. There was a signiﬁcant negative correlation (R=-0.67, p 0.05) between cardiacoutput and contrast savings. Conclusion: Cardiac-output-based, individually computed injection protocols can minimize contrast volume, while consistently achieving a predetermined target enhancement in the coronary arteries. Contrast savings are greatest in patients with lower cardiac-output.

Scientific Sessions the aneurysm. 20 reinterventions were performed in 16 patients (12.5%), 10 percutaneous (distal graft extension (n=2), secondary embolisation of collaterals (n=8)) and 10 operative reinterventions (proximal/distal graft extension (n=4), new stentgrafts (n=4), femoro-femoral crossover bypasses (n= 1), thrombectomy (n=3), augmentation of stentgraft limb with selfexpandable stent (n=1), and combinations of these). Four late conversion operations were performed because of signiﬁcant graft migration (n=1), inﬂammatory aneurysm (n=1) and Type II endoleaks with increasing aneurysm diameters (n=2). MRI and CT demonstrated considerably more and greater endoleaks in group II. Aneurysm shrinkage (transversal diameter) was better in group.i. compared to group II ($ -6.8p10.7 mm [range -50.8 to +6.7 mm] versus -4.4p8.1 mm [range -28.0 to +10.5 mm]. Conclusion: Endovascular aneurysm treatment with Talent stentgrafts demonstrated acceptable long-term results with moderate secondary intervention rates. Preinterventional coil occlusion of all aortic tributaries before stentgrafting reduced the overall endoleak rate and provided an improved aneurysm shrinkage.

B-122

10:39

Long-term follow-up after endovascular treatment of acute aortic emergencies M.B. Pitton, W. Schmiedt, S.C.A. Herber, A. Neufang, B. Dorweiler, C. Düber; Mainz/DE ([email protected]) Purpose: Long-term outcome of endovascular emergency treatment of acute thoracic aortic diseases. Methods and Materials: 37 patients with acute diseases of the thoracic aorta were treated by endovascular stentgrafts: traumatic aortic ruptures (n=9), aortobronchial ﬁstulas due to penetrating ulcer or hematothorax (n=6), acute Type-B dissections with aortic wall hematoma, penetration, or ischemia (n=13), and symptomatic aneurysms of the thoracic aorta (n=9). Besides femoral cut down, two procedures required aorto-femoral bypass grafting prior to stentgrafting for introduction of the stentgraft. Results: 46 stentgrafts were implanted. Stentgraft extension was necessary in seven cases (three primary due to distal migration or aneurysm length; four secondary due to new aortic ulcers or recurrent aorto-bronchial ﬁstulas). 30-day mortality rate is 8.1% (3 of 37), overall follow-up is 29.9p36.6 months (range 0-139 months), and mean survival is 87.9p13.0 months. Traumatic ruptures demonstrated immediate sealing of bleeding and no complications during follow-up. Aorto-bronchial ﬁstulas demonstrated satisfactory follow-up despite necessity of reintervention and graft extension in 3 of 6 cases. In Type-B dissection, two patients died due to mesenteric ischemia despite sufﬁcient restitution of mesenteric blood ﬂow. Moreover, two neurological complications, one paraplegia, one cerebral ischemia, one penetrating ulcer, and one persistent ischemia of the kidney occured. In aneurysms, one cardiac fatality occured (6 h) and ﬁve cases presented endoleaks with aneurysm increase in one. Conclusion: Endovascular treatment is safe and effective for emergency treatment of life-threatening acute thoracic aortic syndromes. Results are encouraging, particularly for traumatic aortic ruptures. Regular follow-up is still mandatory to perform appropriate additional corrections at times.

B-123

10:48

Endovascular treatment of ruptured abdominal aortic aneurysm: 3-years experience G. Guzzardi, R. Fossaceca, P. Brustia, A. Carriero; Novara/IT ([email protected]) Purpose: To report our early experience with endovascular treatment of patients with ruptured abdominal aortic aneurysms (RAAAs). Methods and Materials: Between January 2005 and August 2007, 85 patients with an RAAA presented to our unit. All hemodynamically unstable patients (systolic blood pressure 80 mmHg) were transferred directly to the operating room for open repair. The hemodynamically stable patients underwent computed tomography angiography (CTA) to evaluate anatomic suitability for emergency endovascular treatment (EVAR). 20 patients (19 M; 1 F; mean age: 78.6) underwent EVAR procedure. A total of 18 bifurcated and two aortouni-iliac stent grafts were implanted. The mean dose of contrast medium was 140 ml. Results: Overall successful graft deployment was 100%. The 30-day mortality was 15% (3/20 patients); two patients died from cardiac causes and one patient developed a multiple organ failure. There were no post-operative reinterventions. Patients underwent CTA scan at 1, 6, 12 months and yearly (Eurostar protocol). The mean follow-up is 188.6 days (range: 90-510 days); the overall mortality rate was 25% (5/20). We observed no type I and III endoleaks and two cases of type II endoleak (10%). Conclusion: Endovascular treatment of RAAAs is feasible and the early experience is promising. Large randomized trials are required to conﬁrm the efﬁcacy of the procedure.

168

A

B

C

D

E

F

G

B-124

10:57

Prospective, intraindividual comparison of MDCT and MRI for endoleak detection after endovascular repair of abdominal aortic aneurysms R. Wyttenbach, M. Oberson, A. Fogliata, M. Alerci; Bellinzona/CH ([email protected]) Purpose: To compare MDCT and MRI in the assessment of endoleaks (EL) after endovascular repair of abdominal aortic aneurisms (EVAR). Methods and Materials: Forty-three patients, 42 men and one woman (mean age 72.3 years), with previous EVAR underwent both CT (16 slice MDCT; unenhanced and biphasic contrast-enhanced) and MRI (1.5 Tesla; 2D T1 FFE unenhanced and contrast-enhanced; 3D triphasic contrast-enhanced MRA) examination within one week for endoleak detection (mean time after EVAR 34 months, range 2-84). MRI was performed by using a high relaxivity contrast medium (0.15 mmol/kg at 2.5 ml/sec Gd-BOPTA, Multihance®). For MDCT 120 ml of Ioversol at 4 ml/sec (Optiray® 350) was used. Two blinded, independent observers evaluated MRI and MDCT separately. Consensus reading of CT and MR studies was deﬁned as the standard of reference. Sensitivity, speciﬁcity and accuracy were calculated and Cohen’s k statistics were used to estimate agreement between readers. Results: Twenty-one endoleaks were detected in 18 patients (1 type I, 10 type II and 10 indeﬁnite EL). Sensitivity, speciﬁcity and accuracy for endoleak detection were 100, 96 and 98%, respectively, for reader 1 (90, 81, 85% for reader 2) for MRI and 48, 96 and 74% for both readers for MDCT. Interobserver agreement was excellent for CT (k=0.93) and good for MRI (k=0.75). Conclusion: MR imaging with the use of a high relaxivity contrast agent is signiﬁcantly superior in the detection of endoleaks after EVAR compared to MDCT. MRI may therefore become the preferred technique to follow-up patients after endovascular aneurysm repair.

B-125

11:06

Trans-caval endoleak embolization technique R. Gandini, L. Boi, T. Erdembileg, L. Di Vito, M. Di Primio, G. Angelopoulos, G. Simonetti; Rome/IT ([email protected]) Purpose: To investigate the feasibility, efﬁcacy of alternative technique for treatment of type II endoleak after endovascular repair (EVAR), through transfemoral-caval approach of abdominal aortic aneurysm (AAA) with adhesion to vena cava. Methods and Materials: Eleven-years single centre experience in treatment of type II endoleak after EVAR was reviewed; the data of the 48 treated type II endoleaks were examined to establish the methods of treatment. Among these patients, in the last year, nine patients were treated with direct transfemoral-caval puncture and embolization inside the aneurysm sac. Embolization of endoleak was performed using a combination of coils and N-butyl-cyanoAcrylate (Glubran 2). Coils range in number from 10 to 21 and mean size of 9x60 mm and maximum 1 ml of Glubran 2 mixed with 2 ml of lipiodol was enough to achieved complete occlusion of endoleak. Results: Population of study is nine patients (six men, three women; mean age 76.3 years, range 69-85). The mean time of follow-up was 10 months. Technical success after endoleak treatment using trans-caval endoleak embolization technique (TCEE) was achieved in all nine cases (100%). The follow-up at 10 months has not shown aneurysm-related death or diameter increase of aneurysmal sac and no endoleak recurrence. Conclusion: TCEE represents a new technique of type II endoleak treatment associated with aneurysm enlargement which seems to be safe, feasible and can offer new possibilities for the treatment of this complication.

B-126

11:15

Vessel lumen diameter measurements and stenosis quantiﬁcations of the superﬁcial femoral artery prior to interventional treatment: Comparison of angiography-based algorithms and visual estimation with measurements by intravascular ultrasound in vivo M. Treitl, J. Rieger, S. Wirth, M.F. Reiser; Munich/DE ([email protected]) Purpose: To compare interreader and intermodality differences of vessel lumen diameter measurements and stenosis quantiﬁcations in the superﬁcial femoral artery (SFA) by angiography-based algorithms like catheter tip calibration, caliper calibration or visual estimation by the interventionalist with measurements by intravascular ultrasound (IVUS). Evaluation of their impact on balloon size determination. Methods and Materials: 30 consecutive patients with untreated atherosclerotic SFA stenosis underwent digital subtraction angiography (DSA) and IVUS prior to interventional treatment. All examinations were randomly presented twice to two experienced radiologists. Measurements of native vessel lumen diameter were

Scientific Sessions

B-127

11:24

Endovascular treatment of critical limb ischemia in a diabetic population T. Zander, M. Rabellino, S. Baldi, M. Basaury, I. Zerolo, M. Maynar; Santa Cruz de Tenerife/ES ([email protected]) Purpose: To evaluate the outcomes of percutaneous endovascular procedures in diabetic patients with critical lower extremity ischemia. Methods and Materials: From January 2004 to February 2007, endovascular treatment was performed on 128 limbs in 109 diabetic patients (37 females, 72 males) with limb-threatening ischemia. Of these patients, 22 presented rest pain, while ulcerations were found in 48 and gangrene in 58 patients. On angiography, lesions were detected only in the infrapopliteal sector in 27 limbs, only in the femoropopliteal sector in four and in both sectors in 97 limbs. All limbs were treated using exclusively endovascular techniques. Results: Technical success was obtained in 125 limbs (97.6%). In three patients it was not possible to cross the occluded segment. Eleven patients (12 limbs) were lost in follow-up (1-36 months, mean: 7.9 months). The limb salvage rates were 92.2, 82.8, 78.4 and 76.6% after 1, 3, 6, and 18 months with a major amputation rate of 20%. Of the amputated limbs, 61.5% suffered from gangrene, 30.8% from ulcers and 7.7% from rest pain before the endovascular treatment. Overall mortality was calculated at 4.6% within 30 days of endovascular treatment and 13.7% at 18 months. Conclusion: Endovascular techniques offer an effective and safe therapy modality to avoid limb loss in diabetic patients with critical lower extremity ischemia. For this reason the endovascular treatment should be considered as ﬁrst choice treatment.

B-128

11:33

Short-term intra-arterial administration of taxane for reduction of restenosis after PTA of femoropopliteal arteries: One year follow-up results of the THUNDER trial T. Albrecht1, G. Tepe2, T. Zeller3, S. Heller2, A. Oldenburg1, K.-J. Wolf1, C.D. Claussen2, U. Speck1; 1Berlin/DE, 2Tübingen/DE, 3Bad Krozingen/DE ([email protected]) Purpose: Drug-eluting stents reduce re-stenosis post angioplasty in coronary arteries but initial controlled clinical trials showed little effect in peripheral arteries. The purpose of this prospective blinded multi-centre trial was to assess the effect of two alternative local drug delivery approaches, Paclitaxel coated on angioplasty balloons or dissolved in contrast agent, on re-stenosis after peripheral PTA. Methods and Materials: 154 patients with femoro-popliteal artery stenosis/occlusion were randomized to treatment with Paclitaxel-coated PTA balloons, Paclitaxel admixed to contrast agent or control treatment (standard balloons and contrast medium). Mean lesion length was 6.7p5.0 cm without signiﬁcant differences between sub-cohorts; 36% of lesions included were re-stenoses. The primary endpoint was late lumen loss (LLL) on control DSA performed at six months (quantitative off-site analysis), clinical follow-up was obtained at 12 months. Results: There were no adverse events attributable to Paclitaxel. At six months, the primary endpoint LLL was 1.7p1.8 mm in the control group, compared to 0.4p1.2 mm (p 0.01) for Paclitaxel on balloon and 2.2p1.6 mm (p 0.05) for Paclitaxel in contrast agent. Target lesion revascularization rate (rate of lesions requiring re-intervention due to re-stenosis /-occlusion) at six months was 20/54 (37%) in the control group, 2/48 (4%) in the coated balloon (p 0.01 versus control), and 15/52 (29%) in the contrast agent group. It increased to 26/54 (48%), 5/48 (10%), 18/52 (35%), respectively, at 12 months. Conclusion: Drug-coated balloons have the potential to reduce re-stenosis in femoro-popliteal arteries. Local Paclitaxel administration did not cause adverse events during and up to 12 months after the intervention.

B-129

11:42

Transfemoral and transtibial combined approach in treatment of obstructions extending on popliteal and distal vessels origin M. Stefanini, R. Gandini, V. Pipitone, L. Maresca, L. Di Vito, G. Simonetti; Rome/IT ([email protected]) Purpose: To propose our technique of subintimal recanalization of long femoralpopliteal extending on popliteal and distal vessels’ origin obstruction, by combined trans-femoral and trans-tibial approach. Methods and Materials: In 47 patients with trophic lesions (out of 635 treated for critical limb ischemia), we performed an antegrade trans-femoral and retrograde puncture of ATA (21 cases) and PTA (26 cases) to obtain the recanalization of the occlude vessels. The tibial artery punctures were performed under US guidance as distal as possible. The best distal run-off vessel has been chosen previously by AMR or Doppler US-evaluation. Results: We achieved 95.74% technical success. The subintimal antegrade and retrograde randezvous was achieved by snaring of the wire into subintimal space. Doppler US and TcPO2 follow-up at 12 months showed primary patency of 92%. In 88% of the patients, at 12 months, there has been an improvement within Texas classiﬁcation. All patients had resolution of rest pain and initial healing of the trophic lesions. Conclusion: For the patients who are often poor candidates for surgical by-pass and endovascular failure, we suggest combined antegrade and retrograde recanalization, in case of long obstruction involving the origin of the leg vessels, in order to exactly recanalize the best distal run-off vessel.

B-130

11:51

MR-guided angioplasty of peripheral arterial stenoses with a permeable balloon catheter with delivery of contrast-media doped solutions to the vessel wall and tracking of the distribution G.A. Krombach1, M. Wehner1, A. Perez-Bouza1, L. Kaimann1, S. Kinzel1, T. Plum1, M. Friebe2, R.W. Günther1, C. Hohl1; 1Aachen/DE, 2Castrop-Rauxel/DE ([email protected]) Purpose: To assess the feasibility of MR-guided angioplasty of stenosis in peripheral arteries combined with the delivery of a contrast-medium containing solution and assessment of its distribution within the vessel wall. Methods and Materials: In three pigs, the feasibility of MR-guided atraumatic delivery of a contrast-medium-doped solution (0.05 mmol/ml Gd-DTPA, 3% Evans blue dye) into the vessel wall of the iliac arteries using a balloon catheter (8 mm) with laser perforations was tested. Catheter placement was monitored using realtime imaging with steady-state free precession (SSFP). Prior and after dilatation, the vessel wall was delineated with T1-weighted gradient echo sequence. In ﬁve pigs, bilateral stenoses in the external iliac artery were created by placing a suture around the vessels. The balloon catheter was used to dilate the stenoses and deliver the contrast-medium-doped solution into the vessel wall. In these animals, in addition, contrast enhanced MR-angiography was performed prior and after the interventions to assess the degree of the stenoses. Results: All interventions could successfully be performed under real-time MRguidance. MR angiography performed before and after the intervention demonstrated successful dilatation of nine of the 10 stenoses. In all eight animals, the delivered solution could be depicted in all layers of the arteries, on MR images and post-mortem. Conclusion: MR-guided angioplasty of stenoses with the simultaneous application of a contrast-medium-marked solution to the vessel wall and subsequent tracking of the solution using T1-weighted images is feasible.

A

B

C

D

E

F

G

169

Friday

done with all algorithms and modalities in a standardized and previously determined unaffected segment proximal to the target lesion. Intermodality and interreader differences were calculated. Resulting balloon sizes and quantiﬁcations of diameter and area stenosis were compared. Results: Mean vessel diameter (mm) was 5.72 for IVUS, 6.56 for caliper calibration, 5.99 for catheter tip calibration and 4.67 for visual estimation. Resulting balloon sizes (mm) were 6.0, 7.0, 6.0 and 5.0, respectively. Mean diameter stenosis (%) was 78.8 for IVUS, 81.6 for caliper calibration, 79.7 for catheter calibration and 88.8 for visual estimation. Interreader variability was lowest for IVUS (0.994; p 0.0001) and worst for visual estimation (0.384; p 0.099). Intermodality correlation was best for IVUS and catheter tip calibration (0.881; p 0.0001). Conclusion: Vessel lumen diameter measurements and diameter stenosis quantiﬁcations of the SFA based on DSA equipment calibrated to a catheter tip correlate best with direct intravascular measurements by IVUS. Visual estimation can lead to undersized balloons and overestimation of stenosis grade.

Scientific Sessions 14:00 - 15:30

Room B

SS 310 Spine and shoulder Moderators: M. Abd El Bagi; Ryiadh/SA A. Siemianowicz; Piekary Slaskie/PL

14:00

To evaluate the application of DTI in cervical spondylosis with 3.0 T MR X.S. Meng, J.W. Hou, C.F. Li, Y. Cui, J.H. Wei; Jinan/CN ([email protected]) Purpose: To evaluate the application of the diffusion tensor imaging (DTI) in cervical spondylosis with 3.0 T MR. Methods and Materials: Twenty-one healthy volunteers and 84 patients with cervical spondylosis underwent T2WI and DTI by using a 3.0 T MR scanner of GE Company. Eighty-four patients were divided into four groups according to T2WI. Group A showed that only the dura mater spinalis were compressed, Group B showed that the cervical spinal cords were compressed, but no high signal in it, Group C showed that the cervical spinal cords were compressed with suspected high signal, Group D showed that the cervical spinal cords were compressed with high signal. Results: For healthy subjects, average ADC, FA, L1, L2 and L3 values were (0.784p0.083)x10-3 mm2/s, 0.721p0.027, (1.509p0.145)x10-3 mm2/s, (0.416p0.094) x10-3 mm2/s and (0.411p0.102)x10-3 mm2/s, respectively. Between the control group and group A, there were no statistically signiﬁcant differences in ADC (Q=1.451, p 0.05)and FA (Q=1.999, p 0.05), L1 (Q=1.137, p 0.05), but there were statistically signiﬁcant differences in L2 (Q=3.039, p 0.05) and L3 (Q=2.982, p 0.05). Between the control group and group B, between the control group and group C, and between the control group and group D, there were statistically signiﬁcant differences in ADC, FA, L1, L2 and L3 (p 0.01). Conclusion: ADC, FA, L1, L2 and L3 values were sensitive indices for evaluating cervical spinal cord of cervical spondylosis patients.

B-132

14:09

Delayed gadolinium-enhanced magnetic resonance imaging (dGEMRIC) for the quantiﬁcation of glycosaminoglycans depletion and regeneration in human intervertebral discs S. Vaga, M.T. Raimondi, E.G. Caiani, F. Perona, M. Fornari; Milan/IT ([email protected]) Purpose: 1) to set up and test a new application of the delayed Gadolinium-Enhanced MRI of Cartilage (dGEMRIC) protocol for lumbar intervertebral discs; 2) to develop a software for the regional quantiﬁcation of sulphated glycosaminoglycans (sGAG) in the intervertebral disc in vivo; 3) validation of this methodology by comparing the MRI results with biochemistry-derived sGAG measurements of disc samples. Methods and Materials: 23 patients with lumbar discopathy underwent two MRI scans each, according to the dGEMRIC protocol adapted to the disc before (PRE) and 3.5 hours after (POST) Gd (DTPA)2- injection. A $T1 parametric image of the disc was computed as the difference between T1 images calculated at PRE and POST. A total of 31 tissue samples were obtained at herniectomy and analysed for sGAG content. Linear correlation between biochemical results and local $T1 was calculated. Results: $T1 images resulted in extremely high values within degenerated areas with respect to normal discs. The linear regression analysis resulted in a signiﬁcant correlation (r=0.7, p 0.0001) between MRI-derived and biochemistry-derived measurements. Conclusion: The parameter $T1 was able to correctly quantify the local concentration of sGAG within disc tissue in vivo. This method could lead to a better understanding of the physiopathology of lumbar discs cartilage through accurate in vivo molecular studies, opening a new perspective for the study of the disc regeneration pathways and providing an accurate tool to monitor the effects of several kinds of therapies. Further studies are required to better understand the contrast agent distribution and perfect the protocol.

170

A

B

14:18

Imaging the cervical spine in acute trauma: MRI vs CT S.M. Shaikh; Hyderabad/IN

Musculoskeletal

B-131

B-133

C

D

E

F

G

Purpose: CT is the standard for imaging the cervical spine in trauma. MRI of the cervical spine is indicated if soft tissue or injury is suspected. The purpose of this study is to assess the beneﬁt of cervical spine MRI and CT in the setting of Level I or II trauma. Methods and Materials: We retrospectively reviewed 91 trauma patients with Level I or II trauma patients who had undergone both CT and MRI of the cervical spine within a 5-day interval. Results: There were 91 trauma patients who had undergone both CT and MRI after trauma. Nine of the 91 patients (10%) had positive MRI studies after negative CT scans of cervical spine; ﬁve of these patients (56%) were classiﬁed as Level I traumas, and four (44%) were Level II traumas. Of the nine patients with a falsenegative CT, four were involved in motor vehicle accidents, four had falls, and one was injured while operating farm equipment. Six of the nine patients (67%) had a substantial change in management (i.e. unexpected surgery) based on the MRI ﬁndings. The other three patients (33%) had minor changes in management (i.e. extended placement of a cervical collar). Thus, 6/43 (14%) trauma patients with negative CT who underwent subsequent MRI examinations had signiﬁcant changes in management based on MRI ﬁndings. Conclusion: We believe that MRI is complimentary to CT in the trauma setting to localise soft tissue and bony changes resulting in 14% changes in management.

B-134

14:27

MR imaging in lumbar spinal stenosis: Prediction of surgical outcome S. Hong, J. Lee, J.-S. Yeom, K. Park, J.-Y. Choi, N. Kim, H. Kang; Seoul/KR ([email protected]) Purpose: The purpose of this study was to determine if MR imaging can predict the outcome of surgery in patients with lumbar spinal stenosis. Methods and Materials: We retrospectively reviewed MR images of 118 patients with lumbar spinal stenosis who had undergone spinal surgery. On MR imaging, number of stenotic levels, degree of stenosis, cross sectional area of the dural sac at the single most constricted level, and degree of cauda equina nerve root redundancy were evaluated. To assess patient’s functional disability, we measured the Oswestry disability index (ODI) before and 3 months afer surgery. The change of ODI before and after surgery was calculated and used as a measure of surgical outcome. To ﬁnd signiﬁcant MR predictors of surgical outcome, multiple linear regression analysis was performed. Results: In the multivariate analysis, only the number of stenotic levels was related to surgical outcome; the other variables on MR imaging were not related. At 3 months after surgery, patients who had single level stenosis had greater improvement in the primary outcome measures of ODI than who had multilevel stenosis (mean ODI changes after surgery: single level stenosis, -18.9 vs multilevel stenosis, -9.3) (t=2.221, p=0.028). Conclusion: At MR imaging, the number of stenotic levels is the only predictor of the postoperative improvement in patients with lumbar spinal stenosis. Greater postoperative improvement can be expected in single level stenosis than in multilevel stenosis.

B-135

14:36

Dural ectasia in Marfan syndrome J. Desperramons, J.-P. Pelage, L. Gouya, A. Beauchet, S. Chagnon, P. Lacombe; Boulogne/FR ([email protected]) Purpose: To evaluate the prevalence of dural ectasia in patients with Marfan syndrome (MS). To compare ﬁve different methods to diagnose dural ectasia. Methods and Materials: From January to December 2006, all patients diagnosed with MS undergoing multidetector helical CT of the aorta were included. Non affected patients of similar age and sex were used as a control group. Speciﬁc axial and sagittal reconstructions of the spine were obtained. Four diagnostic methods based on the diameters of the spinal canal and the dural sac at L1 to S1 levels were compared: Villeirs, Ahn, Oosterhoff and modiﬁed Oosterhoff (abnormal ﬁndings in L3 and S1) methods. Visual impression made by experienced readers was also recorded. Results: Thirty-six patients (44.3 y) with MS and 36 controls were included. Dural ectasia was present in MS patients in 47, 78, 44 and 72% of cases using the Villeirs, Ahn, Oosterhof and modiﬁed Oosterhof methods, respectively. In the control group, dural ectasia was diagnosed in 0, 39, 0 and 6% of cases. Using the visual method, dural ectasia was diagnosed in 50% and 3% of MS and control patients, respectively. A good agreement was found between the Villeirs method and the

Scientific Sessions

B-136

14:45

A potential clue in ultrasonographic diagnosis of adhesive capsulitis in shoulder joint: Limited subacromial gliding J. Lee, I.-H. Jeon, C.-H. Kim; Daegu/KR ([email protected]) Purpose: Adhesive capsulitis is characterized by decreased joint volume. In this study, we evaluated the feasibility of an ultrasonographic sign, ‘limited subacromial gliding of supraspinatus tendon’, to acknowledge decreased joint volume. Methods and Materials: Among 378 cases of shoulder MR arthrography in the last ﬁve years, 131 cases were selected. The inclusion criterion was shoulder pain. The exclusion criteria were capsular defect or loosening, joint effusion, synovial thickening, and superior impingement syndrome. The male:female ratio was 88:43 with a mean age of 42.6. Shoulder ultrasonography was performed prior to MR arthrography as contrast media was injected under ultrasonographic guidance. A dynamic ultrasonographic exam of subacromial gliding of supraspinatus tendon was performed and scored from 0 to 3 by the degree of limitation. The injection of contrast media was halted when joint capsule bulged out. The degree of subacromial gliding and the injected volume of contrast media were compared. Results: The cut-off values of subacromial gliding score and injected volume were 1 and 15 ml, respectively. The decreased volume less than 15 ml could be predicted by subacromial gliding limitation score higher than grade 1. The sensitivity, speciﬁcity, positive predictive value, negative predictive value, and accuracy were 89.3, 98.7, 98.07, 92.5, and 94.7%, respectively (p 0.001 in Fisher's Exact Test). Conclusion: The ultrasonographic depiction of the ‘limited subacromial gliding of supraspinatus tendon’ was feasible to predict the ‘decreased joint volume’ which is an important fact of adhesive capsulitis in shoulder joint.

B-137

14:54

High resolution ultrasound (HRUS) in the calcifying tendonitis of the rotator cuff: Where does the calcium go in the resorptive phase? L.M. Sconfienza1, F. Lacelli2, E. Silvestri1, G. Garlaschi1, C. Martinoli1, C. Murolo1, G. Seraﬁni2; 1Genoa/IT, 2Pietra Ligure/IT (io@lucasconﬁenza.it) Purpose: The aim of this study is to describe the ultrasonographic features of calcifying tendonitis of the rotator cuff during the resorptive phase and to demonstrate the calcium deposits migration within the ﬁbres of rotator cuff tendons and into the subacromial-subdeltoid (SASD) bursa. Methods and Materials: 722 patients affected by calcifying tendonitis during the hyperalgic phase were treated in a period of 36 months. Among these, 178 patients showed a calciﬁcation rupture with calcium deposit migration at the sonographic evaluation. Static and dynamic ultrasound evaluations were performed with a high resolution linear array probe (12-5 Mhz). All examinations included a complete assessment of SASD bursa after the injection of saline. Results: During the treatment of patients with calciﬁcation rupture, the sonographic evaluation demonstrated the loss of the calciﬁc wall with limited interruptions, especially on the sub-bursal side (95%), a change in ecogenicity both of calciﬁcation and tendon structure due to calcium deposit diffusion among the tendon ﬁbres (87%), calcium deposit migration in the sub-bursal space and within the SASD bursa. Conclusion: The sonographic evaluation of rotator cuff tendons during the percutaneous treatment of calcifying tendonitis allowed to demonstrate the interruption of the calciﬁc wall, the intratendineous and pericalciﬁc modiﬁcations and the calcium deposit diffusion in the sub-bursal space and the SASD bursa as speciﬁc signs of calciﬁcation rupture. On the other hand, modiﬁcations of calciﬁc structure and bursitis are non-speciﬁc signs of tendonitis.

B-138

15:03

Internal and external rotation combination on shoulder MR-arthrography in clinically suspected SLAP lesion M.-T. Guedez1, E. Llopis2, E. Hernandez1, A. Cruz3, L. Cerezal3; 1Maracai/VE, 2 Valencia/ES, 3Santander/ES ([email protected]) Purpose: To improve diagnosis of SLAP lesions and visualization of the biceps-labral complex, by adding external rotation to MR arthrography in internal rotation. Methods and Materials: We prospectively evaluated 54 patients with clinically suspected SLAP lesion. Patients complained of “dead arm syndrome”, sudden onset of pain and glenohumeral internal rotation deﬁcit. All patients had direct arthro MRI at

1.5 T. Internal rotation sequences were obtained in coronal, axial and sagittal planes; subsequently, coronal and axial T1-weighted fat suppressed sequences with the arm in maximal external rotation were obtained. SLAP lesions were classiﬁed using the extended Snyder classiﬁcation and clockwise arthroscopic classiﬁcation. Results: Twelve of 54 patients did not have SLAP lesions (type I in three, type II in ﬁve and type III in four biceps-labral complex) but anterior capsular detachment (two patients), rotator interval injury (three), partial thickness rotator cuff tear (ﬁve). Surgically conﬁrmed SLAP lesions were found in 42 patients. In all patients external rotation (especially in the coronal plane) improved the visualization of biceps superior labral junction, facilitating differentiation from normal variants, and increasing conﬁdence of radiological diagnosis. In 19%, (nine patients) external rotation allowed diagnosis of SLAP lesions not seen on conventional internal rotation. Conclusion: External rotation position on shoulder MRI improves diagnosis of SLAP lesions. Combined internal and external rotation shoulder MR-arthrography is recommended when SLAP lesions are clinically suspected.

B-139

15:12

Fast MR arthrography using VIBE sequences to evaluate the rotator cuff J.E. Vandevenne1, J.M. Mahachie John2, G. Gelin1, P.M. Parizel3, F.M.H.M. Vanhoenacker4; 1Genk/BE, 2Hasselt/BE, 3Antwerp/BE, 4Duffel/BE ([email protected]) Purpose: Applying prolonged MR sequences to patients in pain after recent shoulder injury often results in degraded MR images due to motion artefacts. To evaluate the accuracy of fast MR arthrography using volumetric interpolated breathhold examination sequences (VIBE) for diagnosing rotator cuff (RC) tears. Methods and Material: Eighty-two patients underwent direct MR arthrography of the shoulder joint using T1-weighted sequences with fat saturation (T1-FS, 5 min) and VIBE (13 sec) in axial and paracoronal planes, applying a phased array shoulder coil in a 1.0 T MR unit. To evaluate for RC tears, two radiologists in consensus scored in a blinded fashion VIBE and T1-FS images separately. Scores included normal, fraying or small partial thickness undersurface RC tears, large partial thickness undersurface RC tears, and full thickness tears with or without geiser sign. Utilizing SAS software, sensitivity, speciﬁcity, negative and positive predictive values of VIBE were calculated per score, using T1-FS as the gold standard. Results: Sensitivity, speciﬁcity and predictive values of VIBE ranged 92-100% for large partial thickness tears and full thickness tears. For diagnosing fraying and small partial thickness tears, values ranged 55-94%. Conclusion: Fast MR arthrography of the shoulder joint using VIBE was shown to be an accurate tool for diagnosing large partial thickness tears and full thickness tears of the RC compared to the standardly used T1-FS. In our experience, we found that VIBE, thanks to its very short acquisition time (13 sec), was very useful especially for patients in pain and for claustrophobic patients.

B-140

15:21

Safety and tolerability of direct contrast-enhanced MR-arthrography: Assessment of infections and pain in 1083 patients N. Saupe1, M. Zanetti1, C.W.A. Pﬁrrmann1, T. Wels2, C. Schwenke3, J. Hodler1; 1 Zurich/CH, 2Rotkreuz/CH, 3Berlin/DE ([email protected]) Purpose: To evaluate prospectively the incidence of infections and pain after direct contrast-enhanced (CE) MR-arthrography. Methods and Materials: MR-arthrography was performed in 1083 patients (609 females, 474 males, mean age 46.7 years) in six different joints (shoulder, hip, elbow, wrist, ankle, and knee). 1013 patients received Magnevist® 2 mmol/l, 70 Dotarem® (concentration: 5 mmol/l). Pain was assessed at four different time points (directly after injection, 4 h, 18-30 h and 6-8 days after injection) versus baseline using a visual scale ranging from 0 to 10. When increased pain persisted after injection, additional diagnostic assessment (laboratory, physician consultation) was performed. Evaluated factors with a potential impact on pain were: time, joint, volume of contrast agent, age and gender. For statistical analysis, ANOVA for repeated measurements was used. Results: Infection was not detected in any patient. Highest pain scores were obtained 4 hours after injection for all joints and all age groups (p 0.0001). Highest pain scores were found in the hip (n=294, delta mean pain value increase versus baseline: 1.15p1.6), followed by the elbow (n=11, 1.09p1.4), knee (n=11, 1.00p1.8), wrist (n=73, 0.86p1.6), ankle (n=19, 0.68p0.9) and shoulder (n=675, 0.49p1.2). Highest pain scores were found in the age group 30 years (p=0.044). Contrast agent, joint, volume of the contrast agent and gender did not signiﬁcantly inﬂuence pain scores. Conclusion: No joint infection was detected after 1083 direct CE MR-arthrographies. Direct CE MR-arthrography is associated with mild pain 4 hours after injection, most distinctive after hip arthrography, and in younger patients.

A

B

C

D

E

F

G

171

Friday

visual or Oosterhof methods (kappa 0.93 and 0.80, respectively). A good agreement was found between the Oosterhof method, and visual or modiﬁed Oosterhof methods (kappa 0.74 and 0.62, respectively). Conclusion: The Oosterhoff method has a good ratio of sensitivity/speciﬁcity. The modiﬁed Oosterhof method that we propose has a better sensitivity and a slightly decreased speciﬁcity.

Scientific Sessions 14:00 - 15:30

Room C

Abdominal Viscera (Solid Organs)

SS 301a Challenges in hepatobiliary diseases Moderators: J.S. Laméris; Amsterdam/NL A. Lubienski; Lübeck/DE

B-141

14:00

Qualitative and quantitative comparison of 3.0 T and 1.5 T MRI for the liver with chronic liver disease M. Ramalho, M. Tsurusaki, V. Heredia, E. Altun, R.C. Semelka; Chapel Hill, NC/US ([email protected]) Purpose: The purpose of this study was to compare signal characteristics and image qualities of MR imaging at 3.0 T and 1.5 T in patients with chronic liver disease. Methods and Materials: Twenty-four patients with chronic liver disease underwent abdominal MR imaging at both 3.0 T and 1.5 T within a 4-month interval. A retrospective study was conducted to obtain quantitative and qualitative data from both 3.0 T and 1.5 T MRI. Quantitative image analysis was performed by measuring signal-to-noise ratios (SNRs). Qualitative image analysis was assessed by grading each sequence on a 3- and 4-point scale, regarding the presence of artifacts and image quality, respectively. The results were compared using Student’s t-test and Wilcoxon signed-rank test. Results: Average liver parenchyma SNRs were signiﬁcantly higher at 3.0 T (p 0.05) than at 1.5 T on T2-weighted HASTE, T1-weighted 2D SGE in phase and second and third phases of 3D-VIBE images; overall image qualities of the 1.5 T non-contrast T1-weighted images were signiﬁcantly better than those of 3.0 T (p 0.01). In contrast, overall image quality of the 3.0 T post gadolinium VIBE sequence was signiﬁcantly better at 3.0 T (p 0.01). Conclusion: The advantage of 3.0 T imaging was high-quality thinner section acquisition of post gadolinium images. At present, the most important limitation is lesser image quality of pre-contrast 2D in-phase and out-of-phase T1-weighted images.

B-142

14:09

Non-alcoholic hepatosteatosis: Correlation of unenhanced CT and liver fat content P. Narayanan, F. Coyle, W. Drake, A.G. Rockall, R. Feakins, G.R. Foster, R.H. Reznek; London/UK ([email protected]) Purpose: The current gold standard for the diagnosis of non-alcoholic hepatosteatosis (NASH) is liver biopsy. The aim of this prospective study was to ascertain (a) the accuracy of unenhanced CT (NECT) in predicting the percentage of liver fat and (b) the value of NECT in identifying those patients in whom biopsy could be avoided. Methods and Materials: IRB approval was obtained. Patients were prospectively recruited into the study and underwent a 2-slice NECT through the liver and a liver biopsy. Six Hounsﬁeld Unit (HU) readings of the liver were taken using a standard circular ROI and the mean was recorded. On histological analysis, the percentage fat content was expressed using the Non-Alcoholic Steatosis (NAS) activity score. Results: To date, 22 patients have attended both NECT and liver biopsy. The range of HU was between 26.3 and 67. The range of fat percentage in the biopsy specimens was between 0 and 82.5. Pearson’s correlation coefﬁcient of the HU and percentage fat values was 0.884 with a p value of 0.001. Conclusion: Our preliminary data from this prospective study shows an excellent correlation between the degree of histopathological fat content and the liver HU. Further work is on-going to conﬁrm these initial results and to try and predict an Hounsﬁeld unit below which the likelihood of clinically relevant NASH is high as this would obviate the need for a biopsy in all patients.

B-143

14:18

CT laparoscopy for detecting small superﬁcial metastases of the liver surface Y. Maetani, H. Isoda, Y. Hirokawa, S. Arizono, K. Shimada, K. Togashi; Kyoto/JP ([email protected]) Purpose: To determine the efﬁcacy of CT laparoscopy for detecting superﬁcial metastasis of the liver surface. Methods and Materials: This is an IRB approved prospective study. From April 1, 2007 to July 1, 2007, a total of 34 consecutive patients (19 male, 15 female, median age 55 years) with various intra-abdominal malignancies underwent preoperative

172

A

B

C

D

E

F

G

CT and comprised the study population. All patients underwent superparaoxideenhanced MR imaging and portal phase enhanced CT obtained by 64 detector multi-slice CT, including CT laparoscopy. CT laparosocopy is a volume-rendering 3D image of the liver; with this modality, one can observe the liver surface in detail. Results: Of the 23 patients who underwent surgery, four patients had seven superﬁcial metastases of the liver surface. None of these metastases could be detected by axial CT imaging or SPIO-enhanced MR imaging preoperatively. In contrast, CT laparoscopy revealed 4 out of 7 lesions in 4 patients. On a lesion by lesion basis, the sensitivity was 57%, the speciﬁcity was 100%, and the positive predictive value was 100%. Conclusion: Our initial experience proves that the CT laparoscopy is a promising method for detecting small superﬁcial liver metastases of the liver surface that can inﬂuence decisions regarding tumor resectability.

B-144

14:27

Intraindividual comparison of gadobenate dimeglumine (Gd-BOPTA) enhanced MR imaging and multiphasic 64-slice CT for the detection of hepatocellular carcinoma (HCC) in patients with cirrhosis: Correlation with explanted liver M. Di Martino, D. Marin, C. Catalano, A. Guerrisi, M. Rossi, S.G. Corradini, R. Passariello; Rome/IT ([email protected]) Purpose: To intraindividually compare gadobenate dimeglumine (Gd-BOPTA) enhanced MRI and 64-slice CT for detection of HCC in patients with cirrhosis. Methods and Materials: Thirty-six consecutive patients with 46 HCC nodules underwent MRI at 1.5 T (Avanto, Siemens) and 64-slice CT (Sensation 64, Siemens) at a mean interval of 14 days (range, 10-20 days). All patients underwent transplantation within 60 days. MR acquisitions comprised unenhanced breath-hold T2W images and volumetric 3D Gd-BOPTA-enhanced (0.1 mmol/kg; MultiHance, Bracco) T1W GRE images acquired at 25, 60, 180 s (dynamic phase) and 120 min (hepatobiliary phase). 64-slice CT was performed with 0.6x64 mm collimation, 3-mm section thickness, 250 mAs, 120 kVp. A triple-phase protocol was started 18, 60 and 180 s after reaching a trigger threshold of 150 HU above baseline CT number of the aorta. Image analysis was independently performed by three observers in two sessions separated by four weeks. Findings were compared directly with explanted liver pathologic results. Diagnostic accuracy was evaluated using the alternativefree response receiver operating characteristic (AFROC) method. Sensitivity and speciﬁcity with corresponding 95% conﬁdence intervals were determined. Informed consent and ethical approval were obtained. Results: The mean area under the AFROC curve for Gd-BOPTA MRI (0.92) was signiﬁcantly higher than that of CT (0.84) (P 0.05).On a lesion-by-lesion basis, the mean sensitivity (77%, 91/138) of Gd-BOPTA MRI was signiﬁcantly higher than that of CT (66%, 91/138) (P 0.05). Both techniques showed an equal mean speciﬁcity (90%, 133/148). Conclusion: Gd-BOPTA-enhanced MRI is signiﬁcantly more accurate and sensitive than 64-slice CT for the diagnosis of HCC in patients with cirrhosis prior to liver transplantation.

B-145

14:36

Automatic postprocessing of multi-breath hold 3D DCI MRI of the liver for precise anatomical alignment M.P. Lichy1, C. Plathow1, C. Chefdhotel2, D. Plathow1, W. Horger3, M.S. Horger1, P. Aschoff1, B. Kiefer3, C.D. Claussen1, H.-P. Schlemmer1; 1Tübingen/DE, 2 Princeton, NJ/US, 3Erlangen/DE ([email protected]) Purpose: To evaluate the clinical value of anatomical correction of multi-breath hold 3D dynamic contrast-enhanced (DCE) MRI of the liver. Methods and Materials: A software prototype for automated retrospective correction of unenhanced, arterial, portal-venous (p.v). and late phase liver DCE-MRI data sets was implemented. Data sets were also used to obtain subtraction images. A total of 33 patients were evaluated: 17 patients (mean age 66) with histological proven HCC, six patients with FNH, ﬁve with adenoma and ﬁve with liver metastases (mamma carcinoma). In all cases, at least one further MRI follow-up was available. Data evaluation was performed by two experienced radiologists for the original and corrected data sets (with and without subtraction). Evaluation included qualitative criterions (image quality, numbers of lesions, off-set of data sets, reporting time) and quantitative assessment of signal-intensities changes by a region-of-interest (ROI) analysis. Results: Signiﬁcant mismatch of initial 3D DCE MRI data sets was observed in 22/33 patients. In all cases, the postprocessing routine of DCE MRI data sets was successful, yielding exact anatomical matches of the corrected 3D data sets. Corrected images were rated equal or superior to original data sets and signal intensity changes of the left/right hepatic artery were found to be the most sensitive parameters to evaluate the function of the correction algorithm.

Scientific Sessions Conclusion: Automated anatomical correction of multi-breath hold 3D DCE liver MR examinations is feasible, enabling the subtraction of multi-phase data sets within clinical routine.

14:45

MRI for quantiﬁcation of hepatic steatosis with histopathologic correlation T. Chen1, R.-C. Chen1, C.-T. Chou2, J.-M. Chou1,C.-S. Li1, W.-T. Chen1, K.-C. Ding1; 1Taipei/CN, 2Changhua/CN ([email protected]) Purpose: To evaluate the performance of T1-weighted dual-phase magnetic resonance imaging for hepatic steatosis quantiﬁcation. Methods and Materials: Forty-four patients were recruited from two medical centers; all had undergone pre-operative surgery to remove liver tumors. A radiologist then assessed the T1-weighted dual-phase images and calculated their mean hepatic and splenic signal intensities (SI) using averages from nine hepatic and three splenic regions of interest. Hepatic SIs from in-phase images were then compared with those of out-phase images, and relative changes in SI (RCSI) were derived using their respective splenic SIs as reference. The resected specimens were assessed by an experienced hepatopathologist for hepatic steatosis and graded into 4 categories, which include 0 (60%). Pearson correlation, simple regression, and one-way ANOVA with Bonferroni correction were used to analyze the results. Results: Nine patients were found to have category 0, 18 had category 1, seven had category 2, and four category 3 hepatic steatosis. Pearson’s correlation revealed a signiﬁcant correlation between their RCSIs and the pathological specimens’ degree of hepatic steatosis (p 0.001), which is also supported by one-way ANOVA (p 0.001) and simple regression (regression coefﬁcient=0.0988, p 0.001). Conclusion: These results support the current opinion that hepatic steatosis is closely associated with RCSI from MR images. While some investigators contend that signiﬁcant steatosis ( 50%) can cause a paradoxical decrease in RCSI, our results demonstrate a relatively linear correlation between the degree of hepatic steatosis and RCSI.

B-147

14:54

Diagnostic performances of MDCT, SPIO-enhanced MRI, and double contrast MRI in predicting the appropriateness of a transplant recipient based on Milan criteria: Correlation with histopathologic ﬁndings D. Lee, S. Kim, J. Lee, H. Park, J. Lee, N.-J. Yi, K.-S. Suh, J. Jang, J. Han, B. Choi; Seoul/KR ([email protected]) Purpose: To retrospectively evaluate diagnostic performances of MDCT, SPIOenhanced MRI (S-MRI), and double contrast MRI (DC-MRI) in predicting the appropriateness of recipients with hepatocellular carcinomas (HCC) for liver transplantation (LT) based on Milan criteria. Methods and Materials: During a 4-year period, 78 patients with 80 HCCs underwent LT. MDCT and DC-MRI with the sequential use of SPIO and gadolinium were performed in all patients. Interval reading for MDCT, S-MRI, and DC-MRI was done. Two radiologists independently recorded their conﬁdence level using 4- and 5-point scales for the presence of HCC and for the appropriateness regarding LT, respectively. Image interpretation was compared with histopathologic results on a lesion-by-lesion basis. Diagnostic performances of the three imaging techniques were compared using JAFROC and ROC analyses. Results: Sixty-nine of 78 recipients were categorized as appropriate and the remaining nine were found to be inappropriate for LT based on Milan criteria. The AZ value in predicting the appropriateness for LT was highest in S-MRI (0.846 for reviewer 1 and 0.778 for reviewer 2), followed by DC-MRI (0.809 and 0.778) and MDCT (0.79 and 0.703). However, signiﬁcant differences were not achieved for both radiologists (p=0.583~1). In terms of detecting HCCs, the reader-averaged ﬁgure-of-merit was highest in DC-MRI (0.764), followed by S-MRI (0.702) and MDCT (0.672). However, the differences were also not signiﬁcant (p=0.11). Conclusion: SPIO-enhanced MRI and double contrast MRI tend to provide better diagnostic performance than MDCT in predicting the appropriateness of transplantation recipients based on Milan criteria and in detecting HCCs.

B-148

15:03

Hypoechoic focal fatty inﬁltration in liver steatosis at US: Fast and deﬁnitive assessment with contrast enhanced ultrasound L. Tarantino1, I.F.M. Sordelli2, V. Nocera1, A. Schiano1, C. Ripa2; 1Frattamaggiore/IT, 2 Naples/IT ([email protected]) Purpose: The standard techniques for diagnosis of hypoechoic focal areas (FFS) at ultrasound (US) in liver steatosis are computed tomography (CT) and/or magnetic resonance (MR) and/or biopsy. We evaluated the role of contrast enhanced US (CEUS) in the diagnosis of FFS.

B-149

15:12

Color-Doppler ultrasonography (CDU), multidetector computed tomography (MDCT) and endoscopic ultrasonography (EUS) in periampullary tumors: Evaluation of vascular involvement of major peripancreatic vessels S. Gusmini, R. Nicoletti, C. Martinenghi, C. Soldati, C. Caborni, M. Salvioni, M. Venturini, S. Rocchetti, G. Balzano, A. Zerbi, P. Arcidiacono, V. Di Carlo, A. Del Maschio; Milan/IT ([email protected]) Purpose: In patients affected by periampullary tumors, preoperative evaluation of vascular involvement is necessary to avoid uncurative surgical treatment. Aim of our study was to assess the performance of MDCT, EUS and CDU in the evaluation of vascular involvement of major peripancreatic vessels in periampullary tumors. Methods and Materials: In two years, 65 patients affected by periampullary tumours were enrolled in a diagnostic protocol to determine vascular involvement of the following peripancreatic vessels: portal vein, superior mesenteric artery and vein, hepatic artery, celiac trunk. Patients underwent CDU, MDCT and EUS examination before surgical evaluation, the gold standard of our study. Vascular involvement was assessed by this pattern: grade 0 (no contact between tumour and vessel), grade 1 (contact between tumour and vessel, bloodless surgical excision), grade 2 (vascular inﬁltration, bloody surgical excision). Results: Surgical treatment was performed in 54 patients.We analysed 270 peripancreatic vessels. Surgical grades 0, 1, 2 were in 224, 18, 28 vessels, respectively. Accuracy of MDCT, EUS, CDU was: 94, 89, 88% in grade 0, 93, 91, 88% in grade1 and 97, 96, 93% in grade 2. MDCT had an higher overall accuracy than EUS (95 vs 92%; p 0.0189) and CDU (95 vs 90%; p 0.0002). Surgical resectability was described in 34 patients, surgical non resectability in 20 patients. MDCT and EUS had a high accuracy in the prediction of resectability: 33/34 were correctly evaluated as resectable. 15/20 patients were correctly evaluated as non resectable. Overall accuracy of MDCT and EUS in the evaluation of resectability was 89%. Conclusion: MDCT is the best imaging technique in the evaluation of vascular involvement of major peripancreatic vessels and in the prediction of resectability in periampullary tumors.

B-150

15:21

Detection and determination of advanced hepatocellular carcinoma: Prospective comparison of MR imaging and PET/CT J. Baek, J. Lee, S. Kim, J. Lee, J. Han, B.-I. Choi; Seoul/KR ([email protected]) Purpose: To determine and compare the diagnostic performance of magnetic resonance (MR) imaging and positron-emission tomography/computed tomography (PET/ CT) in detection and determination of advanced hepatocellular carcinomas (HCC). Methods and Materials: Forty-six patients with HCC underwent MR imaging and PET/CT before liver transplantation. MR-based prediction of HCC differentiation was made using T2-weighted fast spin echo (FSE) and superparamagnetic iron oxide (SPIO)-enhanced T2-weighted gradient-echo (GRE). Explanted liver specimens were sliced to 5 mm-thick sections to determine presence and histologic differentiation of lesions. Overall, 21 dysplastic nodules (DN), 14 well differentiated HCCs, and 44 advanced HCCs were histologically conﬁrmed. For every lesion, detection of FDG uptake by PET/CT, signal increase of T2-weighted FSE and signal drop of SPIO-enhanced T2-weighted GRE were investigated. Sensitivity and speciﬁcity of PET/CT and MR imaging sequences were calculated. McNemar’s test was used to evaluate the difference in sensitivity and speciﬁcity.

A

B

C

D

E

F

G

173

Friday

B-146

Methods and Materials: We retrospectively reviewed clinical and imaging records of 250 patients (204 males; age range: 16-82 years) with 315 FFS (diameter range 0.9-7.0 cm; mean 2.8 cm) in liver steatosis at US. The ﬁnal diagnoses were made by CT and/or MR in 178 patients with 232 nodules, biopsy in 54 patients with 58 nodules; US follow-up in 18 patients with 25 nodules. All patients underwent CEUS. We compared CEUS ﬁndings with CT, MR, biopsy and follow-up results. Results: A ﬁnal diagnosis of FFS was made in 86 out of 315 (27%) cases. 83/86 (96%) cases showed CEUS pattern of FFS (absence of focal lesions and normal distribution of liver vessels during all CEUS phases). In 3/86 cases CEUS ﬁndings were equivocal because of persistence of small hypoechoic images in the target areas. In all three cases, biopsy proved FFS diagnosis, and clinical and US follow-up (12-28 months) showed absence of changes over time. 229/315 focal areas showed hyperechoic or hypoechoic patterns at CEUS. Sensitivity, speciﬁcity, positive predictive and negative predictive values of CEUS for the diagnosis of FFS were 96.5, 100, 100 and 98.7%, respectively. Conclusion: CEUS is a fast and safe technique, highly sensitive and speciﬁc for the diagnosis of FFS. Therefore, for this purpose, it should replace more expensive and invasive imaging techniques.

Scientific Sessions Results: For the detection of advanced HCCs, sensitivity/speciﬁcity of PET/CT, T2-weighted FSE, and SPIO-enhanced T2-weighted GRE were 29.5/94.3%, 52.3/68.6%, and 54.5/82.9%, respectively. According to McNemar’s test, sensitivity differed signiﬁcantly between PET/CT and T2-weighted FSE (p=0.0129), and between PET/CT and SPIO-enhanced T2-weighted GRE (p=0.0074). There was no signiﬁcant difference in sensitivity between MR imaging sequences (p=1). PET/CT showed highest speciﬁcity, but there was no signiﬁcant difference in speciﬁcity between PET/CT and MR imaging sequences. Conclusion: PET/CT showed higher speciﬁcity in detection of advanced HCCs than T2-weighted FSE and SPIO-enhanced T2-weighted GRE though it showed lower sensitivity.

14:00 - 15:30

Room F1

Genitourinary

SS 307 Imaging of the pelvis Moderators: P. Aspelin; Stockholm/SE A. Bianek-Bodzak; Gdansk/PL

B-151

14:00

Magnetic resonance imaging (MRI) in the presurgery assessment of MayerRokitansky-Kuster-Hauser (MRKH) syndrome G. Pompili1, R. Meroni2, G. Franceschelli1, A. Munari1, G. Frontino1, G. Cornalba1, L. Fedele1; 1Milan/IT, 2Bergamo/IT ([email protected]) Purpose: To evaluate accuracy of MRI in women with primary amenorrhea, with suspected MRKH syndrome (congenital absence of both the vagina and uterus and the presence of normal ovaries). Methods and Materials: 35 women (16-27 years, mean 20.2), with primary amenorrhea, were studied with MRI. All patients were examined supine, using a phased array coil. Turbo spin-echo T2-w images (TR 4750-6686, TE 100-120, 4 mm). and T1-w images (TR 470, TE 15, 4 mm) were acquired in three planes. Two experienced radiologists evaluated all examinations in consensus. MRI results were judged on the basis of laparoscopy ﬁndings in 26/35 patients (74%) who underwent a subsequent neovagina surgery construction. Results: MRI identiﬁed bilateral mullerian rudiments in 19/35 (54%) and monolateral in 10/35 (29%) patients; 22 rudiments were cavitary. In all patients, both ovaries were visualized with regular morphology in 17/35 (49%), polycystic in 18/35 (51%); pelvic in 29/35 (83%), extrapelvic in 6/35 (17%). We found associated anomalies of upper urinary tract (monokidney, ptosis) in 6/35 patients. Laparoscopy conﬁrmed MRI data in all patients. MRI wasn’t able to detect mullerian rudiments in 8/26 patients, while it made a correct diagnosis in 18/26 (69%). In nine patients, laparoscopy evaluated rudiments cavitation, conﬁrming MRI data in 7/9 (78%). Sensibility and speciﬁcity were 100% for MRKH diagnosis, while in detecting rudiments were 77 and 100%. In one patient a big cavitated rudiment allowed surgical uterus reconstruction. Conclusion: MRI was useful in the pre-surgery evaluation of MRKH syndrome, is less expensive and invasive than laparoscopy.

B-152

14:09

HistoScanning: A new computer aided diagnostic tool for non invasive detection of malignant ovarian masses. A European clinical study report O. Lucidarme1, J.-P. Akakpo1, B. Lauratet1, S. Granberg2, M. Sideri3, R. Mashiach4, H. Bleiberg5, D. Nir5, P. Autier6, J. Lefranc1, P. Grenier1; 1Paris/FR, 2 Stockholm/SE, 3Milan/IT, 4Tel Aviv/IL, 5Waterloo/BE, 6Lyon/FR ([email protected]) Purpose: Ovarian HistoScanningTM (OVHS) is an innovative diagnosis support technology quantifying characteristic features of backscattered ultrasounds (US). We will present accumulated data from two consecutive studies carried out to evaluate OVHS as aid to ovarian cancer diagnosis. Methods and Materials: Women r18 years scheduled for the surgical removal of r1 ovary were eligible. Exclusion criteria included previous diagnosis of ovarian cancer, pelvic surgery, radiotherapy or chemotherapy for breast or ovarian cancer. 3D trans vaginal US scans (Voluson 730) were performed before surgery and the voxel data were analyzed by OVHS. After OVHS results were locked in the database, pathology reports were made available and OVHS results were compared to histology. A retrospective comparison of OVHS to US and CA 125 was made. Results: 375 ovaries were removed from 264 women undergoing surgery for pelvic mass. Histology reported 234 non-cancerous ovaries (107 normal, 127 benign

174

A

B

C

D

E

F

G

tumors) and 141 cancers (83 adenocarcinomas, 24 borderline, 16 carcinomatosis, nine metastases and nine other cancers). 138/141 cancers and 206/234 noncancerous tissues were correctly identiﬁed (sensitivity 97.8%, speciﬁcity 88%). The false negatives (FN) were metastasis from a breast cancer, primary ovarian cancer and cystadenoﬁbroma with microscopic foci of borderline cancer. There was no FN among 47 FIGO I, II. Ultrasound and CA-125 had a sensitivity/speciﬁcity of 96/62% and 88/79%, respectively. Conclusion: OVHS is highly sensitive for the diagnosis of malignant ovarian masses while maintaining a good speciﬁcity and is a promising tool for the non invasive characterization of pelvic masses.

B-153

14:18

Differential diagnosis of suspected ovarian cancer in pre-menopausal women J.M. Hawnaur, A. Talwalkar; Manchester/UK ([email protected]) Purpose: UK cancer targets require doctors to expedite investigation and treatment of pelvic masses that could be malignant. Management decisions in suspected ovarian cancer are increasingly reliant on imaging ﬁndings. This study aims to raise awareness of alternative diagnoses in pre-menopausal women. Methods and Materials: Clinical, radiological and histological ﬁndings were reviewed in women 50 years with suspected ovarian cancer referred to a gynaecological cancer multidisciplinary team (MDT) in 2003-7. All patients had prior imaging (US, CT or MRI) interpreted as possible cancer by the referring hospital. The ﬁnal diagnosis was established by surgery in most cases; some presumed benign disease was stable or resolved on conservative management. Patients were categorised as having ovarian neoplasm: malignant, borderline or benign, ovarian metastases, other benign or malignant pelvic tumours or inﬂammatory disease/ endometriosis. For epithelial ovarian tumours, benign, borderline or invasive histology was correlated with the patient's age. Results: Of 215 women reviewed, 18% had epithelial ovarian cancer (associated with endometriosis in about 1:5), 29% borderline tumours and 6% cystadenomas. The mean age of patients with cystic ovarian cancer (43 years) was older than those with benign or borderline cystic tumours (34 years). 10% had other primary ovarian tumours, 8% other pelvic malignancy and 4% ovarian metastases. Benign masses, pelvic inﬂammatory disease and endometriosis each accounted for 8-9%. Examples of imaging ﬁndings on US, CT and MRI will be shown. Conclusion: Ovarian cancer is relatively uncommon in young women and radiologists should consider other possible diagnoses when reporting pelvic masses.

B-154

14:27

Role of a computed tomography derived score (PIrad) for the prediction of optimal cytoreduction at primary surgery in patients with ovarian cancer E. Cucci, G. Restaino, G. Garganese, G. Scambia, G. Sallustio; Campobasso/IT ([email protected]) Purpose: Residual disease after primary surgery is one of the crucial prognostic factors in advanced ovarian cancer. We designed this prospective study to assess the role of a CT derived score (PIrad) in the preoperative prediction of ovarian cancer optimal cytoreduction (residual tumor b1 cm). Methods and Materials: 109 patients underwent abdominal helical contrast enhanced CT scan before exploratory laparotomy and were evaluated for potential cytoreduction. CT image evaluation was performed without knowledge of the operative ﬁndings, and included nine parameters: peritoneal and diaphragmatic carcinosis, bowel mesentery involvement (r2 cm), supra- or infrarenal para-aortic lymphnodes (r2 cm), liver involvement, spleen, stomach, or lesser sac localization, frozen pelvis, and ascites. The presence/absence of each sign was assigned a score which produced an overall radiologic resectability predictive index (PIrad) for each patient. Results: Median value of PIrad was 6 (0-15). After laparotomy, 56 patients out of 109 (49.6%) were optimally cytoreduced, while 54 patients (47.8%) were considered to have a non-resectable disease. At a PIrad cut-off value r12, corresponding PPV=87.5% and NPV=51.5%, i.e., if the PIrad was r12, the risk of leaving unexplored a potentially cytoreducible patient was 12.5%, and 48.5% of patients with a PI. Conclusion: A PIrad r12 showed high PPV, acceptable NPV, and high accuracy in identifying patients with advanced ovarian carcinoma unlikely to undergo optimal primary cytoreduction.

Scientific Sessions B-155

14:36

Value of MR imaging for the diagnosis of borderline ovarian tumors: Correlation with pathology M. Bazot, D. Haouy, I. Thomassin-Naggara, A. Cortez, E. Daraï, C. Marsault; Paris/FR ([email protected])

B-156

14:45

The tail sign: New sign for differentiation of hydrosalpinx from other adnexal cysts on three-dimensionally displayed MR imaging A.E. Mahfouz, H. Sherif; Doha/QA ([email protected]) Purpose: To differentiate hydrosalpinx from other cystic adnexal masses on MR imaging. Methods and Materials: Fifty-four women with 60 cystic adnexal masses (36 surgically and histologically proven to be tubal and 24 to be ovarian or paraovarian) were examined by transverse fat-saturated thin-section T2-weighted turbo spin echo (TR/TEeff 3000/90 msec, slice thickness 2.5 mm, no gap). Images were post-processed by volume rendering technique (VRT) to generate rotating threedimensional display of the adnexal cysts. Sectional images and the corresponding VRT images were randomized and reviewed by an experienced blinded reader to differentiate between tubal and ovarian masses based on detection of tail-like end of the cystic mass. Results: On the VRT images, a tail-like end could be identiﬁed in 30 cysts, which proved to be tubal in nature. The remaining 30 cysts (six tubal and 24 ovarian or paraovarian) had no tail-like outward projections (pb0.001). The tail sign had positive predictive value of 100%, negative predictive value of 80%, speciﬁcity of 100%, sensitivity of 83.3%, and accuracy of 96.4% for the diagnosis of hydrosalpinx. Evaluation of the sectional images resulted in corresponding values of 100, 68.5, 100, 67.4, and 81.6%. Conclusion: The tail sign is a speciﬁc sign for differentiation of hydrosalpinx from other adnexal cysts on MR imaging. Three-dimensional VRT images are superior to the corresponding sectional images in demonstration of this sign.

B-157

14:54

Role of diffusion-weighted MR imaging in the assessment of lymphadenopathies in patients with pelvic disease L. Ballesio, S. Savelli, M. Angeletti, M. Di Maurizio, A. Perrone, A. Francioso, L. Manganaro; Rome/IT ([email protected]) Purpose: The aim of the study is to analyze the role of apparent diffusion coefﬁcient value (ADC) in the evaluation of pelvic lymph nodes and to test how this value is useful for characterization of lymphadenopathies. Methods and Materials: We included 39 patients with pelvic disease: 14 endometrial carcinoma, 10 ovarian carcinoma, eight rectum carcinoma, ﬁve prostate carcinoma, two bladder carcinoma, all histopathologically proved. We used: TSE and HASTE T2-weighted, VIBE T1-weighted (with and without fat suppression) sequences before and after administration of gadolinium. Diffusion-weighted MR images were acquired with a b factor of 0, 500 and 1.000 s/mm2 using single-shot

B-158

15:03

Role of MR imaging and 3D sonography in the preoperative work out of deep inﬁltrating endometriosis E. Faiella, R. Del Vescovo, C.C. Quattrocchi, V. Di Giacomo, R.F. Grasso, B. Beomonte Zobel; Rome/IT ([email protected]) Purpose: To compare two different imaging modalities, Magnetic Resonance Imaging (MRI) and 3D-sonography, in order to evaluate the speciﬁc role in the preoperative work-out of deep inﬁltrating endometriosis. Methods and Materials: Thirty-three women (mean age=35) with endometriosis underwent 3D-sonography and MRI followed by surgical and histopathological investigations. 3D-sonography was performed by a gynaecologist and MRI by a radiologist in blind-fashion. The investigators described the disease extension in the following sites: torus uterinus and uterosacral ligaments (USL), vagina, rectovaginal-septum, rectosigmoid, bladder, ovaries. Results were compared with surgical and histopathological ﬁndings. Results: Ovarian and deep pelvic endometriosis were found by surgery and histology in 24 (72.7%) and 22 (66.6%), respectively, of the 33 patients. Sensitivity, speciﬁcity, and the positive and negative predictive values of 3D-sonography for the diagnosis of endometrial cysts were 87.5, 100, 100 and 88.8%, respectively; those of MRI were 96.8, 91.1, 91.1 and 96.8%, respectively. Sensitivity and speciﬁcity of 3D-sonography for the diagnosis of deep inﬁltrating endometriosis in speciﬁc sites were: USL 50 and 94.7%; vagina 25 and 90%; rectovaginal-septum 78.5 and 100%; rectosigmoid 25 and 95%; bladder 25 and 100%. Those of MR imaging were: USL 69.2 and 94.3%; vagina 80 and 95.6%; rectovaginal-septum 77.7 and 100%; restosigmoid 100 and 100%; bladder 83.3 and 100%. Conclusion: MRI accurately diagnoses deep inﬁltrating endometriosis; 3D-sonography accurately diagnoses deep inﬁltrating endometriosis in speciﬁc locations. However, MRI seems to be the imaging technique of choice due to the typical high contrast resolution, multiplanarity and greater ﬁeld of view.

B-159

15:12

Testicular microlithiasis and macrocalciﬁcation: Prevalence of tumors and follow-up in a single center over 10 years I. Amaechi, P. Sidhu; London/UK ([email protected]) Purpose: Testicular microlithiasis (TM) and macrocalciﬁcation (TC) are associated with an increased prevalence of primary testicular tumors. We report our experience of ultrasound (US) follow-up and suggest a strategy for management. Methods and Materials: A retrospective review of patients identiﬁed with TM or TC subject to follow-up US over a 10-year period. US images were reviewed to classify calciﬁcation into TM or TC; TM was sub-divided into limited (LMT 5 microliths/ ﬁeld), classical (CMT 5 microliths/ﬁeld) and ﬂorid (FMT, too numerous to count). Presenting symptoms, tumor-risk factors, US abnormality, age and follow-up period were recorded. Tumor development with pre-existing calciﬁcation was noted. Results: 12,426 US examinations were performed; testicular calciﬁcation was identiﬁed in 326/12,426 (2.6%) patients and 464/24,852 (1.7%) testes; mean age 38.1 years (range 8-82 years). TM was present in 398/464 (85.8%) testes; CTM in 103/464 (22.2%), LMT in 279/464 (60.1%) and FMT in 16/464 (3.4%). TC was present in 66/464 (14.2%) testes. Co-existent testicular abnormalities were identiﬁed in 6/67 (8.0%) with CTM, 13/226 (5.8%) with LMT, 1/8 (12.5%) with FMT and 1/54 (1.9%) with TC. Co-existent testicular tumor was present in 8/67 (11.9%) with CTM, 6/226 (2.7%) with LMT and 5/54 (9.3%) with TC; mean age 36.4 years (range 21-55 years). Mean follow-up was 12.0 months (range 0-123 months). Two non-primary tumors developed. Conclusion: There is a high prevalence of primary testicular malignancy with CTM and TC, lower with LTM. No primary tumors developed. The opportunity for detecting tumors may be missed; follow-up should be undertaken in a younger age group.

A

B

C

D

E

F

G

175

Friday

Purpose: To analyze morphological and vascular MR imaging criteria of borderline ovarian tumors (BOTs), and to compare these criteria with those exhibited by benign ovarian tumors. Methods and Materials: Retrospective study of 72 patients (mean age: 46 years) having resection of ovarian tumors containing foci of borderline morphology from January 1999 to January 2007. A reclassiﬁcation of tumors by a pathologist expert was performed according to borderline ovarian tumor workshop (Bethesda, 2003): 47 BOTs and 25 benign ovarian tumors containing less than 10% of borderline foci. All the patients underwent preoperative MRI, reviewed blindly in consensus by two radiologists assessing morphologic criteria: tumor type; number, gathering or thickening of septations; loculi of different intensities; vegetation or solid portion; signal intensity of solid tissue on T2-weighted MRI; ovarian parenchyma; ascites; and DCE-MRI patterns (type 1: slow, 2: moderate, and 3: intense). Results: At MRI examinations, no difference was noted among borderline and benign tumors for size, tumor type, number, gathering or thickening of septations, loculi of different intensities, solid portion, visualization of normal ovarian parenchyma, and pelvic ascites. A higher incidence of vegetations was found in borderline tumors than in benign ovarian tumors (p=0.03). Among tumors containing solid tissue, a low-signal intensity on T2-weighted MRI was highly indicative of benign lesions (p=0.01). A DCE-MRI type 2 was signiﬁcantly correlated with the presence of BOTs (p=0.02). Conclusion: Vegetations with intermediate signal on T2-weighted MRI and type 2 pattern on DCE-MRI may be a useful criterion for distinguishing borderline from benign ovarian tumors.

echo-planar sequence. ADC maps were reconstructed for all patients and the ADC value of lymph nodes was also calculated. The parameters used for the assessment of metastatic nodes were the size criteria, abnormal signal intensities on T1 and T2w images and ADC values. MR ﬁndings were compared with histologic ﬁndings and sensitivity, speciﬁcity, positive and negative predictive values were calculated. Results: The ADC values were signiﬁcantly greater in benign lymph nodes than in metastatic lymph nodes. When an ADC value of 1.41x10-3 mm2/s was used as a threshold value for differentiating malignant from benign lymph nodes, the best results were obtained with accuracy of 76.9%, sensitivity of 82.6%, speciﬁcity of 68.7%, positive predictive value of 79.2% and negative predictive value of 73.3%. Conclusion: ADC determination on diffusion-weighted MR imaging may be supportive for discriminating metastatic lymph nodes in pelvic disease.

Scientific Sessions B-160

B-162

15:21

Role of 18 F-FDG PET-CT in detection of recurrence of ovarian carcinoma with low tumor markers S.M. Shaikh; Hyderabad/IN Purpose: The aim of this study was to assess the role of 18 F-FDG PET-CT for the detection of recurrence of ovarian carcinoma, especially in patients who did not show any increase of tumor markers. Methods and Materials: We have evaluated 45 patients preliminarily submitted to radical surgery for ovarian cancer who had clinical or diagnostic imaging feature suspicious of a recurrence. All these patients underwent 18 F-FDG PET-CT. 80% of these patients had low tumor markers. PET-CT images were obtained 1 hour after 18 F-FDG injection and contrast CT images were used for attenuation correction. The images obtained were fused for analysis. Patients were prepared with overnight fasting before 18 F-FDG injection and the plasma glucose level had to be less than 140 mg/dl. The patients had to empty their bladder before PET-CT exam. Results: We detected nine patients with recurrence of ovarian cancer with 18 FFDG PET-CT. We found three recurrences at the level of pouch of douglas after surgery, one rib metastasis and ﬁve sub-diaphragmatic metastatic lymphadenopathy (four retroperitoneal and one at the porta). One of the retroperitoneal metastatic lymph-nodes was inﬁltrating the adjacent ureter. The smallest lesion detected by 18 F-FDG PET-CT was 8 mm in diameter. We had only one false positive case due to chronic inﬂammation after the surgery (5 months after radiotherapy). Conclusion: 18 F-FDG PET-CT is the most accurate procedure for the diagnosis of the recurrent ovarian malignancies and should be advised in early follow-up and high risk patients.

14:00 - 15:30

Room F2

Breast

Innovation in breast diagnosis Moderators: G. Rizzatto; Gorizia/IT S. Weum; Tromsö/NO

B-161

Purpose: To evaluate the performance of ultrasound elastography in breast lesions. Methods and Materials: 429 lesions (six centers, 268 benign, 161 malignant, 46.4% b10 mm) were evaluated by B-mode (BI-RADS classiﬁcation) and elastography (EUB 8500 Logos unit, Hitachi, Japan). Diagnosis was obtained by FNA, core or surgical biopsies (84%), follow-up or comparison with previous studies (benign lesions). The elastic score was classiﬁed according to a ﬁve-point color scale (Ueno classiﬁcation, 1-3=benign and 4-5=malignant). Intra- and inter-observer variables were evaluated in each center (kappa, 399 lesions). Results: With B-mode, sensitivity, speciﬁcity, positive predictive value, negative predictive value of BI-RADS classiﬁcation (2-3=benign and 4-5=malignant) were 100, 51.5, 55.3 and 100%, respectively. When category 4 was subdivided into 4 low (one pejorative sign) and 4 high (two pejorative signs), speciﬁcity increased to 91% with a sensitivity of 85.7%. For elastography, sensitivity, speciﬁcity, positive predictive value, negative predictive value were: 72, 89.5, 84.6 and 79.9%, respectively. When the elastographic score 3 (central part of the mass stiffer than the peripheral normal tissue) was considered as malignant, these values were 83.2 (14 false negative results), 80.5, 77.5 and 85.6%, respectively. Intra- and inter-observer reproducibility were kappa=0.8 and 0.73. Conclusion: US elastography is an effective and reproducible US technique for increasing speciﬁcity of B-mode ultrasound. By adding new criteria for benignity, this technique may avoid unnecessary diagnostic procedures especially in breast masses categorized as BI-RADS categories 3 and 4 low.

B-163

SS 302

14:00

Six years of real time breast ultrasound elasticity imaging W.E. Svensson, N. Zaman, N.K. Barrett, G. Ralleigh, K. Satchithananda, S. Comitis, V. Gada, J. Dunn, C. Vajdi; London/UK ([email protected]) Purpose: To report the ﬁrst four years of research and subsequent two years of routine use of elasticity imaging in breast ultrasound. Methods and Materials: Patients appearing for routine breast ultrasound for focal abnormalities had ultrasound strain imaging on a Siemens Elegra (research period - ﬁrst four years) and on a Siemens Antares (audit period - last eighteen months). The strain images are assessed and correlated with subsequent histology or follow-up. Results: 930 strain imaging cases have been collected to date (research 334, audit 596). 237 cancers, 35 intermediate grade lesions, 584 benign pathologies. 74 cases currently have incomplete data. Excluding cysts, lymph nodes and normal breast tissue, using an elasticity:B-mode ratio 0.75 as an indicator for cancers and intermediate lesions, an audit of the Antares-examined lesions over the last eighteen months shows sensitivity=97.5%, speciﬁcity=48%; this compares with earlier research on the Elegra with sensitivity=96.3%, speciﬁcity=53.3% (previously reported at ECR2006). False negatives include two cases each of ADH and DCIS and one radial scar. Four phylloides, six mucinous cancers and 4 pure DCIS were true positives. During the last 18 months (Antares) the number of cancers with an elasticity:B-mode ratio between 1 and 0.75 dropped from 16 to 3.8%. Conclusion: An elasticity:B-mode size ratio 0.75 is still a good indicator to exclude a diagnosis of cancer with the potential to halve benign biopsy rates. Increasing experience with, and improvements in, elasticity imaging are leading to improved diagnostic accuracy.

14:09

Ultrasound elastography: Results of a European multicentric study of 429 breast lesions F. Schafer1, A. Athanasiou2, I. Leconte3, R. Salvador4, S. Ganau Macias5, M. Wallis6, C. Plancher2, A.A. Tardivon2; 1Kiel/DE, 2Paris/FR, 3Brussels/BE, 4 Barcelona/ES, 5Sabadell/ES, 6Cambridge/UK ([email protected])

14:18

Quantitative assessment of breast lesions’ visco-elasticity: Initial clinical experience using supersonic shear wave imaging A. Athanasiou1, A. Tardivon1, M. Tanter1, B. Sigal-Zafrani1, T. Defﬁeux1, J. Bercoff2, J.-L. Gennisson1, M. Fink1, S. Neuenschwander1; 1Paris/FR, 2Aix-en-Provence/FR ([email protected]) Purpose: To prospectively evaluate the accuracy of sonoelastography using Supersonic Shear Wave Imaging technique in the characterization of breast lesions. Methods and Materials: Thirty-six nodules (4-22 mm, 25 benign and 11 malignant lesions classiﬁed as BIRADS category 3=6, category 4=21, category 5=7, category 6=2) in 34 consecutive patients (42-71 years, mean: 59 y) were included with cytological (n=10) and/or histological (n=26) correlations. B-mode ultrasound (US) and Supersonic Shear Wave sonoelastography (innovative technique based on the combination of a remote palpation induced in tissues by the radiation force induced by the ultrasonic beam and an ultrafast imaging sequence capable of catching in real time the propagation of the resulting shear waves) were performed during the same session (same conventional, linear probe). A quantitative lesion elasticity measurement was obtained in terms of Young’s modulus (in kilopascal, kPa units). Results: All breast nodules were detected by Supersonic Shear Wave sonoelastography. The 1 mm resolution facilitated the detection of small or isoechoic lesions. Elasticity values of malignant lesions (Young’s modulus=170p 41.6 kPa) were found to be signiﬁcantly different from those of benign ones (62p21 kPa). Due to different displacement rates of solid versus cystic material, this technique characterized correctly the cystic component of simple or complicated cysts at B-mode US. Conclusion: Supersonic Shear Wave Sonoelastography is an innovative, operatorindependent technique, insensible to patient movements, that enables a quantitative elasticity measurement. It could be a valuable complementary tool for characterizing benign versus malignant lesions.

B-164

14:27

Is sonoelastography helpful in BI-RADS III-IV (probably benign-probably malignant) lesions? N.M. Abdel Razek, L. Ezzat, M. Shaalan; Cairo/EG ([email protected]) Purpose: The goal of the study is to investigate the role of sonoelastography in evaluation of probably benign and probably malignant lesions (BI-RADS category 3 & 4) in sono-mammography and its contribution in patient managment. Methods and Materials: In this study 120 patients with BI-RADS 3 & 4 lesions in sono-mammography were investigated by sonoelastography (using the EUB 7500, Hitachi, Japan) and their elastography scores were analysed. All lesions were histopathologically proven after FNB, CNB, mammotome or operative biopsy. The

176

A

B

C

D

E

F

G

Scientific Sessions

B-165

14:36

Breast ultrasound elastography: Results of 193 breast lesions in a prospective study with histopathologic correlation F.K.W. Schaefer, S. Osterholz, B. Order, D. Fuerst, P.J. Schaefer, N. Hofheinz, J. Hedderich, W. Jonat, I. Schreer; Kiel/DE ([email protected]) Purpose: To evaluate the diagnostic performance of ultrasound elastography in breast masses. Methods and Materials: 193 lesions (129 benign, 64 malignant) were analyzed with the EUB 8500 Logos ultrasonic unit (Hitachi Medical, Japan) and a linear-array transducer of 7.5-13 MHz. Standard of reference was cytology (FNA) or histology (core-biopsy). The elastic score was classiﬁed according a 6-point color scale (Ueno classiﬁcation, 1-3=benign, 4-5=malignant). Conventional B-mode-US ﬁndings were classiﬁed according to the BI-RADS classiﬁcation. Statistical analysis included sensitivity, speciﬁcity, ROC-analysis and kappa-values for intra- and interobserver reliability. Results: For malignant lesions, the mean score in elasticity was 4.1p0.9, for benign lesions 2.1p1.0 (p 0.001). With a best cut-off-point between score 3 and 4 in elastography, sensitivity was 96.9%, speciﬁcity 76%. When a best cut-off-point for conventional US was set between BI-RADS 4 and 5, sensitivity was 57.8%, speciﬁcity 96.1%. Elastography had higher sensitivity, and lower speciﬁcity than conventional US, but for two lesions of elastic-score 1 were false negative whereas no lesion that scored BI-RADS 1-3 was false negative. ROC-curve for elastography was 0.884 (p 0.001) and for conventional US 0.820 (p 0.001). Weighted kappavalue for intra-/interobserver reliability for BI-RADS-classiﬁcation was 0.784/0.634 and for elasticity scores 0.720/0.561. Conclusion: Elastography has no potential to replace conventional B-mode-US for the detection of breast cancer, but can complement conventional US to improve the diagnostic performance.

B-166

14:45

Ultrasound (US) elastography of breast lesions associated with mammographically detected suspicious microcalciﬁcations J. Choi, N. Cho, W.K. Moon; Seoul/KR ([email protected]) Purpose: To evaluate the difference in strain between US lesions associated with malignant and benign microcalciﬁcations detected at screening mammography by using US elastography. Methods and Materials: 50 consecutive women who had been scheduled to undergo US-guided vacuum assisted biopsy due to suspicious microcalciﬁcation cluster (larger than 1.0 cm) detected at screening mammography were examined with a commercialized US elastography. A total of 50 lesions (25 DCIS lesions and 25 ﬁbrocystic changes) were found. Real-time imaging ﬁles were saved as video clips in avi format. Two experienced radiologists who had not performed the examinations analyzed in consensus the randomly ordered video clips without knowledge of the histology or mammographic ﬁndings and provided the elasticity score (1-5) according to the degree of strain induced by light compression. Results: For the elasticity score, the mean was 3.0p1.1 for malignant lesions and 1.3p0.5 for benign lesions (p 0.001). When a cutoff point between 1 and 2 was used, elastography had 96% sensitivity, 72% speciﬁcity, 77% PPV, and 95% NPV. When a cutoff point between 2 and 3 was used, elastography had 64% sensitivity, 100% speciﬁcity, 100% PPV, and 74% NPV. Mean area under the ROC curve was 0.930 (0.860-1.0, 95% conﬁdence interval). Conclusion: Breast lesions associated with malignant microcalciﬁcations tend to have less strain than benign microcalciﬁcations at US elastography. US elastography has the potential to improve the accuracy of gray-scale US for the detection and differentiation of breast lesions associated with screening mammographically detected suspicious microcalciﬁcations.

B-167

14:54

Computer-aided classiﬁcation of breast masses detected at ultrasound according to ACR BI-RADS S.-Y. Kim1, R.-F. Chang2, W.-C. Shen3, W.K. Moon1; 1Seoul/KR, 2Taipei/CN, 3 Chiayi/CN ([email protected]) Purpose: To quantify the US characteristics of breast masses deﬁned in the ACR BI-RADS and to evaluate the accuracy of the CAD system for the classiﬁcations of benign and malignant breast tumors. Methods and Materials: Digital US images of 265 pathology-proven cases including 180 benign and 85 malignant masses were used. Eight US features including shape, orientation, margin (angular and undulation characteristics), lesion boundary, echo pattern (the average gray intensity and the variation), and posterior acoustic features were computed and evaluated by the point-biserial correlation coefﬁcient (r). A CAD system was constructed to integrate all proposed features for predicting the likelihood of malignancy by the binary logistic regression model. Results: On each proposed feature, the mean values of malignant tumors were signiﬁcantly different from benign tumors (p 0.05 for all eight features). The correlation between the features and pathological result showed the highest value in angular characteristic of margin (r=0.7) followed by undulation (r=0.56), shape (r=0.56), orientation (r=0.12), posterior acoustic features (r=-0.32), the average gray intensity (r=-0.35), the variation of the echoes (r=-0.46) and lesion boundary (r=-0.57). For the constructed CAD system, the performance indices, i.e., accuracy, sensitivity, speciﬁcity, PPV, and NPV were 91.7 (243 of 265), 90.6 (77 of 85), 92.2 (166 of 180), 84.6 (77 of 91), and 95.4% (166 of 174), respectively, with an Az value of 0.97. Conclusion: Among the quantiﬁed BI-RADS US features, margin and shape of breast tumors were most important for predicting the likelihood of malignancy by CAD systems.

B-168

15:03

Slip imaging: Reducing ambiguity in elasticity imaging of breast lesions R.F. Adams, M.J. Kadour, R.E. English, V. Parulekar, S. Christopher, J.A. Noble; Oxford/UK ([email protected]) Purpose: To assess how including images of slip (mobility) with elasticity images reduces the ambiguity between benign and malignant lesions and leads to more accurate diagnoses of breast cancer. Methods and Materials: Ultrasound, elasticity and slip images were acquired in 70 women with deﬁnitively diagnosed breast lesions (32 malignant, 38 benign). Slip imaging is a new technique combining the ultrasound and elasticity data to create an image of tissue mobility, speciﬁcally the degree to which one region of tissue moves independently from another. Lesions suspected at mammographic screening and ultrasound assessment were included. However, lesions appearing cystic were only included if they were small ( 15 mm), complex, or challenging to diagnose. The images were evaluated independently by three breast radiologists. Results: Diagnostic accuracy was improved with the addition of slip images, from 75.7% with ultrasound alone, to 78.1% with elasticity imaging, to 80.0% including slip. Speciﬁcity also increased from 74.6 to 75.4 to 82.5%, respectively. The addition of slip images appeared to trade-off sensitivity for speciﬁcity, resulting in a change in sensitivity from 77.1 to 81.3 to 77.1%. Crucially for reducing benign biopsies, false-positives were reduced with elasticity imaging by 3.4%, but by 31.4% with the addition of slip images. Interobserver agreement was also further improved with slip imaging, from 60 to 75.7 to 81.4%. Conclusion: Slip imaging improves the identiﬁcation of benign characteristics, correctly classifying some solid benign lesions which appear malignant in elasticity imaging, thus increasing diagnostic accuracy and the potential for reducing benign biopsies.

B-169

15:12

Initial evaluation of a semi-automated whole-breast ultrasound system in 6244 studies in 4369 asymptomatic women K.M. Kelly1, J.C. Dean2; 1Venice, CA/US, 2Santa Barbara, CA/US ([email protected]) Purpose: Creation and initial evaluation of a Semi-Automated Whole-Breast Ultrasound (SAWBU) System. Methods and Materials: A computer controlled SAWBU system was devised to gather and permanently store a ciné loop of sub-millimeter-spaced compound 2-D ultrasound images from parallel overlapping craniocaudad rows of the entirety of both breasts. This system allows the operator to control the pressure and the angle of incidence of the transducer to maintain the appearance of hand scanning. The ciné loop is evaluated on a 2048x1536 pixel monitor using a computer interface

A

B

C

D

E

F

G

177

Friday

elastography scores and results were correlated to the histopathology. Results: A total of 120 patients participated in the study. Their mean age was 48 yrs. 35% were histopathologically malignant lesions (including DCIS, lobular carcinoma, invasive duct carcinoma and medullary carcinoma). 65% were proven benign lesions icluding ﬁbroadenomata, cysts, scars and papillomas as also sclerosing adenosis. Sonoelastography has increased the sensitivity and speciﬁcity of sono-mammography from 85 and 78% to 95 and 99%; however, as a sole diagnostic test, it has a sensitivity of 66% compared to a sensitivity of 58% of US in diagnosis of BI-RADS 3 & 4 lesions. The difference is not signiﬁcant. It was found that the poor results were obtained with large lesions. Conclusion: In evaluation of BI-RADS 3 & 4 lesions, sonoelastography offers additional information that increased the sensitivity and speciﬁcity of sonomammography in evaluating these probable lesions. The accuracy of the test is increased in small lesions.

Scientific Sessions which allows the reader control of the brightness, contrast, gamma, size and review speed. 3-D reconstructions of areas of interest can be viewed from any perspective. The high-resolution monitor produces smaller, sharper images that ﬁt in the reader’s ﬁeld-of-view without eye movement. Results: Forty-two cancers were found in 6244 asymptomatic, mostly densebreasted women who had both screening mammography and SAWBU. Six of the cancers were found by mammography alone, 22 by SAWBU and 14 by both. Seventeen of the cancers seen only by SAWBU were Stage 1. Eight additional interval cancers presented in less than one year. Of these, six were retrospectively visible by the SAWBU. Conclusion: Properly performed, SAWBU will ﬁnd signiﬁcantly more cancers than those seen by mammographic screening alone in asymptomatic women with radiographically dense breasts.

B-170

15:21

A dedicated CT scanner for breast imaging J.M. Boone, K. Yang, N. Packard, S.-Y. Huang, K.K. Lindfors, W. Monsky; Sacramento, CA/US ([email protected]) Purpose: While mammography is the mainstay of breast cancer screening, more advanced imaging technologies may contribute to better cancer detection, staging, and intervention. Methods and Materials: A prototype breast CT scanner has been designed, fabricated, tested, and evaluated for patient imaging. Off the shelf components, including an X-ray source, ﬂat-panel detector, and motor were utilized; all other components of the breast CT scanner were designed and integrated in our laboratory. The radiation dose levels were adjusted by mA selection (at 80 kV) to be equivalent to two-view mammography, the standard in the United States. Image quality metrics including the MTF and NPS were evaluated. Results: The spatial resolution of this cone beam CT scanner was found to be superior to commercial whole body CT scanners, with a cut-off resolution of approximately 1.8 1/mm. The noise properties of the scanner were reasonably similar to that of commercial systems. As of this time, over 130 patients have been scanned. The tomographic images have isotropic voxel dimensions of approximately 0.25 mm. Conclusion: The good spatial resolution and excellent contrast resolution of breast CT was found to deliver superior mass lesion detection but slightly inferior detection of microcalciﬁcations. The breast CT scanner may play an important role in the diagnosis of breast cancer, possibly as a secondary tool in the breast imaging clinic. It also may provide a solid platform for image guided biopsy, and minimally invasive breast cancer treatment.

14:00 - 15:30

Room G/H

Head and Neck

SS 308 Tumors: Metastases and imaging in therapy Moderators: H.B. Eggesbø; Oslo/NO P. Piñero; Sevilla/ES

B-171

14:00

Level-based lymph node staging in head and neck cancer: Value of repeated CME-based training C.R. Habermann, F. Andree, A. Koops, M.C. Cramer, G. Adam, G. Krupski-Berdien; Hamburg/DE ([email protected]) Purpose: To determine the value of repeated CME-based training focused on level-based lymph node staging (LBLS) in patients with head and neck cancer using computed tomography. Methods and Materials: In a university radiological department, a refresher course focused on LBLS was repeated twice within six months. The courses were held according to CME requirements with compulsory attendance. CTs of the neck were grouped into twenty and evaluated a month prior to the ﬁrst lecture, directly, two and four months after the ﬁrst lecture, and directly and two months after the second lecture by three independent readers. The results were compared with the radiological reports performed in the daily routine regarding the use of LBLS and the accuracy. Results: Correlation among the observers was high (r=0.845). In none of the routine reports prior to the ﬁrst course were the LBLS used. Directly after the ﬁrst course, the allocation to a level was performed in 26.1% in enlarged nodes and in

178

A

B

C

D

E

F

G

13.3% of normal sized nodes. Two and four months later, the use of level classiﬁcation decreased slightly (p=0.56, 0.63). After the second course, the use of level classiﬁcation increased signiﬁcantly to 85.7% in enlarged nodes and in 35% of no enlarged nodes (p 0.001). Another two months later, a signiﬁcant decrease was observed (42.9, 20%, respectively; p=0.03, 0.037). Conclusion: CME signiﬁcantly improves the knowledge and leads to an implementation in daily clinical routine. To enhance the performance, repeated teaching in a particular topic is indispensable.

B-172

14:09

Diffusion-weighted MRI for nodal staging in patients with head and neck squamous cell carcinoma: A comparison with conventional MRI in correlation to histopathology V. Vandecaveye, F. De Keyzer, V. Vander Poorten, E. Verbeken, P. Dirix, S. Nuyts, R. Hermans; Leuven/BE ([email protected]) Purpose: Diffusion-weighted magnetic resonance imaging (DW-MRI) allows tissue characterization based on changes of tissue water properties. The purpose of this study was to compare DW-MRI with conventional MRI for detection of nodal metastases in head and neck squamous cell carcinoma (HNSCC). Methods and Materials: Prior to surgery, 33 patients underwent a 1.5 Tesla MRI-examination consisting of conventional MRI-sequences and an echo-planar DW-MRI (b-value range from 0 to 1000 sec/mm2). The apparent diffusion coefﬁcient (ADC) of the lymph nodes (LN) was calculated for the entire b-value range. For histopathological evaluation, the LN were sectioned in 3-mm slices and were evaluated using a prekeratine-immunostaining. Subsequently, an optimal threshold for the ADC-value between metastatic and benign LN was determined, based on ADC-values in tumoral and non-tumoral LN. With this threshold, sensitivity and speciﬁcity of DW-MRI were calculated per LN and per neck-level and compared to conventional MRI-sequences. Results: Historadiological correlation was possible in 301 LN, of which 87% were smaller than 1 cm. DW-MRI yielded signiﬁcantly lower ADC for metastatic LN than for benign LN and showed a sensitivity of 84% and speciﬁcity of 94% per LN and a sensitivity of 94% and speciﬁcity of 97% per level using a threshold of 0.00094 mm2/sec. TSE-sequences yielded a sensitivity of 46% and speciﬁcity of 96% per LN and a sensitivity of 57% and speciﬁcity of 95% per level. Conclusion: DW-MRI showed higher sensitivity for detection of nodal metastases than conventional MRI in patients with HNSCC.

B-173

14:18

New MRI criteria improve the detection of lymph node metastases in head and neck squamous cell carcinoma: Multivariate logistic regression analysis of MRI features of cervical lymph nodes R.B.J. de Bondt, P.J. Nelemans, F. Bakers, C. Peutz-Kootstra, B. Kremer, P.A.M. Hofman, J.M.A. van Engelshoven, R.G.H. Beeets-Tan; Maastricht/NL ([email protected]) Purpose: New MRI criteria were evaluated to assess the contribution to improvement of the detection of cervical lymph node metastases. Methods and Materials: The local medical ethics committee approved the study. Forty-four consecutive patients (11 women, 33 men, mean age 61.2 years [range: 40-86 years]) with head and neck squamous cell carcinoma (HNSCC) were evaluated by MR imaging. Lymph node characteristics were assessed by two radiologists on size and three new criteria: (1) border irregularity, (2) heterogeneity of signal intensity (SI) on contrast enhanced (CE) T1-WI and (3) heterogeneity of SI on T2-WI. Histological results were the reference standard. Multivariate logistic regression analysis was performed and diagnostic odds ratios (DORs) with 95% conﬁdence intervals (95% CI) and areas under the curve (AUCs) of receiver operating characteristic (ROC) curves were used to determine added diagnostic value of the new criteria. Results: The DORs derived from multivariate analysis associated with border irregularity were 2.61 (95% CI: 1.12-6.08) and 66.2 (95% CI: 20.4-217) for observer 1 and 2, respectively; for heterogeneity on T2-WI, DORs were 2.97 (95% CI: 1.42-6.18) and 22.6 (95% CI: 6.40-80.1). AUCs increased from 0.67 (95% CI: 0.61-0.73) using size only to 0.81 (95% CI: 0.75-0.87) using all four criteria for observer 1 and from 0.68 (95% CI: 0.62-0.74) to 0.96 (95% CI: 0.94-0.98) for observer 2 (p 0.001). Conclusion: New morphologic criteria like border irregularity and heterogeneity of SI on T2-WI in addition to size improved the detection of cervical lymph nodes metastases on MRI in HNSCC.

Scientific Sessions B-174

14:27

Malignant and reactive cervical lymph nodes in computed tomography perfusion (CTP) examination: Can we differentiate them? A. Trojanowska, A. Drop, P. Trojanowski, G. Staskiewicz, L. Grzycka-Kowalczyk, K. Rosinska-Bogusiewicz; Lublin/PL ([email protected])

B-175

14:36

Value of diffusion-weighted MR imaging in the differentiation between benign and malignant cervical lymph nodes K. Holzapfel, S. Duetsch, E.J. Rummeny, C. Fauser, J. Gaa; Munich/DE ([email protected]) Purpose: To evaluate the usefulness of diffusion-weighted MR imaging (DWI) in the differentiation between benign and malignant cervical lymph nodes. Methods and Materials: 45 patients with 65 enlarged ( 10 mm) cervical lymph nodes (metastases from squamous cell carcinoma [n=30], non-Hodgkin lymphoma [n=5], reactive lymphadenitis [n=27], cat scratch lymphadenitis [n=2] and sarcoidosis [n=1]) underwent MR imaging at 1.5 T (Avanto, Siemens) using a head and/or neck circular polarization surface coil. DWI was performed by a single-shot echoplanar MR imaging (SSEPI) sequence (TR 4500 ms, TE 97 ms, matrix 192x144, slice thickness 5 mm, GRAPPA2, b-values 0, 500 and 1000 sec/mm2). Apparent diffusion coefﬁcient (ADC) maps were reconstructed for all patients and ADC values were calculated for each lymph node. Imaging results were correlated with histopathologic ﬁndings after neck dissection or surgical biopsy, ﬁndings in PET-CT or imaging follow-up. Mann-Whitney test was used for statistical analysis and a receiver operating characteristic (ROC) curve analysis was performed. Results: The mean ADC values (x10-3 mm2/sec) were 0.78p0.11 for metastatic lymph nodes, 0.64p0.10 for lymphomatous nodes and 1.24p0.16 for benign cervical lymph nodes. ADC values of malignant lymph nodes were signiﬁcantly lower than ADC values of benign lymph nodes. 95.5% of lesions were correctly classiﬁed as benign or malignant using a threshold ADC value of 0.92x10-3 mm2/sec. Conclusion: Diffusion-weighted MR imaging using a SSEPI sequence allows reliable differentiation between benign and malignant enlarged cervical lymph nodes.

B-176

14:45

Computed tomography perfusion (CTP) studies are useless in the determination of bone inﬁltration by malignant tumors: True or false? A. Trojanowska, A. Drop, P. Trojanowski, J. Klatka, K. Rosinska-Bogusiewicz; Lublin/PL ([email protected]) Purpose: To evaluate if CTP examination would be a valuable tool in determining bone inﬁltration by a malignant tumor. Methods and Materials: We examined 18 patients with oral cavity squamous cell carcinoma (SCC) adjacent to the mandible. In all patients we performed standard contrast-enhanced computed tomography (CECT) of head and neck region (100 ml of contrast medium, ﬂow 1 ms/s, delay 100 s) in 1.25 mm slices, reconstruction kernels - bone and soft tissue. Afterwards CTP examination was performed at the level of tumor, in 2.5 mm slices. Basic perfusion parameters were obtained (rBF, rBV, MTT, PS) and perfusion maps were automatically created. All patients underwent surgery ca 10 days after imaging, with fragmental, marginal, or no mandible

B-177

14:54

Comparison of F18-FDG PET/CT scan and bone scintigraphy in detecting bone metastasis in head and neck tumors N. Al-Bulushi1, H. Almeer1, A. Alrifai1, A. Alsugair1, M. Abouzied2; 1Riyadh/SA, 2 Buffalo, NY/US ([email protected]) Purpose: To evaluate the efﬁcacy of FDG PET/CT and bone scan for the diagnosis of bone metastases in patients with head & neck cancer. Methods and Materials: 319 patients with head & neck cancer have been identiﬁed in our database that had 18 F FGD-PET/CT study. 156 patients (44 females and 12 males; age range 15-100 years) have met our inclusion criteria: biopsy-proven head and neck cancer, bone scan and PET/CT within 30 days. Comparison was done on a lesion-by-lesion analysis; MRI, MDCT, and the clinical course of the patients were our references. Results: PET/CT identiﬁed 213 lesions in 18 patients, in whom PET/CT also detected distant metastases in 12 patients involving liver, lung, adrenal glands and lymph nodes. Bone scan identiﬁed 198 lesions in 16 patients and missed 2 patients with biopsy- and radiology-conﬁrmed bone metastases. PET/CT was true positive in 18 patients, and true negative in 138 patients, while bone scan was true positive in 16 and true negative in 134 patients. No false positives or false negatives were found with PET/CT. On the other hand, bone scan had 2 false positives and 2 false negatives. The overall sensitivity, speciﬁcity and accuracy of PET/CT and bone scan were 100 vs 88%,100 vs 98% and 100 vs 96%, respectively. Conclusion: PET/CT is a powerful tool in assessing the visceral as well as the skeletal metastases. It is not necessary to add bone scan to PET/CT for the staging purposes of head and neck carcinoma.

B-178

15:03

Accuracy of PETCT in diagnosis and exclusion of recurrent head and neck squamous cell carcinoma A. Nasoodi, R. De Freitas, C. Fahy, B. Primrose, S. Hughes; Belfast/UK ([email protected]) Purpose: Head and neck squamous cell carcinoma (HNSCC) is the commonest head and neck cancer. We evaluate the accuracy of PETCT in both the diagnosis and the exclusion of HNSCC recurrence. Methods and Materials: A retrospective study was conducted on treated HNSCC patients with clinical suspicion of recurrence who underwent PETCT scanning between 11/2002 and 09/2004. Patient notes were reviewed to establish the sensitivity and speciﬁcity of the PETCT ﬁndings. The gold standard was pathologically conﬁrmed recurrence, obvious clinical progression or no clinical progression after 18 months follow-up. Results: A total of 45 scans in 40 patients were divided into those showing recurrence (positive PETCT group) and those showing a normal appearance (negative PETCT group). The mean interval from completion of treatment to scan date was 15.0 months (3-44 months) in the positive PETCT group and 15.6 months (3-73 months) in the negative PETCT group. In the negative PETCT group the mean duration of follow-up after a scan was 23.1 months. There were 21 positive scans in 19 patients in the positive PETCT group. There were no false positives. There were 24 negative scans in 21 patients in the negative PETCT group, including two false negative scans in a single patient. In diagnosing recurrence of HNSCC, PETCT sensitivity was 91.3%, speciﬁcity was 100%, the positive predictive value was 100% and the negative predictive value was 91.6%. Conclusion: In our series, a positive scan always represented a true recurrence, and in almost all cases a negative scan excluded recurrence.

A

B

C

D

E

F

G

179

Friday

Purpose: Last year we presented initial results concerning differentiation between malignant and non-malignant cervical lymph nodes based on computed tomography perfusion (CTP) studies. The aim of the present paper is to prove on a larger population of patients that CTP is a reliable and sensitive method for depicting malignant nodes. Methods and Materials: Forty-one patients, with head and neck squamous cell carcinoma (SCC) of different origins (ﬂoor of mouth, nasopharynx, oral tongue, larynx) and suspected for lymph node metastases, underwent standard contrastenhanced computed tomography examination followed by CTP (8 cm coverage) at the level of most suspected lymph nodes. Based on CTP basic parameters (rBF, rBV, MTT, PS), suspected lymph nodes were depicted. Finally, 36 patients underwent operation and a total number of 108 nodes were separately removed and sent for detailed histological analysis. Results of CTP examination were compared with histology; sensitivity and speciﬁcity of the method was caclulated. Results: The results conﬁrmed our earlier ﬁndings that there was a distinctive difference between malignant and reactive lymph nodes in terms of CTP values. Malignant lymph nodes showed remarkable hyperperfusion (rBF=148 l/100 g/min, rBV=5.2 ml/100 g, MTT=4.8 s) in comparison to reactive nodes (rBF=88 ml/100 g/min, rBV=3.1 ml/100 g, MTT=6.8 s) and this difference was statistically signiﬁcant (p 0.05). Conclusion: CTP is a feasible and widely available tool; as such, it may be used in differentiation between malignant and reactive lymph nodes, especially in doubtful cases.

removal. The extent of removal was decided based on standard CECT studies and intra-operative ﬁndings. Surgical specimens underwent histological analysis. Results were compared with CTP ﬁndings. Results: Nine patients underwent mandibular removal (six fragmental and three marginal). In eight cases inﬁltration was conﬁrmed on histology. In this group, only in two cases CTP ﬁndings were suggestive of mandibular invasion (elevated values of rBF, rBV and PS). Based on CECT studies it was possible to determine mandibular invasion in ﬁve cases (three medullary invasions and two cortical invasions), which was also conﬁrmed on histology. Conclusion: This study shows no evidence that CTP examination may be a useful adjunct in determination of possible mandibular inﬁltration by SCC. Conventional CECT and especially bone kernel reconstructions show better results in deﬁning cortical and medullary inﬁltration.

Scientific Sessions B-179

14:00 - 15:30

15:12

Tumor R2* as a prognostic indicator of acute radiotherapeutic response in human cervical cancer B. Wu, W. Peng, Y. Gu; Shanghai/CN ([email protected]) Purpose: To test the prognostic potential of tumor R2* with respect to radiotherapeutic sensitivity with blood oxygenation level dependent (BOLD) imaging using clinical 1.5 T MR scanner. Methods and Materials: BOLD imaging was performed in a cohort of 40 patients with cervical cancer. The baseline air induced R2* and carbogen (95% O2, 5% CO2) induced R2* of tumor were quantiﬁed using gradient echo MR sequence, recorded with ROI of whole tumor area in the maximal slice. The histogram of all R2* values of ROI was generated and a proper threshold of value was presumed. The areas of R2* value lower and higher than this threshold (Arealow and Areahigh) in the histogram were calculated automatically with volume analysis function in workstation, therefore came to the relative ratio Z of this two portion. The Z value was set with the equation: Z=Areahigh/Arealow. This procedure was performed repeatedly before and during therapy. With the same presumed threshold, we compared the ratios before and after therapy and correlated with subsequent tumor volume decrease in response to ionizing radiation. Results: All patients had good tolerance to carbogen breathing and BOLD MR imaging. There was a discrepancy in tumor R2* values between air- and carbogen gas breathing. The Z value also showed signiﬁcant discrepancy before and after therapy according the paired t-test (t=2.909, p 0.05). There was a linear correlation about Z value and tumor shrinkage volume. Conclusion: Quantitation of tumor R2* provides completely noninvasive prognostic indicator of a potential acute radiotherapeutic response.

B-180

15:21

Usefulness of magnetic resonance volumetric evaluation in predicting response to induction chemotherapy in patients with advanced head and neck cancers M.G. Mack, M. Baghi, R. Knecht, M. Hambek, S. Bisdas, T. Vogl; Frankfurt a. Main/DE ([email protected]) Purpose: To analyze the value of magnetic resonance volumetry (MRV) before neoadjuvant chemotherapy to predict the response to therapy in advanced head and neck cancers. Methods and Material: 50 patients with advanced head and neck cancer who underwent MR volumetry pre and post neoadjuvant chemotherapy consisting of docetaxel, cisplatin and 5-Fluorouracil (TPF) were included in this study. The tumor outlines were traced and digitized on each pretreatment MR slice for each of the 50 patients. Considering the magniﬁcation factor of the scan and MR slice thickness, the tumor volume was calculated in cubic centimeters (cc). To evaluate the efﬁcacy of volumetry, pre treatment volume has been compared with remission. Results: 45 (90%) patients demonstrated a tumor down-staging after chemotherapy. 14 (28%) patients showed a complete histologic remission (CR), 31 (62%) patients showed a partial remission (PR). Pre and post treatment MR tumor volumes were signiﬁcantly less in patients whose tumor were completely down-staged (CR) than in patients that were not completely down-staged or even showed no change (NC) or progression (PD) after chemotherapy (p 0.004). Pretreatment tumor volume of patients was signiﬁcantly different between patients whose tumor was completely down-staged (CR) and those that were not completely down-staged, were stable or progressive (p 0.00023). The median pretreatment tumor volume for patients with CR was 11.89 cc. Conclusion: We conclude that MRI tumor volume analyses can be a useful parameter to predict the response to neoadjuvant chemotherapy in SCCHN.

180

A

B

C

D

E

F

G

Room I

Vascular

SS 315 CT angiography Moderators: L. Boyer; Clermont Ferrand/FR J. Ferda; Plzen/CZ

B-181

14:00

Comparison of different concentration of contrast medium for aortic enhancement D. Wu; Shanghai/CN ([email protected]) Purpose: To prospectively evaluate moderate versus high concentration of contrast medium for aortic enhancement. Methods and Materials: 20 male volunteers without cardiovascular disease were studied, randomly delivered into 4 groups with different protocols, each containing ﬁve persons. Contrast medium was administered via antecubital vein with moderate (300 mg iodine per ml) or high (370 mg iodine per ml) iodine concentration using a power injector. Group 1 A, moderate contrast medium of 90 ml was injected at 3 ml/sec; Group 1B, high contrast medium of 73 ml was injected at 3 ml/sec; Group 2 A, moderate contrast medium of 90 ml was injected at 5 ml/sec; and Group 2B, high contrast medium of 73 ml was injected at 5 ml/sec. Time-density curve of aorta was taken using same-slice-dynamic-scan at the lever of ﬁrst hepatic hilum. The enhancement peak (Pmax) and time-to-peak (Tmax) of aorta was calculated. Statistical analysis was performed with SPSS (T-test). Results: The Pmax were 250p22.1, 207p18.7, 395p13.6 and 303p24.7 Hu for protocols 1 A, 2 A, 1B and 2B, respectively. The t values were 3.12 and 6.67 intragroup A and B, respectively, and the p values 0.014 and 0.0002, respectively. The Tmax were 32p2.8, 27p1.8, 28p2.8 and 22p2.2 sec for protocols 1 A, 2 A, 1B and 2B, respectively. The t values were 3.21 and 4.00 intragroup A and B, respectively, and the p values 0.012 and 0.0039, respectively. Conclusion: When total iodine dose and injection rate were ﬁxed, higher concentration of contrast medium may not increase the enhancement peak of aorta than was moderate concentration of contrast medium.

B-182

14:09

Multi-detector CT angiography of the aorta: Do we beneﬁt from high concentration iodine and bolus tracking? A.S. Ernst, P. Friedrich, M. Jeltsch, M.H.K. Hoffmann, H.-J. Brambs, A.J. Aschoff; Ulm/DE ([email protected]) Purpose: To evaluate the beneﬁt of a contrast medium with high-concentration iodine for Multi-Detector CT Angiography (MDCTA) of the aorta using automatic bolus tracking. Methods and Materials: 100 consecutive MDCTA examinations of the entire aorta including pelvic axis were evaluated. Scans were performed on a 40 channel MDCT scanner (Philips Brilliance 40). A contrast medium with a high iodine concentration of 400 mg/ml (Imeron 400, Bracco) was administered by power injector; a saline chaser followed. Scanning was initiated by automated bolus tracking. Time-density distributions calculated from the ascending aorta downstream to the common femoral arteries were graded as from poor to excellent based on contrast density, homogeneity, and timing of the scan acquisition. Results: Evaluated aortic lesions included aneurysms (n=62) and aortic dissections (n=30). Mean (pSD) overall arterial density of MDCTA was 306p66 HU (range: 129-476 HU). Averaged arterial contrast density over the complete scan range was rated excellent in 73 cases ( 350 HU; n=25, 276-350 HU; n=48) while contrast homogeneity was rated excellent in 49 cases and good in 22 cases. Automatic bolus tracking permitted accurate scanning during peak contrast enhancement in 87 cases. The 13 mistimed examinations comprising 12 MDCTA studies started early and one started late. In 8 out of 12 early scans the contrast density was still considered excellent ( 350 HU; n=1) or very good (276-350 HU; n=7). Conclusion: Combination of high concentration contrast medium and automatic bolus tracking allows reliable MDCTA of the entire aorta with excellent and homogenous contrast distribution in different aortic pathologies.

Scientific Sessions B-183

14:18

Determination of optimal scan delay for routine abdominal dual source CT with body weight adapted application of contrast material A. Tschugunow, K. Juergens, M. Puesken, W.L. Heindel, J. Wessling; Muenster/DE ([email protected])

B-184

14:27

Peripheral multi-detector CT angiography (MDCTA): Inﬂuence of a compact bolus injection at high iodine concentration on arterial contrast enhancement and image quality T. Albrecht1, B.C. Meyer1, A.J. Aschoff2, J. Skrok1, T. Nentwig1, B.L. Schmitz2, K.-J. Wolf1, S. Klein2; 1Berlin/DE, 2Ulm/DE ([email protected]) Purpose: A compact bolus injection can increase arterial enhancement in peripheral MDCTA, but timing is more critical to avoid bolus overriding or venous overlap. This study compared a compact bolus injection protocol utilizing Iomeprol-400 (mg I/mL) with a standard injection protocol using Iomeprol-300 at equivalent total iodine dose for peripheral MDCTA. Methods and Materials: Sixty-four patients randomized to receive 134 mL Iomeprol-300 (n=32; 40.2 gI; 1.2 gI/s) or 100 mL Iomeprol-400 (n=32; 40.0 gI; 1.6 gI/s) at 4 mL/sec (injection duration: 33.5 and 25 s, respectively) underwent peripheral MDCTA on 40- or 64-slice scanners (table speed 48 mm/s) using bolus-tracking. Arterial contrast density (HU) was measured intermittently from the abdominal aorta to the plantar arteries; vascular opaciﬁcation was assessed qualitatively from the pelvic arteries to the feet. Cases of venous overlap or bolus overriding were recorded. Results: Arterial contrast density was higher with Iomeprol-400 in all arterial segments analysed; mean contrast density from the aorta to the plantar arteries was signiﬁcantly (p 0.0017) higher with Iomeprol-400 (279Âp57 HU vs 234Âp32 HU). A trend towards better qualitative assessment of arterial opaciﬁcation was noted for Iomeprol-400. Venous overlap occurred more frequently with Iomeprol-300 (18/32 patients) than with Iomeprol-400 (11/32) but did not interfere with image interpretation. Slight overriding of the contrast bolus was seen in one (Iomeprol300) and two (Iomeprol-400) patients. Conclusion: The more compact bolus achieved with Iomeprol-400 compared to Iomeprol 300 at identical injection rates and iodine doses leads to greater arterial contrast enhancement and fewer cases of venous contamination without increased risk of bolus overriding in peripheral MDCTA.

B-185

14:36

Identiﬁcation of vulnerable plaque: A clinical, radiological and surgical study L. Saba, R. Sanﬁlippo, G. Caddeo, R. Montisci, G. Mallarini; Cagliari/IT ([email protected]) Purpose: The so-called “Vulnerable plaque” identiﬁes a plaque at high risk of rupture. Our purpose was to evaluate MDTCA plaque characteristics of vulnerability in a cohort of 156 patients who underwent carotid endarterectomy studied using MDCTA. Methods and Materials: Between January 2004 and March 2007, 156 patients

B-186

14:45

Evaluation of symptomatic atherosclerotic carotid plaques: Relationship between MDCT-assessed plaque volume and brain infarcts P.J. Homburg, S. Rozie, T.T. de Weert, C. de Monyé, D.W.J. Dippel, A. van der Lugt; Rotterdam/NL ([email protected]) Purpose: Atherosclerotic plaque volume (PV) can be assessed by multidetector computed tomography angiography (MDCTA). We hypothesize that PV is related to the presence and type of brain infarctions on CT. Methods and Materials: We studied 100 patients who had cerebrovascular symptoms in the carotid artery territory with MDCTA. Patients with a likely cardioembolic stroke etiology were excluded. CT brain images were reviewed for the presence of recent and old infarcts. Infarcts were subdivided into cortical and lacunar infarcts. We assessed PV with a custom-made software tool. Severity of stenosis on MDCTA was measured according to the NASCET criteria. Mann-Whitney U test was applied for statistical analysis. Results: Atherosclerotic disease was present in 62%. Brain infarcts were present in 49%. PV was signiﬁcantly larger in the symptomatic carotid artery of patients with a brain infarct than in patients without an infarct (p=0.001). PV was not signiﬁcantly larger in the symptomatic carotid artery of patients with a cortical infarct than in patients with a lacunar infarct (p=0.09). Stenosis was more severe in patients with than without infarcts (p=0.001). There was no signiﬁcant difference in severity of stenosis between patients with cortical and lacunar infarctions (p=0.11). Conclusion: The severity of atherosclerotic disease (PV and stenosis) in the symptomatic carotid artery of patients is related to the presence of brain infarcts. No relation was found with the type of brain infarct. Measurement of PV may prove to be a useful diagnostic tool in the evaluation of stroke risk.

B-187

14:54

Comparative study between 64 multi-detector row CT spinal angiography and DSA in diagnosing spinal cord vascular malformation Y.G. Gao, K.L. Li, X.D. Du, Y.Y. Yang, S. Shen; Beijing/CN ([email protected]) Purpose: To evaluate the clinical application value of 64 multi-detector row CT spinal angiography in the patients of spinal cord vascular malformation. Methods and Materials: Six children were examined by digital subtraction angiograph and diagnosed with Cobb syndrome based on the segmental cutaneous hemangioma and perispondylic arteriovenous hemangiomas and spinal arteriovenous malformation (AVM). Four children were diagnosed with simple spinal arteriovenous malformation (AVM) by DSA. A subsequent spinal CT angiography by 64 multi-detector row CT (GE LightSpeed VCT) was done. The parameters of the CT scan protocol were: 64x0.625 collimation, pitch 0.8, 680 mAs, 120 kv, rotation time 0.35 s, mean scan time 15 s. Volume rendering, maximal intensity projection, multiplanar reformation were used to depict the lesion. The ﬁndings of CT angiography were compared with those of DSA. For the purpose of comparison of CTA with DSA in detecting the spinal arteries, statistical signiﬁcance was determined using C² test. Results: Spinal arteries that included vertebral artery, thyro-cervical trunk, costocervical trunk, intercostal arteries and lumbar arteries were revealed clearly to supply blood to the spinal AVMs, dermatome and paravertebral hemangiomas by CTA. There was no statistically signiﬁcant difference in detecting spinal arteries between CTA and DSA (p 0.05). Moreover, angioma of vertebral body that was difﬁcult to be revealed and located by DSA could be diagnosed by CT examination in Cobb syndrome.

A

B

C

D

E

F

G

181

Friday

Purpose: The considerable reduction of scan duration in Dual Source CT (DSCT) is associated with a higher risk of insufﬁcient vessel and parenchyma enhancement due to “off-time scan”. We prospectively determined an adequate scan delay for routine abdominal DSCT with body weight adapted contrast application. Patients and Methods: 56 patients underwent abdominal CT with a 64-Scanner (Deﬁnition, Siemens; 120 kV, 210 mAs; Care Dose®, 0.5 s rotation time, 24x1.2 mm, pitch 1.4). Contrast material was adapted to patients’ body weight (518 mg I/kg), ﬂow of 2.5 ml/s. All patients were prospectively randomized into four groups with varying scan delay and direction (group A: delay 75 s, cranio-caudal; group B: 85 s, cranio-caudal; group C: 85 s, caudo-cranial; group D: 95 s, cranio-caudal). Two blinded radiologists performed evaluation for diagnostic acceptability. CT values (HU) were obtained in different segments of the aorta, inferior vena cava, iliac veins, portal vein, hepatic veins and liver, spleen and pancreas. Statistical analysis was performed by using the Wilcoxon test. Results: Diagnostic acceptability of protocols B and C were rated equally good and signiﬁcantly/substantially superior to protocol A (p=0.004) and D (0.008), respectively. Contrast enhancement in the aorta and portal vein peaked at 75 s (154 and 167 HU, respectively). Contrast enhancement in the hepatic and iliac veins peaked at 85 s independently of scan direction but was substantially lower at 75 s. Liver parenchyma enhancement was lowest at 95 s. Conclusion: Our data suggest an optimal scan delay for routine abdominal DSCT of 85 s independently of scan direction.

were studied with MDCTA to assess carotid arteries and afterwards underwent carotid endarterectomy (CEA). The following features were studied: the type of the plaque (fatty, mixed and calciﬁed), plaque morphology (regular versus irregular) and presence of ulcerations; we also evaluated stenosis degree (according to the NASCET criteria). Plaque characterization and stenosis degree measurement were performed by two experienced radiologists in consensus. CEA were performed by two experienced vascular surgeons. We correlated MDCTA data with surgical results and with the presence/absence of ischemic events. Statistical analysis was used to determine if an interaction existed between presence of ischemic episodes and speciﬁc plaque characteristics as conﬁrmed by surgery. Results: An history of symptomatic ischemic episode was present in 98 patients (63%). We found a positive correlation between presence of ulcerations and symptomaticity (p 0.05), irregular morphology and symptomaticity (p 0.05) and presence of fatty plaque and symptomaticity. Concordance between MDCTA and surgical results was very high. Conclusion: Fatty plaques, ulcerations, and irregular plaque morphology are statistically signiﬁcant risk factors for ischemic episodes. These plaque characteristics individuate a high risk, “ vulnerable” plaque.

Scientific Sessions Conclusion: From the comparative study between CTA and DSA on examination of spinal feeding arteries, it was proved that CTA had a high sensitivity in revealing the feeding arteries for spinal cord vascular malformation.

B-188

15:03

The role of CT angiography in preoperative assessment of deep inferior epigastric perforator ﬂaps in breast reconstruction A. Dimopoulou, R. Acosta, A. Liss, A. Magnusson; Uppsala/SE ([email protected]) Purpose: The use of abdominal perforator ﬂaps has widely increased in breast reconstructive surgery. The purpose of this study was to visualize inferior epigastric artery perforators with computed tomography angiography (CTA) and achieve a detailed vascular mapping of deep inferior epigastric perforator (DIEP) ﬂaps. Methods and Materials: From March to December 2006, 51 breast cancer patients had breast reconstruction with a DIEP ﬂap. All patients had previously undergone CTA of the abdomen with focus on the inferior epigastric artery and its perforators in the rectus muscle. The CTA and 3D reconstructed images were then demonstrated for the plastic surgeons and the most suitable perforators were identiﬁed. Surgery time, complications and surgical failure were monitored for these patients, the control group being 51 breast cancer patients who had undergone the same operation without preoperative CTA assessment. Results: The mean surgery time in the CTA group was signiﬁcantly lower than in the control group with a p value of 0.001. Fewer complications (hematoma, infection, superﬁcial necrosis and revision of the anastomosis) occurred in the CTA group but the differences between the groups were not statistically signiﬁcant. In the CTA group all ﬂaps were successful. Conclusion: The use of CTA in the preoperative assessment of DIEP ﬂaps is essential, giving a detailed anatomical and vascular map of the area, helping the radiologist target the best perforator, and therefore enabling the surgeon to individually customize the ﬂap for each patient. It signiﬁcantly reduces surgery time and minimizes the number of complications.

B-189

15:12

Active bleeding with MDCT: Can we say more? A. Alcala-Galiano, P. Díez, P. Martín, A. Hilario, E. Salvador, E. García-Hidalgo, S. Borruel; Madrid/ES ([email protected]) Purpose: Multidetector-row CT (MDCT) allows assessment of active extravasation of traumatic and non-traumatic hemorrhage in hemodynamically stable patients, but it is often unfeasible to indicate the exact vessel of origin and whether bleeding is arterial or venous, both facts essential information for endovascular treatment. We try to assess if there are imaging signs that help clarify this issue. Methods and Materials: Forty patients with suspected active bleeding that received non-operative management were imaged with the ER 6-slice MDCT and underwent angiography in 2006-2007. We reviewed the studies to identify ﬁndings correlated with the source of bleeding on CT and compared them to angiographic ﬁndings. Results: MDCT showed active extravasation in 17 cases, while angiography demonstrated only 14 cases. Prophylactic embolization was performed in 12 cases albeit with negative result. The location of active bleeding as determined by extravasation of contrast media shown on MDCT included bronchial tree (3), liver (10), spleen (6), kidneys (5), retroperitoneum (3) and pelvis (10). Three cases presented signs of acute mesenteric ischemia. Conclusion: MDCT effectively demonstrated the site of active extravasation (attenuation 85 UH, signal ﬂare phenomenon, pooling of contrast); however, in most cases it could only suggest the exact vessel of origin. We found no speciﬁc signs to discern the arterial or venous origin of bleeding. The ﬁndings in angiography do not always duplicate those on CT and many times transarterial embolization is performed even if the study does not demonstrate active hemorrhage. Hemodynamic stability should be the only determinant for non-operative management.

B-190

15:21

Automatic bone and plaque subtraction in runoff-CTA using dual energy acquisition B.C. Meyer, T. Werncke, K.-J. Wolf, T. Albrecht; Berlin/DE ([email protected]) Purpose: To evaluate the accuracy of automatic bone subtraction using Dual Energy acquisition in CT-Angiography of the pelvic, thigh and calf arteries. Methods and Materials: Dual energy-CTA of peripheral arteries was performed in 15 patients using a Somatom Deﬁnition (Siemens, Germany) with the following parameters: collimation 64x0.6; 80 kV, IqualRef=350 mAs; 140 KV, IqualRef=80 mAs, rotation time 0.5 s, pitch: 0.7; slice 1.5 mm. 100 mL Iomeprol 400 (Bracco, Italy)

182

A

B

C

D

E

F

G

was injected at 4 mL/s followed by 50 mL saline. Bolus tracking was used. Separate datasets were calculated for each voltage and used to generate automatically bonesubtracted images (ABS) using a dedicated workstation (MMWP, VE25 A, Siemens). In addition a merged dataset from both dual energy acquisitions resembling a routine 120 kV CT-acquisition was generated and used for standard manual bone subtraction (MBS). Operator time for bone removal was measured. Images were assessed for effectiveness of bone subtraction and presence of luminal erosions by two readers (consensus). Results: A total of 372 arteries were analysed. Residual bone fragments (ribs: 37%, patella: 45%) were only observed with ABS and did not interfere with image interpretation. Above the knee, luminal erosions occurred in 4.1% of arteries (aorta: 0.8% and internal iliac: 3.3%) with ABS and in 1.6% (internal iliac) with MBS (NS). Below the knee, 38% and 40% were eroded with ABS and MBS, respectively (predominantly anterior tibial artery). Operator time for bone removal was 6p2 min for ABS and 12p3 min for MBS (P 0.01). Conclusion: Automatic bone removal with dual energy CTA is more time effective than conventional CTA and manual bone removal and provides similar results.

14:00 - 15:30

Room K

Computer Applications

SS 305 Image viewing and PACS Moderators: J. Fernandez-Bayó; Sabadell/ES R.W.R. Loose; Nuremberg/DE

B-191

14:00

Open-source platform for remote teleradiology in mobile PET-CT unit O. Ratib, A. Rosset, J. Heuberger, J. Willi, M. Velazquez; Geneva/CH ([email protected]) Purpose: With the deployment of a mobile PET-CT unit we needed to implement a two-way teleradiology system allowing technicians and physicians in the mobile unit to communicate with other members of the team located at the homebase hospital for supervision and image interpretation. Methods and Materials: An Open Source image analysis software called OSIRIX developed in our institution was adapted to allow two-way video conferencing and image sharing using iChat protocol from Apple. The software provides all necessary image rendering, fusion and processing tools required for adequate PET/CT image display and interpretation. Images were sent directly from the scanner to the OSIRIX workstation located in the trailer, and a two-way encrypted communication over Internet allowed this unit to be linked to other OSIRIX workstations located in the hospital. While ﬁxed Internet connection was available at each location, wireless communication over conventional high-speed phone connections were also tested for performance comparison. Results: Direct link between the mobile unit and the hospital was found to be extremely efﬁcient and critical for supporting the mobile team in performing PET/ CT studies in different locations where technical and clinical expertise was not always available. Multi-way video conferencing with image sharing was found to be extremely fast and convenient. DICOM ﬁles could also be sent to the hospital for image processing and interpretation by the homebase staff. Conclusion: The video/conferencing extension of our open source OSIRIX software provides a cost effective solution for supervision and support of mobile imaging units by homebase staff and specialists.

B-192

14:09

A research PACS system with extensible by end-users image processing and scientiﬁc visualization capabilities K.P. Karolemeas1, T.G. Maris2, N. Gourtsoyiannis2; 1Athens/GR, 2Iraklion/GR ([email protected]) Purpose: The increasing need for customized medical image processing in radiology laboratories demands post-processing workstations with functionality that can be easily extended by end-users. This implies the development of research-driven image processing techniques. Furthermore, these techniques should be directly accessible for application even by non-researchers through a seamless integration with a mainstream RIS-PACS solution. Methods and Materials: The cooperation of a University Hospital with an independent software vendor resulted in the presented Research PACS system, which facilitates the development (in Matlab or Java), real-time modiﬁcation and

Scientific Sessions

B-193

14:18

The value of grid technology in building regional and national medical image archives J. Crawford1, S. Ohlsson2; 1London/UK, 2Stockholm/SE ([email protected]) Purpose: To devise new technological approaches which will help to reduce the cost of long-term medical image storage, whilst promoting automated management and migration of the data, less chance of losing images, and faster access to archived images when needed. Methods and Materials: ‘Grid computing’, typically associated with the capture of idle computational cycles on distributed servers, offers a novel approach. ‘Storage grids’ use intelligent software to enable PACS systems to gain access to pools of distributed storage across a network of physical locations. Three different approaches that use ﬂavours of ‘grid computing’ will be discussed. Results: Medical imaging storage grids can be used to create virtual storage networks that can i) optimise storage resources, ii) enable PACS to access remote image ﬁles as if they were local, iii) automatically create replicas of images for availability and disaster recovery, iv) perform information lifecycle management, and v) seamlessly migrate data to new media. Conclusion: By designing regional and national scale image archives around storage grids, opportunities are created to manage costs, whilst delivering a better and more secure service to radiologists.

B-194

14:27

DICOM-CDs for communicating radiological images: Needs for a composite CD containing the radiological history of the patient improve user accessibility to data A. Nitrosi1, F. Nicoli1, G. Borasi1, P. Pattacini1, A. Botti1, M. Bertolini1, S. Marzani1, G. Modigliani2; 1Reggio Emilia/IT, 2Cinisello Balsamo/IT ([email protected]) Purpose: DICOM CDs are usually in use for data interchange between hospitals and/or practitioners. Typically for an oncology patient the diagnosis and the follow-up need a great number of examinations. These result in serious complaints by practitioners in analyzing a great number of DICOM CDs. Therefore our Institution started a process to test the production of CD containing the radiologic history of the patient. Methods and Materials: For the oncology patient who wants to obtain an external consultation we burn DICOM CD containing the last study performed and the key images from previous automatically selected studies using prefetching rules. These rules were developed by a group of experts from Radiology and Oncology Departments, and then brought into a comment process for our Institution PACS vendor. Results: First feedbacks from practitioners working outside our Institution are greatly positive. Almost all appreciate the capability to take a complete purview of the evolution of patient disease in a shorter time limiting misleading interpretation. We evaluate an increasing number of our CD requests. In future we will be able to evaluate the number of out-patient examinations directly requested by external pratictioners. Conclusion: Diagnoses and treatment need a complete and updated knowledge of clinical informations. The IHE-Portable Data for imaging proﬁle is the standard that should be commonly adopted, but in addition it could be very useful to adopt shared rules for compose DICOM CD containing the key images from previous selected studies. Improvement of quality of patient care using historical DICOMCDs is possible.

B-195

14:36

Volume rendering and maximum-intensity projection techniques in sixteen-row multidetector computed tomography angiography. A comparative study A. Mishra, E.F. Ehtuish; Tripoli/LY ([email protected]) Purpose: To compare the accuracy of volume rendering (VRT) and maximum-intensity projection (MIP) techniques in sixteen-row multidetector computed tomography (MDCT) angiography and correlate with operative ﬁndings. Methods and Materials: 16-row MDCT angiography scans were performed in 85 renal and 15 hepatic transplant patients between 10th October, 2004 and 30th August, 2007. 80-100 cc of injection Iohexol (300 mgI/mL) was injected for renal studies in doses of 2 ml/kg body weight for hepatic studies. Saline bolus chase was used in all the studies. 1.25 mm contiguous axial sections were acquired and submillimeter sections reconstructed with 50% overlap. Post-processing included VRT, MIP and multiplanar reconstruction techniques. The raw data set was anonymized for the two blinded trained radiologists in all patients and they performed VRT and MIP reconstructions, which were compared with surgical ﬁndings. Interobserver agreement between the two reviewers and between MDCT angiographic post-processing techniques and surgical ﬁndings was quantiﬁed by using weighted K statistics. Results: MDCT angiography scans were technically satisfactory in all the patients and showed clear delineation of the renal and hepatic arteries. Discordance between the two post-processing techniques was seen in 62% patients and especially for submillimeter accessory arteries. Correlation between MIP and operative ﬁndings was excellent for renal and hepatic arteries (K=0.843). Agreement between VRT and surgical ﬁndings was good (K=0.710). Conclusion: MIP is superior to VRT in delineating vascular anatomy accurately and should be the standard ﬁrst post-processing technique in MDCT angiography study.

B-196

14:45

Semi-manual quantitative right and left ventricular analysis: Comparison with automatic delineation E.E.J.G. Coche1, M. Vembar2, M. Walker2, A. Vlassenbroek1; 1Brussels/BE, 2 Cleveland, OH/US ([email protected]) Purpose: To evaluate an automatic detection boundary algorithm to measure right (RV) and left ventricle (LV) volumes and ejection fraction (EF) on MDCT images. Methods and Materials: Fifteen enhanced MDCT studies performed with 16x1.5 mm collimation, 2 mm reconstruction interval and retrospective ECG synchronisation were analysed. End diastolic volumes, end systolic volumes, stroke volumes for RV and LV and corresponding EF were computed independently at several months interval by semi-automatic and automatic detection boundary algorithm by a blinded observer on a dedicated workstation. The measurements were repeated by the same observer at one month interval. Reproducibility of measurements was performed with intra-class correlation coefﬁcient (ICC). Measurements with the two techniques were compared using Bland-Altman plots and paired student’s t test. Results: Ventricular analysis with automatic delineation was successful in 13/15 patients (86%). Reproducibility of measurements was better for automatic (ICC:1) than for semi-manual measurements (ICC 0.95) for both ventricles. End diastolic volumes, end systolic volumes obtained with the semi-manual method were signiﬁcantly lower for the RV compared to the automatic method: mean difference: 11, 4, 8, 7 mL, respectively (p 0.05). EF was signiﬁcantly increased with the semimanual method: mean difference: 8.9% (p 0.05). Lower values were obtained also for the same parameters for the LV with no signiﬁcant difference except for the end-diastolic volume: mean difference 3.9 mL (p 0.05). Conclusion: Measurements of ventricular volumes appear more robust with the automatic delineation algorithm. Concordance of measurements obtained by both methods was better for the LV than for the RV.

B-197

14:54

Semi-automatic multi planar reformatting of the coronary arteries from computed tomography angiography data C.T. Metz, M. Schaap, T. van Walsum, N.R.A. Mollet, W.J. Niessen; Rotterdam/NL ([email protected]) Purpose: Multi planar reformatted images (MPRs) are useful for the diagnosis of coronary artery disease in computed tomography angiography data. However, generating an MPR requires laborious manual annotation of the vessel centerline. A new method is introduced to minimize user interaction for MPR generation to one mouse click per coronary artery. Methods and Materials: Coronary centerlines are determined using Dijkstra’s shortest path algorithm incorporating intensity and second order image informa-

A

B

C

D

E

F

G

183

Friday

application of image post-processing techniques as an integral part of the life-cycle of examinations. Intermediate and ﬁnal results can be visualized using 2D and 3D graphs. Furthermore, stacks of intermediate or resulting images can be visualized using 3D technologies such as volume rendering. Collaboration of geographically dispersed research teams is supported by providing the full functionality of the system to remote users through the internet. The Research PACS system can be deployed in parallel with a central RIS-PACS solution by any vendor. Results: The implementation of this new Research PACS system helped the researchers at the University Hospital to accelerate the development of new image processing techniques and integrate their work with the regular examination work-ﬂow. Geographically dispersed teams can now reduce traveling, eliminate image exchange and increase members’ productivity. Conclusion: Since the implementation of the Research PACS system, the researchers of the University Hospital experience increasing productivity locally and remotely. The incorporation of generated images, measurements and other information in publications has been signiﬁcantly accelerated.

Scientific Sessions tion. The starting point for Dijkstra’s algorithm is determined using a novel aorta detection method while the endpoint is provided by the user. MPRs are automatically generated based on the centerline found. Centerline determination and MPR generation can be done in real-time. The proposed method has been tested on 24 coronary arteries in CTA data (Sensation 64/Deﬁnition, Siemens, Germany) of six patients who were referred for coronary intervention. Each centerline is quantitatively evaluated by comparing it with the average of two centerlines annotated by trained observers. MPRs are visually inspected by these two observers. Results: Visually correct MPRs were automatically generated in 22 out of 24 cases. For these 22 centerlines the mean distance to the average manually annotated centerlines was 0.58 mm (p0.28 mm) compared to an interobserver variability of 0.47 mm (p0.09 mm). Conclusion: The presented method provides an effective and efﬁcient way to create visually correct MPRs for the diagnosis of coronary artery disease in CTA data. The accuracy of the method is very close to the variability between trained observers.

B-198

15:03

3-way multimodality image registration for PET/CT and MRI fusion in clinical practice J. Heuberger, A. Rosset, O. Ratib; Geneva/CH ([email protected])

B-200

15:21

Development of a new resolution enhancement technology for medical liquid crystal displays M. Hasegawa1, Y. Tanaka1, S. Hayashi1, Y. Nishi1, N. Kimura1, K. Ichikawa2; 1 Ueda/JP, 2Kanazawa/JP ([email protected]) Purpose: Liquid crystal displays (LCDs) do not have enough resolution properties that meet the resolution of modality devices such as ﬂat-panel detectors (FPDs) and computed radiography systems (CRs) that have already achieved about 4-67 megapixels (MP). A method of a new resolution enhancement technology that used independent sub-pixel driving (ISD) technology and an anti reﬂection-coated panel for medical LCDs is introduced. Methods and Materials: Each pixel of monochrome LCDs consists of three subpixels equivalent to RGB sub-pixels of color LCDs. Driven each sub-pixel by pixel value corresponding to detailed information recorded in the FPD or CR image, 3X resolution enhancement is achieved. Thus the new technology realized 9MP LCD out of a 3MP LCD and 15MP LCD out of a 5MP LCD. We evaluated their sharpness and noise characteristics for comparison. DICOM viewing software that implements computer algorithms for displaying corresponding to the ISD was developed accordingly and perceptual evaluations of clinical images were performed by radiologists. Results: Spatial frequency was improved obviously according to very ﬁne subpixel width of 69 μm for the 9MP LCD and of 55 μm for the 15MP LCD. The anti reﬂection-coated panel reduced image noise and improved the resolution signiﬁcantly. In perceptual comparisons, images displayed on the newly developed LCD were faithful to their originals compared to the sub-sampled images on the conventional LCDs. Conclusion: This new technology was considered effective for the accurate depiction of the ﬁne anatomical structures such as the micro calciﬁcation of the mammography.

Purpose: Hybrid PET/CT scanners provide co-registered images but it is often required to correlate these images with images obtained from MRI or from CT studies acquired at different times. Using CT images with high-deﬁnition anatomical landmark it is possible to obtain accurate fusion of PET images with a third modality such as MRI by co-registering CT images with MRI images. Methods and Materials: We implemented an interactive image registration tool based on identiﬁcation of anatomical landmarks on CT and MRI image sets to align PET images obtained from the PET/CT studies with MRI studies. This tool was implemented on OSIRIX software, an open source multimodality image processing platform designed for display and 3D navigation through multimodality images. OSIRIX being widely distributed outside the radiology department, it allows clinicians and other users to easily correlate ﬁnding obtained from different imaging studies. Results: The 3-way image fusion tool was tested in clinical applications such as co-registration of images in brain tumors, bone and liver metastases and even gated cardiac studies. Users with little or no experience in image processing were able to rapidly match images from different imaging modalities for accurate localization of focal abnormalities. Conclusion: With clinical decision-making and image interpretation relying on ﬁndings observed on different imaging modalities, a convenient 3-way registration tool can provide clinicians and non radiologists with a convenient way to review and localize focal abnormalities. Open source OSIRIX platform provides a cost effective solution to this problem that can be widely distributed across a hospital.

14:00 - 15:30

B-199

32 directions DTI of cervical spinal cord in patients with multiple sclerosis M. Ukmar, E. Makuc, G. Garbin, X. Santarelli, A. Bratina, M.A. Cova; Trieste/IT ([email protected])

15:12

OSIRIX: Open-source extension to PACS for multi-modality image analysis A. Rosset, J. Heuberger, D. Bandon, O. Ratib; Geneva/CH ([email protected]) Purpose: Conventional PACS workstations lack display and analysis tools required for rendering and processing of multi-modality and multidimensional imaging modalities. We elected to expand conventional legacy PACS with OSIRIX, an open-source software platform. It provides the necessary tools for navigating and processing very large datasets of multimodality imaging devices such as PET/CT. It can be implemented outside radiology departments and provide a solution for increasing needs and requirements of clinicians, surgeons and other specialists. Methods and Materials: OSIRIX software developed on Apple Macintosh platform was integrated to an existing RIS/PACS combination in our hospital. This platform offers advanced 3D, 4D and 5D rendering tools as well as image navigation, reformatting and quantitative analysis tools. It was implemented in numerous locations across the whole medical enterprise, in particular in conference rooms, clinical wards and physicians’ ofﬁces. Results: The integration with our legacy PACS required appropriate integration of institutional security and conﬁdentiality control mechanisms and constraints. The ﬂexibility and performance of the OSIRIX software allowed rapid deployment and adaptation to different clinical needs and requirements. Over 60 units are being implemented in different settings. Users can conveniently import, review and process studies in different locations for surgical planning, and clinical decision making during clinical conferences and review boards. Selected users can also upload results from 3D processing back to the PACS for archive. Conclusion: Open source component can provide cost effective and efﬁcient ways for deploying advanced PACS workstations outside radiology in specialized clinical wards and conference rooms in large academic hospitals.

184

A

B

C

D

E

F

G

Room L/M

Neuro

SS 311a Advanced spinal MRI Moderators: M. Sasiadek; Wroclaw/PL A. Urbanik, Krakow/PL

B-202

14:00

Purpose: To evaluate the usefulness of diffusion tensor imaging (DTI) of the normal appearing cervical spinal cord in patients with multiple sclerosis (MS). Methods and Materials: Ten healthy volunteers and 30 patients with MS (10 with relapsing remitting course, 10 with primary progressive and 10 with secondary progressive course) have been evaluated. In all of them a conventional, including sagittal TSE T2, STIR and SE T1 weighted sequences after the administration of paramagnetic contrast media, and a DTI MR study were performed by using a 1.5 T magnet. The DTI images were obtained on axial plane along 32 directions using a b value of 1000 mm2/s. For DTI evaluation the region of interest included the spinal cord from C1 to C5. The fractional anisotropy (FA) was evaluated by using a dedicated software at the level of normal appearing spinal cord on the conventional imaging. Results: Considering the FA of the normal appearing spinal cord, a statistically signiﬁcant difference (p 0.05) between controls and patients was obtained. Conclusion: DTI could be a useful tool in the evaluation of normal appearing spinal cord in patients with MS allowing a better comprehension of MS pathology and moreover could better explain the clinical disability of the patients.

B-203

14:09

Diffusion tensor imaging (DTI) in HIV-positive patients at 3.0 T MRI: Assessment of regional differences in the axial plane within normalappearing cervical spinal cord C. Mueller-Mang1, T. Mang1, C.M. Plank1, A. Rieger1, E.-M. Law2, M.M. Thurnher1; 1 Vienna/AT, 2New York, NY/US ([email protected]) Purpose: Up to date HIV-associated Myelopathy (HIVM) is usually detected only

Scientific Sessions

B-204

14:18

Spinal cord intramedullary tumor microstructure assesment using high B value diffusion tensor imaging and ﬁber tractography for characterization: A study of 22 cases R.M.S.V. Vadapalli1, V. N2, M. Naidu1, P. Ravuri1, R. Mangasree1; 1Hyderabad/IN, 2 Bangalore/IN ([email protected]) Purpose: To assess the role of DTI indices and tractographic characteristics of intramedullary spinal cord tumors using 3 T MRI: A study of 22 cases. Methods and Materials: 22 consecutive patients in the age range of 25-52 and M:F ratio of 2:1 and 15 normal controls were included in the study after informed consent. Clinical categories included Astrocytomas (n=9), Ependymomas (n=7), Hemangioblastomas (n=2), Metastases (n=4). All patients were examined on a 3.0 T/1.5 T (17/5) using EPI DTI sequence (in addition to routine post contrast tumor protocol) with a multichannel coil with sequence parameters as follows: TR9772 TE 88 msec B values of 0 and 2000, 3000, 25 directions, matrix 256/256, SENSE factor of 2. The data was then processed using ﬁber tracking software. Results: A rim of high FA value zone was seen bordering the cord tumor interface in ependymomas and haemangioblastomas (n=9) with sharp narrow transition zone of FA values with a mean of 0.57 called as High FA rind sign. Astrocytomas (n=9) showed absent High FA rind all round with a wide transition zone of FA value loss at the cord and tumor interface with a mean of 0.21 with an advancing edge pattern. Metastases (n=4) showed showed no advancing edge pattern. Cord edema showed higher ADC and FA values than the tumor matrix (n=6) with Ino interruption of high FA rind. All results were correlated with surgery and histopathology. Conclusion: High resolution DTI helps to assess tumor microstructure with an emerging role to characterize intramedullary spinal cord tumors.

B-205

14:27

MR diffusion tensor imaging in evaluation of morphological changes of acute spinal cord injury in rats N. Chen; Beijing/CN ([email protected]) Purpose: To observe the morphological changes of acute spinal cord injury in rats with MR diffusion tensor imaging (DTI) and ﬁber tracking. Methods and Materials: Prior to and one week after injury, 12 adult Sprague-Dawley rats suffering from left spinal cord hemisection at the level of T6 were subjected to T1/T2-weighted images and DTI by using 3.0 T MR scanner. After MR scan, the rats were sacriﬁced for histopathologic examination. The ADC and FA values of the spinal cord were calculated and compared with the histopathologic ﬁnding. Results: The ADCs of normal rat spinal cord at T6 level were 1.00p0.15x10-3 and 1.01p0.17x10-3 mm2/s for the left and right sides, and the FA values of the corresponding area were 0.59p0.11 and 0.60p0.08, respectively. There were no statistically signiﬁcant differences between the ADCs and FA with the left and right sides. After injury, the ADC of the injured spinal cord increased to 1.65p0.45x10-3 mm2/s and FA decreased to 0.30p0.17. There were statistically signiﬁcant differences between pre- and post-operation. Compared with histologic ﬁnding, the FA/ADC are corresponding with the morphological changes of axons including decreased numbers of axons, axonal swelling, and demyelination. The ADC and FA values of the contralateral uninjured rats were not signiﬁcantly different. Conclusion: DTI can provide useful information of spinal cord damage in experimental spinal cord injury.

B-206

14:36

Can MR diffusion tensor imaging predict the function recovery of spinal cord injured in vivo in the rat? N. Chen, L. Kuncheng; Beijing/CN ([email protected]) Purpose: To evaluate the value of DTI in predicting the outcome of function recovery in spinal cord injured rat. Methods and Materials: 12 adult SD rats suffered from left spinal cord hemisection at the level of T6 underwent DTI that had single-shot echo planner imaging sequence at 3.0 T MR scanner on Day 0 (before surgery) and every week thereafter up to eight weeks, and the status of animal function recovery was evaluated using the BBB scale before the MRI scan. The ADC and FA of the injured area of spinal cords were measured and compared with results of BBB scores. Results: FA values of the injured area decreased signiﬁcantly one day after injury compared with pre-operation, and gradually increased up to three weeks; the BBB score gradually increased simultaneously during this period. The FA value correlated well with BBB score change. From the 4-8 weeks after injury, the FA values increased or decreased irregularly; however, the BBB score gradually increased. There was no signiﬁcant correlation between FA and the BBB score. Only one day after injury, the ADC increased signiﬁcantly. There was no signiﬁcant correlation between ADC and the BBB at any time. Conclusion: The FA value correlated with the function status only at the acute stage of spinal cord in vivo in rat, but it cannot predict the prognosis of functional recovery of spinal cord. Therefore, DTI appears to be the inevitable tool for predicting the prognosis of spinal cord injury.

B-207

14:45

MR diffusion tensor imaging and ﬁber tracking in acute spinal cord injury in humans N. Chen, L. Kuncheng; Beijing/CN ([email protected]) Purpose: To evaluate the spinal cord structural integrity after injury using DTI. Methods and Materials: Nineteen patients with acute thoracic spinal cord injury and eight healthy volunteers underwent T2WI and DTI at thoracic spinal cord using 3.0 T MR scanner. Regions of interest were placed at the upper and lower thoracic cord levels for the healthy subjects and on the area with abnormal T2WI signal intensity in the patients with spinal cord. Additionally, we used ﬁber tracking at the injured area in ten patients. The sensitivity (SE) and speciﬁcity (sp) of T2, ADC, and FA maps in detecting spinal cord abnormality were statistically evaluated. Results: For the healthy subjects, the average ADC and FA values were 0.98p0.12 and 0.78p0.16x10-3 mm2/s, respectively. Sixteen patients had decreased FA (0.66p0.067), and three had increased FA values; only two patients had increased ADC values (1.13p0.15). There was a statistically signiﬁcant difference in the FA values between volunteers and patients (p=0.12). FA had a much higher sensitivity (SE=84.2%) and speciﬁcity (sp=100%) in spinal cord abnormalities detection compared with T2WI (SE=43.9%, sp=100%) and ADC (SE=16.1%, sp=82%). Spinal cord ﬁber tracking showed the neuroﬁbra discontinuity partly (n=6) or completely (n=4) in the injured area. Conclusion: FA has the highest sensitivity and speciﬁcity in the detection of acute spinal cord abnormalities. Spinal cord ﬁber tracking is a useful tool to display the integrity of neuroﬁbra on the injured spinal cord.

B-208

14:54

The diagnostic value of three dimensional dynamic contrast-enhanced MR angiography in spinal vascular malformations Y. Yang, K. Li, Y. Liu, W. Shi, P. Yang, H. Zhang, X. Zhi; Beijing/CN ([email protected]) Purpose: To explore the diagnostic value of three dimensional dynamic contrastenhanced MR angiography (3D-DCE-MRA) in spinal vascular malformations. Methods and Materials: 20 patients with spinal vascular malformations underwent 3D-DCE-MRA with a 3.0 T MR system. The 3D-DCE-MRA was obtained with a 3D Flash sequence after Gd-DTPA was injected with a high pressure syringe. Digital subtraction angiography (DSA) was performed on all patients within one week. 3DDCE-MRA images were analyzed in some aspects, including type of spinal cord vascular malformation, vascular nidus or ﬁstula, feeding artery, and draining vein by two experienced neuroradiologists. We compared the results with those of DSA. Results: Of 20 patients conﬁrmed by DSA, 10 cases had intramedullary arteriovenous malformations (AVMs), eight cases had spinal dural arteriovenous ﬁstulas (SDAVFs), and two cases had perimedullary arteriovenous ﬁstula (PMAVFs). All 20 patients were correctly diagnosed with 3D-DCE-MRA. Its sensitivity and the coincidence rate to DSA were 100%. The majority of ﬁstulas were revealed on 3D-DCE-MRA. The coincidence rate with DSA is 85.7% in SDAVFs and 100% in

A

B

C

D

E

F

G

185

Friday

in the late stage of disease using conventional MRI. The purpose of this study was to compare DTI metrics of the cervical spinal cord in asymptomatic HIV positive patients with those measured in healthy volunteers, and to assess if DTI is a valuable diagnostic tool in the early detection of HIV-associated spinal cord abnormalities. Methods and Materials: MRI of the cervical spinal cord was performed in 20 asymptomatic HIV-positive patients and in 20 healthy volunteers on 3.0 T MR scanner. All patients and volunteers underwent axial DTI using pulsed gradient, spin-echo, double-shot echo-planar imaging. Average fractional anisotropy (FA) and mean diffusivity (MD) were calculated within regions of interest at bilateral anterior, lateral, and posterior white matter regions of the cord. The statistical signiﬁcances of the computed DTI metrics were assessed using a paired, doubletailed student’s t test. Results: No patient had abnormalities on conventional T2-weighted MR images. FA was lower in HIV-positive patients in the lateral (mean 0.57) and posterior regions (mean 0.65) at C3-4 and C4-5 levels of the cervical spinal cord. MD was higher compared with controls but did not reach statistical signiﬁcance. Conclusion: Asymptomatic HIV-positive patients demonstrate changes in DTI metrics measured in the cervical spinal cord compared to healthy volunteers in the absence of spinal cord abnormalities on conventional MR sequences. These changes are predominant in the lateral and posterior regions and may be markers of HIVM.

Scientific Sessions PMAVFs. 45 feeding arteries for vascular nidus of ﬁstulas were detected by DSA. The sensitivity and speciﬁcity of 3D-DCE-MRA in diagnosing feeding arteries were 77.8% and 91.4%. 3D-DCE-MRA visualized all perimedullary draining veins, and was superior to DSA for wide-ranging draining veins. Conclusion: 3D-DCE-MRA is a safe, convenient and reliable method for detecting spinal vascular malformations and diagnosing the type. In feeding arteries, it is inferior to DSA, but can direct selective catheter angiography more conveniently.

B-209

15:03

3 T MR neurography of sciatic nerve and lumbosacral plexus using DWIBS sequence in patients with spinal and extra spinal sciatica: A study of 41 patients R.M.S.V. Vadapalli, R. Mulukutla; Hyderabad/IN ([email protected]) Purpose: To evaluate the role of MR neurography of sciatic nerve and lumbosacral plexus at 3 T using diffusion weighted imaging with back ground suppression (DWIBS). Methods and Materials: 15 normal volunteers and 41 patients with clinical sciatica with age range of 24-57 years, M:F ratio of 1:2 were recruited for the study after informed consent. All patients were imaged on a 3 T Achieva XR series (Phillips Medicals) system with following protocol: T1/T2 sagittal, DWIBS STIR EPI sequence with following parameters: TR 5000 msec, TE min-70 msec, EPI factor-47, half scan factor-0.645, TI-180 msec, slice orientation axial, coronal, number of slices60, thickness/gap-3 mm/0 mm; 28 cm FOV coverage from L2 till S5, B value-1000. All the images were processed with reverse video on and off on the workstation generating 3D MIP plexograms and ADC values. Results: All normal controls and patients showed excellent visualization of the lumbosacral plexus with its branches (n=56, 41+15). Sciatic nerve in is intra and extra pelvic course was visualized in all controls with a mean ADC of 0.87. Low ADC with nerve edema with mean ADC of 0.45 in 19 patients is due to disc-related compression. Pelvic tumor compression of sciatic nerve at greater sciatic notch (n=4), Schwannoma (n=2) with ADC values of 0.34 (60% reduction). Extrinsic compression by adenomyotic uterus (n=4), endometriosis (n=2), post-radiation lumbar plexopathy in a carcinoma cervix (n=5), NonHodgkin’s Lymphoma (n=3) with reduced ADC values of 0.45, 0.42, 0.32, 0.39, respectively distal to site of compression. Conclusion: DWIBS neurography provides optimal visualization and evaluation of lumbosacral plexus and intra and extra pelvic compression of sciatic nerve.

B-210

15:12

Upright MRI of disk herniation: Status and changes of its size under physiological loading M. Di Fabio, C. Bultrini, R. Masi, R. De Amicis, A. Splendiani, M. Gallucci; L’Aquila/IT ([email protected]) Purpose: Though the clinical symptomatology of patients with lumbar disk herniation worsen under loading conditions, there is nowadays in the literature a lack of studies performed on large series describing the possible changes in disk herniation size depending on the upright position. The radicular compression may depend on many factors which are not necessarily related only to the disc itself. Alterations of canal and foramina, ligamentous weakness, hydrodynamic changes of the dural sac can lead to pain in presence of a pathological disc. Methods and Materials: Thirty disk herniations (17 males and 12 females, age ranging between 20 and 60 years) were submitted to MRI in both the upright and recumbent positions. The examinations were subsequently evaluated by two neuroradiologists who focused their attention on the measurements of size changes of the disc herniation, of lordosis angle, canal size and of the thickness of intersomatic discs. Results: In 13 cases of disc herniation, there is a modiﬁcation between upright and recumbent position; in 23 we observe an increase of lordosis angle. In 24 cases there is a reduction of lumbar-sacral angle and in 18 we observe a reduction of thickness of intersomatic discs. Conclusion: 45% of herniations increase in upright positions; 100% of increasing herniations show residual hydration. In patients with modiﬁed disk herniation, we have hypomobility of the spine and increase of lumbar-sacral angle. All patients with high intensity of signal herniation have a reduction of thickness of intersomatic discs.

186

A

B

C

D

E

F

G

14:00 - 15:30

Room N/O

Neuro

SS 311b Brain tumors Moderators: P.C. Maly Sundgren; Ann Arbor, MI/US T.A. Yousry; London/UK

B-211

14:00

High-resolution contrast-enhanced, susceptibility-weighted MR imaging at 3 T in patients with brain metastasis tumors: Correlation with histopathologic ﬁndings T. Gu; Beijing/CN ([email protected]) Purpose: To demonstrate susceptilility effects (SusE) in various types of brain metastasis tumors with 3 T high-resolution (HR)-contrast-enhanced (CE)-susceptibility-weighted (SW)-MR imaging and to correlate SusE with histopathology. Methods and Materials: Sixteen patients with brain metastasis tumors, scheduled for tumor extirpation, underwent high-ﬁeld (3 T)MR imaging. In all of the patients, an axial T1-spin-echo (SE) sequence and an HR-SW imaging sequence before and after IV application of a standard dose of contrast agent was obtained. The frequency and formation of intralesional SusE in all the images were evaluated and correlated with type of metastasis tumors. Contrast enhancement of the lesions was assessed in both sequences. Results: Metastasis tumors easy to bleed demonstrated either high or medium frequency of SusE in 90% of the patients. The type of metastasis with low blood supply demonstrated low frequency of SusE or no SusE. Correlating between intralesional frequency of SusE and histopathology showed that the type of metastasis was signiﬁcantly different. Contrast enhancement was equally visible in both SW and SE sequences. Side-to-side comparison of tumor areas with high frequency of SusE and histopathology revealed that intralesional SusE reﬂected conglomerates of increased tumor microvascularity. Conclusion: 3 T HR-CE-SW-MR imaging shows both intratumoral SusE not visible with standard MR imaging and contrast enhancement visible with standard MR imaging. Because frequency of intratumoral SusE correlates with metastasis type, this novel technique is a promising tool for noninvasive differentiation of metastasis brain tumors and evaluation therapy effects by radiotherapy and chemotherapy.

B-212

14:09

MR/PET of the human brain: Feasibility of simultaneous imaging for scientiﬁc and clinical applications H.-P.W. Schlemmer1, B. Pichler1, D.W. Townsend2, C. Nahmias2, W.-D. Heiss3, C.D. Claussen1; 1Tübingen/DE, 2Knoxville, TN/US, 3Cologne/DE ([email protected]) Purpose: Simultaneous MR/PET imaging of the human brain by implementing a novel PET detector technology. Methods and Materials: PET detectors consist of a 12x12 LSO (Lutetium-oxyorthosilicate) scintillation crystal matrix read out by a 3x3 APD (avalanche photodiodes) array with 5x5 mm² active surface of individual diodes. The PET detector ring (inner diameter 35.5 cm, FOV 19 cm) is formed by 32 radially arranged cassettes (5 LSO-APD blocks each) and built into a clinical 3 Tesla MRI scanner (Trio, Siemens) ﬁtting around the standard bird cage transmit/receive head coil. Phantom studies were performed using a Hoffman brain phantom ﬁlled with 37 MBq [18 F]Fluoride. First human MR/PET brain imaging was conducted in 3 patients with history of cancer who underwent a clinical PET/CT for restaging after intravenous injection of 370 MBq of [18 F]Fluoro-2-deoxyglucose ([18 F]FDG). PET data were acquired for 30 min and the MR examination was performed simultaneously including MR imaging, time-of-ﬂight MR angiography, diffusion-weighted MRI and 2D proton MR spectroscopic imaging. Results: The overall performance of each imaging modality was maintained during simulateous scanning revealing image qualities comparable to stand-alone systems without any distortions or artifacts. The MR/PET scan enabled accurate spatial and temporal co-registration of morphologic and functional data in different brain regions by matching the FDG uptake and MR signal characteristics. Conclusion: Simultaneous high-ﬁeld MR and PET imaging is feasible with the novel PET detector technology maintaining the overall performance of each imaging modality. This opens a new window for molecular imaging of the human brain.

Scientific Sessions B-213

14:18

Perfusion weighted MRI in differential diagnosis between primary central nervous system lymphomas and high grade gliomas L. Rizzo1, S. Greco Crasto2, D. Sardo1, D. Gned1, D. Lanfranchini1, A. Veltri1, C. Fava1; 1Orbassano/IT, 2Turin/IT ([email protected])

B-214

14:27

Multi-center comparison of cerebral blood volume and vascular permeability measurements with tumor volume measurements following anti-angiogenesis therapy with Avastin in human gliomas L. de Villiers, M. Law, S. Chheang, J. Babb, N. Peccerelli, G. Fatterpekar, N. Leeds, M.V.V. Aragão, M. Fowkes, A. Narayana, R. Aiken, G. Johnson; New York, NY/US ([email protected]) Purpose: Dynamic Susceptibility Contrast MRI (DSC MRI) is emerging as a biomarker to assess effectiveness of anti-angiogenic therapies in treatment of brain tumors. Our purpose is to compare relative cerebral blood volume (rCBV) and vascular permeability measurements with tumor volume in predicting therapeutic response in high-grade gliomas treated with Avastin® (Bevacizumab). Methods and Materials: Eighteen patients (nine females, nine males, median age 55, range 15-69) were treated with 1-4 cycles of Avastin®/Irinotecan® (CPT11). Histologic diagnoses were: anaplastic oligodendroglioma (n=2), anaplastic astrocytoma (n=3), GBM (n=11), anaplastic ependymoma (n=1), pontine glioma (n=1). Patients had baseline DSC MRI prior to administration of Avastin and were followed clinically and radiographically with conventional and DSC MRI. Mixed model regression was used to compare pre- and post-treatment levels of each response measure. Results: There was a statistically signiﬁcant reduction in both rCBV and T1 size following treatment with Avastin®/Irinotecan®. The pretreatment rCBV and T1 rates of change (per 100 days) correlated signiﬁcantly with time (p=0.03 and 0.005), while only the post-treatment rCBV demonstrated signiﬁcant rate of change (p=0.001), suggesting that rCBV may reﬂect the effects of Avastin better than tumor volume. However, when the change in rate from pre- to post- treatment was considered, rCBV and T1 demonstrated signiﬁcance (p=0.001, 0.02). Conclusion: rCBV can be used as a biomarker to determine therapeutic response to Avastin®. This may provide an earlier indicator than conventional imaging and clinical response in the assessment of novel therapies in the treatment of gliomas.

B-215

14:36

Biopsy targeting in gliomas: Which imaging technique should be used? M.-A. Weber1, M. Henze1, J. Tüttenberg2, B. Stieltjes1, S.E. Combs1, M. Vogt-Schaden1, H. Remme1, M. Meissner1, H.-U. Kauczor1, M. Essig1; 1 Heidelberg/DE, 2Mannheim/DE ([email protected]) Purpose: Because of glioma heterogeneity, biopsies should be targeted to the most anaplastic part. To assess which functional MRI- or PET-technique should be preferred for biopsy guidance, we evaluated whether they identify similar target areas. Methods and Materials: Twenty-three glioma patients were assessed within one day by MRI and 18 F-ﬂuorothymidine (FLT)-PET. MRI was performed on a 1.5 T broadband transmit/receive system using a double-resonant birdcage coil. The MRI protocol comprised sodium (23Na)-MRI (3D-radial projection imaging), proton spectroscopic imaging (1H-MRSI, point-resolved spectroscopy), and dynamic contrast-enhanced (DCE) MRI and dynamic-susceptibility-weighted (DSC) perfusion MRI after a single dose each of gadobenate-dimeglumine. Image analysis

B-216

14:45

SPECT/CT or SPECT/MR image fusion for the evaluation of primary or metastatic brain tumors and assessment of tumor grade V.S. Dekan, G. Trufanov, G. Romanov, V.A. Fokin, B. Martynov; St. Petersburg/RU ([email protected]) Purpose: Glial neoplasms comprise the majority of primary intracranial tumors. There is some confusion among pathologists on the proper system for tumor grading. The purpose of this study was to evaluate whether the brain SPECT with CT and MR image fusion could be an efﬁcient tool in non-invasive determination of anaplastic activity, grading and clear localisation of primary glial brain tumors. Methods and Materials: Seventy-nine persons (mean age 48 years) with brain neoplasms, distinguished by CT or MRI and 15 patients with non-neoplastic lesions were included in the study. The subjects were given 99mTc-methoxyisobutylisonitrile (MIBI) early (one hour) and delayed (four hours) brain SPECT studies. In all cases there was histological veriﬁcation of tumor type and grade. In all cases image fusion with CT or MR images was used. Results: In 46 patients with high grade gliomas (21 with glioblastoma and 25 with anaplastic astrocytomas) early SPECT images showed signiﬁcally high uptake of MIBI in tumor tissue with high “tumor-to-background” rate (4.5 and higher). In 33 patients with grade II astrocytomas delayed uptake in the site of tumor was distinguished in all cases. All patients with benign lesions had no evidence of pathological tracer uptake. Conclusion: Keeping in mind that MIBI uptake is a characteristic of mitochondrial activity in malignant tumors, the uptake of the tracer is a very sensitive sign of malignancy of tumor lesion. SPECT/CT or SPECT/MR image fusion increases the accuracy of surgical stereotaxic approach.

B-217

14:54

Left frontal lobe tumors: What the neurosurgeon needs to know C.-V. Salvan, J.L. Ulmer, W.M. Mueller, D.L. Daniels, R.W. Prost; Milwaukee, WI/US ([email protected]) Purpose: To improve the presurgical planning in left frontal lobe tumors (LFLT) using modern functional brain mapping. To provide information to the neurosurgeon regarding best anatomic and spatial relationships of LFLT to eloquent brain systems, presurgical. Methods and Materials: BOLD functional MRI (fMRI) and Diffusion Tensor Imaging (DTI) were performed in 10 patients (22-63 y; 6 F, 4M; all right-handed), previously diagnosed with LFLT by MRI. Lesion border proximity of less or equal to 5 mm to speech-language, premotor-SMA, cortico-bulbar and cortico-spinal cortex and/or eloquent white matter tracts were considered at risk for operative injury to speciﬁc functional brain systems and for a corresponding neurologic deﬁcit. Results: Combined fMRI and DTI can provide better estimation of the proximity of LFLT borders to eloquent brain systems sub-serving language-speech, motor and premotor functions. Twice as many functional systems were at risk, compared to that afforded by fMRI alone. Neurosurgery resection strategies and trajectories were inﬂuenced not just by the morphological characteristics i.e. size and location of the tumor, but mainly by the proximity of brain functions at risk, operative strategy and approach being changed in seven of the patients. Conclusion: Integration of fMRI and DTI, combined with morphologic imaging, enhance the estimation of the proximity of functional brain systems to LFLT, and not only. Presurgical mapping inﬂuences Neurosurgeon attitude, surgical outcomes as well as the quality of life of each patient.

A

B

C

D

E

F

G

187

Friday

Purpose: To investigate the usefulness of Perfusion Weighted MR Imaging (PWI) in differential diagnosis between primary central nervous system lymphomas (PCNSL) and high grade gliomas (WHO III, IV). Methods and Materials: Thirty-six untreated patients (11 PCNSL and 25 high grade gliomas, average age 59.7 years) were examined using a 1.0 T MR scan with morphological and dynamic susceptibility contrast MRI. The relative Cerebral Blood Volume (rCBV) ratios were obtained by using a dedicated software (functool 2). Intensity-time curves of each tumor were analysed, and relative regional cerebral blood volume ratios were determined (rrCBV[tumor/contralateral normal white matter]) to discover whether these parameters can be used to separate PCNSL from high grade gliomas. Statistical analysis was performed with the Student’s t test where a P-value less than 0.05 was considered statistically signiﬁcant. Results: The rCBV ratio was 1.29p0.33 in PCNSL and 3.48p1.32 in high grade gliomas (P 0.0001). Conclusion: rCBV ratios for the two groups can help discriminate between PCNSL and high grade gliomas. PCNSL showed an rCBV ratio lower than high grade gliomas. An rCBV ratio value of 2 represents a valid threshold by which the two groups can be distinguished.

comprised identiﬁcation of ‘hot spots’, i.e. those tumor areas with highest values on parameter maps, e.g. maximum tumor perfusion or highest metabolite ratios of choline-containing compounds/N-acetyl-aspartate within tumor tissue. Then co-localization of these ‘hot spots’ was assessed. Results: FLT-PET, DSC- and DCE-MRI, and MRSI evidenced glioma heterogeneity in all 18 high-grade gliomas but not in ﬁve patients with histologically proven grade II gliomas. Tumor areas with increased thymidine-uptake and highest choline, both suggestive of increased tumor proliferation, were co-localized. Also, microcirculation was elevated in these tumor areas compared to other contrast-enhancing tumor regions as demonstrated by DSC- and DCE-MRI. 23Na-MRI was unable to depict these target areas but depicts the highest signal within necrotic tumor areas. Conclusion: Both imaging techniques that depict microcirculation and techniques that visualize proliferation identify similar target areas. Because of the shortest acquisition and post-processing time, we prefer the use of DSC perfusion imaging.

Scientific Sessions B-218

14:00 - 15:30

15:03

Tumor volume estimation from MR images in patients with malignant glioma. Results from the American College of Radiology imaging network (ACRIN) 6662 trial B. Ertl-Wagner1, J.D. Blume2, D. Peck3, J. Udupa4, I. Schmalfus5, B. Herman2, A. Levering4; 1Munich/DE, 2Providence, RI/US, 3Detroit, MI/US, 4Philadelphia, PA/US, 5 Gainesville, FL/US ([email protected]) Purpose: Reliable assessment of tumor growth in patients with malignant glioma poses a common problem both clinically and for assessing the response to novel therapeutic agents. We aimed to evaluate two software systems in their ability to estimate volume change of tumor and/or edema on MRI of patients with malignant glioma. Methods and Materials: Twenty patients with histology-proven malignant glioma from different sites were included. MR images were assessed with two software systems representative of the two fundamental segmentation methods, single image fuzzy analysis (3DVIEWNIX-TV) and multi-spectral image analysis (Eigentool) and with a manual method by 16 independent readers (eight MR-certiﬁed technologists, four neuroradiology fellows, four staff neuroradiologists). Enhancing tumor volume and tumor volume plus edema were assessed separately by each reader with each software system at two time points. Intraclass correlation coefﬁcients (ICC), variance components, and prediction intervals were estimated. Results: There were no signiﬁcant differences in the average tumor volume change over time between the two software systems (p 0.05). In addition, both software systems were much more reliable and yielded smaller prediction intervals than manual measurements. There were no signiﬁcant differences between the volume changes determined by radiologists/neuroradiologists or technologists. Conclusion: Semi-automated software systems are reliable tools to serve as outcome parameters in clinical studies and as a basis for therapeutic decision making in malignant gliomas, whereas manual measurements are less reliable and should not be the basis for clinical or research outcome studies. Supported by the American College of Radiology Imaging Network (Grant number CA80098-05).

B-220

15:12

Neuroimaging of cerebral and facial involvement in Erdheim-Chester disease A. Vendrely, F. Bonneville, J. Savatovsky, J. Haroche, J.-C. Piette, D. Dormont, J. Chiras; Paris/FR ([email protected]) Purpose: To study and describe the MR features of intracranial, orbital and facial bones lesions in Erdheim-Chester disease. Methods and Materials: We retrospectively reviewed the cerebral MR images of 25 histologically proven cases of Erdheim-Chester disease with intracranial, facial bones and/or orbital lesions on MRI. The locations (meninges, brain, pituitary stalk, orbits, facial bones and sinuses), size and number of lesions, and gadolinium enhancement were analyzed. Results: Intracranial lesions were present in 11/25 patients (44%): 7/11 (63%) had meningeal involvement with a single (n=2) or multiple (n=3) focal mass lesions mimicking meningioma, or diffuse dural thickening (n=2). 3/11 (27%) presented with an isolated intra-axial lesion, and only 1/11 patients (9%) demonstrated both meningeal and intra-axial involvement. Among those patients with intracranial lesions, 4/11 (36%) also had hypothalamic and pituitary stalk lesions. Facial bone involvement was observed in 22/25 patients (88%) and bilateral lesions of maxillary sinuses were the most frequent pattern (n=9/25, 36%). Orbital involvement was present in 7/25 patients (28%), with four cases of bilateral intraconal nodular lesions. All patients with an orbital lesion also haboured intracranial and sinuses involvement. Conclusion: This study demonstrates the frequent association between intracranial masses (meningeal and intra-axial), orbital lesions and maxillary sinuses involvement in Erdheim-Chester disease.

188

A

B

C

D

E

F

G

Room R1

Chest

SS 304 Screening/Pulmonary nodule Moderators: G. Bastarrika; Pamplona/ES J. Votrubová; Prague/CZ

B-221

14:00

Screening prior asbestos workers with low-dose computed tomography H.C. Roberts1, D. Patsios1, H. Bayanati1, A. Pereira1, Z. Dong1, M. Johnston2; 1 Toronto, ON/CA, 2Halifax, NS/CA ([email protected]) Purpose: We established a screening program for prior asbestos workers using low-dose computed tomography (LDCT), and report on the ﬁrst 424 study participants. Methods and Materials: We enroll individuals with asbestos exposure 20 years ago and/or known pleural plaques. Following informed consent, LDCT was performed (40-60 mA, 120 kV, 1.25 mm). Parenchymal nodules are followed according to lung cancer screening recommendations; morphology and location of pleural plaques are noted in detail. Results: Between 03/2005 and 04/2007, we enrolled 424 individuals: 416 males, 32-83 years old (mean 61 years). 329 (78%) were current or former smokers. 308 (73%) had pleural plaques, which were mostly calciﬁed (83%), ﬂat (86%), and symmetric (91%), and most commonly involved the costal ( 97%) and diaphragmatic ( 88%) pleura. Reversely, "abnormal" plaques were lobulated (14%), asymmetric with right-sided dominance (2%), or associated with effusions (1%). 11% had diffuse pleural thickening. 312 (75%) had pulmonary nodules, 103 nodules in 86 subjects were r5 mm, resulting in ﬁve CT-guided biopsies. Three lung cancers have been conﬁrmed: two Stage 1 non-small cell carcinoma, and one extensive small cell carcinoma; two suspicious nodules remain under surveillance. Two MPMs and one malignant peritoneal mesothelioma were found, resulting in an overall malignancy rate of 1.4-1.9%. Conclusion: We developed a screening program using LDCT in prior asbestosexposed individuals, and assessed the appearance of normal and “abnormal” pleural plaques. Both MPM and lung cancers were found. We expect to learn more about the “early mesothelioma” with continued screening and observation of changes in pleural plaques.

B-222

14:09

Early lung cancer detection using spiral computed tomography (CT) and ﬂuorescence bronchoscopy G. Serra-Tosio, L. Felix, C. Brambilla, J.-F. Timsit, M. Coulomb, G. Ferretti; Grenoble/FR ([email protected]) Purpose: To report the preliminary results of a lung cancer detection for high risk volunteers using CT and autoﬂuorescence bronchoscopy. Methods and Materials: From 2001 to 2006, 233 volunteers (188 men, 58.7 years, 39.5 pack-years), were enrolled in a prospective cohort study and followed-up annually for ﬁve years. They were considered at high risk of developing lung cancer and distributed in three groups: patients with history of resected non-small-cell lung cancer (n=81), patients with previous head-and-neck cancer in remission (n=63), and current or former smokers with respiratory symptoms (n=89). They received low-dose CT scan, autoﬂuorescence bronchoscopy, and cytological analysis of bronchial aspiration ﬂuid. Results: 29 lung cancers were detected in 22 patients to date. CT scan showed 23 cancers (79.3%), bronchoscopy 13 (44.8%) with lung cancer depicted by both techniques in seven cases (24.1%). There were 14 prevalence cancers in 13 patients (13/233, 5.6%) and 15 incidence cancers in 12 patients. 26 of these tumors were primary lung cancers (17 squamous cell carcinomas, seven adenocarcinomas, two large cell carcinomas), one a recurrence and two pulmonary metastases. 70% of these lesions (20 of 29) were stages 0 or 1. 72.4% (21 of 29) were potentially curatively treated (four cryotherapies, 17 surgeries). Bronchoscopy detected 2 head-and-neck cancers. Conclusion: Both techniques are complementary as bronchoscopy disclosed six cancers that were invisible at CT. The rates of detection by CT alone were 4.7% (11 of 233) on baseline and 2.2% (9 of 409 examinations) on annual CT-screenings. The prevalence rate of detection was 5.6%.

Scientific Sessions B-223

14:18

Localized ground-glass opacities in a CT screening for lung cancer L. Felix, G. Serra-Tosio, S. Lantuejoul, C. Brambilla, J.-F. Timsit, M. Coulomb, G. Ferretti; Grenoble/FR ([email protected])

B-224

14:27

Natural course of pure ground-glass opacity nodules: 5-year follow-up H. Kim, K. Lim, S. Lee, J. Lee, J. Zo, J. Lee; Goyang-si, Gyeonggi-do/KR ([email protected]) Purpose: To investigate natural course of pure ground-glass opacity (GGO) nodules for 5-year follow-up period. Methods and Materials: From May 2001 to December 2002, we encountered 70 patients with pure GGO nodules. Nodules disappeared in 11 and 25 subjects were lost to follow-up. Thirty-four patients with nodules retained for more than six months or surgically removed were included. Clinical ﬁndings (age, sex, CEA level) and size of nodules at initial presentation and changes of nodules in follow-up CT were assessed. HRCT scans for nodules were done. Results: A total 37 nodules were analyzed (two nodules in three, each). The subjects were 17 men and 17 women of average age 50 (range, 35-68 years). Fifteen were ever-smokers and 19 were never smokers. CEA levels ranged from 0.5 to 4.5 ng/ml (average, 2.42 ng/ml). The median follow-up period was 52 months ranging from seven to 71 months. The size of lesion at the time of ﬁrst presentation was 4-15 mm in largest diameter. During follow-up, the size of lesion showed no change in 31 and increase only in two. In 11 patients, the nodules showed no change for more than ﬁve years. Two patients underwent pulmonary resection, one of them was resected for growing nodule after 34-month follow-up. One patient had adenocarcinoma and the other had two bronchioloalveolar carcinomas. Conclusion: Most of the pure GGO nodules never change in size for median 52month follow-up and may cause overdiagnosis in CT screening. Further long-term follow-up studies are necessary.

B-225

14:36

Persistent pure nodular ground-glass opacity in the lung: Comparison of malignancy and benignity C. Park1, J. Goo1, H. Lee1, C. Lee1, K. Kim2, J.-G. Im1; 1Seoul/KR, 2Ilsan/KR ([email protected]) Purpose: To compare clinical and thin-section CT ﬁndings of malignancy and benignity showing persistent pure nodular ground-glass opacities (PNGGOs), and to investigate the possibility of differentiating between them based on those features of persistent PNGGO. Methods and Materials: We evaluated pathologically proven 99 persistent PNGGOs smaller than 3 cm in 67 patients from two tertiary hospitals. Fifty-two nodules were benign and 47 were malignancies. We compared the clinical and thin-section CT ﬁndings of malignant and benign nodules. Results: Patients with malignancy were signiﬁcantly older than patients with benign lesions (57.83p8.57 years vs 52.19p10.71, p=0.005). Size of the malignant lesions was larger than that of benign lesions (10.37p5.74 mm vs 7.93p5.01 mm,

B-226

14:45

Increase in density is associated with malignancy in solid nodules detected at baseline screening of the Dutch-Belgian lung cancer screening trial (NELSON) D. Xu1, R.J. van Klaveren2, G.H. de Bock1, A. Leusveld1, M.D. Dorrius1, Y. Zhao1, Y. Wang1, H.J. de Koning2, E.T. Scholten3, J.A. Verschakelen4, M. Prokop5, M. Oudkerk1; 1Groningen/NL, 2Rotterdam/NL, 3Haarlem/NL, 4Leuven/BE, 5 Utrecht/NL Purpose: To evaluate whether increases in attenuation or changes in morphology, shape or margin in indeterminate solid pulmonary nodules were associated with malignancy. Methods and Materials: 372 purely intra-parenchymal nodules between 50500 mm3 without pleural-, ﬁssure- or vascular attachment and with 1-year of follow-up were included. They were classiﬁed based on morphology (spherical or non-spherical), shape (round, polygonal or irregular) and margin (smooth, lobulated, spiculated or irregular). The mean CT density of all voxels within the margin of the nodule was automatically generated in Hounsﬁeld units (HU) by the Lungcare© software. Results: Of the 372 nodules, 16 (4%) were malignant and 356 (96%) benign. Baseline median CT density was 41.9 HU (range, -195.1 - 192.1 HU) for benign and -2.2 HU (range, -127.5 - 75.0 HU) for malignant nodules (p=0.06). Median change in density for benign nodules was -0.2 HU (range, -93.1 - 175.2 HU) and 12.8 HU (range, -29.5 - 45.1 HU) for malignant nodules (p 0.05). For benign nodules, median volume was 82.8 mm3 (range 50.3-491.0 mm3) and for malignant nodules 274.5 mm3 (range 64.4-498.0 mm3) (p=0.000). Compared with benign nodules, malignant nodules were more often non-spherical, irregular, lobulated or spiculated (p 0.0001). Conclusion: Increase in nodule attenuation is associated with malignancy in solid pulmonary nodules. Malignant nodules were larger at baseline screening and more often non-spherical, irregular, lobulated or spiculated compared to benign nodules. With regard to morphology, shape and margin change over time, there is no signiﬁcant difference between benign and malignant nodules.

B-227

14:54

The added value of consensus double reading in baseline phase of a lung cancer screening Y. Wang1, R.J. van Klavern2, H.J. van der Zaag-Loonen1, Y. Zhao1, A. Leusveld1, D. Xu1, H.J. de Koning2, E.T. Scholten3, J.A. Verschakelen4, M. Prokop5, M. Oudkerk1; 1 Groningen/NL, 2Rotterdam/NL, 3Haarlem/NL, 4Leuven/BE, 5Utrecht/NL ([email protected]) Purpose: To assess the added value of consensus double reading compared with single reading in the baseline phase of a lung cancer screening trial. Methods and Materials: This study was institutional review board approved. Informed consent was obtained from all participants. Double reading was performed on 7074 baseline low-dose CT scans from April 2004 to March 2006. Based on size, follow-up recommendations for small (15-50 mm3), intermediate (50-500 mm3) and large ( 500 mm3) nodules were an annual repeat scan, 3-month repeat scan or referral to a pulmonologist for work-up and diagnosis, respectively. The differences between single reading and consensus double reading in nodule identiﬁcation, follow-up recommendations and malignancy detection were compared. Results: 8114 nodules were identiﬁed by at least one reader, of which 5790 (71%) were small, 2230 (27%) intermediate and 124 (2%) large. The second reading identiﬁed 23% (1519/6625) more nodules, of which 0.7% (11/1519) was large. Consensus double reading upgraded the follow-up recommendations in 1.3% (94/7074) and downgraded it in 1.3% (90/7074). In total, 67 participants were diagnosed as malignant. The sensitivity for malignancy detection ranged from 97% (95% CI: 93-100%) for single reading to 99% (95% CI: 95-100%) for double reading and no

A

B

C

D

E

F

G

189

Friday

Purpose: To study the frequency and natural history of localized ground-glass opacities (GGOs) in a French screening program for lung cancer. Methods and Materials: 280 patients at high risk for lung cancer (221 men; mean age, 58.6 years), divided into four groups (history of lung cancer (n=83), history of head and neck cancer (n=63), symptomatic cigarette smokers (n=88), asymptomatic cigarette smokers (n=46)), were included in a prospective trial with annual low-dose CT for lung cancer screening (median follow-up, 24 months). We recorded all non-calciﬁed pulmonary nodules detected at baseline screening, classiﬁed nodules according to density and determined the frequency and malignancy rate for each type. We retrospectively reviewed all localized GGOs, analyzed CT characteristics on initial diagnostic CT scan and changes during follow-up (median 29.1 months). Results: Nodular GGOs represented 10.2% of the 362 nodules detected at baseline screening and the malignancy rates were 4.6% for solid nodules and 5.4% for nodular GGOs. A total of 75 localized GGOs were detected in 37 patients, 41 (54.7%) were present on baseline and 34 (45.3%) appeared on annual CT. During follow-up, 41 (54.7%) were persistent and 32 (42.7%) disappeared. The resolving localized GGOs were signiﬁcantly more often lobular GGOs (p=0.006), polygonal in shape (p=0.02), mixed (p=0.003) and larger (p 0.0001) than non-resolving localized GGOs. Conclusion: Localized GGOs are frequent and many of them disappeared on follow-up. CT characteristics of resolving GGOs and malignancy lesions may be similar. That is why a short-term CT follow-up is recommended.

p=0.026). Lesion size 8 mm or larger was the optimal threshold for differentiating malignancy and benignity with balanced sensitivity and speciﬁcity. Malignant lesions were more likely to show pleural retraction, bubble lucency or air bronchogram (p=0.035). Signiﬁcant associations between malignancy and patient age (odds ratio, 1.061), lesion size 8 mm or larger (odds ratio, 2.784), and the presence of pleural retraction, bubble lucency or air bronchogram (odds ratio, 4.414) were found. With these characteristics, the differentiating performance of multiple logistic regression models was relatively good (AUC, 0.762). Conclusion: There was a statistically signiﬁcant difference between benignity and malignancy manifesting as persistent PNGGOs with regard to the patient age and the lesion size, and the presence of pleural retraction, bubble lucency or air bronchogram. These features could be useful in differentiating malignancy from benignity manifesting as persistent PNGGOs.

Scientific Sessions signiﬁcant difference was found (p=0.27). The speciﬁcity remained constant at 81%. Consensus double reading led to an earlier referral of three cancer cases but later referral of one cancer case. Both readings missed one malignancy. Conclusion: The added value of consensus double reading is low at baseline lung cancer screening.

B-228

15:03

Clinical utility of a virtual non-enhanced image of a dual-energy CT in the evaluation of solitary pulmonary nodule: Initial experience E. Chae1, J.-W. Song1, J. Seo1, B. Krauss2, K.-S. Song1; 1Seoul/KR,2Forchheim/DE ([email protected]) Purpose: To know the clinical utility of a virtual non-enhanced image of a dualenergy CT (DECT) in evaluation of solitary pulmonary nodule (SPN). Methods and Materials: Forty-nine patients with SPN were scanned using SOMATOM Deﬁnition (Siemens) at 140/80 kV. Non-enhanced (Scan0) and 3-mindelayed scan (Scan3) after contrast were obtained. Weighted average imageScan0 (Real Nonenhanced, RN), weighted average imageScan3 (Real Enhanced, RE), virtual non-enhanced imageScan3 (Virtual Nonenhanced, VN), and iodine imageScan3 (Virtual Enhanced, VE) were produced. CT numbers on VN vs RN and on VE vs degree of enhancement (CT number on RE-CT number on RN) were compared. On VN, margin characteristic, number and size of calciﬁcations within SPN and lymph nodes, noise, and diagnostic quality were compared with those on RN. Results: CT numbers on RN vs VN and VE vs degree of enhancement showed good agreements (intraclass correlation coefﬁcients, 0.68 and 0.70, respectively). Margin characteristics of VN and RN showed a strong agreement (weighted Kappa, 0.91). On VN, 61.9% (13/21) and 91.3% (42/46) of calciﬁcations within SPN and lymph nodes were detected compared to RN. On VN, sizes of calciﬁcations were smaller than those on RN. VN showed similar noise in 16.3% but more noise in 83.7%. VN showed similar diagnostic quality in 77.6% but inferior or unacceptable quality in 24.4%. Conclusion: Acquisition of VN using a DECT provides information about degree of enhancement and detection of calciﬁcation without additional radiation dose.

B-229

15:12

Solitary pulmonary nodules: A meta-analysis of the diagnostic utility of alternative imaging tests P. Cronin, B.A. Dwamena, A.M. Kelly, R.C. Carlos; Ann Arbor, MI/US ([email protected]) Purpose: To assess the clinical utility of diagnostic tests (dynamic contrast-enhanced computed tomography, magnetic resonance imaging, positron emission tomography and single photon emission computed tomography) for identifying malignancy within a solitary pulmonary nodule (SPN), and to create a nomogram or “look-up” table using clinical data and non-invasive radiology (positive) test results to estimate post-test probability of malignancy by combining the results of one of the four tests to aid clinician management. Methods and Materials: Studies that examined CT, MRI, PET and SPECT for the evaluation of SPN, enrolled at least 10 participants and were published between January 1990 and December 2005 were used. Two reviewers independently abstracted data and assessed study quality. Study-speciﬁc and overall positive likelihood ratios for each diagnostic test conﬁrming a diagnosis of malignancy and negative likelihood for each diagnostic test excluding a diagnosis of malignancy within a SPN were calculated. Results: 44 of 242 articles were included. Positive likelihood ratios for diagnostic tests were: CT 3.91 (95% conﬁdence interval 2.42, 5.40), MRI 4.57 (3.03, 6.1), PET 5.44 (3.56, 7.32) and SPECT 5.16 (4.03, 6.30). Negative likelihood ratios were: CT 0.10 (0.03, 0.16), MRI 0.08 (0.03, 0.12), PET 0.06 (0.02, 0.09) and SPECT 0.06 (0.04, 0.08). Conclusion: From a clinical perspective, differences in performance for all tests were negligible; therefore, the clinician may conﬁdently use any of the four tests presented in further evaluating an SPN. Given the low cost and prevalence of the technology, SPECT appears to be the leading choice for additional testing in SPN evaluation.

B-230

15:21

CT-guided percutaneous core biopsy of small (b1-cm) pulmonary nodules: Diagnostic accuracy and complication rate M.C. Freund, C. Wolf, T. Schmid, M. Hackl, B. Zelger, W.R. Jaschke; Innsbruck/AT Purpose: To assesss the safety and efﬁcacy of percutaneous core biopsy of pulmonary nodules one cm or smaller. Methods and Materials: CT-guided percutaneous core biopsy was performed in

190

A

B

C

D

E

F

G

72 patients aged 25-76 years with 76 small lung nodules. The nodules measured 0.4-1.0 cm (mean 7.8p1.3); all measurements were performed electronically using a commercially available software on a standard desktop PC. CT-guided core biopsy was performed in local (33x) or general anaesthesia (49x). An automated biopsy gun with detachable 17G coaxial cutting needle was used with mean number of 7p3 biopsy attempts. In all patients the needle tract was obliterated with ﬁbrin glue. Results: Core biopsy samples were adequate for diagnosis in 67 (88.2%) of 76 lesions. Diagnosis was malignancy in 32 (49%) and benign ﬁndings in 35 (52%) of 67 lesions. Findings were nondiagnostic in 9 (11.8%) of 76 lesions. Eleven biopsy procedures performed initially in local anaesthesia had to be repeated in general anaesthesia. Perilesional hemorrhage was observed in all patients and obscured in 58 lesions (78%) the delineation during biopsy; no severe hemoptysis was noted. Nine (12.5%) of 72 patients had pneumothorax, with four (5%) requiring CT-guided percutaneous chest tube placement. One patient died due to myocardial infarction one day after CT-guided biopsy. Conclusion: CT-guided percutaneous core biopsy of small (b1 cm) pulmonary nodules is a safe and effective technique to obtain adequate samples for histologic examination.

14:00 - 15:30

Room R2

Cardiac

SS 303 Advances in cardiac MR Moderators: E. Di Cesare; L’Aquila/IT C. Saraiva; Lisbon/PT

B-231

14:00

Safety and efﬁcacy of MR imaging at 1.5 T in patients with pacemakers and deﬁbrillators G.D.E. Papini, F. Secchi, A. Esseridou, P. Lupo, R. Cappato, S. Tiralongo, F. Sardanelli, G.P. Cornalba; Milan/IT ([email protected]) Purpose: Implantable intracardiac devices, i.e. pacemakers (PMs) and implantable cardioverter deﬁbrillators (ICDs), have been regarded as an absolute contraindication to MR-imaging. We tested MR safety and efﬁcacy in patients with PM or ICD. Methods and Materials: After IRB approval, patients with PM/ICD and clinical indication to MR were enrolled. PM-dependant patients and those with devices implanted before 2000 were excluded. Before MR, we collected data about battery and leads and disabled all the therapies. MR exams were performed at 1.5 T. After MR, all devices were checked again. Markers of myocardial necrosis were sampled immediately before and 3 h after MR (only in-patients). Results: We performed 24 exams in 24 patients (17 males; age 63.6p13.8 years) for a total of 12 PMs and 12 ICDs. Body segments: heart (n=15); abdomen (n=2); brain (n=2); spine (n=4), breast (n=1). Two cardiac exams were interrupted due to claustrophobia (n=1) or light reaction to Gd-agent (n=1). After MR, the 24 devices showed neither modiﬁcations nor alert warnings and no signiﬁcant increase of myocardial necrosis markers was observed in the eight in-patients. During the follow-up, all 24 patients did not present any symptom of device dysfunction. Of the 22 completed exams, non-cardiac (n=9) were fully diagnostic while 13 cardiac were fully diagnostic (n=1), with artefacts but partly diagnostic (n=9), with artefacts and nondiagnostic (n=3). Conclusion: MR imaging was safe in all PM/ICD patients, diagnostic in 19/22 exams (86%). Under controlled conditions, MR imaging can be performed in patients with PM/ICD implanted since 2000.

B-232

14:09

MRI-related heating at 1.5 Tesla of a non-ferromagnetic catheter designed for interventional intracardial ablation of arrhythmia A. Koops, B. Lutomsky, M. Steinke, R. Karst, J. Graessner, S. Willems, G. Adam; Hamburg/DE Purpose: To assess MRI-related heating of an intracardial electrophysiological (EP) ablation catheter designed for interventional MRI. Methods and Materials: The non-ferromagnetic ablation catheter was evaluated in torso phantoms ﬁlled with 45 l of polyacrylic-acid gel (PAA) and highly viscous hydroxyl-ethyl-cellulose (HEC) gel, in isocenter position and gradually off-center in 4-cm steps. The steerable distal catheter portion was measured in straight and 90º bent conﬁgurations. Additional measurements were obtained in a ﬂow model of a PAA-ﬁlled tube at 0.5, 0.1 and 0.05 l/min. All tests used a 1.5 T MR system with a

Scientific Sessions and areas rich in extracellular matrix without signiﬁcant cellular interaction. Targed ApoE-micelles demonstrate localization to core areas of atherosclerotic plaque with speciﬁc co-localization with macrophages. Conclusion: We demonstrated the speciﬁc molecular imaging of macrophages using ApoE derived peptide-containing gadolinium mixed-micelles. Further studies are ongoing to investigate molecular nature of the interaction of ApoE micelles with macrophages.

B-233

Purpose: At our institution MSPMRI is performed at 1.5 T with a sensitivity of 90% and a speciﬁcity of 74% for stenoses 70%. Shortcomings at 1.5 T are low SNR and image contrast. Increased ﬁeld strength and a highly concentrated contrast agent should provide optimized conditions to overcome these shortcomings. Aim of our study was to investigate the diagnostic capability of MSPMRI under optimized conditions. Methods and Materials: 57 consecutive patients (age 62.3p11.0 years) with suspected coronary artery disease were examined at 3 T (Magnetom Trio, Siemens, Erlangen, Germany). MMRSPI was assessed using a 2D sr gradient echo sequence in short axis orientation (TR 1.9 ms, TE 1.0 ms, ﬂip 12°, 0.1 mmol Gadobutrol/kg bw (Gadovist, Schering, Berlin, Germany), 140 μg adenosine/kg bw/min). Myocardial function was assessed using cine SSFP sequences in standard angulations. For the determination of myocardial viability, delayed enhancement ir gradient echo sequences were acquired in standard angulations. Perfusion images were assessed visually and semiquantitatively (upslope, peak signal intensity and myocardial perfusion reserve index). Standard of reference was invasive coronary angiography performed in all patients. Results: Stress induced malperfusion was found in 43 patients and myocardial infarction in 27 patients. Sensitivity for hemodynamically relevant coronary artery disease (stenoses 70%) was 97%, speciﬁcity 81%. The myocardial perfusion reserve index was signiﬁcantly reduced in malperfused myocardium with 1.3p0.2 compared to normal myocardium with 2.6p0.7. Conclusion: Our results indicate improved sensitivity and speciﬁcity of MSPMRI for hemodynamically relevant coronary artery disease at 3 T compared to our experience at 1.5 T.

14:18

Diagnostic accuracy of single breath-hold, contrast enhanced whole heart cine MRI using k-t BLAST in measuring LV volumes and function S. Takase, H. Sakuma, K. Inoue, M. Nagata, Y. Mikami, N. Ishida, K. Takeda; Tsu, Mie/JP ([email protected]) Purpose: To determine if single breath-hold whole heart cine MRI accelerated with k-t BLAST can provide accurate measurement of LV function as compared with conventional cine MRI, and to determine the value of gadolinium contrast administration in this approach. Methods and Materials: Sixteen patients with suspected coronary artery disease were studied. Whole heart k-t BLAST cine MR images on 3D axial imaging planes were acquired in 20 seconds by using an 1.5 T MR imager and a balanced TFE acquisition (TR/TE=3.2/1.6 ms, FOV=35 cm, acquisition matrix=160x160x25) before and after administration of Gd-DTPA (0.1 mmol/kg). Conventional balanced TFE cine MR images were obtained on continuous short axis imaging planes. Two observers determined the LV volumes and ejection fraction. Results: Administration of contrast medium signiﬁcantly improved blood contrast on whole heart k-t BLAST cine MRI (CNR=76.1p28.0 before contrast, 162.6p59.2 after contrast, p 0.001), allowing a better delineation of LV endocardial borders. LV volumes measured by k-t BLAST cine MRI with contrast showed an excellent agreement with the results by conventional cine MRI (r=0.98 for LVESV, r=0.95 for LVEDV and r=0.97 for LVEF, p 0.001). The measurement bias between two methods was found to be minimal (8.0p10.3 ml, p=0.57 for LVEDV; 2.9p6.2 ml, p=0.82 for LVESV and -0.1p2.7%, p=0.97 for LVEF) Conclusion: Blood contrast on whole heart k-t BLAST cine MRI was substantially improved if images were acquired after administration of extracellular gadolinium contrast medium. Whole heart k-t BLAST cine MRI can provide accurate measurements of LV volumes and function in a markedly reduced imaging time as compared with conventional cine MRI.

B-234

14:27

Apolipoprotein-E (ApoE) gadolinium nanoparticles for speciﬁc magnetic resonance imaging of atherosclerosis V. Amirbekian1, S. Amirbekian2, E. Vucic3, K.C. Briley-Saebo3, J.S. Aguinaldo3, Z.A. Fayad3; 1Boston, MA/US, 2Atlanta, GA/US, 3New York, NY/US ([email protected]) Purpose: We tested the hypothesis that incorporation of Apolipoprotein-E derived peptide (P2 A2-or-ApoE), which binds LDL-receptor, into Gadolinium mixed-micelles (previously shown to be an effective MRI contrast-agent) will shift the properties of untargeted micelles from untargeted non-cellular to targeted cellular particles for assessment of atherosclerosis with the potential of speciﬁc macrophage imaging. Methods and Materials: After a pre-injection scan using in vivo MRI of the aorta, atheroslceortic ApoE-/-mice (n=5) were injected with ﬂuorescent-labeled ApoEMicelles (0.038 mmol-Gd/kg). A post-contrast enhanced MRI-scan was performed at 24 h using a T1W-black-blood sequence. For controls, atherosclerotic mice (n=6) were injected with ﬂuorescent untargeted micelles and were imaged as described. Another control group (n=6) was injected with Gd-DTPA and imaged at one and 24 h post-injection. After MRI, the aortas were removed and stained using various techniques for imaging using confocal laser-scanning microscopy. Results: Standardized relative to muscle (%NER), injection of untargeted and targeted ApoE-micelles resulted in a signiﬁcant enhancement of the vessel wall of atherosclerotic mice. %NER: 62p5 for untargeted micelles, and 113p5 for targeted ApoE-micelles. With GdDTPA, transient lower enhancement was observed in the vessel one h post. No signiﬁcant enhancement of the vessel wall of ApoE-/- or WT mice was observed 24 h post. Confocal-ﬂuorescence-laser-scanning imaging demonstrates the localization of untargeted micelles primarily to perivascular areas

B-235

14:36

Myocardial magnetic resonance stress perfusion imaging at 3 T using a 1-molar contrast agent B. Klumpp, C. Doesch, A. Seeger, J. Doering, T. Hoevelborn, U. Kramer, M. Fenchel, A.E. May, M. Gawaz, C.D. Claussen, S. Miller; Tübingen/DE ([email protected])

B-236

14:45

Cardiac 31P-MRS compared to echocardiographic ﬁndings in patients with hypertensive heart disease without overt systolic dysfunction: Preliminary results T. Burkhard, C. Herzog, S. Linzbach, I. Spyridopoulos, F. Huebner, T. Vogl; Frankfurt a. Main/DE ([email protected]) Purpose: To evaluate changes in high energy phosphate (HEP) metabolism in patients with hypertension and diastolic dysfunction assessed in electrocardiography and tissue Doppler but with normal LVEF 55%. Methods and Materials: 20 patients (16 men and 4 women, mean age 57p13 years) were studied with phosphorus magnetic resonance spectroscopy and echocardiography. MRS was performed at 1.5 Tesla using an ECG-gated CSI sequence with nuclear Overhauser effect. According to electrocardiograph ﬁndings, 12 patients were found to have a diastolic dysfunction whereas eight patients were identiﬁed as normal, serving as control group in the following statistical analysis. All patients had normal systolic function (LVEF 55%). Statistical analysis was done by using meanpSD for description of the data, Spearman correlation and 2-tailed student's-t test for independent samples. Results: No differences were found in weight, BMI, age, LVEF, end-diastolic volume, cardiac output, heart rate and BNP levels between patients and control group (all p 0.05). Myocardial mass at end-diastole correlated signiﬁcantly to PCr/ATPratio (r=-0.66; p=0.04) in patients and control group. Myocardial PCr/ATP-ratio in patients was signiﬁcantly decreased compared to controls (1.21p0.22 versus 1.54p0.24; p=0.006). Conclusion: Cardiac 31P-MRS could be a non invasive means for detecting early states of heart failure in hypertensive patients.

A

B

C

D

E

F

G

191

Friday

15 min lasting SSFP sequence for applying a standardized whole-body averaged SAR of 4.0 W/kg. Temperature was recorded with a ﬁber-optic thermometer at eight catheter locations. Results: The maximum temperature increase at the tip of the straight catheter in isocenter and +4, +8, and +12 cm ex-center positions were 0.7, 1.6, 4.5 and 8.9 ºC for the PAA, and 0.2, 6.4, 11.4 and 15.2 ºC for HEC gel. With the catheter in isocenter and a the distal catheter bent 90º (consecutively 8 cm ex-centered tip), a maximum increase of 1.8 ºC was found, compared to 2.4 ºC of an isocentered tip with similar bending and an 8 cm ex-centered shaft. In the ﬂow model no temperature increase above 0.2 ºC was found. Conclusion: Substantial passive heating of the EP catheter can occur in the MR environment in a clinically unlikely worst-case scenario, as in experiments of right atrial ablation no tip position of more than 4-8 cm ex-center was observed. In consideration of blood ﬂow, no critical temperature increase can be simulated.

Scientific Sessions B-237

14:54

Contrast-enhanced MRI in the acute phase after experimentally induced myocardial ischemia and reperfusion: Can it determine the area at risk? K.F. Kreitner, M. Vosseler, N. Abegunewardene, K.-H. Schmidt, N. Hoffmann, W.G. Schreiber, G. Horstick; Mainz/DE ([email protected]) Purpose: To analyze early (EE) and late enhancement (LE) 24 hours after experimentally induced ischemia and reperfusion in a pig model. Methods and Materials: The LAD of 12 house swines was occluded after the origin of the second diagonal branch for 30 minutes (group 1, n=6), and for 45 minutes (group 2, n=6), respectively. 24 hours after occlusion and reperfusion, the animals underwent contrast-enhanced MRI (ce-MRI) using 0.2 mmol/kg BW of Gd-DTPA and a 2D-Turbo-Flash sequence that was carried out 2 (early enhancement, EE) and 15 minutes (late enhancement, LE) after contrast application. After repeat LAD occlusion at the same location, Evans-blue was applied for delineation of the area-at-risk. The left ventricle was then sectioned, creating short-axis cuts for further documentation and planimetry of the area-at-risk. Finally, a TTC-staining was performed. Results: Based on the TTC-staining, group 1 showed more than 90% of vital myocardial tissue in the area-at-risk. There was a good correlation between the area-at-risk as delineated in the Evans-blue-staining and the EE (r2=0.86) and LE (r2=0.96). There was no correlation between EE and LE and TTC-staining. In group 2, the reperfused myocardium proved to be 80% avital. There was a moderate to good correlation between EE (r2= 0.67) and LE (r2= 0.87) and TTC-staining. Conclusion: Ce-MRI overestimated non-viable myocardium in group 1, while good agreement was present in group 2. However, ce-MRI enabled depiction of the area-at-risk. A reversible capillary leakage after ischemia and reperfusion may explain the underlying pathomechanism.

B-238

15:03

Multimodality MR imaging in detecting coronary microembolization: An animal experiment correlating with MR features and pathological ﬁndings H. Jin, M.-S. Zeng; Shanghai/CN ([email protected]) Purpose: This study was aimed at evaluating the value of multimodality MR imaging in detecting coronary microembolization (CMB). Methods and Materials: CMB was induced in 11 swines by injection of 50,000, 120,000 and 150,000 microspheres (42 μm) into the left anterior descending coronary artery in two, six and three swines, respectively, and coronary angiography (DSA) was performed before and after CMB. Multimodality MR imaging was scanned before and after CMB, including six hours (acute stage) and one week later. We analyzed the features of ﬁrst-pass perfusion imaging and delayed contrast enhancement and evaluated the systolic function. Finally, the animals were sacriﬁced for pathologic analyses (NBT). Results: No abnormality was found on DSA before and after CMB. Hypoenhanced regions were only shown on ﬁrst-pass perfusion imaging of the swines with an injection of 150,000 microspheres. Delayed enhanced areas were found in all the swines during acute stage, and only existed in the swines injected with 150,000 microspheres one week later, which was consistent with NBT pathologic ﬁndings. In all 11 swines, systolic function of anterior wall and anterior septal was impaired during acute stage and improved slightly one week later, and left ventricular ejection fraction decreased during acute stage and recovered gradually accompanying the dilation of ventricular cavity. Conclusion: It is difﬁcult to show CMB through DSA. Multimodality MR imaging is useful for detecting CMB and its manifestations may vary with different doses of microspheres. Follow-up scans may demonstrate the procedure of ventricular remodeling to a certain extent.

B-239

15:12

Is delayed myocardial enhancement imaging in addition to MRangiography useful in patients with peripheral artery disease? A longitudinal study of cardiac events A. Seeger, J. Doering, U. Kramer, F. Grimm, M. Fenchel, B. Klumpp, A.E. May, C.D. Claussen, S. Miller; Tübingen/DE ([email protected]) Purpose: Aim of the study was to assess the diagnostic performance of cardiac MRI as a supplement to a state-of-the-art MR-angiography (MRA) in the case of vascular risk patients. Methods and Materials: 87 consecutive patients (66p10 years) with symptomatic PAD were examined within the clinical indication of MRA at a 1.5 T system. Afterwards delayed cardiac enhancement (DE) imaging was performed using an IR turboFLASH 2D sequence. A follow-up was carried out two years later with regard to cardiac events (cardiac death, myocardial ischemia, heart insufﬁciency, coronary revascularisation).

192

A

B

C

D

E

F

G

Results: In total, 40/87 patients had myocardial infarctions shown in DE (46%). In 25 patients (29%), the myocardial infarction was already known whereas in 15 patients (17%) an occult progressing infarction was diagnosed. Follow-up data could be obtained after two years in 82 patients. 15 patients had a major cardiac event in the follow-up period; 10 (67%) of them already showed DE in the MRI. In the group with occult progressing infarctions, cardiac events occurred in 40% (6/15 patients, cardiac death n=1, ischemia n=4, heart insufﬁciency n=1), in patients with known infarction in 17% (4/23 patients, cardiac death n=1, ischaemia n=3) and in 11% of patients without myocardial scars (5/44 patients, cardiac death n=1, ischemia n=2, heart insufﬁciency n=2). Conclusion: The high prevalence of myocardial infarctions and the relevance to the prognosis show that cardiac MRI is a meaningful supplement to MRA in risk patients with PAD. Important additional information as to the risk stratiﬁcation and best targeted therapy can be obtained.

B-240

15:21

Cardiac MRI at 7-Tesla: Initial experiments in pigs H.H. Quick, F. Breuckmann, K. Nassenstein, L. Schäfer, S. Maderwald, M.E. Ladd, J. Barkhausen; Essen/DE ([email protected]) Purpose: The purpose of this study was to perform cardiac MRI on a 7-Tesla wholebody system in a pig model to evaluate potential advantages and disadvantages speciﬁcally associated with cardiac MRI at high ﬁelds. Methods and Materials: Two fully anaesthetized and ventilated minipigs (25 and 27 kg) were placed feet ﬁrst supine inside an 8-channel transmit/receive head coil (Rapid Biomedical, Wurzburg) such that the thorax was completely covered by the sensitive region of the RF coil. Scanning was performed on a 7-Tesla whole-body MRI system (Magnetom 7T, Siemens, Erlangen). Cardiac function along standard views (short and long axis, 4-chamber, 2-chamber, LVOT) was evaluated using ECG-triggered TrueFISP retro and Cine FLASH sequences. Myocardial tagging was performed in conjunction with FLASH sequences. Image quality was visually assessed for signal homogeneity and myocardium-to-blood contrast. Results: The animals could successfully be examined under stable cardiac conditions over the full length of the experiments (8 hours). While TrueFISP images were degraded by banding artifacts and signal inhomogeneities, the FLASH sequence provided excellent imaging quality with good signal homogeneity and with high myocardium-to-blood contrast. The achieved spatial resolution was 1.1x1.1x4 mm3. Tagging RF pulses with the FLASH sequence provided high-contrast tagging grids on the myocardium that persisted over the full length of the cardiac cycle. Conclusion: This study can be considered as an initial step towards cardiac imaging in high-ﬁeld MRI. While TrueFISP images were severely degraded by artifacts, the FLASH sequence provided excellent image quality, contrast, and spatial resolution for evaluation of cardiac function.

14:00 - 15:30

Room R3

Interventional Radiology

SS 309 Thoracic interventions Moderators: B.V. Czermak; Innsbruck/AT M. Köcher; Olomouc/CZ

B-241

14:00

Risk factors involved in the development of pneumothorax during radiofrequency ablation in lung neoplasms N.A.N. Mohammed, N.N.N. Naguib, M.G. Mack, T.J. Vogl; Frankfurt a. Main/DE ([email protected]) Purpose: To evaluate retrospectively the risk factors affecting the incidence of pneumothorax and necessity of intercostal chest tube placement during radiofrequency ablation of lung neoplasms. Methods and Materials: This retrospective study was applied for 124 ablation sessions for lung tumors (10 primary lesions and 72 metastatic lesions) in 82 patients (46 males, 36 females; range 37-82 years) in the time period between March 2005 and July 2007. The exclusion criteria for ablation therapy were, lesions with maximal diameter 5 cm, or lesions more than ﬁve in number. A bipolar electrode needle was used under CT guidance. Four cases required two simultaneous electrodes. Results: The incidence of pneumothorax was 11.3 % (14 of 124 sessions). Risk factors involved in the development of pneumothoraces include: male sex (71.4%), chronic obstructive airway disease (COPD) (64%), anatomical location of the lesion,

Scientific Sessions

B-242

14:09

Transpulmonary chemoembolization (TPCE) as a treatment for nonresectable lung metastases T.J. Vogl, K. Eichler, T. Lehnert, R. Hammerstingl, S. Lindemayr; Frankfurt a. Main/DE ([email protected]) Purpose: To evaluate tumor response with volumetric assessment for tumor sizes after treating unresectable lung metastases with transpulmonary chemoembolization (TPCE) in palliative intention. Methods and Materials: From 2001 to 2007, 93 patients (41 males/52 females; mean: 59.9 years) suffering from 527 unresectable lung metastases were treated with TPCE. Treatment was repeated between 1 and 10 times (mean: 3.2). They featured a mean of 6.5 metastases of different origins: colorectal carcinoma (n=40), breast cancer (n=12), renal cellular carcinoma (n=7), thyroideal cancer (n=4), cholangiocellular carcinoma (n=2), sarcoma (n=5), and others (n=23). Tumor feeding arteries were selectively probed after puncturing the femoral vein. Five to 10 ml Lipiodol, 5-10 mg Mitomycin C and microspheres (Spherex) were administered by balloon protection. At 4-week intervals, diagnosis and follow-up were accomplished using unenhanced and contrast-enhanced computed tomography. Results: All patients tolerated the treatments well without any complications or side effects. In 43.0% (n=40) moderate-high Lipiodol uptake was found, while 56.9% (n=53) of the treated tumors showed a low storage. In 24.7% (n=23) of embolized metastases, tumor volume was resolved to 56.3% on average. In 29.0% (n=27) tumor volume remained unchanged and in 46.2% (n=43) volume increased to 122.8% on average. According to the RECIST criteria “partial response” was recorded in 23 cases, “stable disease” in 27 cases and “progressive disease” in 43 cases. Mean survival according to the Kaplan-Meyer analysis was 715 days. Conclusion: According to these ﬁndings, TPCE might be a well-tolerated procedure for the palliative treatment of nonresectable lung metastases.

B-243

14:18

Circumferential ground glass opaciﬁcation with a margin of at least 3 mm on CT predicts complete ablation in RFA of lung metastases E.M. Anderson, A.R. Gillams, W.R. Lees; London/UK ([email protected]) Purpose: Radiofrequency ablation (RFA) of lung tumours is gaining acceptance but the optimal treatment endpoints remain contentious. During treatment, a halo of ground glass opaciﬁcation (GGO) develops around the tumour. We assessed the ability of GGO to predict complete ablation. Methods and Materials: Thirty-six metastases (mean size 2 cm [0.5-4.2]) were ablated in 24 patients (11 female, mean age 67 yrs [41-81]). Ablation was performed with a cool-tip electrode, active tip 3 cm, powered by a 200 W generator (Valleylab, CO, USA) using CT ﬂuoroscopic guidance. Twenty-one treatments were performed under conscious sedation and three under general anaesthesia. Un-enhanced CT scans, maximum collimation 3 mm, obtained immediately, 1 day and 1 month after ablation were reviewed by a single reader, blinded to any subsequent follow-up. The locations of maximal and minimal GGO were recorded. A threshold of 3 mm was considered as an inadequate margin. Recurrence was diagnosed by a second reader if there was enlargement of part/whole of the ablation zone or new nodular enhancement and the location was recorded. Results: Recurrence occurred in 12/36 (33%). Circumferential GGO was predictive of complete ablation with no recurrence in 21/24 (88%) tumours. A 3 mm margin was predictive of recurrence in 10/12 (83%) tumours (chi square=17.4, p 0.0001). There was excellent agreement (Kappa=0.727) between the location of recurrence and the site of inadequate margin after treatment. Conclusion: A GGO margin of 3 mm is an excellent predictor of complete ablation.

B-244

14:27

Percutaneous cryoablation for normal lung of rabbits: Radiologic and pathologic correlation G.-Y. Jin, Y. Kim, Y. Han; Jeon-ju/KR ([email protected]) Purpose: To assess CT and histologic ﬁndings according to variable freezing time during cryoablation in normal lung of rabbit. Methods and Materials: We performed percutaneous cryoablation composed of two freezing cycles (freezing-thawing-freezing; 10-5-10 minutes) for group 1 and one cycle of variable freezing time for group 2 (10, 15, and 20 minutes). We measured the longest and shortest diameters (Dl and Ds, respectively) on all CT images of the ablated zone which were taken after cryoablation, and calculated means and standard deviation of Dl and Ds. Within 72 hours after cryoablation, we sacriﬁced all rabbits. Results: Ablated zone on CT images is deﬁned by consolidation and ground glass opacity. In group 1, mean Dl and Ds [mm] of the ablated zone were 28.1p3.9 and 16p1.6. In group 2, mean Dl and Ds were 27p3.2 and 10.9p3.5 for 10 minutes, 27.4p7.2 and 12.7p3.1 for 15 minutes, and 28.7p6.3 and 14.4p3.8 for 20 minutes. On histologic ﬁndings, the ablated zone was composed of necrotic, hemorrhagic, and inﬂammatory areas, in order from center to outside. In group 1, each area was well differentiated. In group 2, each area was poorly differentiated and intermingled except for freezing of 20 minutes, which was similar to group 1. Conclusion: A total freezing time of at least 20 minutes and two more cycles may be important factors for making sufﬁcient area of ablated zone and uniform necrotic area during percutaneous cryoablation.

B-245

14:36

Thermal ablation of the lung: In vivo experimental comparison of microwave and radiofrequency L. Crocetti, E. Bozzi, A. Sbrana, R. Lencioni, C. Bartolozzi; Pisa/IT ([email protected]) Purpose: To compare feasibility, safety and effectiveness of microwave (MW) ablation and radiofrequency (RF) ablation in a lung rabbit model. Methods and Materials: Twenty New Zealand White rabbits were submitted to either MW ablation (n=10, group A) or RF ablation (n=10, group B) of lung tissue. The procedures were carried out with a prototype MW ablation device (Vivawave; Covidien) and a commercially available RF ablation system (Cool-tip; Covidien). Pathology assessment was performed three days (group A, n=5; group B, n=5) and seven days (group A, n=5; group B, n=5) after the procedure. Results: Technical success was achieved in 9/10 rabbits in both groups. One rabbit of group B died 24 hours after the procedure because of massive hemothorax. Complications included pneumothorax (group A, n=4; group B, n=4), abscess (group A, n=1; group B, n=1), and thoracic wall burn (group A, n=4). The diameter of ablation zones was 12.1p3.2 mm in group A and 14.8p4.9 mm in group B (p=0.2). At histopathology, zones of thermal injuries in group A and B were similar, with septal necrosis, edema, hemorrhage and peripheral lymphocytic inﬁltrate. Thrombosis of small vessels surrounding the ablation zone was extensively depicted in group A specimens and focally present in group B specimens. No severe damage of adjacent organs was observed. Conclusion: Feasibility and safety of MW ablation are similar to those of RF ablation in a lung rabbit model. MW and RF ablated zones are similar in pathologic appearance. MW ablation produces a greater damage to peripheral small vessels inducing thrombosis.

B-246

14:45

Transpulmonary chemoembolization (TPCE) as a palliative treatment for non resectable primary tumors of the lung S.G. Lindemayr, T. Lehnert, K. Eichler, H. Korkusuz, R. Hammerstingl, S. Zangos, T.J. Vogl; Frankfurt a. Main/DE ([email protected]) Purpose: Purpose of this study was to evaluate tumor-response with volumetric assesment for tumor sizes after treatment of unresectable primary lung tumors with TPCE in palliative intention. Methods and Materials: Between 2002 and 2007, 30 patients suffering from unresectable primary lung tumors were treated with transpulmonary chemoembolization. Treatment was repeated between one and 26 times with a mean of 4.2 sessions. Mean age was 63.6 years, female to male ratio was 8:22. Patients suffered from different primary lung tumors: adenocarcinoma (n=7), pleuramesothelioma (n=2), squamous cell carcinoma (n=2), SCLC (n=2) and NSCLC (n=17). Tumor-feeding-arteries were selectively probed after puncture of a femoral vein and 5-10 ml Lipiodol and 5-10 mg Mitomycin C and microsphere particles (Spherex) were administered during balloon-protection. At intervals of four weeks, diagnosis

A

B

C

D

E

F

G

193

Friday

as basal lesions showed incidence (71.4%), crossing a major lung ﬁssure along the way to the lesion was associated with incidence (28.6%), central intraparenchymal lesion (85.7%) in comparison to peripheral subpleural lesions (16.3%). Multiple pleural puncture by multiple electrodes was also recognized (1/4 cases in which multiple electrodes were needed). 14.3% of pneumothoraces patients were treated conservatively; in 57.1% of pneumothoraces immediate complete evacuation by intercostal catheter and manual evacuation was successful, while chest tube placement was indicated in 28.6% (4/14 pneumothoraces patients). Conclusion: Pneumothorax is a controllable complication of radiofrequency ablation of lung neoplasms and essential risk factors have to be ruled out.

Scientific Sessions and follow-up were accomplished by using unenhanced and contrast-enhanced computed tomography. Results: All patients tolerated the repeated treatments without adverse effects; neither complications nor inﬂuences on labaratory parameters have been observed. In 40.0% (n=12) moderate to high Lipiodol uptake was found while 60.0% (n=18) of the treated tumors showed a low storage. In 20.0% (n=6) of embolized lesions, tumor-volume was resolved to 36.7% on average. In 46.6.% (n=14), tumor-volume remained unchanged and in 33.3% (n=10), volume increased by 1010.3 % (27 ml) on average. Conclusion: According to these ﬁndings, transpulmonary chemoembolization (TPCE) might be a well tolerated procedure for the palliative treatment of non resectable lung metastases.

B-247

14:54

Role of 64-slice CT in pre-embolization evaluation of hemoptysis: Our initial experience S.S. Hedgire, P. Gupta, J. Sathenjeri Rajendra, V. Ilayaraja, T. Kalyanpur, M. Cherian, P. Mehta, R. Karwal; Coimbatore/IN ([email protected]) Purpose: Role of MDCT in demonstrating systemic and bronchial supply in hemoptysis and its relevance to trans-arterial embolization. Methods and Materials: Eleven consecutive patients referred for therapeutic embolization were evaluated with light-speed VCT scanner (GE Medical systems, Milwaukee, WI). CT aortic angiography was performed following administration of 70 cc nonionic intravenous contrast media at the injection rate of 4 ml/s. The angiography was performed using the Smart Prep technique by placing the ellipse region of ineterest (ROI) on the descending aorta. The abnormal vessels were retrogradely reconstructed from the area of parenchymal abnormality in the lungs to the aorta using the GE Advantage Workstation 4.3. The origin and number of vessels were localized to plan therapeutic embolization. Results: A total of 11 patients referred for therapeutic embolization underwent CT aortic angiography. The total number of abnormal vessels were found to be 10. A normal CT angiogram was seen in one patient. The culprit vessel was found to be the bronchial artery in all the cases. Aberrant origin of bronchial artery was seen in two cases of which one was seen to arise from the subclavian artery and the other from the arch of aorta. Excellent angiographic correlation was seen in all 10 cases that underwent a subsequent digital substraction angiogram and therapeutic embolization. Conclusion: 64-slice CT aortic angiography is an excellent modality to study the systemic and bronchial vessels in hemoptysis. It not only helps in planning transarterial embolization but also in considerably reducing the ﬂuoroscopic time.

B-248

15:03

Bronchial artery embolisation, an effective procedure for treatment of massive hemoptysis in pulmonary tuberculosis K. Taori, B.K. Saha; Nagpur/IN ([email protected]) Purpose: Pulmonary tuberculosis is very common in developing countries in Asia including great prevalence in the Indian subcontinent. One of the dreadful complications of pulmonary tuberculosis is massive hemoptysis. This study is about treating patients of massive hemoptysis by bronchial artery embolization. Methods and Materials: Bronchial artery embolization was done in 101 patients (85 male, 16 female) with massive hemoptysis between 14 and 73 years of age (mean age 46 years). Pre-procedure chest radiograph was done in each case. Bronchial artery embolization was performed on the side with the greater abnormality on the chest radiograph. Gel foam was used as embolising agent. These were introduced through a 4 F visceral hook catheter. Pre- and post-procedure angiographic ﬁlms were obtained. Results: Embolization was performed in 100 of 101 patients. Bronchial arteries (n=76) as well as nonbronchial arteries like intercostal arteries (n=98), internal mammary (n=36), lateral thoracic (n=25), costocervical trunk (n=3) and thyrocervical trunk (n=3) responsible for haemoptysis were selectively embolized. The average number of arteries embolized per patient was 2.7. Out of the 100 patients immediate control of hemoptysis was achieved in 98 patients within a period of 24-48 hours with a procedure success rate of 98%. Twelve patients had rehaemoptysis within 30 days with re-bleed rate of 12%. Out of these 12 patients, eight patients were re-embolized and the remaining four patients were managed conservatively. Conclusion: Bronchial artery embolization is an effective procedure for treatment of massive hemoptysis in pulmonary tuberculosis.

194

A

B

C

D

E

F

G

B-249

15:12

Accuracy of CT-guided transthoracic needle biopsy (TNB) of lung lesions: Results of 828 consecutive procedures A.M. Priola, S.M. Priola, A. Cataldi, G. Garofalo, P. Ricca, V. Marci, C. Fava; Turin/IT ([email protected]) Purpose: Aim of this study was to evaluate the efﬁcacy of CT-guided TNB and to analyze variables affecting diagnostic accuracy. Methods and Materials: From November 2002 to August 2006, a total of 822 patients (828 procedures) were prospectively considered. At time of TNB the onsite pathologist assigned to each specimen a semiquantitavive score (0 bloody sample without other cells; 1 aspeciﬁc benign or inﬂammatory cells; 2 malignant cells without histotype characterization; 3 established benign or malignant histotype. 0-1 inadequate samples; 2-3 adequate samples). Factors that could affect diagnostic accuracy were classiﬁed as lesion and procedure related: lesion size was considered the average lesion diameter in two axial planes. All variables were analysed by Chi-square and Student’s-t-test for correct or incorrect diagnosis. A p-value less than 0.05 was considered statistically signiﬁcant. Results: Most of the procedures were performed by needle aspiration biopsy (93%), while in a minority of cases (7%) a true-cut biopsy was requested. A single sample was obtained in 69% of cases (mean 1.3 specimens for procedure; range 1-4). 72% of samples obtained higher score (89% malignancies). A deﬁnitive diagnosis was made in 83.5% of procedures, while a score of 0-1 was assigned in 137 cases. The diagnostic accuracies for benign and malignant lesions were 67 and 92%, respectively, with overall diagnostic accuracy of 83.3%. The variables affecting diagnostic accuracy were ﬁnal diagnosis (benign 67%, malignant 92%; p 0.001) and lesion size (lesions 1.5 cm 68%, 1.5-5 cm 87%, 5 cm 78%; p 0.05). Other variables did not affect diagnostic accuracy. Conclusion: In consecutive cases of CT-guided TNB ﬁnal diagnosis and lesion size affect diagnostic accuracy.

B-250

15:21

Factors affecting risk of pneumothorax in CT-guided transthoracic needle biopsy (TNB) of lung lesions: Results of 892 consecutive procedures A.M. Priola, S.M. Priola, A. Cataldi, G. Garofalo, F. Pazè, S. Durando, V. Marci, C. Fava; Turin/IT ([email protected]) Purpose: Aim of this prospective study was to estimate the risk of pneumothorax in patients undergoing CT-guided TNB and to determine which factors affect its occurrence. Methods and Materials: Between November 2002 and April 2007, 892 TNB were performed in 870 patients. Risk factors for pneumothorax were classiﬁed in three groups: patient, lesion, and procedure-related. Lesion depth was measured as the length of the aerated lung traversed from the surface of the pleura to the edge of the lesion. Lesion size was considered the average lesion diameter in two axial planes. All variables were analysed by Chi-square and Student’s-t-test for occurrence of pneumothorax. A p-value 0.05 was interpreted as statistically signiﬁcant. Results: Pneumothorax occurred in 228/892 procedures (25.6%); a chest tube was required in 22 cases (2.5%). Increased lesion depth was the most signiﬁcant predictor of pneumothorax (p=0.002): the mean depth of lesions from the pleural surface was 27.4 mm in patients with pneumothorax and 17.2 mm in patients without pneumothorax. The pneumothorax rate increased with the increasing distance of the lesion from the pleural surface. Among lesion-related variables, chest wall invasion (p 0.03) and lesion size (p=0.03) showed correlation. The greater frequency of pneumothorax was seen in smaller lesions: for lesions 1 cm or smaller, the pneumothorax rate was 35%. Among patient and procedure variables, age (p 0.03) and number of cutting specimens (p=0.01) were associated with an increased risk of pneumothorax. Other variables did not affect risk of pneumothorax. Conclusion: In consecutive cases of CT-guided TNB the length of lung parenchyma crossed during the biopsy is the predominant risk factor for pneumothorax.

Scientific Sessions

A

B

C

D

E

F

G

195

Saturday

Saturday, March 8

Scientific Sessions room A 2nd level

room B 2nd level

room C 2nd level

room E1 entr. level

room E2 entr. level

room F1 entr. level

room F2 entr. level

room G/H lower level

room I lower level

07:00

07:00

07:30

07:30

08:00

08:00

08:30 CC 516 MRI Neuro: The brain as the centre of the world (p. 32)

09:00

09:30

CC 517 Multidetector CT Made Easy The guts: Growing opportunities (p. 32)

NH 5 New Horizons Session The further minimisation of minimally invasive therapy (p. 33)

CC 518 Breast: From Basics to Advanced Imaging Breast lesions at mammography: How to characterise? (p. 34)

3

E 520a Foundation Course: Interventional Radiology Liver tumors (p. 34)

E3 520b Interactive Teaching Session Prostate cancer: Meeting clinical needs by MRI at diagnosis and on follow-up (p. 35)

08:30 SF 5a Special Focus Session Nephrogenic systemic ﬁbrosis: Current situation and future direction (p. 35)

EF 1 EFOMP Workshop State-of-the-art of high ﬁeld MRI instrumentation (p. 36)

WS 521 Image Guided Breast Biopsy: How to do it Hands-on Workshop Lecture

10:30

10:30 EM 1 ESR meets Germany Computer aided diagnosis (CAD) (p. 41)

11:00

11:30

SS 610 Musculoskeletal Hand, knee, feet (p. 198)

SS 601a Abdominal Viscera Functional imaging of the abdomen: Are we there yet? (p. 200)

3

E 620 Foundation Course: Interventional Radiology Musculoskeletal (p. 42)

SS 609a Interventional Radiology Spine (p. 202)

SS 602 Breast Intervention (p. 204)

EF 2 EFOMP Workshop Clinical applications, safety issues and regulations for very high ﬁeld MRI (p. 42)

SS 613 Physics in Radiology Radiation dose in CT (p. 206)

12:00 Honorary Lecture 1 (p. 43)

13:00

11:30

12:30 SY 3 GE Healthcare Satellite Symposium (p. 537)

SY 4 Bracco Satellite Symposium (p. 538)

Workstation Face-Off Session

13:00

13:30

13:30

14:00

14:00 SS 701a Abdominal Viscera Magic bullets for hepatic lesions (p. 219)

SS 710 Musculoskeletal Arthritis (p. 217)

14:30

15:00

SS 701b GI Tract Colon and rectum (p. 221)

SY 6 Bayer Schering Pharma Satellite Symposium (p. 539)

E3 720 Interactive Teaching Course: Useful Signs in Imaging Signs in abdominal imaging (p. 43)

SS 709a Interventional Radiology Radiofrequency ablation (p. 223)

SY 7 Siemens Satellite Symposium (p. 540)

SS 708 Head and Neck New approaches to examination techniques (p. 225)

SS 713 Physics in Radiology New approaches in MRI (p. 227)

15:30

14:30

15:00

15:30

16:00

16:00

16:30

registration

EPOS™ - scientiﬁc exhibition

17:30

technical exhibition

17:00

CC 816 MRI Musculoskeletal: A step-by-step approach (p. 44)

CC 817 Multidetector CT Made Easy The chest: A high resolution country (p. 44)

EM 2 ESR meets General Practitioners Manifestations of atherosclerosis: A common challenge in diagnosis and treatment (p. 45)

RC 810 Musculoskeletal Osteoporosis (p. 46)

E3 820 Interactive RC 807 Teaching Course: Genitourinary Useful Signs in What to do with all Imaging these small renal Signs in masses? genitourinary (p. 47) imaging (p. 47)

RC 802 Breast MRI of the breast (p. 48)

RC 811 Neuro Basic MR (p. 49)

RC 815 Vascular Non-invasive imaging of the lower limb venous system (p. 49)

16:30

17:00

17:30

18:00

18:30

19:00

196

11:00

12:00

12:30

18:30

09:30

10:00

10:00

18:00

09:00

19:00

A

B

C

D

E

F

G

Scientific Sessions room K lower level

room L/M 1st level

room N/O 1st level

room R1 1st level

room R2 1st level

room R3 1st level

room U 2nd level

room X 2nd level

room XI, Z & Y 2nd level

room Z 2nd level

07:00

07:00

07:30

07:30

08:00

08:00

08:30

09:00

09:30

08:30

HS 1 Hospital Management Symposium Management

SF 5b Special Focus Session Comprehensive evaluation of the heart (p. 37)

MC 519 Molecular Imaging Molecular imaging in cardiovascular diseases (p. 38)

RC 504 Chest Pulmonary embolism update (p. 38)

RC 501 Abdominal and Gastrointestinal Imaging in pancreatitis (p. 39)

RC 507 Genitourinary Imaging of the pelvic ﬂoor (p. 40)

09:00

09:30

10:30

11:00

10:00

HS 2 Hospital Management Symposium Information Technology

10:30 SS 611 Neuro New vascular MR techniques (p. 208)

11:30

SS 604 Chest Reporting and CAD (CXR and CT) (p. 210)

SY 2 Bracco Satellite Symposium (p. 536)

SS 603 Cardiac Imaging of the right heart (p. 212)

SS 609b Interventional Radiology Liver: Embolisation (p. 215)

WS 621 Image Guided Breast Biopsy: How to do it

WS 622 Experience Vascular Procedures Using Simulators

WS 623 Tips and Tricks in Radiofrequency Ablation

11:00

11:30

12:00

12:30

12:00 HS 3 Hospital Management Symposium Finance.

SY 5 Siemens and Bayer Schering Pharma Joint Satellite Symposium (p. 539)

13:00

12:30

13:00

13:30

13:30

14:00

14:30

14:00

SS 712 Pediatric Fetal imaging (p. 229)

SS 715 Vascular MR angiography (p. 231)

SS 711 Neuro Grey and white matter (p. 234)

SS 704 Chest Pulmonary embolism (p. 236)

15:00

SS 703 Cardiac CT/MR in cardiovascular risk assessment (p. 238)

SS 709b Interventional Radiology Musculoskeletal intervention (p. 240)

WS 721 Image Guided Breast Biopsy: How to do it

WS 722 Experience Vascular Procedures Using Simulators

WS 24B1 Test-Driving your Workstation

WS 723 Tips and Tricks in Radiofrequency Ablation

15:00 WS 24B2 Test-Driving your Workstation

15:30

16:00

16:30

17:00

TF 1 RTF Radiology Trainees Forum Highlighted Lectures (p. 50)

SF 8a Special Focus Session Fat radiology: Imaging in obesity and metabolic syndrome (p. 51)

RC 808 Head and Neck Imaging after treatment of head and neck tumors (p. 51)

RC 803 Cardiac How to perform a complete CT/MR examination of the heart (p. 52)

RC 801 Abdominal and Gastrointestinal Imaging of the acute abdomen (p. 53)

14:30

SF 8b Special Focus Session Novel image analysis and modelling techniques to improve diagnosis and therapy of cancer (p. 54)

RC 809 Interventional Radiology Percutaneous interventions in musculoskeletal disease (p. 53)

15:30

16:00

16:30

17:00

17:30

17:30

18:00

18:00

18:30

18:30

19:00

19:00

A

B

C

D

E

F

G

197

Saturday

10:00

Scientific Sessions 10:30 - 12:00

Room B

SS 610 Hand, knee, feet Moderators: B. Forgács; Rotherham/UK C. Masciocchi; L’Aquila/IT

10:30

Cost-effectiveness of general practitioner direct access to magnetic resonance imaging for assessment of patients with suspected internal derangement of the knee: A randomised controlled trial S.D. Brealey, D. King; York/UK Purpose: To compare the cost-effectiveness of general practitioner referral to Magnetic Resonance Imaging (MRI) with referral to an orthopaedic specialist for patients with suspected internal derangement of the knee (e.g. meniscal or ligamentous injuries). Methods and Materials: A pragmatic multi-centre randomised trial design and concomitant economic evaluation. Eligible patients were referred to see an orthopaedic specialist in secondary care and randomised to have MRI or not before the hospital appointment. To cover events up to and including arthroscopy, patients were followed up from random allocation at 6, 12, and 24 months. Results: 553 patients were recruited from 163 general practices across the United Kingdom from November 2002 to October 2004. Patients randomised to MRI improved mean SF-36 physical functioning scores by 2.45 (95% conﬁdence interval -0.56 to 5.45) compared with direct referral to orthopaedics, which was not statistically signiﬁcant (p=0.110). Patients randomised to MRI improved mean Knee Quality of Life questionnaire (KQoL-26) physical functioning scores by 3.28 (0.68 to 5.88), which was statistically signiﬁcant (p=0.014). There were no other signiﬁcant outcomes. Early MRI is associated with a higher NHS cost, by £294 ($581; _435) per patient (£31-£573), and a larger number of quality-adjusted life years, by 0.05 (0.025-0.118). Mean differences in cost and QALYs generated an incremental cost per QALY gained of £5,840 ($11,538; _8,642). Conclusion: GP access to MRI brought about small but statistically signiﬁcant beneﬁts in patients’ knee-related quality of life and represents a cost-effective use of health service resources.

B-252

10:39

Which oblique plane is more helpful in diagnosing the anterior cruciate ligament tear? Y. Yoon, J. Kwon, I. Lee; Seoul/KR ([email protected]) Purpose: To evaluate the diagnostic role of additional use of oblique imaging for ACL tear and compare the usefulness of oblique coronal and oblique sagittal imaging. Methods and Materials: From January to December 2006, 101 patients who met the inclusion criteria were enrolled in this study: a) no history of previous surgery, b) presence of preoperative knee MRI with both conventional and two sets of oblique coronal and oblique sagittal PD weighted images along the course of ACL, c) less than 180 days interval between MR and surgery. Three radiologists evaluated the MR images by using four methods: conventional MR only (method A), conventional MR and additional oblique coronal MR (method B), conventional MR and additional oblique sagittal MR (method C), conventional MR and additional oblique coronal and oblique sagittal MR (method D). The status of ACL (normal, partial tear, complete tear) was recorded. Sensitivity and speciﬁcity for ACL tear of each method were calculated with arthroscopic ﬁnding as standard reference, and values were statistically analyzed by C2 test. Results: Arthroscopy revealed 10 partial tears and 30 complete tears of ACL. Sensitivity and speciﬁcity for ACL tear were: A, 97/75%, B: 98/87%, C: 98/87%, D: 98/93%. Speciﬁcities of methods B, C, and D are statistically signiﬁcantly higher than that of method A. Conclusion: The role of additional oblique imaging for ACL was improving speciﬁcity. Any one of both oblique imaging techniques is enough, and no further improvement in diagnostic efﬁcacy was achieved by simultaneous use of both.

198

A

B

10:48

Osteochondrosis dissecans of the knee: The evolution of MR morphology during treatment P.D. Palczewski, P. Strzelczyk, M. Golebiowski; Warsaw/PL ([email protected])

Musculoskeletal

B-251

B-253

C

D

E

F

G

Purpose: To assess the changes in MR morphology of osteochondrosis dissecans (OCD) lesions in response to treatment and feasibility of MRI as a control study in patients with treated OCD. Methods and Materials: Eleven patients (six women and ﬁve men, aged 11-24 years, average age 17 years) with lesions affecting 12 knee joints were followed for approximately 24.2 months with repeated clinical examination and MR imaging. All MR studies were performed on a 1.5 Tesla unit. MR artrography was performed in three patients. The evolution of MR morphology of lesions was assessed with a 5-point scale and correlated with the results of clinical assessment with a modiﬁed Lysholm scale. Results: In 10 patients, a good clinical result was achieved. In seven knee joints of patients from this group, the features of the healing of OCD lesion were observed in MRI and in four knees, the morphology of OCD lesion was stable. In ﬁve MR examinations, a transient deterioration was observed, which in four cases preceded healing and in one case stabilization in MR morphology. Conclusion: Transient deterioration may be observed in MR studies prior to the healing of OCD lesions irrespective of the improvement in clinical examination. The follow-up of patients with OCD should rely mainly on clinical examination, with MR studies reserved for late control or patients with clinical signs of deterioration.

B-254

10:57

Iliotibial band friction syndrome in patients with medial compartment knee osteoarthritis V.J.N. Vasilevska1, U. Szeimies2, A. Staebler2; 1Skopje/MK, 2Munich/DE ([email protected]) Purpose: The purpose of this retrospective study was to assess the frequency of MR imaging signs of iliotibial band friction syndrome (ITBFS) in patients with advanced medial compartment knee osteoarthritis. Methods and Materials: Coronal and sagittal PDw fatsat MR images (1.5 T, SL 2.5-3 mm, CP and 8-chanel coil) of 128 patients with advanced medial compartment knee osteoarthritis (41 male, 87 female, median age 63 years) within 4-year-period were analyzed. Graded criteria: cartilage thickness, meniscus degeneration, bone marrow edema, knee effusion. Control group had only medial meniscus degeneration. Lateral compartment in both groups was normal. MRI signs - ﬁbrovascular tissue between lateral epicondyle and ITB - were evaluated in both groups. Results: In the group with severe medial compartment knee osteoarthritis, MRI signs of ITBFS were present in 95/128 patients (74.2%) (moderate 91 (71.1%) and severe ITBFS in 4/128 (3.12%)). Complete cartilage loss in medial compartment was in 69 patients (53.9%), subtotal loss of cartilage in 59 patients (46.1%), with statistically signiﬁcant difference between cartilage thicknesses on medial and lateral compartments (p 0.01). Eighty-nine patients (69.53%) had complete disintegration of the meniscus. In the control group only moderate ITBFS was present (27.7%). There was a statistically signiﬁcant difference between both groups for the presence of ITBFS (pb0.01). Conclusion: MRI signs of ITBFS were often present in patients with severe medial compartment knee osteoarthritis. Severe damage of the cartilage combined with severe degeneration of the meniscus (which lead to joint space narrowing and varus knee deformity) may be one of the causes for ITBFS.

B-255

11:06

Contrast-enhanced MRI for the assessment of bone viability in scaphoid nonunion R.R. Schmitt, S. Fodor, S. Froehner, K. Megele, G. Christopoulos, H. Krimmer; Bad Neustadt a.d. Saale/DE ([email protected]) Purpose: Unenhanced MRI displays the osseous viability by means of high-signal intensity of normal fatty marrow in T1- and T2-weighted sequences. The aim of this study was to evaluate the use of intravenous gadolinium for assessing the viability of the proximal fragment in scaphoid nonunion. Methods and Materials: 88 patients suffering from symptomatic scaphoid nonunion were included in this prospective study. MRI sequence protocol: coronal PD-w FSE fat-suppressed, sagittal-oblique T1-w SE unenhanced and contrast-enhanced with fat-suppression, sagittal T2*-w GRE. Score of the contrast enhancement: 0=none, 1= focal, 2=diffuse. Intraoperatively, the osseous viability was assessed by scoring bleeding points on the osteotomy sites of the proximal scaphoid fragments: 0=missing, 1=moderate, 2=good. Imaging and surgical ﬁndings were compared. Results: Viability states of the proximal fragments were found in surgery/unen-

Scientific Sessions hanced MRI/contrast-enhanced MRI with these frequencies: normal viability in 42/67/53 cases, partial osteonecrosis in 29/20/21 cases, and complete osteonecrosis in 17/1/14 cases. By assessing surgical ﬁndings as the standard of reference, statistical analysis for unenhanced MRI/contrast-enhanced MRI was as follows: sensitivity 6%/77%, speciﬁcity 100%/99%, PPV 100%/93%, NPV 83%/95%, and accuracy 83%/94%. Most obvious was the very low sensitivity (p 0.001) of unenhanced MRI in detecting osteonecrosis. Conclusion: The presence of bone-marrow edema does not allow to rule out osteonecrosis of the proximal fragment in scaphoid nonunion. Bone viability is more accurately assessed by means of contrast-enhanced MRI. Thus, application of intravenous gadolinium is recommended for MRI of scaphoid nonunion.

B-256

11:15

Finger ﬂexor annular pulley reconstruction: Assessment by US A. Klauser, R. Arora, P. Soegner, T. De Zordo, G. Feuchtner, M. Gabl; Innsbruck/AT ([email protected])

B-257

11:24

Midcarpal instability: A diagnostic role for dynamic carpal ultrasound? A.P. Toms, A. Chojnowski, J. Cahir; Norwich/UK ([email protected]) Purpose: The aim of this study was to perform a retrospective review of dynamic ultrasound (US) examinations in patients with a diagnosis of midcarpal instability and US ﬁndings of a triquetral clunk. Methods and Materials: Four patients were identiﬁed (three male, one female). The case notes, plain radiographs, MRI and dynamic US for each patient were reviewed. Digital video ﬁles recording the dynamic US of the triquetral clunks were analysed for the following features of abnormal triquetral motion: direction and speed of triquetral clunk, amount of anteroposterior translocation, and ﬂexion or extension during the triquetral clunk. Results: Five different triquetral clunks were recorded in four patients. In four out of ﬁve cases the clunk occurred during ulna deviation of the wrist, and in one patient it occurred during radial deviation. Anteroposterior translocation was anterior (range 3.4-4.7 mm) in three of the clunks and posterior (range 1-10 mm) in two. The degree of ﬂexion or extension varied between 1 and 16°. The snapping phase of the clunk lasted between 0.17 and 0.25 seconds. A palmar sag of the proximal carpal row during ulna deviation was also noted in one patient. Conclusion: Dynamic US can be used to conﬁrm the diagnosis of midcarpal instability by differentiation of triquetral catch-up clunks from other causes of a wrist clunk or click. Quantiﬁcation of carpal motion with dynamic US may prove to be useful in further classiﬁcation of midcarpal instability.

B-258

11:33

High ﬁeld MR imaging at 7 T of ankles after marathon: Preliminary results P.C. Kokulinsky, O. Kraff, S. Maderwald, J.M. Theysohn, M.E. Ladd, J. Barkhausen, S.C. Ladd; Essen/DE ([email protected]) Purpose: To present a ﬁrst 7T protocol for imaging the ankle of marathon runners and compare ﬁndings to 1.5 T. Methods and Materials: Six marathon runners were examined at 1.5 T (Siemens Sonata) and at 7T (Magnetom 7T) within three hours of each other. Imaging was preceded by a full-length marathon three days before. Sequences used at both ﬁeld strengths were STIR, PD and T2 TSE, MEDIC, DESS, and CISS. 7T was also

B-259

11:42

Dynamic assessment of plantar plate: The role of high resolution ultrasound L.M. Sconﬁenza, F. Lacelli, V. Raimondi, G. Grillo, E. Silvestri, G. Garlaschi; Genoa/IT (io@lucasconﬁenza.it) Purpose: To describe the normal and pathologic appearance of plantar plate, assessed with dynamic high resolution ultrasound (D-HRUS). Methods and Materials: We used D-HRUS to assess 22 patients (age range 38-71) with an orthopaedic diagnosis of plantar plate tear. An MRI control was performed in all cases. Moreover, we performed D-HRUS on 10 healthy volunteers to understand the correct scan technique of the plantar plate and to determine the correct dynamic manoeuvres that allow to depict the normal ultrasound anatomy of this structure. Results: Normal plantar plate is a triangular structure, with homogeneous echogenic appearance; it can be easily assessed with plantar longitudinal scan on the metatarsophalangeal joint (MTPJ) and passive ﬂexion/extension of foot ﬁngers. The dynamic scan allows to depict the relations among the plantar plate and the ﬂexor tendons and the articular cartilage. In our study, D-HRUS has detected a plantar plate tear in 21/22 patients (14 on MTPJ II, ﬁve on MTPJ III, two on MTPJ II and III). They appear like small hypoechoic defects of the distal insertional ﬁbres and can be divided into partial (eight patients) and full thickness tears (13 patients). In all patients, MRI conﬁrmed the diagnosis of plantar plate rupture but in nine cases it has not been able to differentiate a partial from a full thickness tear. Conclusion: HR US allows to assess the plantar plate in detail and to detect its lesions. In particular, the dynamic analysis is able to differentiate a partial from a full thickness tear.

B-260

11:51

The plantar fasciitis: A percutaneous ultrasound (US) guided approach to its treatment M. D’Auria, L.M. Sconﬁenza, D. Morcaldi, F. Paparo, E. Silvestri, G. Garlaschi; Genoa/IT (io@lucasconﬁenza.it) Purpose: To describe the technical approach to US-guided percutaneous treatment of plantar fasciitis. Methods and Materials: Thirty-one patients with clinical and US diagnosis of plantar fasciitis unresponsive to medical therapy were treated. The procedure is performed by two expert radiologists. After the injection of a small amount of local anaesthetic, we perform a dry-needling on its insertional portion and on periostium to produce local phenomena of hyperaemia. Then the needle is retracted to reach the perifascial soft tissues and a small amount of steroid (1 ml of Triamcinolone acetonide 40 mg/ml) is injected. After the treatment, a orthotic soft arch support to relieve the hind part of the foot is suggested. All the patients were followed-up clinically after the treatment for 4-6 months. Results: In 89% of patients we have observed a complete disappearance of symptoms after 2-3 weeks from the treatment. In 8% of patients we have observed a relative worsening of symptoms in the ﬁrst days after the treatment, followed by a progressive reduction and a disappearance of symptoms in about three weeks. In 3% of patients we have observed no signiﬁcant response to the therapy; in these patients, MR showed a remarkable calcaneal bone edema at the fascial insertion. Conclusion: US-guided treatment of plantar fasciitis is a non invasive and successful technique that allows to be more precise when injecting the steroid, avoiding the fascia. The dry-needling performed on plantar fascia and on periostium produces a local hyperemia that can be compared to a surgical debridement.

A

B

C

D

E

F

G

199

Saturday

Purpose: To describe ﬁndings in reconstructed A2/A3 ﬁnger pullies and to assess the distance between ﬂexor tendons and phalanx (TP) postoperatively by using dynamic US. Methods and Materials: Fifteen patients with reconstructed A2/A3 ﬁnger pullies underwent US and clinical examination using the Buck Gramcko score at a mean follow-up time of 48 months. The TP distance was evaluated dynamically in extension and forced ﬂexion at the level of the A2 and the A3 annular pullies as an indicator of tendon bowstringing by using US. Further US ﬁndings as tenovaginitis and ganglion cysts were recorded. Results: Mean TP distance was 2.6 mm (range 1.5-3.7) at the A2 level and 4.1 mm (range 3.5-5.7) at the A3 level. No increased TP distance under forced ﬂexion was found in 14 patients, 13/14 showed excellent clinical outcome measured by the Buck Gramcko score. Signs of tenovaginitis and ganglion cysts were found in 1/14 patients with good clinical outcome. Increased TP distance (1.5 mm) under forced ﬂexion was seen in 1/15 patient only presenting with a fair clinical result. Conclusion: An increased TP distance postoperatively is found in reconstructed A2/ A3 ﬁnger pullies, but no further increase of TP distance under forced ﬂexion in cases of excellent clinical outcome. US measurement of TP distance under forced ﬂexion allows proving tendon bowstringing, which contributes to fair clinical results.

performed in two healthy controls. Pathologic ﬁndings were compared between 1.5 T and 7T. Results: Presumed bursal and/or peritendineal ﬂuid (found in all runners) could be best appreciated with 7T MEDIC. In one case a cartilage defect, visible at 1.5 T, could be better depicted at 7T (best: CISS). Other ﬁndings better seen at 7T included a presumed ligament rupture, osseous avulsions, and osteophytes. In one subject, however, 1.5 T STIR showed a bone bruise in the anterior compartment of the distal tibia not visible at 7T in any sequence. Conclusion: Pathologies were seen in each of the marathon runners, in contrast to the two controls with normal ankles. A beneﬁt of 7T, as shown in this admittedly small study, might lie in the improved depiction of inﬂammatory pathologies of soft tissue (increased contrast) and cartilage defects (increased spatial resolution). The prospective beneﬁt of 7T for athletes, however, remains to be determined.

Scientific Sessions 10:30 - 12:00

Room C

SS 601a Functional imaging of the abdomen: Are we there yet? Moderators: F. Arredondo; Guatemala City/GT M.A. Müller; Zurich/CH

10:30

Application of whole body diffusion weighted MR imaging for diagnosis and staging of malignant lymphoma S. Li, H.D. Xue, F. Sun, Z.Y. Jin; Beijing/CN ([email protected]) Purpose: To evaluate the clinical impact of whole body diffusion weighted MR imaging (WB-DWI) for diagnosis and staging of malignant lymphoma. Methods and Materials: Thirty-one patients (15 males, 16 females, aged 13-76 years) with suspected lymphadenopathy (18 malignant lymphomas, ﬁve lymph node metastases and eight benign lymphadenopathies) were enrolled. WB-DWI was performed by using short TI inversion recovery diffusion weighted EPI sequence with free breathing and build-in body coil. Axial T2w images scans of the same slice were used as reference. The results of WB-DWI were compared with pathological examination and other imaging modalities. The mean ADC values of different kinds of lymph nodes were compared. Results: WB-DWI was positive in all 18 cases with lymphoma, ﬁve cases with metastatic lymph nodes and four of eight benign lymphadenopathies. The sensitivity, speciﬁcity and accuracy of WB-DWI were 100 (18/18), 30.8 (4/13) and 71.0% (22/31). The mean ADC values (×10-3) of lymphomatous, metastatic and benign lymph nodes were 0.87p0.17, 0.98p0.09 and 1.20p0.10 mm2/s. There was a statistically signiﬁcant difference between benign lymph nodes and other groups (p 0.01). When an ADC value of 1.08×10-3 mm2/s was used as a threshold value for differentiating malignant from benign lymph nodes, the best results were obtained with a sensitivity of 87.8% and speciﬁcity of 91.3%. Sixteen of 18 cases (88.9%) were accurately staged in accordance with clinical staging. Conclusion: WB-DWI is a sensitive, but less speciﬁc method for diagnosis of lymphoma. It is difﬁcult to differentiate lymphomatous from metastatic lymph nodes. But it is a valuable imaging modality for staging of patients with malignant lymphoma.

B-262

10:48

Apparent restricted diffusion in cirrhotic liver likely reﬂects changes in perfusion A. Furlan1, R. Girometti1, K.T. Bae2, C. Zuiani1, M. Bazzocchi1; 1Udine/IT, 2 Pittsburgh, PA/US ([email protected])

Abdominal Viscera (Solid Organs)

B-261

B-263

10:39

Diffusion weighted MR imaging in patients with chronic hepatitis: Grading and response to therapy A.A.A. Abdel Razek, E. Omran, M. Elbendary, K. Zalata, M. Shady; Mansoura/EG ([email protected]) Purpose: To evaluate the role of diffusion weighted MR imaging in patients with chronic hepatitis. Methods and Materials: Fifty ﬁve patients (32M, 23 F aged 18-38 ys; mean 27 ys) with chronic hepatitis underwent diffusion weighted MR imaging of the liver and biopsy. Follow-up after therapy was done for 15 of these patients. Fifteen volunteers underwent the same MR study. Diffusion MR weighted imaging was done on 1.5 Tesla unit (Symphony, Siemens) using a single shot spin echo type of echo planar imaging. The parameters used were TR/TE: 4200/38 msec, B value=0, 250, and 500 sec/mm2. The mean ADC value of the liver was calculated. Percutaneous liver biopsy was obtained in all patients with a mean delay of 15 days. The patients were divided based on Metavir ﬁbrosis score. Statistical analysis was done. Results: The mean ADC value was signiﬁcantly lower in patients with chronic hepatitis compared to volunteers (p 0.03). The mean ADC value decreased signiﬁcantly with increasing ﬁbrosis scoring. There was statistically signiﬁcant difference in the mean ADC value with mild and advanced scores of chronic hepatitis (p 0.04) and in patients before and after treatment (p 0.02). The mean ADC values of the liver were signiﬁcantly correlated with ﬁbrosis score (r=0.77, p 0.01). Selection of 1.11X10-3 mm2/sec for ADC value as a threshold value to predict cirrhosis revealed 87% accuracy, 92% sensitivity and 90% speciﬁcity. Conclusion: Apparent diffusion coefﬁcient value can be used for diagnosis of patients with chronic hepatitis and for assessment of the response to therapy.

Purpose: Diffusion-weighted (DW) MR imaging has been used to characterize tissue composition of the liver parenchyma. Although reductions in apparent diffusion coefﬁcient (ADC) values have been associated with cirrhotic livers, it remains to be determined whether reduced ADC is related to restricted diffusion of water molecules or perfusion changes in cirrhosis. The purpose of this study is to assess the effect of b-values on the association of ADC with hepatic cirrhosis. Methods and Materials: 26 cirrhotic patients (histologically proven liver ﬁbrosis: Ishack 6) and 23 healthy volunteers underwent DW MRI of the liver on a 1.5 T system equipped with a phased array surface coil. Single-shot spin-echo echo-planar DW images were acquired in a single breath hold using b-values of 0, 250, 800 sec/mm2. A pair of ADC values were calculated in each subject: one using images acquired with small (0-250 sec/mm2) b-values, and the other using images acquired with large (250-800 sec/mm2) b-values. Differences in mean ADCs between healthy and cirrhotic livers were tested with Mann-Whitney U test. Results: ADCs computed with low b-values were signiﬁcantly lower in cirrhotic than non-cirrhotic livers (mean 1.79p0.61 versus 2.21p0.61x10-3 mm2/sec, p=0.025). No signiﬁcant difference between cirrhotic and non-cirrhotic livers (mean 0.80p0.36 versus 0.66p0.40x10-3 mm2/sec, p=0.288) was noted, however, with ADCs that were calculated using high b-values. Conclusion: Cirrhotic livers showed lower ADC values than non-cirrhotic livers only when small b-values (0-250 sec/mm2) were used. This suggests that apparent restricted diffusion of cirrhotic livers is strongly inﬂuenced by changes in perfusion.

B-264

10:57

Detection of liver metastases: Diffusion-weighted respiratory triggered EPI versus dynamic gadolinium-enhanced MR imaging M. Bruegel, K. Holzapfel, J. Gaa, K. Woertler, S. Waldt, C. Ganter, E.J. Rummeny; Munich/DE ([email protected]) Purpose: To compare the value of diffusion-weighted respiratory triggered singleshot echo-planar imaging (DW-EPI) and dynamic gadolinium-enhanced MR imaging for detection of hepatic metastases. Methods and Materials: Fourty-four patients with extrahepatic primary malignancies underwent MR imaging at 1.5 T (Avanto/Siemens) including DW-EPI, T2-weighted turbo spin-echo and dynamic gadolinium-enhanced 3D gradient-echo sequences. Images were divided into three sets: The ﬁrst set consisted of DW-EPI images acquired with b-values of 50, 300 and 600 s/mm² and an apparent diffusion coefﬁcient map. The second set included T2-weighted and dynamic gadoliniumenhanced images and the third set comprised all images. Image sets were reviewed independently by three blinded observers who identiﬁed and classiﬁed lesions with a ﬁve-point conﬁdence scale. Results were correlated with surgical/histopathological ﬁndings and follow-up imaging. Accuracy of each technique was measured with free-response ROC analysis. Results: 171 lesions with a mean size of 16.6 mm (range, 3-84 mm) were evaluated (102 metastases, 69 benign lesions). Mean accuracy values for DW-EPI and the gadolinium-enhanced image set were 0.90 and 0.87, respectively (P=0.18). Combined reading of both image sets resulted in a signiﬁcantly higher accuracy value of 0.95 (P 0.05). Mean sensitivities for detection of all metastases and for small metastases ( 10 mm) were 0.91 and 0.85 with DW-EPI, 0.87 and 0.82 with the gadolinium-enhanced image set, and 0.93 and 0.90 with combined reading of both image sets, respectively. Conclusion: DW-EPI was as accurate as dynamic gadolinium-enhanced MR imaging in conjunction with T2-weighted images. Combined reading of both image sets signiﬁcantly improved lesion detection, especially with respect to subcentimeter metastases.

B-265

11:06

Respiratory-triggered echo-planar diffusion-weighted MR imaging for the differentiation between benign and malignant focal liver lesions with a diameter of up to 10 mm K. Holzapfel, M. Eiber, M. Bruegel, E.J. Rummeny, J. Gaa; Munich/DE ([email protected]) Purpose: To evaluate the usefulness of diffusion-weighted MR imaging (DWI) using a respiratory-triggered single-shot echo-planar imaging (SSEPI) sequence in the differentiation between benign and malignant focal liver lesions with a diameter of up to 10 mm.

200

A

B

C

D

E

F

G

Scientific Sessions

B-266

11:15

Diffusion-weighted echo-planar MRI in normal vs pathologic pancreas parenchyma: A promising additional tool for differentiation of pancreatic diseases H. Hoff, K. Ullrich, I. Platzek, R. Grützmann, A. Engel, M. Laniado, C. Stroszczynski; Dresden/DE ([email protected]) Purpose: To investigate if diffusion-weighted echo-planar imaging (DW-EPI) may distinguish between normal pancreas parenchyma and pancreas disease. Methods and Materials: Forty-four patients (carcinoma n=22, symptomatic pancreatitis n=17) were examined before resection. Seventeen patients without any known pancreatic diseases were used as a control group. Diffusion weighted sequence and T1- and T2- weighted sequences for anatomical information were acquired. The axial DW-EPI sequence was performed using a matrix of 144x192, a ﬁeld-of-view of 375x500 (pixel size 2.6x2.6 mm) and a section thickness of 2.6 mm. Apparent diffusion coefﬁcient (ADC) maps were analysed using a dedicated user interface. In the group with pancreas disease the abnormal parenchyma was detected by using the T1- and T2-weighted images. The region of interest (ROI) was transferred exactly to the ADC-map and the coefﬁcients were registered. In the control group the ROI was set to the head of the pancreas followed by a similar registration of the ADCs. For statistical comparison of results a t-test was performed with p=0.05. Results: In the control group the ADC was 0.18x10-3 mm2/sec (standard deviation 0.07). In the group with pancreatic disease the ADC was 0.74x10-3 mm2/sec (standard deviation 0.66). This difference proved to be statistically signiﬁcant (p=0.001). Conclusion: Diffusion-weighted echo-planar imaging clearly differentiated between normal pancreatic tissue and pancreas disease. It may be an interesting additional tool for differentiation of pancreatic diseases as also for the follow-up of chemotherapy.

B-267

11:24

Magnetic resonance elastography: A new quantitative tissue characterization parameter for differentiating benign and malignant primary liver tumors S.K. Venkatesh1, M. Yin2, N. Takahashi2, J.F. Glockner2, J.A. Talwalkar2, P.A. Araoz2, R.L. Ehman2; 1Singapore/SG, 2Rochester, MN/US ([email protected]) Purpose: To evaluate the potential value of magnetic resonance elastography (MRE) for differentiation of benign and malignant primary liver tumors. Methods and Materials: MRE was performed in 39 patients with 45 liver tumors (hepatocellular carcinoma-17, cholangiocarcinoma-10, hemangioma-10, focal nodular hyperplasia-ﬁve and hepatic adenoma-three). The MRE studies were performed on a 1.5 T scanner with a modiﬁed phase-contrast, multi-slice gradient-echo sequence. The tumors were identiﬁed by the conventional MR signal intensities and post gadolinium enhanced sequences and the MRE slice was targeted to the tumor. Stiffness maps were generated by an automated process using local frequency estimation (LFE) algorithm. Mean shear stiffness of the tumor was calculated with a manually speciﬁed region of interest including most of the tumor. The stiffness of the surrounding non-tumor bearing hepatic parenchyma was also calculated. The cut-off value for stiffness of normal liver was 2.93 kiloPascal (kPa).

Results: Malignant liver tumors had signiﬁcantly higher mean shear stiffness (12 kPa) than benign liver tumors (2.8 kPa, p 0.05), ﬁbrotic liver (6 kPa, p 0.05) and normal liver parenchyma (2.3 kPa, p 0.05). Fibrotic liver was signiﬁcantly stiffer than normal liver but the stiffness values overlapped both benign and malignant tumors. There was no signiﬁcant difference in stiffnesses of benign tumors and normal liver. Using a cut-off value of 5 kPa, all malignant tumors could be accurately (100%) differentiated from benign tumors and normal liver parenchyma. Conclusion: MRE is a promising non-invasive technique and provides quantitative information which complements conventional MRI for better differentiation of benign and malignant primary liver tumors.

B-268

11:33

MR elastography of liver ﬁbrosis: Comparison of spin-echo and echoplanar imaging L. Huwart1, N. Salameh1, L. ter Beek2, E. Vicaut3, R. Sinkus3, B. Van Beers1; 1 Brussels/BE, 2Best/NL, 3Paris/FR ([email protected]) Purpose: The purpose of this study was to prospectively compare the performance of MR elastography using echo-planar and spin-echo imaging for staging liver ﬁbrosis. Methods and Materials: Twenty-four patients who had liver biopsy for suspicion of chronic liver disease had MR elastography performed with both spin-echo and echo-planar sequences. At histology, the ﬁbrosis stage was assessed according to the METAVIR scoring system (from F0 to F4). Results: The examination time was about 20 minutes using spin-echo, and only 2 minutes using echo-planar imaging. The hepatic signal-to-noise ratios were similar on the spin-echo and echo-planar images (34.10p8.14 versus 31.85p7.22, p=0.33). The elasticity measurements obtained with the spin-echo and echo-planar sequences correlated substantially (Spearman’s correlation coefﬁcient: r=0.83, p 0.01). The Spearman’s coefﬁcients between the elasticity measurements and the METAVIR ﬁbrosis stages were, respectively: r=0.91 (p 0.01) for spin-echo and r=0.84 (p 0.01) for echo-planar imaging. These correlation coefﬁcients did not differ signiﬁcantly (p=0.17). Bland and Altman analysis showed that the mean difference between spin-echo and echo-planar elasticity was -0.10p0.48 kPa. Conclusion: The results of this preliminary study suggest that echo-planar MR elastography can be performed much faster than spin-echo MR elastography and gives results of similar accuracy for staging liver ﬁbrosis.

B-269

11:42

Enhancement patterns at contrast-enhanced ultrasonography (CEUS) as a prognostic factor for resectable ductal adenocarcinoma of the pancreas M. D’Onofrio, R. Malago, T. Gubello, G.A. Zamboni, N. Faccioli, F. Principe, R. Pozzi Mucelli; Verona/IT ([email protected]) Purpose: To retrospectively detemine whether enhancement pattern of ductal adenocarcinoma of the pancreas at CEUS is related to patients’ prognosis after surgery. Methods and Materials: CEUS of 42 resected ductal adenocarcinomas of the pancreas were retrospectively reviewed. The lesion enhancement at CEUS was scored as: 1=poorly vascularized (presence of avascular areas) or 2=well vascularized (absence of avascular areas). All the lesions underwent pathological examination assesing tumor differentiation as: 1=undifferenziated (poorly differentiated) or 2=differentiated (moderately and well differentiated). The results of CEUS enhancement and pathology were compared by using Spearman’s correlation test. Survival analysis was made by Kaplan-Meier method. Results: The diameters of the lesions ranged between 1.2 and 4.3 cm (mean=2.68 cm; SD=0.66). There were 30 differentiated and 12 undifferentiated adenocarcinomas at pathology. CEUS revealed poor vascularization in nine tumors, while 33 adenocarcinomas were well vascularized. Moderate correlation was found between vascular pattern at CEUS and tumor differentiation (r=0.70; p 0.0001). At CEUS 29/30 (97%) differentiated adenocarcinomas were well vascularized while 8/12 (67%) undifferentiated adenocarcinomas were poorly vascularized. Survival rate in group 1 of CEUS was signiﬁcantly lower than that of group 2 (p=0.0011). Survival rate in group 1 of pathology was not signiﬁcantly lower than that of group 2 (p=0.2892). Conclusion: Correlation between vascular pattern of ductal adenocarcinoma at CEUS and tumor differentiation is not always present. The vascular pattern of enhancent of ductal adenocarcinoma seems to be a prognostic indicator.

A

B

C

D

E

F

G

201

Saturday

Methods and Materials: 75 patients with focal liver lesions (36 cysts, 29 metastases, 18 hemangiomas and four HCCs) underwent MR imaging at 1.5 T (Avanto, Siemens) using a body-phased array coil. DWI was performed by a non breathhold SSEPI sequence (TR=d, TE=69 ms, matrix 144x192, slice thickness 5 mm, GRAPPA 2, b-values 50, 300 and 600 sec/mm2) combined with navigator echo technique (PACE). Lesion-to-liver contrast-to-noise ratio (CNR) was determined at b=50 and 600 sec/mm2 and the ratio of these CNRs (CNR50/600) was calculated for each lesion. Additionally, ADC values were obtained. Imaging results were correlated with histopathologic ﬁndings and imaging follow-up. Mann-Whitney test was used for statistical analysis and a receiver operating characteristic (ROC) curve analysis was performed. Results: The mean ADC values (x10-3 mm2/sec) were 1.01p0.14 for HCCs, 1.10p0.21 for metastases, 1.80p0.20 for hemangiomas and 2.51p0.43 for cysts. CNR50/600-ratios were 0.97p0.08 for metastases, 1.02p0.22 for HCCs, 1.43p0.18 for hemangiomas and 2.22p0.53 for cysts. CNR50/600-ratios and ADC values of hemangiomas and cysts were signiﬁcantly higher than those of HCCs and metastases. 97.2% of lesions were correctly classiﬁed as benign or malignant using a threshold ADC value of 1.53. Conclusion: DWI using the respiratory-triggered SSEPI sequence allows reliable differentiation between benign (cysts and hemangiomas) and malignant focal liver lesions, even when the diameter of lesions is 10 mm or less.

Scientific Sessions B-270

B-272

11:51

Whole-body FDG-PET/CT with integrated CT-perfusion for liver metastases evaluation. Initial results on technical feasibility P. Veit-Haibach, V. Treyer, J. Soyka, K. Strobel, M. Pérez Lago, K. Mende, T.F. Hany; Zurich/CH ([email protected]) Purpose: To assess the technical feasibility of the integration of a CT-perfusion protocol into a full diagnostic clinical whole-body PET/CT for staging purposes without additional use of contrast media. Methods and Materials: 16 patients (eight male, eight female, mean age: 61 years) with different liver metastases underwent the integrated PET/CT-perfusion protocol. An attenuation-CT was acquired ﬁrst; the PET were acquired consecutively. The position of the liver metastases was deﬁned on the non-cePET/CT. For CT-perfusion, 40 ml of contrast media were applied (4 ml/s, 10 sec duration). The perfusion parameters were acquired consecutively for 35 sec (1 sec rotation time). Afterwards, another bolus of contrast media was applied (50 ml, 3 ml/s) to assure full diagnostic ceCT-data for clinical staging. A routine abdominal ceCT was performed after another 70 sec. Blood volume, mean transit time, surface permeability product and blood ﬂow were calculated with a commercially available CT-perfusion software. Results: Acquisition of CT-perfusion within the diagnostic PET/CT protocol was feasible in all but one patient (perfusion parameters could not be calculated due to breathing artefacts and lesion size). The integrated PET/CT-perfusion protocol provided full diagnostic images for clinical whole-body staging in all patients. Conclusion: PET/CT with integrated CT-perfusion for liver metastases is technically feasible in cases with adequate lesion size and provides perfusion parameters and a clinically full diagnostic PET/CT-image without additional contrast media. The integrated PET/CT-perfusion protocol might serve as a valuable tool to assess tumor perfusion and the corresponding viability for therapy response studies.

10:30 - 12:00

Room F1

Interventional Radiology

SS 609a Spine Moderators: D. Bilecen; Basle/CH A. Dimopoulou; Uppsala/SE

B-271

10:30

Technical results, extraosseous cement leakages, and complications in 500 vertebroplasty procedures M.B. Pitton, S.C.A. Herber, C.M. Ahlers, B. Röhrl, P. Drees, B. Böhm, K. Oberholzer, C. Düber; Mainz/DE ([email protected]) Purpose: To analyze technical results and complications in 500 procedures. Methods and Materials: Technical results after vertebroplasty were documented with CT and complications and secondary fractures were analyzed. Results: 500 vertebroplasty procedures have been performed in 251 patients (osteoporosis, n=217; tumour, n=34). 1.96p1.29 (range 1-10) vertebrae were treated per patient (1.33p0.75 interventions per patient, 1.01p0.10 interventions per vertebra). Amount of cement was 4.5p1.9 ml. 30-day mortality was 0.4% (1 of 251 patients), caused by a pulmonary cement embolism. The procedure-related morbidity was 2.8% (7/251): one acute coronary syndrome 12 h after procedure, one missing patellar reﬂex in a patient with a cement leak near the neuroforamen because of osteolytic destruction of the respective pedicle, one medullary conus syndrome after a fall of the patient during night after vertebroplasty, two patients with an inadequate depth of conscious sedation, and two cases with additional fractures (one pedicle fracture, one rib fracture). The overall CT-based cement leak rate was 55.4% and included leakages predominantly into intervertebral disc spaces (25.2%), epidural vein plexus (16.0%), through the posterior wall (2.6%), into the neuroforamen (1.6%), into paravertebral vessels (7.2%), and combinations of these and others. During follow-up (15.2p13.4 months) the secondary fracture rate was 17.1%, including comparable numbers for vertebrae at adjacent and distant levels. The presence of intradiscal cement leaks was not associated with increased adjacent fracture rates. Conclusion: CT-guided vertebroplasty is safe and effective for treatment of vertebral compression fractures. However, even small volumes of pulmonary PMMA embolism might be associated with fatal outcome.

202

A

B

C

D

E

F

G

10:39

Symptomatic vertebra plana: A primary indication to vertebroplasty A. Pedicelli, M. Rollo, R. Polito, M. Piano, C. Colosimo, L. Bonomo; Rome/IT ([email protected]) Purpose: To evaluate the efﬁcacy of vertebroplasty in treating symptomatic vertebra plana, considered a relative contraindication to the procedure and absolute contraindication for kyphoplasty. Methods and Materials: Under ﬂuoroscopy guidance we treated 20 cases of vertebra plana (18 osteoporotic compression fractures and 2 tumor fractures, between Th6 and L2; mean age: 72 ys) out of 150 levels treated in 80 patients. The procedure was always performed by mono trans-pedicular access followed by PMMA injection (range volume of cement 1-4 mL). All patients underwent preliminar MRI a/o CT examination, immediate control after procedure and follow-up (mean duration, 6 months; range, 1-18 months). Results: Immediately and at follow-up, signiﬁcant reduction of mean VAS score for pain was observed in all cases compared to VAS score before treatment (pb0.0002). No major complications during needle insertion or cement injection occurred. In 12/20 cases a different rate of asymptomatic intra-discal cement leakage occurred. Posterior or peri-vertebral leakage never occurred. In presence of an intravertebral cleft this was ﬁlled by PMMA and partial vertebral height restoration (average 2.5 mm) was achieved in ﬁve cases and resulted stable in 3/5 cases at follow-up. In three cases a new fracture adjacent to the treated level occurred between 1 and 3 months. Conclusion: In agreement with other authors, our preliminary results show that vertebroplasty is safe and effective as a treatment for symptomatic vertebra plana and should not be refused in this group of patients. Clinical evaluation and preoperative MRI are basic for the correct indication to the procedure.

B-273

10:48

Two years follow-up after percutaneous vertebroplasty of osteoporotic vertebral fractures: Height gain of vertebral bodies and stabilization of vertebral geometry U. Koch, N. Morgen, S.C.A. Herber, P. Drees, J. Heine, C. Düber, M.B. Pitton; Mainz/DE Purpose: To investigate geometrical stability and preservation of height gain of vertebral bodies after percutaneous vertebroplasty during two years follow-up. Methods and Materials: Patients with osteoporotic vertebral compression fractures resistant to analgetic drugs were treated with polymethylmethacrylate (PMMA) vertebroplasty. Mean cement volume was 5.1p2.0 ml. Vertebral geometry was documented by sagittal and coronal reformations from MSCT: anterior and posterior vertebral heights, endplate angles, and compression index (CI=anterior/posterior height). Patients were assigned to two groups: severe compression with CI 0.75 (group I) and moderate compression with CI 0.75 (II). Results: 83 vertebral bodies in 30 patients (7 male, 23 female, age 70.7p9.7 years, range 40-82) were treated with vertebroplasty and followed for 24 months. Group.i. showed a greater beneﬁt compared to group II with respect to peri-interventional anterior height gain (2.84p2.15 vs 0.73p1.88 mm, p 0.001), reduction of endplate angle (-4.89p4.53 vs -0.98p2.48°, p 0.001), and compression index (0.13p0.12 vs 0.02p0.06, p 0.001). Compared to Group II, group.i. demonstrated some preserved anterior height gain at two years (1.23p2.98 mm, p 0.005), and improved endplate angles (-4.55p5.41°, p 0.001) and compression indices (0.11p0.15, p 0.01). In group II, the vertebral heights returned to and were ﬁxed at the pre-interventional levels. In both groups, measurements did not change signiﬁcantly between 12 and 24 month follow-ups. Conclusion: Vertebroplasty preserved vertebral height and geometry during midterm follow-up. In cases with severe vertebral compression, signiﬁcant height gain and improvement of endplate angles were achieved peri-interventionally and were preserved to some extent during mid-term follow-up.

B-274

10:57

Incidence of symptomatic vertebral fractures in patients after percutaneous vertebroplasty J. Hierholzer, K. Westphalen, H. Fuchs, C. Baumann; Potsdam/DE ([email protected]) Purpose: To evaluate the incidence of vertebral compression fractures (VCF) in patients previously treated by percutaneous vertebroplasty (VTP). Methods and Materials: 316 patients with 486 treated VCF were included in this study according to the inclusion criteria. Patients were kept in regular follow-up using standardized questionaires before, one day, seven days, six months, one year and further on, on a yearly basis after VTP. The incidence of secondary symptomatic VCF was calculated and anatomical distribution with respect to previous fractures characterized.

Scientific Sessions Results: Mean follow-up was eight months (6-56 months) after VTP. 52/316 (16.4%) patients (45 female, seven male) returned for treatment of 69 secondary VCF adjacent to (35/69, 51%) or distant from (34/69, 49%) previously treated levels. Adjacent secondary VCF occurred signiﬁcantly more often as compared to distant secondary VCF. Of the total 69 secondary VCF, 35/69 occurred below and 27/69 above pretreated VCF. Of the 65 sandwich-levels generated, in 7/65 (11%) secondary VCF were observed. Secondary VCF below pre-treated VCF occured signiﬁcantly earlier in time as compared to VCF above and as compared to sandwich-body-fractures. No major complication occurred during initial or follow-up intervention. Conclusion: VTP is an effective treatment for pain associated with VCF. Complications are rare and risk of secondary fractures after VTP is not convincing in the light of the natural course of the underlying disease. Secondary VCF do occur in individuals after VTP but the rate found in our study remains below the level expected from epidemiologic studies.

B-275

11:06

Purpose: To prospectively follow-up geometrical changes of vertebral bodies following vertebroplasty and kyphoplasty. Methods and Materials: Seventy-ﬁve patients (23 male, 52 female, age 70.1p9.2 years, range 27-92) with 153 osteoporotic fractures were treated with vertebroplasty (n=100) or kyphoplasty (n=53) and prospectively followed-up for 12 months. Vertebral geometry was analyzed using reformatted CT images: anterior, posterior and lateral heights as also endplate angles. Vertebrae of each treatment group were assigned to group.i. with severe compression (compression index (CI) 0.75). Results: Kyphoplasty demonstrated better post-interventional anterior height gain (1.7p3.3 vs 1.1p1.8 mm) and respective improvements of the endplate angles (-1.37p6.02° vs -1.54p3.13°). At 12 months, anterior heights decreased to the pre-interventional levels. In group I, both techniques showed an improved post-interventional anterior height gain compared to group II (kyphoplasty: height gain 3.9p2.9 (I) vs 1.1p3.1 mm (II), endplate angle -0.6p7.0°(I) vs -1.6p5.7°(II); vertebroplasty: 2.1p1.9 (I) vs 0.7p1.6 mm (II), -3.7p4.2°(I) vs -0.9p2.3°(II)). At 12 months, anterior height gain was preserved to some extent with both techniques in group I, whereas group II returned to baseline height levels, (kyphoplasty height gain 3.60p2.4 (I) vs -0.6p3.6 mm (II), endplate angle -3.1p7.2°(I) vs -0.7p4.5°(II); vertebroplasty: 1.5p2.8 (I) vs -0.5p2.5 mm (II), -3.6p4.9°(I) vs 0.0p3.9°(II)). Conclusion: Kyphoplasty provided a greater initial height gain than vertebroplasty. Both methods demonstrate a comparable stable vertebral height at 1-year follow-up. Irrespective of the treatment modality, vertebrae with severe compression fractures showed a greater anterior height gain and a greater beneﬁt of endplate angles.

B-276

11:15

Multiple level vertebroplasty: Feasibility, safety and efﬁcacy G.C. Anselmetti1, P. Carpeggiani2, A. Manca1, G. Chiara1, D. Regge1; 1 Candiolo/IT, 2Pisa/IT ([email protected]) Purpose: To evaluate feasibility, safety and efﬁcacy of multiple levels vertebroplasty. Methods and Materials: Among 1838 patients treated with 6168 vertebroplasties, we retrospectively reviewed 358 patients treated at ﬁve or more levels (up to 12) in one session (total of 2234 vertebrae). Indication for multiple levels treatment was the presence of multiple vertebral fractures associated with severe, invalidating and diffused poorly localizable back-pain. Some prophylactic vertebroplasty such as sandwich vertebra or adjacent vertebrae considered at risk of subsequent fractures were also included in the treatment planning. Patients were fully monitored during the treatment; all procedures were performed in local anesthesia, under ﬂuoroscopy control, using thin needles (13-15 Gauge) and monolateral approach (transpedicular in lumbar vertebrae, parapedicular in thoracic vertebrae) with placement of the needle in the anterior midline. Alteration in oxygen saturation and other vital parameters during the procedure were observed. Average time for the procedures was between 45 and 60 minutes. The average amount of cement injected per level was 1.5-2 cc for thoracic vertebra and 2-3 cc for lumbar vertebra, achieving a full ﬁlling of the central portion of the vertebral body. Results: All treatments were well tolerated with no clinically relevant complications. No relevant bone cement leakages were detected by post-procedural CT. All patients reported back-pain reduction (VAS score dropped from 8.2p1.4 to 1.5p1.8, p 0.0001 Wilcoxon rank test). Subsequent fractures were observed in 37 cases (10.3%) and were treated with vertebroplasty. Conclusion: Multiple levels vertebroplasty performed by an experienced physician with thin needles is feasible, safe and effective.

11:24

Percutaneous vertebroplasty (PVP) for osteolytic breast cancer metastases: Cement leaks in 59 procedures C.G. Trumm, T.F. Jakobs, C. Glaser, M.F. Reiser, R.-T. Hoffmann; Munich/DE ([email protected]) Purpose: This retrospective study aimed to assess the results of CT-ﬂuoroscopy (CTF) guided PVP of painful osteolytic spinal metastases from breast cancer (BC), focussing on the frequency and clinical impact of polymethylmetacrylate (PMMA) leaks. Methods and Materials: In 48 months, 147 patients with multiple myeloma or metastases of the spine underwent PVP. 53/147 patients (52 female, one male; 63p12 years) suffered from painful osteolytic BC metastases. 86 vertebrae were treated in 59 sessions under CTF (single slice, 4- or 16-row MDCT). In the preprocedure CT (3 mm slices, cor./sag. MPR) degree of osteolytic destruction (OD) was differentiated regarding (I) the vertebral cross-sectional area (CSA; b25/b50/b75/b100% involvement), the cortical border (0/b25/b50/b75/b100% involvement) of (II) the spinal canal (CB-SC) and of (III) the outer vertebral circumference (CB-OC). The postprocedure CT was used to detect local PMMA leaks differentiating intradiscal (ID), intraspinal (IS), paravertebral (PV), intercostovertebral/posterolateral (ICV-PL), and vascular leaks (basivertebral vein [BV]; segmental vein [SV]). Results: Ratio of OD was 11.6/51.2/29.1/8.1% regarding the CSA, 52.3/23.3/11.6/9.3/3.5% regarding the CB-SC, and 52.3/26.7/19.8/1.2/0% regarding the CB-OC. No major complications occurred. ID/IS/PV/ICV-PL leaks were seen in 24.4/20.9/20.9/11.6% of 86 vertebrae. Ratio of vascular leaks was 17.4 (BV) and 18.6% (SV). Pain level (VAS) decreased from 6.4p1.1 (24 h before) to 3.4p1.6 (mean follow-up: 9.2 months after PVP). Conclusion: PVP of osteolytic BC metastases can be performed safely under CTF even with posterior wall involvement. In our series, it showed a high success rate regarding pain relief while cortical and vascular PMMA leaks had no clinical impact.

B-278

11:33

Percutaneous management of aggressive vertebral haemangiomas: A series of 32 cases A. Gangi, X. Buy, T. Moser, J.-L. Dietemann; Strasbourg/FR ([email protected]) Purpose: To evaluate the results of percutaneous management of aggressive vertebral haemangiomas (AVH) with vertebroplasty, sclerotherapy and surgery. To establish the role of each modality and combination therapy. Methods and Materials: From 1991 to 2007, 33 AVH in 31 patients were treated for painful AVH with characteristic imaging features on contrast enhanced MR. Neurological symptoms were associated in nine patients (radicular pain 4, motor deﬁcit 5). Vertebroplasty alone was performed in 22 levels in which the AVH was limited to the vertebral body or had minimal paravertebral extension. Sclerotherapy utilizing Ethibloc® followed by vertebroplasty was performed in eight levels in which the AVH extended into the epidural space. Vertebroplasty was combined with scheduled laminectomy for three levels when motor deﬁcit required spinal cord decompression. Pain, neurological symptoms, mass effect on MRI and consolidation were evaluated on follow-up. Results: Technical success was 100%. Pain resolution was noted in 28 patients. Neurological recovery was noted in all eight patients with neurological symptoms. No secondary vertebral collapse occurred. Symptomatic recurrence of the AVH occurred in two patients, after 10 and two years, requiring additional sclerotherapy. One cauda equina syndrome following sclerotherapy was successfully treated by urgent laminectomy. Five minor cement leaks remained asymptomatic. Conclusion: Vertebroplasty is the mainstay of managing AVH, allowing for direct embolization and vertebral consolidation. Sclerotherapy is combined with vertebroplasty for occlusion of the epidural and paravertebral components of the haemangioma. The role of surgery is limited to laminectomy in cases of evolving neurological deﬁcit post percutaneous treatment.

A

B

C

D

E

F

G

203

Saturday

Vertebroplasty and kyphoplasty of vertebral compression fractures: Height gain of vertebral bodies during one year following PMMA cement augmentation B. Röhrl, N. Morgen, S.C.A. Herber, M. Sadick, C. Düber, M.B. Pitton; Mainz/DE ([email protected])

B-277

Scientific Sessions B-279

11:42

Percutaneous nucleoplasty: Our experience in 215 patients S. Marcia, S. Marini, U. Codella, G. Mallarini; Cagliari/IT ([email protected]) Purpose: Nucleoplasty is a minimal invasive percutaneous procedure that uses radiofrequency energy to ablate nuclear material and creates small channels within the disc. The purpose of our study paper was to describe our experience in 215 patients. Methods and Materials: We studied 215 consecutive patients that underwent to percutaneous nucleoplasty procedure (131 males and 84 females). We performed on all patients with local anaesthesia; in the majority of cases we used ﬂuoroscopic guide and in a few cases CT guide. In order to evaluate annulus integrity we always used discography before performing nucleoplasty. A Visual Analog Pain Scale (VAS) was administered during each exam. We considered a reduction of at least two points on the VAS scale as a positive result. Results: Overall success rate in pain reduction was 78, 81, 84 and 87%, respectively, at 1 month, 3 months, 6 months and 1 year. We treated L2L3 in 2% of patients, L3L4 in 14%, L4L5 in 46% and L5S1 in 38%. Conclusion: Our data indicate that nucleoplasty is an optimal therapy for symptomatic patients with contained herniated disk. To avoid failure of outcome a correct patient selection is extremely important.

Methods and Materials: Preoperative and operative histories of all 120,550 women diagnosed with breast cancer in the UK NHSBSP between April 1994 and March 2006 were derived from different NHSBSP databases. Results: Non-operative diagnosis rate (for all cancers) has improved from 48.8% in 1994/5 (when only nine units achieved the then minimum standard of 70%) to 94% in 2005/6 (when only seven units failed to achieve the national target of 90%). In 2005/6, 2790 (17.8%) of the 15,688 women having surgery needed two or more operations. In 2001/2 when the non-operative diagnosis rate was 87%, 2377 (23.8%) of the 9969 women having surgery needed two or more operations. Extrapolation of this trend backwards to 1994/5, when the non-operative diagnosis rate was only 48.8%, suggests that 62% of women required two or more operations. Conclusion: Non-operative diagnosis rates in the NHSBSP have improved dramatically across the whole of the UK since the mid-1990s. Using more recent data on operation rates from the NHSBSP and Association of Breast Surgeons at British Association of Surgical Oncology national audit of the surgical management of screen-detected breast cancer, we have demonstrated the impact that radiologically driven non-operative diagnosis has had on the women attending for screening. This justiﬁes the importance of striving to achieve a non-operative diagnosis.

B-282

10:39

Quantitative approach for diagnosis of cystic lesions of pancreas: Diffusion-weighted MRI and 1H MRS study Z.Z. Rozhkova, J.Z. Lifshits, E.O. Ipatova, V.A. Sychov; Kiev/UA

Breast cancer screening in The Netherlands: Trends in utilization and cost of diagnostic imaging and biopsies following positive screening mammography L.E.M. Duijm1, J.H. Groenewoud2, J. Fracheboud2, M.L. Plaisier3, R.M.H. Roumen3, M. van Beek1, T.F. Le Maire1, H.J. de Koning2; 1Eindhoven/NL, 2 Rotterdam/NL, 3Veldhoven/NL ([email protected])

Purpose: On the basis of the analysis of DW MR-images and 1H MRS data of the main metabolites in cystic pancreatic tissues we propose quantitative approach for differentiating intraductal mucinous tumors (IMT) and mucinous cystoadenomas/ carcinomas from other lesions of pancreas. Methods and Materials: Twenty-ﬁve patients with mucin-producing tumors, pseudocysts, cystoadenomas, polycystic disease, and pancreatic abscesses are studied by MRI and 1H MRS with 1.5 T SIGNA EXCITE (GE). DW images (EPI, b=700 and 1000 s/mm2) were obtained, and ADC coefﬁcients calculated. 1H MR spectra are recorded with 2DCSI: TR/TE=1500/50, 70, 135 ms. In the spectra the signals from water, lipids, choline-containing compounds (Cho) and betaine were obtained. Results: All cystic lesions have high signal intensity on T2wIs. IMTs and mucinous cystoadenomas/carcinomas have high signal intensity at 500, 700, and 1000 s/mm2, and have decreased ADC values, which potentially represents diffusion restriction, indicative of the presence of mucinous products. Simple cysts, complex separate cysts, pseudocysts, and abscesses did not show signs of restriction for all values of b. From the analysis of spectra T1 and T2 values for the main metabolites and for water the following data are calculated: in the normal pancreatic tissue the mean values of T1 (n)=652 ms, T2 (n)=58 ms, in the abscesses T1 (abs)=907 ms, T2=83 ms, and in the IMT the mean value of T1 (tum)=763 ms, T2=69 ms. In IMTs and mucinous cystoadenomas/carcinomas the concentration of Cho-and betaine-compounds increases, and water content decreases. Conclusion: The relation between the mucinous content of pancreatic cystic lesions and diffusion restriction appraised on DW images was found.

Purpose: To assess the utilization and costs of diagnostic follow-up services of screen-positive women in a biennial breast cancer screening program for women aged 50-75 years. Methods and Materials: We included all 2,062 recalled women with suspicious screens between 1 January 2000 and 1 July 2005 (158,997 screens). Data were collected on any diagnostic examinations, interventional procedures and surgical consultations with two-year follow-up. We used 2006 national reimbursement rates for costs related to diagnostic imaging and pathology procedures, and cost prices for outpatient surgical consultation and diagnostic lumpectomy. Results: Per 100 referrals, we observed the largest increase in utilization of diagnostic procedures for axillary ultrasound (from 3.9 (2000) to 33.5 (2005)) and for stereotactic core biopsy (from 2.1 (2000) to 26.8 (2005)). Use of surgical biopsies decreased from 34.7 (2000) to 4.6 (2005); surgical consultation numbers ﬂuctuated between 269.8 (2000) and 309.7 (2004). Surgical biopsies accounted for 54.1% (_121,648) of total diagnostic costs in 2000 (_224,752), but decreased to 9.9% (_13,296) of total diagnostic costs in the ﬁrst half year of 2005 (_134,774). Meanwhile, costs of radiologic imaging increased from 23.7% (_53,157) to 43.8% (_59,023), of percutaneous biopsies from 9.9% (_22,263) to 27.2% (_36,680), and of consultations from 12.3% (_27,684) to 19.1% (_25,776). Workup costs per screen-detected breast cancer varied between _1,223 (2002) and _1,647 (2003), with a mean of _1,501. Conclusion: Increased costs for imaging procedures, percutaneous biopsies and consultations were compensated by decreased costs for surgical biopsies; expenses per screen-detected breast cancer remained fairly stable over the years.

10:30 - 12:00

B-283

B-280

11:51

Room F2

Breast

SS 602 Intervention Moderators: P. Petrow; Paris/FR M. Sentís; Sabadell/ES

B-281

10:30

Non-operative diagnosis in the UK national health service breast screening programme (NHSBSP) and impact on number of operations: An historical perspective M.G. Wallis1, O. Kearins2, S. Cheng2, G. Lawrence2; 1Cambridge/UK, 2 Birmingham/UK ([email protected]) Purpose: Achieving an accurate non-operative diagnosis is a cornerstone of best breast practice as it allows the appropriate planning of optimal management prior to any surgical intervention and with the full co-operation/collaboration of the patient.

204

A

B

10:48

Preoperative axillary ultrasound (US) and US-guided needle localization in the patients with breast cancer undergoing sentinel lymph node biopsy (SLNB) W.K. Moon, N. Cho; Seoul/KR ([email protected])

C

D

E

F

G

Purpose: To evaluate the role of preoperative axillary US and US-guided intervention in patients with breast cancers who can be candidates for SLNB. Methods and Materials: During the last six months, 191 patients (aged 24-77) underwent preoperative US, US-guided needle localization of the LN with the thickest cortex, and a subsequent SLNB including an injection and identiﬁcation of a radiotracer intra-operatively. The correspondence of the localized LN and SLN identiﬁed by radiotracer was recorded. They were separately submitted to and reported on by pathologists. The localized LN was classiﬁed according to its cortical thickness (grades 1-5). Results: No radiotracer uptake was found in 4% (7/191). Of these technical failure cases, all localized nodes, classiﬁed as grade 3 (2.5-3.5 mm), 4 ( 3.5 mm with intact fatty hilum), 5 ( 3.5 mm with a loss of fatty hilum) at US, proved to be positive at frozen biopsy. Of the localized nodes, 91% (168/184) corresponded to the SLN. Overall accuracy of the SLN status was 96%, sensitivity 85%, speciﬁcity 100%, and false negative (FN) rate 5% (7/144) for 54 patients with axillary LN

Scientific Sessions metastases. Of these FN cases, localized nodes were positive metastasis in two cases. The PPV of the grades were 7% (3/43) in grade 1 ( 1.5 mm), 14% (1/80) in grade 2 (1.5-2.5 mm), 48% (23/48) in grade 3, 70% (7/10) in grade 4 and 100% (10/10) in grade 5. Conclusion: Preoperative axillary US and US-guided intervention can detect the LNs which cause technical failure (7/191) and FN (2/191) results at SLNB.

B-284

10:57

Prospective randomized comparison of intratumoral versus peritumoral injection of radiocolloid for sentinel node identiﬁcation and biopsy in nonpalpable breast cancer A. Pérez de Tudela, A. Mariscal, S. Vizcaya, J. Julián, M. Solà, L. Catro; Badalona/ES ([email protected])

B-285

11:06

Ultrasound (US)-guided vacuum-assisted removal of intraductal single mass in patient with pathologic nipple discharge: Is it an alternative to open surgery? J. Chang, W.K. Moon, N. Cho; Seoul/KR ([email protected]) Purpose: To evaluate the utilization of US-guided vacuum-assisted device for complete removal of intraductal single mass in patient with pathologic nipple discharge for a therapeutic procedure. Methods and Materials: Between 2004 and 2006, 46 patients (median age 43, range 23-80 years) with pathologic nipple discharge underwent US-guided, 11-gauge, vacuum-assisted complete removal of intraductal single mass followed by at least six months with US. During the US examination, an intraductal single mass triggering nipple discharge was identiﬁed with compression by transducer. The median lesion size was 10 mm (range 4-35 mm). Histologic results, medical records, and US ﬁndings were reviewed to track the clinical outcome. Results: After vacuum-assisted removal, histologic ﬁndings revealed intraductal papilloma in 26 (57%), ﬁbrocystic change in 10 (22%), ﬁbroadenoma in two (4%), atypical papilloma in six (13%), and carcinomas in two (4%). Of the 38 patients with benign histology, the discharge disappeared in 26 (68%) and persisted (n=7) or recurred (n=5) in 12 (32%) at a mean follow-up of 22 months (range 7-39 months). In seven (58%) of the 12 with persistent or recurrent nipple discharge, subsequent surgical excision was required at a median period of nine months (range 1-21). There were no signiﬁcant differences in lesion size, location and histology between asymptomatic and symptomatic groups. Conclusion: US-guided vacuum-assisted complete removal of intraductal single mass triggering pathologic nipple discharge eliminated the sign in 68% (26/38) of the patients at a mean follow-up of 22 months. The discharge recurred 6-19 months after the procedure in 13% (5/38) of the patients.

B-286

11:15

Vacuum-assisted breast biopsy of non-mass calciﬁcations: How many calciﬁcations are necessary for diagnosis? S.-H. Mun, G.-W. Park, E.-Y. Ko, J.-H. Shin, S.-J. Kim, B.-K. Han; Seoul/KR Purpose: To analyze the number of samples and specimen characteristics contributing to diagnostic accuracy in performing 11-gauge vacuum-assisted breast biopsy of non-mass calciﬁcations.

B-287

11:24

Lobular neoplasia: Malignancy underestimation rate of core-needle breast biopsy in relation to radiologic and pathologic features V. Londero, A. Linda, C. Zuiani, E. Vianello, M. Bazzocchi; Udine/IT ([email protected]) Purpose: To estimate malignancy underestimation rates (MUR) of core-needle biopsy (CNB) demonstrating lobular neoplasia (LN), as a function of lesion radiologic and pathologic features. Methods and Materials: From a series of 3920 consecutive CNB, 35 LN (14 atypical lobular hyperplasia-ALH- and 21 lobular carcinoma in situ-LCIS) were identiﬁed. Twenty-two biopsies were performed under stereotactic-guidance with a vacuumassisted device (VAB-11G) and 13 under US-guidance with an automated-gun (14G). Twenty-eight patients underwent surgical excision (SE) and seven imaging follow-up ( 24 months). We calculated MUR of CNB as a function of (1) histologic type, (2) biopsy probe and guidance, (3) lesion diameter and (4) level of radiologic suspicion. Chi-square and Fisher exact tests were used for statistical evaluation (signiﬁcant: p 0.05). Results: 13/35 lesions proved to be malignant (seven invasive lobular carcinomas and six DCIS), with an overall MUR of 37%. (1) MUR in case of diagnosis of ALH and LCIS were 7.1 (1/14) and 57.1% (12/21), respectively (p=0.003). (2) MUR of stereotactic-guided VAB and US-guided CNB were 22.7 (5/22) and 61.5% (8/13), respectively (p=0.053). In addition, 75% of carcinomas underestimated by US-guided CNB were invasive compared to 20% of those underestimated by stereotactic VAB. (3) MUR was 31.2% (10/32) for lesions 20 mm and 100% (3/3) for lesions 20 mm (p=0.043). (4) MUR was 20% (2/10) for BI-RADS3 lesions and 44% (11/25) for BI-RADS4 lesions (p=0.259). Conclusion: MUR is higher in case of LCIS, US-guided CNB, large and suspicious lesions. Nevertheless, absence of the above-mentioned features do not spare the need of SE in case of percutaneous diagnosis of NL.

B-288

11:33

MR-guided stereotactic breast biopsy using a low-cost mixed ferromagnetic-nonmagnetic coaxial system with small caliber needles (16-18 gauge): Feasibility, efﬁcacy and safety E. Belloni, P. Panizza, F. De Cobelli, S. Gusmini, I. Fedele, I. Sassi, C. Losio, N. Verardi, A. Del Maschio; Milan/IT ([email protected]) Purpose: To assess feasibility, efﬁcacy and safety of MR-guided stereotactic breast biopsy (MR-SBB) using a low-cost mixed ferromagnetic-nonmagnetic coaxial system with small caliber needles for suspicious lesions detected only by breast magnetic resonance (BMR). Methods and Materials: MR-SBB was performed on 40 lesions (mean diameter 10 mm, range 3-30 mm) in 39 female patients. To localize the lesion, a coronal T1W-TSE sequence was performed after 1 mmol/kg gadolinium iv injection. The biopsy device was composed by a nonmagnetic 14-16 gauge coaxial needle and a ferromagnetic 16-18 gauge biopsy needle, whose costs are lower than fully MR-compatible and/or vacuum-assisted devices. 4-10 specimens were obtained from each lesion. Results: The procedure was well tolerated in all cases. Minor bleeding was documented in three biopsies out of 40 (7.5%). All the specimens were considered sufﬁcient by the pathologist and yelded diagnosis of benignity in 25 and malignancy

A

B

C

D

E

F

G

205

Saturday

Purpose: The study investigates the differences between two approaches for radiolabeled tracer administration by the deep route (intratumoral vs peritumoral) in sentinel lymph node detection and biopsy in patients with non-palpable breast cancer. Methods and Materials: This prospective, randomized study included patients with non-palpable breast cancer and no axillary lymphadenopathy undergoing primary surgical treatment. FANB node-positive patients and those with locally advanced carcinomas, multicentric carcinomas, or prior axillary surgery, radiotherapy, or primary chemotherapy were excluded. The patients were divided into two groups: group A (intratumoral route, n= 35) and group B (peritumoral route, n= 45). Results: Intratumoral injection yielded a sentinel node detection rate of 97% as compared to 84.4% for peritumoral injection. Tracer migration to the sentinel node failed in eight cases (seven for peritumoral injection and one for intratumoral). In 22.7% of cases, sentinel nodes were found in an extra-axillary location (11 by intratumoral and 17 by peritumoral detection). Conclusion: Axillary and extra-axillary sentinel node detection were not statistically different between intra- or peritumoral injection approaches. However, the identiﬁcation rate for the intratumoral route was higher and the difference approached statistical signiﬁcance. Nevertheless, intratumoral tracer injection can be used simultaneously for sentinel node detection and radioguided occult lesion localization of the primary lesion, a fact that confers a deﬁnitive advantage to the use of this approach.

Methods and Materials: From January 2004 to December 2006, 320 patients underwent vacuum-assisted breast biopsy for non-mass calciﬁcations. We divided the patients into two groups according to the biopsy results and ﬁnal diagnoses: accurate diagnosis group and inaccurate diagnosis group. We compared total number of retrieved calciﬁcations, number of cores containing calciﬁcations, and number of total sample cores on the specimen mammogram between two groups. We evaluated the most important specimen characteristics to get accurate diagnostic result at vacuum-assisted breast biopsy using logistic regression analysis, and accessed cut-off point of specimen number using ROC analysis. Results: A total of 195 patients had surgery after biopsy (n=99) or followed up benign biopsy results with mammography for more than two years (n=96). As compared with the ﬁnal diagnosis, 170 cases were diagnosed accurately at vacuum-assisted breast biopsy and 25 cases were inaccurate. Number of cores containing calciﬁcations was most important in obtaining specimen for accurate diagnosis (odd ratio 9.8, C.I. 2.013~48.000) (p 0.05). The cut-off point of number of cores containing calciﬁcations was 6 (AUC 0.776) (p 0.001). Conclusion: For accurate diagnosis of non-mass calciﬁcations of breast, six or more cores containing calciﬁcations should be obtained at 11-gauge vacuumassisted breast biopsy.

Scientific Sessions in 15 cases. The 15 malignant lesions underwent surgical intervention; the ﬁnal diagnosis was malignancy in all cases. Of the 25 benign lesions, 20 underwent follow-up BMR (mean 7 months, range 4-10 months), which demonstrated no lesion changes; ﬁve lesions underwent surgery because of high clinical concern and turned out to be benign in three and malignant in two cases (ﬁnal diagnosis of in-situ lobular carcinoma and well-differentiated tubular carcinoma). Sensitivity, speciﬁcity, positive and negative predictive values were, respectively, 88, 100, 100 and 92%. Conclusion: MR-SBB performed with a low-cost ferromagnetic-nonmagnetic coaxial system with small caliber needles may be considered a feasible, effective and safe diagnostic procedure.

11:42

3.0-T MRI-guided focused ultrasound ablation versus MRI-guided needlewire placement for the pre-operative localization of non-palpable breast tumors A.C. Schmitz1, M.A.A.J. van denBosch1, V. Rieke1, F.M. Dirbas1, K. Butts-Pauly1, W.P.T.M. Mali2, B.L. Daniel1; 1Stanford, CA/US, 2Utrecht/NL Purpose: To compare the accuracy of MRI-guided focused ultrasound ablation (MRI-gFUS) versus MRI-guided needle-wire placement (MRI-gNW) for the localization of non-palpable breast tumors. Methods and Materials: In this experimental study, 16 turkey breasts were used. In each turkey breast an artiﬁcial non-palpable tumor was created by injecting an aqueous gel containing a 1:200 dilution of gadopentetate dimeglumine. Both localization techniques were performed in a 3.0 T MRI unit (Signa, GE Medical Systems). MRI-gFUS localization (n=8) was performed with the ExAlblate 2000 system (Insightec). Multiple ablation spots were created around the non-palpable tumor in the coronal plane, each spot 5 mm apart, to create a visible and palpable template. MRI-gNW localization (n=8) was performed by placement of a 20GMRI compatible wire centrally through the tumor under MRI guidance. After both localizations, the tumor was surgically excised. MR images were used to evaluate tumor-free margins (neg/pos), minimum tumor-free margin (mm) and the excised tissue volume (cm³). Results: One case of positive margin was found in the FUS-group (12.5%) and two cases were found in the NW-group (25%) (p=0.5). The mean minimum tumor-free margin pSD in the FUS-group was signiﬁcantly larger (4.7p2.9 mm) than the NWgroup (0.9p1.4 mm) (p=0.004). The mean volume pSD of excised tissue was slightly higher in the FUS-group (44.8p9.0 cm³) than in the NW-group (39.5p10.7 cm³), but there was no signiﬁcant difference (p=0.3). Conclusion: The results of this experimental study indicate that MRIgFUS is as accurate as MRIgNW for localization of non-palpable breast tumor in turkey breasts. One of the main advantages of FUS over NW is that FUS is completely non-invasive.

B-290

11:51

Interdisciplinary consensus on the uses and technique of MR-guided vacuum assisted breast biopsy (VAB): Results of a European consensus meeting S.H. Heywang-Köbrunner1, R. Sinnantaby2, A. Lebeau3, A. Lebrecht4, P.B. Britton2, I. Schreer5; 1Munich/DE, 2Cambridge/UK, 3Hamburg/DE, 4Mainz/DE, 5 Kiel/DE ([email protected]) Purpose: Quality assurance of MR guided vacuum-assisted breast biopsy (VAB). Methods and Materials: A consensus was achieved based on the existing literature and the experience of an interdisciplinary group comprising European specialists in breast imaging and VAB (Ethicon Endosurgery, Norderstedt). Results: The imaging work-up must be completed according to existing standards before an indication for MR guided VAB is established. The procedure should be reserved for lesions demonstrable only by MRI. Acquisition of 24 cores (11 Gauge) should be routinely attempted, the objective being the sufﬁcient removal of small lesions for diagnostic purposes. Following biopsy the patient should be re-scanned to demonstrate the biopsy site and its relationship to the lesion and documentation made as to the likely accuracy of the sampling. All cases require discussion in a regular interdisciplinary conference and a documented consensus concerning further work-up or therapy. Regular audit and review of all MR guided VAB results and subsequent outcomes are recommended. Conclusion: This consensus includes protocols for the establishment of an indication for MR guided VAB, performance indicators, interdisciplinary interpretation and therapeutic recommendation, documentation and follow-up. It does not replace ofﬁcial recommendations for percutaneous biopsy.

A

B

C

D

E

F

G

Room I

Physics in Radiology

SS 613 Radiation dose in CT Moderators: N. Marshall; London/UK P. Vock; Berne/CH

B-291

B-289

206

10:30 - 12:00

10:30

Z over-scan dose reduction in cone-beam CT scanners S. Glasberg, G. Shechter; Haifa/IL ([email protected]) Purpose: We present an invention that solves the problem of excessive dose in spiral scans due to z-over scan of wide-cone CT scanners. Methods and Materials: CT scanners with larger coverage have a decreasing dose utility in spiral scans. The lowered dose utility is a result of an increase in the unused portion of irradiation cone at the start and at the end of scan (known as z-over scan). The over scan is needed for reconstruction of the voxels at the edges of region of interest with all useful rays. Yet, the cost is irradiation of body parts that are not reconstructed. The invention we present consists of a collimator with moving blades such that the unused part of the beam is blocked throughout scan. The motion of blades is synchronized with couch velocity and position to ensure that only rays contributing to the reconstruction will be transmitted through the patient. We provide analytical model for the DLP as a function of the spiral-scan parameters and of the blade acceleration. We show numerical simulations of the dose reduction in practical design, and provide initial demonstration of concept with an engineering prototype. Results: We ﬁnd that for a typical cardiac spiral scan the DLP to patient can be reduced by about a third keeping the same X-ray power per slice. Conclusion: A conceptually simple and effective means to reduce the z-overscan in spiral scans of wide-cone CT scanners was presented and is likely to ﬁnd use in future scanners.

B-292

10:39

Practice in multi-slice computed tomography: How do we perform? S.J. Golding1, Y.L. O2, C. Alvey1, E. Benjamins2, K. Chlapoutakis3, E. Fraile Moreno4, J. Hansen5, S. Meeson1, A. Meijer2, T. Niemann6, R. Patel1, L. Ruiz Lopez4, I. Salmeron4, N. Theocharopoulos3, G. Bongartz6, A. Calzado4, J. Damilakis3, J. Geleijns2, A. Jurik5, V. Alonso Ferreira4, C. Cordeiro Alves4; 1 Oxford/UK, 2Leiden/NL, 3Iraklion/GR, 4Madrid/ES, 5Aarhus/DK, 6Basle/CH ([email protected]) Purpose: To evaluate current patterns of practice in Multi-Slice Computed Tomography (MSCT). Methods and Materials: We know that practice of CT varies between centres and changes under the inﬂuence of new technology. As part of a programme in safety and efﬁcacy of MSCT, a pan-European research consortium has conducted a survey of practice in centres possessing CT systems with at least 16 slices. Data were collected from examinations on adults in six departments in different countries using standardized data collection. Productivity over 12 months was collected, with age and gender distribution, anatomic regions, referring specialities and examination protocols. 10% of examination reports were further analysed to document clinical indications and outcomes of examination in terms of deﬁnite diagnoses. Results: 69,033 examinations were collected from 33,149 patients. Annual productivities ranged from 7,550 to 17,344 examinations, reﬂecting type or size of hospital. Use of MSCT rose with age, being highest over 45. Malignant disease was most common (27.5%), followed by respiratory, gastrointestinal and vascular disease. Clinical justiﬁcation practices varied, reﬂecting imaging strategies and availability. Examination protocols differed signiﬁcantly; this prompted reconsidered practice by participants. Most (92%) studies generated deﬁnite reports. The commonest uses were diagnostic (55%) and follow-up (29%). Repeated examination was a notable ﬁnding. 8% of patients had over six studies yearly; this may inﬂuence protection, for example by reducing exposure in follow-up studies. Conclusion: There remain signiﬁcant variations in practice in CT and this appears to be further opportunity for radiation constraint by optimising examination protocols and exposure.

Scientific Sessions B-293

10:48

A novel robust and dose efﬁcient tube current modulation for ECG-gated cardiac spiral CT imaging R. Raupach1, H. Bruder1, C. Becker2, B. Schmidt1, T.G. Flohr1; 1Forchheim/DE, 2 Munich/DE ([email protected])

B-294

10:57

Establishment of a biological phantom model for dose estimation in cardiac CT M.A. Kuefner1, K. Anders1, J. Hamann1, S.A. Schwab1, S. Achenbach1, W.A. Bautz1, M. Lobrich2, M. Uder1; 1Erlangen/DE, 2Homburg a.d. Saar/DE ([email protected]) Purpose: Evaluation of the biological inﬂuence of various factors including heart rate, tube current time product, and pitch on dose reduction in different examination protocols or scanners is difﬁcult. Therefore, it was the aim of this study to establish a biological phantom model for dose estimation and for assessment of DNA damages in cardiac CT. Methods and Materials: Experiments were performed with a 64-slice and a dualsource CT scanner. Blood samples of healthy individuals were irradiated in a 32-cm CTDi phantom using a standard and a low kV protocol (120 or 100 kV, 800 eff. mAs or 320 mAs/rot, with or without pulsing). Determination of DSBs is based on the phosphorylation of the histone variant H2 AX after formation of DSBs. Blood lymphocytes were isolated, incubated with a speciﬁc Gamma-H2 AX antibody, and DSBs were counted by immunoﬂuorescence microscopy. DLP was determined using a dosimeter equipped with an ionisation chamber. Results: Using the applied protocols, number of DSB was comparable for the 64-slice and dual source CT scanner. A strong correlation was found between the number of DSBs and DLP (R=0.81). By using ECG-pulsing for dose reduction DLP and number of DSBs was diminished up to 30%. A 100 kV protocol led also to a reduction of DLP and number of DSBs. Conclusion: The established biological phantom model using Gamma-H2 AX immunoﬂuorescence microscopy is reliable and suitable for dose estimation and comparison of different scanning protocols and scanners in cardiac CT. Further experiments on this ﬁeld will be performed.

B-295

11:06

Monte Carlo simulation to assess organ radiation dose from coronary CT angiography (CTA) exams using patient-based voxelized models E. Angel1, N. Yaghmai1, J.J. DeMarco1, C.H. Cagnon1, J.G. Goldin1, A.N. Primak2, D.M. Stevens3, D.D. Cody3, C.H. McCollough2, M.F. McNitt-Gray1; Los Angeles, CA/US, 2Rochester, MN/US, 3Houston, TX/US ([email protected]) Purpose: Coronary CTA exams use low pitch values to accomplish retrospective cardiac gating, resulting in increased radiation dose. This study aims to estimate organ dose to breast and lung tissue from coronary CTA exams. Methods and Materials: Nineteen voxelized models of female anatomy were created from clinical CT data. Patient size was deﬁned as the patient’s trans-axial perimeter. Breast tissue was contoured by a radiologist and glandular and lung

B-296

11:15

Optimized tube current selection for consistent image quality and dose reduction in MDCT cardiac imaging W. Qi, X. Du, J. Li, Y. Shen; Beijing/CN ([email protected]) Purpose: To introduce and clinically evaluate a bolus-tracking scan (BTS) based tube current (mA) selection method in MDCT coronary artery (CTA) imaging to obtain precise and consistent image quality and to optimize dose utilization for patient populations. Methods and Materials: We ﬁrst established correlation between the image noises of bolus tracking (BT) and CTA scans at ﬁxed mA using our standard cardiac scan protocol with 64-row MDCT. This correlation was then converted to form a formula to determine the patient-dependent mA for getting a desired and consistent CTA image noise based on the measured BTS image noise for individual patient. Subsequently, 100 cardiac patients were scanned with individually adjusted mA selected using the formula for achieving noise of 28 HU which was determined to be clinically adequate based on analyzing existing CTA images. Results: The average noise for CTA images of 100 patients with BTS-based mA selection was found to be 28.5 HU and the standard deviation of image noise was about 2 HU, which was much smaller than the 5 HU for the group of patients with ﬁxed mA. The lowest CTDIvol for the BTS-based mA group was only 25% that of the ﬁxed mA group, and the average image scores from the two groups showed no statistical difference. Conclusion: Adjusting mA individually based on bolus-tracking scan noise analysis is an accurate and practical method to obtain consistent image quality for different cardiac patients and provides dose reduction for smaller patients compared to the use of ﬁxed mA.

B-297

11:24

Novel tube current modulation algorithms for reducing radiation dose to breasts in thoracic CT E. Angel1, F. Noo2, J.J. DeMarco1, C.H. Cagnon1, N. Yaghmai1, J.G. Goldin1, M.F. McNitt-Gray1; 1Los Angeles, CA/US, 2Salt Lake City, UT/US ([email protected]) Purpose: To develop new tube current modulated (TCM) acquisition algorithms designed to speciﬁcally reduce dose to radiosensitive breast tissue in thoracic CT. Methods and Materials: A voxelized model of female anatomy was created from clinical CT data. Breast tissue was contoured by a radiologist and glandular tissue automatically segmented. A previously validated Monte Carlo model of a multidetector CT scanner was modiﬁed to simulate three novel thoracic TCM algorithms, a conventional TCM exam, and a ﬁxed tube current exam. The novel algorithms were designed to speciﬁcally decrease breast dose while maintaining comparable image quality. The three novel algorithms are: (A) tube current reduced over the patient’s anterior with ﬁxed tube current in the remainder of the rotation, (B) tube current reduced over two arcs centered over the glandular breast tissue with ﬁxed current in the remainder of the rotation, and (C) current reduced over two arcs and conventional TCM used for the remainder of the rotation. Radiation doses were calculated for glandular breast tissue using each of the simulated exams, and compared to the simulated ﬁxed tube current scan. Results: Compared to the dose from a ﬁxed tube current scan, the glandular breast dose for each novel TCM algorithm was reduced by: (A) 10%, (B) 16%, and (C) 43%. The conventional TCM simulation was 34% less than a ﬁxed tube current scan. Conclusion: The most effective TCM algorithm for breast dose reduction was (C) which incorporates conventional TCM and a reduction in tube current in two arcs over the glandular tissue.

A

B

C

D

E

F

G

207

Saturday

Purpose: To improve the robustness of an ECG-correlated tube current modulation with maximum dose reduction for cardiac spiral CT, also in the event of arrhythmia and ectopic heart beats. Methods and Materials: To reduce radiation dose for cardiac spiral imaging, the exposure should be signiﬁcantly lowered outside the cardiac phase of interest. However, the exact time-stamps can only be determined retrospectively. We present a novel approach for on-line control of the high dose sectors. R-peak estimation was based on the following ﬁve criteria: median, minimum, maximum, least square linear ﬁt (trend analysis) and normalized standard deviation from the trend line. The system reacted also to delayed R-peaks and extra-systoles. The algorithm was implemented on a SOMATOM deﬁnition dual source CT and applied to 50 coronary CTA examinations performed without beta-blockers, including ectopic ECGs. The heart rate varied between 48 and 90 bpm. The coincidence of the prediction and the required sectors for typical clinical parameters and the dose reduction by means of the tube current modulation were determined. Results: Coincidence between high dose sectors and reconstructed data segments could be achieved in all cases. Therefore, image quality was maximized for all exams, including those with ectopic beats. The reduction of radiation dose compared to a non-modulated scan averages to 39% where the maximum reduction was 52%. Conclusion: The presented tube current modulation for cardiac spiral imaging proves to offer a robust approach to signiﬁcantly reduce radiation dose without compromising image quality in case of arrhythmia or ectopic beats.

tissues automatically segmented. Previously validated Monte Carlo models of 16- and 64- slice scanners were used, taking into account source spectra and path, ﬁltration, and collimation. Simulated CTA exams used clinical protocols for an average size patient with ﬁxed mA, and accounted for z-axis over-ranging. Organ doses were calculated for breast and lung tissues, and compared to published doses from typical chest CT. Results: Average glandular breast dose was 47 mGy (range:7-73 mGy). Average lung dose was 46 mGy (range:22-63 mGy). For both organs, dose was greater for smaller patients. Dose from the 16-slice scanner was higher than the 64 by approximately 10%, a difference not attributable to pitch or mA. Typical chest CT contributes approximately 21 mGy to breast and 24 mGy to lungs. Conclusion: Coronary CTA delivers relatively high breast and lung doses. Higher organ dose for smaller patients reinforces the necessity of size-dependent scan parameter adjustment, as automated tube modulation is often not used in this application. Female breast and lung doses from coronary CTA can be twice as much as for conventional chest CT.

Scientific Sessions B-298

11:33

Measurements of CT dose proﬁles with a new liquid ionization chamber O. Langner1, R. Kranzer2, Y. Kyriakou1, W.A. Kalender1; 1Erlangen/DE, 2Freiburg/DE ([email protected]) Purpose: A new type of liquid ionization chamber (LIC) developed for application in radiotherapy was investigated with respect to its performance characteristics for measuring dose in X-ray CT. Methods and Materials: The LIC has an active volume of 1.7 mm³ and 0.35 mm depth (PTW, Freiburg, Germany). The LIC was evaluated using different voltage bias levels. Single scan dose proﬁles were measured in air and in standard CTDI head phantom of 160 mm diameter for different slice collimations and tube voltages. Full widths at half maximum (FWHM) of the dose proﬁles were evaluated. Results were compared to common TLD measurements of dose proﬁles in air and in the phantom. CTDI were calculated from both the LIC and TLD measurements and compared to CTDI measurements with a standard pencil-shaped ionization chamber. Results: Highest sensitivity and SNR of the LIC were obtained using a bias voltage of 800 V. Measurements of dose proﬁles with the LIC were found in good agreement with TLD results; maximum deviations of FWHM between LIC and TLD were 11% in air and 4% in the phantom. CTDI calculations from LIC measurements were in good agreement with CTDI obtained by TLD measurements. CTDI measurements with the standard 100 mm chamber were within 10% of the CTDI resulting from LIC and TLD measurements. Conclusion: The small-volume LIC offers the most favourable geometry for measuring dose proﬁles and provides better sensitivity than standard air-ﬁlled ionization chambers. LIC appears to offer great potential for measuring CTDI in CT.

B-299

11:42

Dose assessment for a multicenter study on diagnostic performance of cardiac 64-slice CT J. Geleijns1, A. Calzado2, M. Dewey3, R.M.S. Joemai1, R. Mather4, A. de Roos1, M. Salvado5; 1Leiden/NL, 2Madrid/ES, 3Berlin/DE, 4Tustin, CA/US, 5Reus/ES ([email protected]) Purpose: An international multicenter study (CorE64; 9 centers; 400 patients) has been performed to establish the diagnostic accuracy of coronary CT angiography (CCTA) with 64-slice CT scanners. In this ancillary study we aim at documenting the dosimetric performance of the CT scanners and at calculating organ and effective dose. Methods and Materials: Cardiac CT was performed with 64-slice CT scanners (Aquilion64, Toshiba Medical Systems). Different acquisition protocols were used for patients of different size and gender. CTDIw was measured at all scanners. The Aquilion64 scanner was modeled in a Monte Carlo algorithm (EGS4). Organ and effective dose calculations were performed for six patients of different gender and size. Six segmented voxel phantoms were prepared. Doses were assessed for CT radiography, calcium scoring, dynamic bolus tracking, and CCTA. Results: Small differences were observed for CTDIw (8.6p0.5 mGy/100 mAs). Effective doses for the complete examination (six voxel phantoms) was 17p2 mSv, variations are attributable to size but not to gender. Effective dose for ECG gated CTA was 15p2 mSv. The three organs contributing most to effective dose were lung, stomach and breast. Conclusion: Consistent dosimetric performance was observed for all scanners. The goal of the multicenter study to achieve a similar range of effective doses for females and males has been realized. The Monte Carlo algorithm allowed for deﬁning a realistic acquisition geometry for cardiac CT with the patient positioned off-center. Effective doses are similar to those found in other studies but when compared to other studies, some organ doses show deviations up to a factor of 2.

B-300

11:51

DSCT versus 64-slice spiral CT: Comparison of radiation dose H.-T. Qi1, R. Merges2, J. Chen2, W. Qin1; 1Jinan/CN, 2Beijing/CN Purpose: To optimize the cardiac CTA scanning protocol on DSCT and 64-slice MSCT and to compare the radiation dose between these two scanners for cardiac CTA. Methods and Materials: 40 patients were enrolled for cardiac CTA on DSCT (120 kVp; 550 mA) and 20 patients were enrolled for 64-slice MSCT (120 kVp; 770 mA) separately. Image quality was graded using a 5-point scale and three cardiovascular radiologists assessed and compared the image quality from these two CT systems independently. Estimation of effective radiation dose was performed based on the DLP. Results: The average estimated effective dose of 64-slice MSCT was 11.7p0.27 mSv for low heart rates (50-70 bpm). Because pitch (0.2-0.5) changes adaptive to the heart rate on DSCT, the average estimated effective dose of DSCT was

208

A

B

C

D

E

F

G

12.2p1.73 mSv, 9.46p0.92 mSv, 7.98p1.59 mSv, 6.91p0.71 mSv and 5.94p1.25 mSv with the corresponding heart rate of 50-60 bpm, 60-70 bpm, 70-80 bpm, 80-90 bpm and more than 90 bpm. The image quality with heart rates above 70 bpm was equivalent to that with heart rates below 70 bpm (t=1.16, P 0.05). Conclusion: The radiation dose of DSCT has no difference to that of 64-slice MSCT for CTA examination at low heart rates ( 70 bpm). DSCT dose is far lower at higher heart rates ( 70 bpm) without decrease in image quality.

10:30 - 12:00

Room L/M

Neuro

SS 611 New vascular MR techniques Moderators: X. Leclerc; Lille/FR M.A. Lucic; Novi Sad/RS

B-301

10:30

Perfusion territory imaging of intracranial branching arteries using continuous artery-selective spin labeling M. Helle1, K. Alfke1, D.G. Norris2, O. Jansen1; 1Kiel/DE, 2Nijmegen/NL ([email protected]) Purpose: Continuous Artery-Selective Spin Labeling (CASSL) is a technique to image perfusion territories of cerebral arteries. In this study we demonstrate the feasibility of this method to selectively label blood of individual branching intracranial arteries in vivo. Methods and Materials: In CASSL an oscillating motion of the labeling gradient about the axis of the artery combined with frequency modulation allows labeling of the desired artery. Computer simulations were used to optimize key labeling parameters. MR images were acquired from eight healthy volunteers with a standard transmit/receive head coil on a clinical 1.5 T Philips Intera MR system. Labeling was applied to small arteries with diameter less than two mm that are branching from middle and anterior cerebral arteries. Results: In agreement with the computer simulations, selective labeling clearly showed well delineated perfusion territories of A2/A3 and M2/M3 segments with high grey/white matter contrast. Angles less than or equal to 10° between selected artery and labeling gradient provided good results for well separated arteries without adjacent vessels. For arteries in immediate vicinity of other small vessels or branching points it is recommended to use angles between 12° and 15° to ensure a proper labeling of the selected artery only. Conclusion: CASSL is a technique suitable to selectively label blood of intracranial arteries in vivo, e.g. in the cerebral arteries distal to the Circle of Willis. Perfusion territories of single small intracranial arteries may provide important information about border zones of perfusion and probably about collateral ﬂow.

B-302

10:39

Perfusion heterogeneity in the infarct core and ischemic penumbra: Utilizing the novel “dynamic histogram analysis” in perfusion weighted MRI M. El Koussy, P. Schenk, A. Spreng, L. Remonda, L. Kappeler, R. Wiest, G. Schroth, J. Slotboom; Berne/CH ([email protected]) Purpose: Measurement of perfusion heterogeneity in the infarct core, ischemic penumbra and healthy brain using a novel analysis method, named Dynamic pixel-intensity Histogram Analysis (DHA) allowing for pixel-intensity-histogrammodel-parameter ﬁtting of the regions-of-interest (ROIs)deﬁned in dynamicsusceptibility-contrast-enhanced perfusion-MRI (DSCE) difference MR-image time-series and ﬁnding a possible role for the DHA model to predict outcome of ischemic parenchyma. Methods and Materials: We prospectively analysed patients with acute ischemic stroke, who were scanned with MRI within 24 h after ictus and received conservative therapy (no thrombolysis). The patients received a late MRI follow-up. Ischemic penumbra was deﬁned in the stroke-MRI as the area with diffusion-perfusion mismatch. The DHA method was applied on DSCE-difference images. For each difference image of the time series, pixel-intensity histograms of different ROIs were computed and ﬁtted using the Levenberg-Marquardt algorithm. Histogram centerposition- and width-parameters as also the width versus position were computed as a function of time during bolus-passage. Results: Twenty-ﬁve cases met the inclusion criteria. The asymmetry indices pointed to extremely low perfusion heterogeneity in the infarct core. The perfusion in the penumbra was more heterogeneous than in the healthy parenchyma. The

Scientific Sessions parameter “width versus position” was the most sensitive for differentiating infarct core from ischemic penumbra. Conclusion: Assessment of the perfusion heterogeneity in the ischemic brain adds further insight into the pathophysiology of perfusion disturbance in stroke. DHA is a novel, histogram-based, robust, reproducible method of analysis.

B-303

10:48

Delay insensitive singular value decomposition algorithm for brain perfusion analysis using shifted artery curves and preconditioning techniques J. Wang1, O. Masoud1, B. Frake1, T. Naoko2, O. Miwa2; 1Minnetonka, MN/US, 2 Tochigi/JP ([email protected])

B-304

10:57

Sensitivity of susceptibility weighted imaging (SWI) towards BOLD-signal changes induced with low CO2 concentrations J. Sedlacik1, A. Deistung1, S. Witoszynskyj1, R. Alexander2, H.-J. Menzel1, C. Kutschbach1, W.A. Kaiser1, J.R. Reichenbach1; 1Jena/DE, 2Vancouver, BC/CA ([email protected]) Purpose: Breathing carbogen (5% CO2/95% O2) dramatically increases cerebral blood ﬂow and BOLD-signal. However, 5%CO2 causes fast and deep breathing which discomforts patients and generates motion artifacts in MRI. As BOLD-related signal changes yield information about tissue perfusion and oxygenation which allow, for example, the estimation of tumor sensitivity on chemo- and radiotherapy, the aim of the study was to investigate signal changes in the brain with susceptibility weighted imaging (SWI) with various CO2-concentrations below 5%. Methods and Materials: SWI data of 15 volunteers were acquired using a fully velocity-compensated 3D gradient-echo sequence with high spatial resolution at 1.5 T (TE/TR/FA=40 ms/57 ms/20°, FoV=256x192x72 mm3, matrix=512x256x38). Carbogen was mixed with pure oxygen using a cPAP-respirator to obtain 0/1.67/3.33/5% CO2. Signal changes normalized to pure oxygen (0% CO2) were calculated for ROIs in veins, ventricles and segmented gray/white matter. Results: The signal of cortical veins increased by 3.5p3.8/10.3p4.5/22.7p8.8% with 1.67/3.33/5% CO2, respectively. A simple paired t-test revealed signiﬁcance for all three signal changes. No signiﬁcant changes were obtained in the ventricles. Weak signiﬁcant signal changes up to 2.0p1.3%/0.8p1.1% were observed in segmented gray/white matter only for 5% CO2. As expected, motion artifacts were more severe at 5% CO2 compared to all lower CO2-concentrations and resulted in higher standard deviations of the observed signal changes. This was also conﬁrmed by the subjects who reported distinctly much less discomfort at CO2-concentrations below 5%. Conclusion: We recommend a carbogen mixture with 3.3% CO2/96.6% O2 for perfusion-like measurements, which is well tolerated and still induces signiﬁcant signal change in the venous vasculature.

11:06

Demonstration of the central cerebral veins by susceptibility-weighted imaging (SWI) C.M. Plank, G. Kasprian, C. Mueller-Mang, P.C. Brugger, D. Prayer; Vienna/AT ([email protected]) Purpose: To investigate the ability of SWI to demonstrate the normal course of the basal central vein, internal cerebral vein, and variations thereof. Methods and Materials: 50 consecutive patients (8-60 years, mean 48.5) who underwent cerebral MRI for different indications, whose ﬁndings suggested absence of any venous pathology were included. All patients were examined on a 3 T scanner and in addition to standard sequences an SWI sequence (TR 28, TE 20, FA 15, FOV) was performed. Images were evaluated concerning the main feeders of the basal central vein, the internal cerebral vein and variations of the feeders. Results: Deep cerebral veins and their feeders could be identiﬁed in each case (100%). Evaluation included the proof of the internal cerebral vein, its feeding veins (V septi pellucidi ant, and terminalis), to its conﬂuence to the vena Galeni, of the main oriﬁcing veins (V septi pellucidi and thalamostriata sup), basal vein and ﬁrst main feeder (anterior cerebral vein) to its oriﬁce into the vena Galeni or Sinus rectus. Variations of the feeders of the internal cerebral vein were depicted in 10 cases; in ﬁve cases small subtle deep venous malformations were found. Conclusion: The main deep vessels responsible for the drainage of the brain and their variations are readily and reliably identiﬁable on SWI. Variations and subtle malformations of deep cerebral veins are depicted accurately. Thus SWI holds the potential to become an important non-invasive modality for investigation of venous malformations.

B-306

11:15

Dural sinus disease: Comparison of MR venography (MRV) and CTV S. Xia, J. Qi; Tianjin/CN ([email protected]) Purpose: To compare the value of CTV and MRV in demonstrating disease of cerebral vein. Methods and Materials: 21 patients from June 2005-January 2007 underwent cranial CTV and MRV imaging. CT studies were performed on GE Lightspeed 16 scanner 1.25 mm thick and with 0.6 mm gap. Helical multidetector CT venography with bolus power injection of contrast material and combined use of two-dimensional and three-dimensional reformations (maximum intensity projection, integral display, and volume rendering). MR imaging was performed on 1.5 Tesla Philips scanner with T1 and T2 weighted (T1 and T2W), FLAIR and MRV in multiple planes. Results: Of the 21 patients, 12 patients were diagnosed as having cerebral sinovenous thrombosis on both CT venography and MR venography. The total number of sinuses involved was 21. Five patients with meningioma involved superior (four cases) and transverse sinus (one case). Four patients with multiform glioblastoma inﬁltrated superior sinus. All of the relationships between tumor and sinus were clearly shown by the CTV and MRV. Among several techniques of CTV, 2D reformation can show more lesions of the sinus than 3D SSD and VR. Conclusion: CT venography is as accurate as MR venography for diagnosing cerebral sinus disease. Due to rapid scan, CT scan remains the primary ﬁrst diagnostic modality used in patients suspected with sinus disease.

B-307

11:24

Time-of-ﬂight magnetic resonance venography at 3 T: Comparison of a 12versus 32-channel head coil F.L. Giesel1, V.M. Runge2, E.L. Lowther2, B.J. Correll2, M. Essig1; 1Heidelberg/DE, 2 Temple, TX/US ([email protected]) Purpose: MR evaluation of intracranial circulation provides valuable diagnostic information that can inﬂuence patient management. This study investigated signal improvement in time-of-ﬂight (TOF) MR venography using a prototype 32-channel head coil (CHC) compared to a standard 12-CHC, at 3 T. Methods and Materials: Ten volunteers were imaged on a Siemens 3 T Tim Trio. A high-resolution protocol acquired with the 32-CHC was compared to the standard 12-CHC TOF. Scans were acquired with TR/TE/tip angle=21/4.76/60° and TA=3.35 min. The 12-CHC utilized a parallel imaging factor (IPAT)=2 and voxel dimensions=1x1x2.5 mm 3; the 32-CHC employed IPAT=4 and voxel dimensions=0.7x0.7x2 mm3. Without the change in coils, these parameters would have reduced SNR from a nominal value of 1 to 0.31. Single-slice axial, coronal, and sagittal scans were acquired to investigate actual change in SNR between the two head coils. Each scan was performed twice to determine SNR. ROI measurements were performed in the major dural sinuses. Results: The 32-CHC demonstrated statistically signiﬁcant improvement in SNR in each ROI, when compared to the standard 12-CHC (170-200%). Accounting for the

A

B

C

D

E

F

G

209

Saturday

Purpose: We propose a new method that can better deal with delays in contrast arrival to different areas of the brain for perfusion analysis of datasets acquired from CT and MR, compared with standard singular value decomposition (sSVD), with or without tissue curve shifting, and the block-circulant SVD (bSVD) methods. Methods and Materials: Our method performs a shift of the artery curve so that the time of arrival of contrast to the shifted curve is always before those of all tissue curves. In addition, we apply a novel pre-conditioning step to stabilize the deconvolution calculation. The method was tested on over 80 real CT datasets from various sources and on 157,696 simulated curves. Results: Experiments using simulated curves show that the algorithm is invariant to positive and negative delays. Compared to bSVD, the algorithm runs at least 4X faster and generates MTT values within 2.3p2.3% of bSVD. Tests on real CT datasets with artiﬁcially induced delay of up to p5 seconds show that the computed perfusion values remain within 2p1.9% (5p5.8% for bSVD) of the values calculated using the original dataset without induced delay. Furthermore, our algorithm is much more stable than sSVD, with or without curve shifting, in that the ﬂuctuation of computed MTT values due to the singular value cut-off is about 7p2.3X smaller than that of the sSVD algorithm Conclusion: The proposed deconvolution algorithm using a combination of artery shifting and a preconditioning technique correctly handles contrast delay and has advantages over other SVD-based algorithms.

B-305

Scientific Sessions change in SNR due to different voxel dimensions, each ROI revealed an increase in SNR (p 0.001; nominal value: coronal superior sagittal sinus: 12-CHC: 29.7, 32-CHC: 55.3; axial superior sinus: 12-CHC: 56.9, 32-CHC: 113.3; transverse sinus: 12-CHC: 45.6, 32-CHC: 77.5). Conclusion: The 32-CHC provided a statistically signiﬁcant SNR increase in TOF imaging at 3 T. Detailed depiction of venous sinus structure obtained with this high-resolution protocol may improve diagnosis in venous thrombosis and other pathologies.

B-308

11:33

A comparison between susceptibility-weighted imaging (SWI), T2-weighted fast spin-echo and gradient-echo sequences for the evaluation of patients with familial cerebral cavernous malformations L.C.H. da Cruz Jr, E.L. Gasparetto, R.C. Domingues, J.P.M. de Souza, F.S. Domingues, R.C. Domingues; Rio de Janeiro/BR ([email protected]) Purpose: To compare the sensitivity of susceptibility-weighted images (SWI) with T2-weighted fast spin-echo (FSE) and gradient-echo (GRE) sequences in assigning the number of cerebral cavernous malformations (CCM) in patients with the familial form of the disease. Methods and Materials: Fifteen patients with familial CCM underwent MR imaging at 1.5 T scanner with protocol including T2-weighted FSE and GRE, as also SWI. SWI combines a long echo time, high-resolution, fully ﬂow-compensated, 3D GRE sequence with ﬁltered phase information in each voxel. Two neuroradiologists independently counted the number of lesions in each sequence. The number of CCMs seen in each sequence was compared with analysis of variance, followed by nonparametric Wilcoxon matched pairs test. Results: The difference between the number of lesions seen in the three sequences was statistically signiﬁcant (P 0.0001). The number of CCMs was higher on SWI than on T2-weighted GRE (P=0.001) and FSE (P=0.001) sequences. Regarding the Zabramski's classiﬁcation of the CCMs, there were 14 (2.3%) type 1 lesions, 29 (4.9%) type 2, 32 (5.4%) type 3 and 267 (44.9%) type 4 lesions. A substantial number of CCMs (n=253; 42.5%), which were seen only on the SWI sequence, could not be categorized based on Zabramski's classiﬁcation. Conclusion: The sensitivity of SWI in assigning the number of CCMs in the familial form of the disease is signiﬁcantly higher than T2-weighted FSE and GRE sequences. Also, because the number of lesions seen on SWI was higher than T2-weighted GRE images, we suggest a modiﬁcation on the Zabramski’s classiﬁcation.

B-309

11:42

In-vivo visualization of ﬂow in intracranial aneurysms with 4D ﬂowsensitized MR imaging S. Meckel1, A.F. Stalder2, F. Santini1, K. Schefﬂer1, E.-W. Radü1, M. Markl2, S.G. Wetzel1; 1Basle/CH, 2Freiburg/DE ([email protected]) Purpose: Intra-aneurysmal ﬂow patterns and wall shear stress (WSS) are assumed to play a major role in development and rupture of cerebral aneurysms. To date, these hemodynamic aspects have been extensively studied in vitro using geometric realistic aneurysm models. In this study, the feasibility of in vivo ﬂow-sensitized 4D MR imaging for analysis of intra-aneurysmal hemodynamics was evaluated. Methods and Materials: Five unruptured intracranial aneurysms in three patients were examined. RF-spoiled ﬂow-sensitive time-resolved 3D gradient echo imaging with prospective cardiac triggering at 3 T (Allegra, Siemens, Germany) was applied. Other scan parameters: VENC: 90 cm/s, spatial resolution: 1.0x0.7x1.2 mm3, temporal resolution: 44.8 ms. Postprocessing using a commercial 3D visualization software (EnSight, CEI, USA) included quantiﬁcation of ﬂow-velocities, timeresolved 2D vector graphs and 3D particle traces, vortex core segmentation, and WSS calculation. Flow patterns were analyzed in relation to aneurysm geometry and aspect ratio (aneurysm depth/neck width). Results: The spatial and temporal evolution of vortical ﬂow differed markedly among all aneurysms. Particularly unstable vortical ﬂow was demonstrated in a wide-necked parophthalmic aneurysm (high aspect ratio). Relatively stable vortical ﬂow was observed in aneurysms with lower aspect ratio. Except for a wide-necked cavernous aneurysm (low aspect ratio), WSS was reduced in all aneurysms relative to the parent artery and showed a high spatial variation. Conclusion: In vivo ﬂow-sensitized 4D MR imaging can be applied to visualize and quantify blood ﬂow patterns (e.g. vorticity), and to estimate WSS inside cerebral aneurysms. Flow patterns, distribution of ﬂow velocities, and WSS seem to be determined by the aneurysm geometry.

210

A

B

C

D

E

F

G

B-310

11:51

Time-resolved three-dimensional magnetic resonance digital subtraction angiography without contrast material M. Hori1, N. Shiraga1, M. Shimada1, A. Yamaguchi1, K. Miyamoto1, K. Suzuki1, H. Terada1, K. Goto1, S. Isono1, Y. Masao2; 1Tokyo/JP, 2Tochigi/JP ([email protected]) Purpose: A novel MR angiography (MRA) of three-dimensional (3D) magnetic resonance digital subtraction angiography (MRDSA) without contrast material, which is essentially 3D true steady state free precession (SSFP) with selected inversion recovery (IR) pulse using multiple cardiac phase acquisitions with a short increment delay is introduced for demonstrating normal arteries in the brain, as a feasibility study before clinical use. Methods and Materials: A series of MRA images using 3D MRDSA without contrast material were acquired from 10 healthy volunteers in sagittal plane. The imaging time of 3D MRDSA for each volunteer was up to approximately ﬁfteen minutes because multiple data acquisition with varied delay time after IR pulse was needed. We evaluated visualization of normal cranial arteries (internal carotid arteries, middle cerebral arteries, anterior cerebral arteries, posterior cerebral arteries, basilar arteries and vertebral arteries) on 3D MRDSA and compared MRDSA and conventional MRA method, 3D TOF-MRA, on a workstation. Results: In all volunteers, 3D MRDSA containing hemodynamic information were successfully imaged. Compared with 3D TOF-MRA, 3D MRDSA without contrast material clearly visualized the branches of the cranial arteries (p 0.01). However, pseudo-stenotic lesions were sometimes seen on 3D MRDSA images. Conclusion: 3D MRDSA is a simple and easy method to obtain hemodynamic information, and it may play an important role in assessing cranial arteries without contrast material.

10:30 - 12:00

Room N/O

Chest

SS 604 Reporting and CAD (CXR and CT) Moderators: K. Marten; Wuerzburg/DE M.G. Skilakaki; Athens/GR

B-311

10:30

Do radiologists still need to report chest X-rays? T. Bosemani, P. Mehrotra, J. Cox; Durham/UK ([email protected]) Purpose: The aim of our study was to assess the ability of clinicians of different grades and specialities to correctly evaluate a series of chest X-rays. Methods and Materials: Sixty clinicians of different grades and from different specialities were randomly recruited for the study. They were asked to interpret 15 chest X-rays (normal ﬁlms n=5 and abnormal ﬁlms n=10) within 30 minutes. Nonparametric statistical tests were performed to assess for differences in the ability of different groups of doctors to interpret chest X-rays. Results: Senior doctors (consultants, staff grades and registrars, n=30) attained signiﬁcantly higher scores than junior doctors (SHOs, F2 and F1 doctors, n=30, p 0.0001). Specialists were regarded as radiology consultants and registrars and chest consultants and registrars, as these doctors are generally skilled in interpreting chest X-rays. Specialist doctors achieved signiﬁcantly higher scores than non-specialist senior doctors (p 0.0001). Finally, radiologists had signiﬁcantly higher scores than all other consultants (p 0.0001). Furthermore, radiologists had a signiﬁcantly higher score than medical consultants (p 0.0001). Conclusion: The accurate interpretation of CXRs is an essential skill for all hospital specialities. These results highlight that clinicians who are regarded as specialists in reporting CXRs can interpret CXRs more accurately than ‘non-specialists’. We suggest that in order to avoid reporting inaccuracies all CXRs should be reported by a radiologist. In addition, as senior doctors scored signiﬁcantly better than junior doctors, we suggest that all CXRs should be reviewed by an experienced doctor particularly during an emergency admission in order to avoid serious errors in patient management.

Scientific Sessions B-312

10:39

Computer-aided detection (CAD) of pulmonary nodules in chest CT: Do radiologists beneﬁt equally from CAD? J.E. Roos, D.A. Olsen, D.S. Paik, S. Napel, G.D. Rubin; Stanford, CA/US ([email protected])

B-313

10:48

Comparison of performance of radiologists and a novel CAD system in the detection of pulmonary nodules on digital chest radiographs E. Kotter1, T.A. Bley1, T. Baumann1, U. Saueressig1, G. Pache1, M. Treier1, O. Schäfer1, U. Neitzel2, M. Langer1; 1Freiburg/DE, 2Hamburg/DE ([email protected]) Purpose: To evaluate the performance of a newly approved computer aided detection system in detecting pulmonary nodules on digital chest X-rays. Methods and Materials: The sensitivity of three radiologists and of a CAD system for the detection of pulmonary nodules from ﬁve to 15 mm size on digital chest radiography of 117 patients was compared. The reference standard was established by consensus reading of CT scans by two experienced radiologists. CT scans and chest radiographs were performed within four weeks. Results: In 42 patients, 66 pulmonary nodules with a mean nodule diameter of 7.5 mm (SD 2.2 mm) were included in the statistics. 75 of the 117 patients did not have nodules of 5-15 mm size. 288 false-positive detections of the CAD system were found with an average of 2.46 false-positives per image. Sensitivity of the CAD system was 39.4% (95% CI 11.8%) as compared to 18.2-30.3% (95% CI 9.3-11.1%) of the three radiologists. Substantial agreement for detection of CT proved nodules was found among the three radiologists (kappaN 0.64-0.73) with only moderate agreement among the radiologists and the CAD performance (kappaN 0.45-0.52). Conclusion: The CAD system’s diagnostic senstivity in detecting pulmonary nodules of 5-15 mm size outperformed that of the radiologists. It may be used for assisting the radiologist in the detection of lung nodules on digital chest radiographs. The incremental value of the tested CAD system when used as a second opinion to increase the radiologist’s diagnostic performance in reading digital chest radiography has yet to be deﬁned.

B-314

10:57

Is the computer-aided detection of lung lesions also effective in chest radiography? M. Uffmann1, M. Weber1, M. Scheerder2, E. Boorsma2, C. Schaefer-Prokop2; 1 Vienna/AT, 2Amsterdam/NL ([email protected]) Purpose: To evaluate the performance of chest-radiography-based computer-aided detection (CAD) system in the detection of pulmonary lesions. Methods and Materials: The study group consisted of 72 patients, of which 46 patients had 56 malignant pulmonary lesions (39 metastases and 17 T1 bronchogenic carcinomas). Though lesion conspicuity was generally low, study inclusion was based on a possible identiﬁability on the CXR. For all patients a chest CT provided gold standard. Two experienced chest radiologists and two ﬁrst-year residents independently read the images in randomized fashion (a) with CAD (xLNA

B-315

11:06

Part-solid pulmonary nodule detection in MDCT images: Evaluation of a CAD system A.K. Jerebko1, M. Salganicoff1, T. Itoh2, S. Periaswamy1, M. Wolf1, S. Yu1, V. Anand1, S. Park1, L. Bogoni1, M. Ujita2; 1Malvern, PA/US, 2Tokyo/JP ([email protected]) Purpose: To design a CAD system for detection of part-solid (PSN) and solid pulmonary nodules (SN). Methods and Materials: A data set of 60 MDCT images was acquired and randomly split into 26 training and 34 test images. Ground truth nodule locations, sizes and classiﬁcation (pure ground glass nodule (GGN), PSN or SN) were assigned by an expert thoracic radiologist. A CAD algorithm consisted of candidate generation module, feature extraction and classiﬁcation steps. Features characterizing nodule shape, size, density (HU) and texture were designed to reﬂect both PSN and SN properties. Classiﬁcation algorithm for false positive reduction was trained and calibrated using cross-validation approach on the training set. Only PSN and SN were used to train classiﬁcation algorithm. GGN were set aside and not used as either true or false positive examples. Finally the test set was used to obtain the sensitivity and speciﬁcity estimates. Results: The computer aided detection algorithm correctly identiﬁed 30 out of 35 PSN: 86% sensitivity and 11 out of 14 SN: 79% sensitivity in the test set. These accuracy estimates were conﬁrmed on a larger test set that contains predominantly SN: sensitivity for SN 210/242=87%, PSN 42/52=81%. Median false positive rate was 2.5 per patient. Incidental GGN detections were not counted as either true or false positives, as they were out of scope of intended use of the CAD system under evaluation. Conclusion: The CAD system trained with PSN and SN demonstrated high sensitivity with reasonable false positive rate.

B-316

11:15

Incidental major non-cardiac ﬁndings on coronary computed tomography. A meta-analysis N. Flor, C. Mazzocchi, F. Sardanelli, G. Cornalba; Milan/IT (ﬂ[email protected]) Purpose: The presence of incidental signiﬁcant non-cardiac ﬁndings has been reported in patients who underwent a CT study of coronary arteries. We searched for these papers and pooled their results. Methods and Materials: Starting with a search in PubMed for “non-cardiac ﬁndings AND CT coronary OR CT angiography” on September 12, 2007, we found seven papers (published from 1994 to 2006) useful for our aim, three of them reporting the use of EBCT and four of them the use of MDCT. Absolute number and percentages of patients with major non-cardiac ﬁndings (i.e., with immediate impact on patient management) were obtained. 95% conﬁdence intervals were calculated using the binomial exact distribution. Results: On a total of 5,530 patients, 484 major ﬁndings were reported to have major incidental non-cardiac ﬁndings, as follows: pulmonary nodules (n=204), consolidation and inﬁltrates (n=80), thoracic aneurysm (n=63), liver mass (n=46), large pleural effusion (n=24), hilar adenopathy (n=12), tubercoloma (n=9), pulmonary embolism (n=8), large mucous plug (n=6), breast abnormalities (n=5), suspicious esophageal thickening (n=4), interstitial diseases (n=4), thyroid nodules (n=3), bronchial carcinoma (n=2), dissecting aorta (n=2), sarcoid (n=2), sclerotic bone lesions (n=2), mediastinitis (n=2), pleural mass (n=1), large pulmonary mass (n=1), pneumothorax (n=1), extra pleural mass (n=1), large hiatal hernia (n=1), ascites (n=1). Conclusion: Major incidental non-cardiac ﬁndings on CT coronary studies were reported with a mean incidence of 484/5,530 (8.75%; 95% CI, 8-9.5). Radiologists should be always involved in interpreting and reporting cardiac CT.

A

B

C

D

E

F

G

211

Saturday

Purpose: To assess the impact of the radiologists’ individual performance on the magnitude of the incremental beneﬁt of CAD over free search. Methods and Materials: CAD was applied to 20 chest MDCT scans (patient mean age 64 years) with 115 non-calciﬁed r4 mm nodules. Three readers R1/R2/R3 (9, 14, 24 years of experience reading chest CT) independently analyzed the CT scans. They performed a free search for lung nodules followed by interpretation of CAD detections and assigned a conﬁdence score (0-5) to each detection. CAD output was ﬁltered by their previous free search and up to 60 CAD candidates were available for interpretation. Free-response receiver operating characteristic (FROC) curves were created following free search and CAD evaluation. The curves were compared across readers. Results: R1/R2/R3 achieved a sensitivity after free search of 72/67/63% at a false positive (FP) rate/patient of 0.5. Following assessment of CAD, a net increase in sensitivity of 3/13/9% at the same FP rate/patient was observed. R1 with the highest sensitivity during free search was outperformed by R2 after adding CAD; R3 with the lowest sensitivity during free search beneﬁted more than R1. All readers showed a strong rejection of FP CAD candidates (99.3/98.7/99.9%). In contrast, R1/R3 rejected signiﬁcantly more TP CAD candidates (54/52%) than R2 (17%, pb0.05 for R1/R2 and R2/R3 comparison, chi square test). Conclusion: The magnitude of the beneﬁt of CAD used as second reader in the detection of lung nodules on chest CT may not reﬂect the individual performance during free search.

Enterprise, Philips), (b) with greyscale-reversal (GSR) and (c) without additional tools. Presence, location of lesions and reader’s conﬁdence were recorded. Analysis included logistic regression with repeated measures (p 0.05). Results: Both sensitivity and speciﬁcity were higher for expert readers (p 0.05). Sensitivity for unexperienced readers was improved by CAD (56%) compared to GSR and normal reading (both 52%). Expert readers took advantage from GSR (68%) compared to CAD and normal reading (both 63%). Speciﬁcity for expert readers was improved by CAD (65%) compared to GSF (59%) and normal reading (58%). Speciﬁcity for unexperienced readers was highest with normal reading (29%) and decreased by using CAD (22%) or GSR (17%). Conclusion: CAD has the potential to improve the sensitivity of unexperienced readers for the detection of malignant lung lesions but possibly at the expense of a decreased speciﬁcity. Expert readers take more advantage of GSR than of CAD with respect to their sensitivity, but may improve their speciﬁcity by CAD.

Scientific Sessions B-317

11:24

Volume measurement of pulmonary nodules using two types of volumetric nodule-sizing software: A comparison of reproducibility of volumes of purely intraparenchymal, juxtavascular, and pleural-based nodules A.L.M. Leusveld1, P.M. van Ooijen1, D.M. Xu1, Y.R. Zhao1, R.J. van Klaveren2, H. van de Zaag-Loonen1, Y. Wang1, M. Oudkerk1; 1Groningen/NL, 2Rotterdam/NL ([email protected]) Purpose: To evaluate the reproducibility of semi-automatic, computer-assisted volume measurements of three different types of pulmonary nodules: purely intraparenchymal, juxtavascular and pleural-based. Methods and Materials: We randomly selected 150 solid nodules, which were found with low-dose CT screening. Fifty intra-parenchymal nodules, 50 pleuralbased nodules and 50 juxtavascular nodules, all larger than 50 mm³, were included. A beta version of Siemens Lungcare software (Siemens, Forchheim, Germany) was used, providing two different algorithms for volume measurement. Two subsequent measurements were performed for each algorithm. The reproducibility of both algorithms was calculated and compared for the three types of nodules. The Bland-Altman method for assessing measurement agreement was used to calculate the 95% limits of agreement. Results: Measurement was successful in 146 nodules. No difference in volume was seen for 120 (82% of 146 nodules) nodules. Twenty nodules (14%) showed a discrepancy of measurement in one algorithm and six (4%) had a discrepancy in both. The 95% limits of agreement of intra-parenchymal nodules were from -2 to 2% for the ﬁrst algorithm; the second algorithm generated no signiﬁcantly variant values. The 95% limits of pleural-based nodules were from -8 to 11% for the ﬁrst algorithm and -15 to 12% for the second. For juxtavascular nodules the 95% limits of agreement were from -8 to 8% for the ﬁrst algorithm; no signiﬁcantly variant values were generated by the second. Conclusion: Pleural-based nodules show the largest variability in measurements, followed by juxtavascular and purely intra-parenchymal nodules. Different types of software may vary in performance in nodule measurement.

B-318

11:33

Segmentation of solid and non solid lung nodules: A phantom study A. Larici1, L. Menchini1, F. Maggi1, F. Molinari1, A. Matkovich2, L. Bonomo1; 1 Rome/IT, 2Milwaukee, WI/US ([email protected]) Purpose: To evaluate the performance of an automated tridimensional (3D) software in the segmentation of solid (S) and non solid (NS) artiﬁcial lung nodules. Methods and Materials: A chest phantom containing 28 artiﬁcial nodules (14 S, 14 NS) in three different locations (seven juxtapleural, 12 juxtavascular and nine intraparenchymal) underwent three 64-row CT scans at 20, 40 and 100 mAs, respectively. Each dataset was reconstructed with three slice thicknesses (0.625, 1.25, 2.5 mm) and ﬁve different kernels (Standard, Detail, Lung, Bone and Bone Plus; GE healthcare). Using an automatic 3D software (LungVCAR, GE Healthcare), segmentation of each nodule at different slice thickness, dose and reconstruction ﬁlter was performed. Results: A total of 45 reconstructions and 630 segmentations (210 per slice thickness) were performed. Regardless of dose and kernels, all S nodules were segmented at 0.625 and 1.25 mm-slice thickness, except for two juxtavascular at 2.5 mm. Regardless of kernels, the software did not segment six (three juxtavascular; three juxtapleural) NS nodules in 18% of the cases (38/210; 18 at 20 mAs, 12 at 40 mAs, 8 at 100 mAs) at 0.6 mm; in 19% (41/210; 18 at 20 mAs, 13 at 40 mAs, 10 at 100 mAs) at 1.25 mm; in 30% (64/210; 22 at 20 mAs, 20 at 40 mAs, 22 at 100 mAs) at 2.5 mm. Conclusion: Juxtapleural and juxtavascular NS artiﬁcial nodules represent the majority of the non segmented ones. Unsuccessful segmentations decrease using higher dose and lower slice thickness, regardless of kernel.

B-319

11:42

Subtraction imaging for improved detection of change in ground glass nodules in chest computed tomography M. Staring, J.P.W. Pluim, B. de Hoop, S. Klein, M. Prokop; Utrecht/NL ([email protected]) Purpose: To demonstrate that image subtraction improves detection of change in pulmonary ground glass nodules identiﬁed on chest CT. Methods and Materials: We recruited 33 participants with 37 ground glass nodules from a lung cancer screening trial. Each participant had at least one follow-up scan (86 scans total; 2-4 scans per participant). Pairs of scans of the same nodule were presented in random order, and four observers with varying experience in chest CT were asked to rate growth and density change between the two images (increase,

212

A

B

C

D

E

F

G

no change, decrease). The experiment was repeated with a new random sequence, where additionally subtraction images (after data registration) were provided for each pair of nodules. An experienced chest radiologist established a reference standard using all available information. Weighted kappa statistics Kw were used to assess inter-observer agreement and agreement with the reference standard. Results: The reference standard established a regression over time in 5/37 ground glass nodules and no change in 16/37 nodules. In 16/37 nodules the size increased, and in 8/16 nodules density increased as well. When the subtraction image was available, average inter-observer Kw improved from 0.46 to 0.53 for size change and from 0.36 to 0.50 for density change. Average agreement with the standard of reference improved from Kw=0.53 to 0.63 for size change and from 0.48 to 0.57 for density change. Conclusion: Subtraction imaging improves the detection of subtle changes in pulmonary ground glass nodules and decreases intra-observer variability.

B-320

11:51

Impact of patient pose differences on subtle interval change detection by temporal subtraction in chest radiographs. A phantom study J. von Berg1, D. Manke1, C. Schaefer-Prokop2, U. Neitzel1; 1Hamburg/DE, 2 Amsterdam/NL ([email protected]) Purpose: Detection of subtle interval changes in chest radiographs may be improved by temporal subtraction images. For the quality and diagnostic use of the subtraction images an accurate registration of both images is of utmost importance. Methods and Materials: Pa radiographs (Philips Digital Diagnost) of a chest phantom were obtained with and without a 3-cm piece of foam plastic to simulate a perihilar low density interval change. Left and right rotations along the main body axis and A-P inclination were varied systematically. Subtraction images were calculated with and without a novel non-rigid 2D image registration technique. Quality of the subtraction images was evaluated by three radiologists and three non-radiologists who determined the cut-off between visibility of the interval change and degree of distraction through subtraction artefacts. Results: The interval change was hardly detectable in a side-by-side comparison of the two images but clearly contrasted in the subtraction image. The maximum rated angle to maintain detectability was 4.1° on average (A-P alone only 1.7°) without additional compensation software. With image registration software the average angle was 6.4° (A-P alone 4.2°). The difference between both conditions was signiﬁcant for each of the three directions (t-test, N=6). Conclusion: Subtle interval change detection by temporal subtraction is limited by the pose difference (here an angle of 4.1°). Non-rigid image registration is able to effectively maintain detectability for additional 2.3° of rotation.

10:30 - 12:00

Room R2

Cardiac

SS 603 Imaging of the right heart Moderators: M. Gardarsdottir; Reykjavik/IS F. Sardanelli; Milan/IT

B-321

10:30

Tricuspid anular displacement: A new indicator of right ventricular function on MDCT scans D. Delhaye, M. Rémy-Jardin, J.-B. Faivre, V. Deken, A. Duhamel, J. Rémy; Lille/FR ([email protected]) Purpose: To evaluate whether the degree of tricuspid annular displacement, known to correlate closely with right ventricular ejection fraction (RVEF) at echocardiography, is an accurate indicator of RV function at CT. Methods and Materials: Systolic displacement of the tricuspid annulus toward the RV apex, referred to as tricuspid annular plane systolic excursion (TAPSE), was measured in a population of 31 consecutive patients who underwent an ECG-gated MDCT angiogram of the chest with dual-source CT (Deﬁnition, Siemens Medical Systems). The study group included patients with normal (Group 1; n=22) and altered (Group 2; n=9) RVEF secondary to underlying respiratory diseases. TAPSE was measured on 4-chamber views of the cardiac cavities; RVEF was calculated by mean of a dedicated cardiac software (Argus, Siemens). Results: The mean (pSD) values of TAPSE and RVEF were 19.42p3.40 mm and 53.27p7.75%, respectively, in Group 1 and 12.89p3.40 mm and 32.77p9.10%,

Scientific Sessions respectively, in Group 2. A good correlation was found between TAPSE values and RVEF (r=0.65; p 0.0001; Pearson test). ROC curve analysis showed that TAPSE was a highly sensitive and speciﬁc indicator of RV dysfunction (area under the ROC curve: 0.90; p 0.0001; 95% conﬁdence interval: 0.77-1.03). The optimal ROC-derived cutoff was 15 mm to detect altered RVEF (sensitivity: 78%; speciﬁcity: 91%). RVEF can be predicted on the basis of TAPSE measurements by the equation RVEF=(1.79xTAPSE)+15.98. Conclusion: TAPSE measurements provide an accurate and rapid estimation of RV function on ECG-gated MDCT angiograms of the chest.

B-322

10:39

Right ventricular late enhancement in contrast-enhanced cardiac MRI: An important feature in the differential diagnosis of cardiac disease P. Hunold, O. Bruder, T. Schlosser, K. Nassenstein, M. Jochims, J. Barkhausen; Essen/DE ([email protected])

B-323

10:48

Right ventricular delayed enhancement in patients with pulmonary arterial hypertension (PAH) F.A. Calabrese, M. Francone, I. Iacucci, F. Vasselli, D. Vizza, I. Carbone, C. Catalano, R. Passariello; Rome/IT ([email protected]) Purpose: PAH is a progressive disease characterized by raised pulmonary vascular resistance, which results in diminished right-heart function. CMR has emerged as the primary non-invasive diagnostic tool for comprehensive assessment of RV function, quantiﬁcation of myocardial mass and detection of ﬁbrosis. The aim of this study is to identify presence and distribution of DE areas in patients with PAH and to correlate DE with RV mass and afterload. Methods and Materials: Twenty-ﬁve selected patients with clinical diagnosis of PPH were referred to our Department for CMR evaluation from a reference national center. In all cases hemodynamic parameters derived from diagnostic right heart catheterization were available. An MRI study that included short axis and horizontal long axis cine-SSFP imaging and DE acquisitions after contrast-administration (0.1 mmol/kg Gd-BOPTA) was acquired in all cases. After data acquisition, ventricular mass index (VMI), ejection fraction (EF), end-diastolic, end-systolic and stroke volume index were determined. Presence, location and extent of DE were also analyzed and ﬁbrosis was correlated with RV mass and pulmonary arterial pressure (PAP). Results: DE was demonstrated in 18/25 (72%) patients and predominantly conﬁned to RV insertion points (7/18; 38%) and interventricular septum (11/18; 61%); in two cases concomitant LV involvement was found. Extent of ﬁbrosis correlated positively with RV mass (RVm) (r=0.59; p=0.038) and PAP (r=0.64; p 0.01) and inversely with RV EF (r=-0.69; p 0.05). Conclusion: DE was observed in most patients with PPT. Our data suggest that likelihood of DE seems to be correlated with RVm and pulmonary haemodynamics suggesting a potential correlation between ﬁbrosis and afterload.

10:57

Value of regional and global right ventricular function analysis using cardiac magnetic resonance compared to the Task Force criteria for the diagnosis of arrhythmogenic right ventricular dysplasia/cardiomyopathy A.A. Azarine1, X. Lyon2, F. Hidden-Lucet1, A. Redheuil1, P. Ou1, E. Mousseaux1; 1 Paris/FR, 2Lausanne/CH ([email protected]) Purpose: The aim of the study was to evaluate the value of segmental and global functions of the right ventricle (RV) in CMR as a single test or in combination with other Task Force (TF) criteria for ARVD/C. Methods: CMR was in 143 patients suspected of ARVD/C. Segmentation of the RV for wall motion analysis was established. The kinetics of each segment of the RV were independently assessed visually by two experienced operators, using at least two different views, blinded to clinical data and results of the other imaging techniques. End-diastolic/systolic ventricular volumes were manually evaluated on short axis slices. Results: Based on all exams performed, except CMR, 28 patients fulﬁlled criteria for ARVD/C according to the TF. Using CMR as a single test for detecting ARVD/C, the presence of major anomalies alone/associated with minor anomalies was 75%/93% sensitive and 94%/74% speciﬁc. Quantitative RV assessment did not improve diagnosis accuracy. Combining the other TF criteria with segmental RV kinetics criteria improves the sensibility/speciﬁcity of CMR to 89%/95%. Normal RV regional kinetics could exclude ARVD/C in 98% of cases. Using the combination of clinical data with CMR as an alternative technique to other conventional cardiac imaging modalities changed the ﬁnal diagnosis in 11% of patients. Conclusion: The presence of either major or minor anomalies of RV segmental contraction has a good sensitivity and speciﬁcity for ARVD/C in particular when combined with other criteria of the TF, and normal segmental RV motion is a strong predictor for the absence of the disease.

B-325

11:06

Right ventricle kinetic analysis by a radial steady state approach on patients with ARVC suspicion C. Bultrini, S. Scipione, S. Battisti, E. di Cesare, C. Masciocchi, M.V. di Fabio; L’Aquila/IT ([email protected]) Purpose: Aim of our study was to compare conventional kinetic analysis to a radial steady state approach. Methods and Materials: 83 patients with clinical suspicion of ARVC (42 males; 41 females; mean age 39) underwent MRI performed with a 1.5 T GEHD and a 8-channel phased array dedicated coil. A steady state parallel imaging and sequences with 8 mm slice thickness, 1.6 mm planar resolution lasting 7 s were employed. We acquired standard scans (short axis plane through the entire ventricle and long axis 4-chamber, 2-chamber and ROVT plane). Subsequently, we employed a rotational sequence formed by 12 planes with a 10° inclination and centered on a line passing through the center of the left ventricle and the interventricular septum. The resulting slices were perpendicular to the right wall. Right ventricular kinesis was examined by two different observers who assigned a score (0-4) from normal to aneurysmatic to all the predeﬁned right ventricular segments. Discrepancies between the two observers were re-examined by a third observer. Results: Scores obtained by the rotational scan showed signiﬁcant differences with respect to standard examinations, particularly in segments IIC and IB where it shows an increment in diagnostic accuracy of 17% and 30%, respectively, with respect to long axis planes. Conclusion: We propose rotational analysis in suspicion of ARVC instead of both 4- and 2-chamber and ROVT plane as the rotational scan better deﬁning kinetic alterations at the outﬂow tract and at the apical region (behind the Moderator: band) in association with acquisition time reduction.

B-326

11:15

Quantitative assessment of right ventricular function and size in patients with mitral regurgitation using 64-section multi-detector row CT: Comparison with real-time three-dimensional echocardiography and magnetic resonance imaging Y.-K. Guo, Z.-G. Yang, Y. Li, X.-C. Zhang, Y. Wu, T.-M. Zhang; Chengdu/CN ([email protected]) Purpose: To prospectively assess right ventricular (RV) function and size in patients with mitral regurgitation using retrospectively ECG-gated 64-section multi-detector row CT (64-MDCT) and to compare the results with those of real-time three-dimensional echocardiography (RT3DE) and magnetic resonance (MR) imaging. Methods and Materials: Forty-eight consecutive patients with mitral regurgitation were enrolled and underwent retrospectively ECG-gated 64-MDCT, RT3DE and MR imaging for assessing the ventricular function. Right ventricular volumes and ejec-

A

B

C

D

E

F

G

213

Saturday

Purpose: Aim of this study was to review cases of right ventricular (RV) late enhancement (LE) in cardiac MRI and to assess the possible diagnostic beneﬁt of the LE technique. Methods and Materials: In two institutions, 5676 contrast-enhanced cardiac MRI studies were performed on a 1.5 T scanner. After a functional study, LE data sets were acquired 8-15 min after 0.2 mmol/kg gadolinium using a segmented inversion-recovery TurboFLASH sequence (TR=8 ms; TE=4 ms; TI=200-260 ms; slice thickness=8 mm). All cases of non-ischemic right ventricular LE were reviewed, different patterns of LE were related to the underlying pathology as stated by clinical and other diagnostic features. Results: A total of 1905 (34%) patients presented with LE of the left or right ventricular myocardium. In 10 of 1905 cases (0.5%), LE was detected in the RV free wall: In ﬁve patients the underlying pathology was arrhythmogenic RV cardiomyopathy (ARVC/D), LE probably representing ﬁbrous tissue. Myocarditis was found in two patients. Endomyocardial ﬁbrosis was found in two patients. In one patient, RV involvement in acute cardiac sarcoidosis was responsible for LE. In many more cases, RV LE might have been suggested, but was denied because of possible artifacts. The major difﬁculty in detecting RV LE and differentiating it from artifacts corresponds to the thin wall and the heavy trabeculation. Conclusion: This study suggests an additional diagnostic value of LE in a variety of RV myocardial diseases which are, however, rare. Technical improvements are mandatory to increase the diagnostic conﬁdence in RV free wall imaging.

B-324

Scientific Sessions tion fraction were measured with dedicated cardiac analysis software on 64-MDCT and RT3DE, and compared with values measured on MR imaging which served as the reference standard. Dimension ratios for the RV and left ventricle (LV) were also measured and compared. Difference of functional parameters and size among groups was tested with a parametric analysis of variance test. Agreement between modalities was assessed with Bland-Altman analysis and linear regression. Results: No signiﬁcant differences were revealed in calculated RV volumes, EF and RV/LV dimension ratio of three modalities. Both 64-MDCT and RT3DE measurements resulted in high correlation (r 0.91) compared with MR imaging. RV/ LV dimension ratio was signiﬁcantly different among the three groups, according to the degree of mitral regurgitation (p 0.05). The variability in the 64-MDCT measurements was lower than that in MRI and RT3DE values. Conclusion: 64-MDCT could assess accurately the right ventricular function and size in patients with mitral regurgitation, and measurements correlated and agreed with those obtained with RT3DE and MR imaging. RV/LV dimension ratio measured by 64-MDCT correlated well with the degree of mitral regurgitation.

B-327

11:24

Impact of right ventricular contrast enhancement on the accuracy of right heart function analysis at coronary CT angiography H. Lee, M.R. Aho, P. Suranyi, J.M. Kerl, P. Costello, U.J. Schoepf; Charleston, SC/US ([email protected]) Purpose: To investigate the relationship between right ventricular (RV) contrast attenuation and the accuracy of RV functional measurements at coronary CT angiography (cCTA). Methods and Materials: Forty patients without valvular insufﬁciency on echocardiography underwent contrast enhanced 64-slice cCTA. Functional data sets were reconstructed in 10% increments across the R-R cycle. Functional analyses of both ventricles were performed using dedicated software. Stroke volumes were computed for each ventricle. The deviation of the RV stroke volume from the left ventricular stroke volume was calculated. The level of attenuation was recorded at three regions of interest in the RV (below the pulmonary valve, mid-ventricular, and inferior RV). The accuracy of RV functional measurements was correlated with the level of RV contrast attenuation using regression analysis. Results: There was a statistically signiﬁcant correlation between the accuracy of RV functional analyses and the level of contrast attenuation in the inferior RV (R=0.83, p 0.01). In studies with low RV attenuation ( 250 HU, n=20), the mean deviation of RV stroke volumes from left ventricular stroke volumes was 27.0p16.5 ml (33.7p18.8%). At low attenuation levels RV function was typically underestimated. In studies with high RV attenuation ( 250 HU, n=20) the deviation of RV stroke volumes from left ventricular stroke volumes was signiﬁcantly lower 4.3p8.7 ml (12.3p8.4%) (t-test, p 0.01). Conclusion: The accuracy of RV functional analysis at cCTA depends on the level of contrast attenuation in the inferior RV. If RV function is clinically important, there should be at least intermediate contrast attenuation in the RV to enable accurate functional analysis.

B-328

11:33

Intra-individual comparison of 0.5 and 1.0 M contrast agents in MR angiography of chronic thromboembolic pulmonary hypertension N. Woodhouse1, J.M. Wild1, E.J.R. van Beek2, P.D. Grifﬁths1, D. Kiely1; 1 Shefﬁeld/UK, 2Iowa City, IA/US ([email protected]) Purpose: To investigate the role of high-molarity MR contrast agents in chronic thromboembolic pulmonary hypertension (CTEPH). Methods and Materials: Patients referred for MR angiography for investigation of pulmonary hypertension were imaged twice within a 48-hour period using a 1.5 T MR system (Eclipse, Philips, Cleveland, USA). Administration was performed following administration of the same dose of Gd-DTPA (Magnevist, Schering, Berlin, Germany) and Gd-BT-D3OA (Gadovist, Schering, Berlin, Germany). High resolution MRA scans were acquired during 20 seconds of arrested inspiration using a 3D RF-spoiled gradient echo sequence. Bolus timing was determined with a small test bolus. Scan and inject delays were set to ensure that the centre of k-space coincided with the bolus peak. Time-resolved MRA was acquired during a breath hold of 22 seconds with a 3D RF-spoiled gradient echo sequence acquired at a rate of one frame every 2.5 seconds. Results: There was greater separation of pulmonary arterial and aortic peaks in the Gd-BT-DO3 A examinations, with a faster wash-out of contrast resulting in less contamination of images by cardiac motion, pulmonary venous and systemic vasculature. Filling defects and intra-luminal webs were better visualized. Parametric maps of time-resolved examinations revealed better separation of pulmonary arterial and pulmonary venous circulation using Gd-BT-DO3 A.

214

A

B

C

D

E

F

G

Conclusion: The results contradict previous studies, and show that (with adequate bolus timing) there is beneﬁt of the tighter bolus proﬁle and higher molarity of GdBT-DO3 A in pulmonary MR angiography.

B-329

11:42

Noninvasive Qp/Qs ratio measurement with phase-contrast cine MRI in patients with atrial septal defect: Comparison with heart catheterization L. Rizzo1, A. Chiribiri2, M. Marocco2, F. Orzan2, M. Allasia1, F. Solitro1, C. Fava1; 1 Orbassano/IT, 2Turin/IT ([email protected]) Purpose: Blood ﬂow can be quantiﬁed noninvasively by phase-contrast cine MRI (PC-MRI) in adults. Little is known about the accuracy of this measurement in patients with known atrial septal defect. Methods and Materials: In 25 patients with a type-II atrial septal defect, blood ﬂow rate in the great vessels was determined by 1 Tesla PC-MRI, and the ratio of pulmonary to aortic ﬂow (Qp/Qs) was compared with Qp/Qs by oximetry requiring heart catheterization, performed before an attempt of percutaneous closure of the defect. The interval between measurements was 11p18 days (range 1-78 days). Results: Mean Qp was 9.2p3.3 l/min with PC-MRI, and 8.6p4 l/min with catheterization (p=0.57). Mean Qs was 4.8p1.6 l/min with PC-MRI, and 4.5p1.3 l/min with catheterization (p=0.46). Qp/Qs ratio was 1.99p0.72 with PC-MRI, and 1.96p0.96 with catheterization (p=0.92). The correlation analysis showed a good overlap between measurements (Qp: r=0.65, p=0.0004; Qs: r=0.64, p=0.0005; Qp/Qs ratio: r=0.68, p=0.0002), also conﬁrmed by regression analysis (R2 = 0.42, p 0.001 for Qp; R2 = 0.41, p=0.001 for QS; R2=0.46, p 0.001 for Qp/Qs ratio), and by the Bland-Altman statistical analysis for method comparison. The interobserver variability was low. Conclusion: Determination of Qp/Qs by PC-MRI is quick, safe, and reliable compared with oximetry. This noninvasive method could be used to quantify the interatrial shunt in patients with known atrial septal type-II defect, in order to demonstrate hemodynamic importance of the defect, and to reduce the need for time-consuming diagnostic heart catheterization.

B-330

11:51

Detection of myocardial ﬁbrosis in patients with suspicion of arrhythmogenic right ventricular cardiomyopathy in contrast-enhanced magnetic resonance imaging P. Klimeczek, R. Banys, B. Laskowicz, M. Pasowicz; Krakow/PL ([email protected]) Purpose: We hypothesized that both RV and LV in ARVD will demonstrate presence of regions with increased signal in delayed enhancement because of the diffused ﬁbrotic nature of the disease process. Methods and Materials: We examined 27 patients (25M, 2 F) between 11 and 52 (32p13.5) years with clinical, echocardiographic and electrocardiographic ﬁndings of ARVD: CMR examinations were performed using 1.5 T MR device - Magnetom Sonata Maestro Class (Siemens, Germany). LV function, RV ejection fraction evaluation and delayed enhancement (DE) study were performed. Delayed enhancement images were obtained approximately 15 min after a peripheral bolus injection of 0.2 mmol/kg gadolinium-dimethyltriaminepentacetic acid (Gd-DPTA) using a multislice inversion recovery fast gradient echo sequence triggered every heartbeat interval with average of two acquisitions (IR multislice Turbo-Flash). Results: Left ventricular function as assessed by MRI was normal in 18 patients; in nine patients we found moderate dilatation of LV and decreased EF with global mild hipokinesia. In ﬁve patients among 27 we found delayed enhancement in LV (p 0.05) which was located predominantly in interventricular septum (four patients). RV function was mild to moderately impaired in all of 27 patients. Delayed enhancement areas located in RV were detected in 12 patients (44%, p 0.02).The area of delayed enhancement also showed contractility impairment on cine imaging. Conclusion: We found that ﬁbrosis of the RV and LV in patients with suspicion of ARVD/C can be noninvasively visualized using MRI. In advanced RV cardiomyopathy left ventricle may be involved in process of morphologic changes.

Scientific Sessions 10:30 - 12:00

Room R3

Interventional Radiology

SS 609b Liver: Embolisation Moderators: V. Bérczi; Budapest/HU T.J. Popiela; Kraków/PL

B-331

10:30

Long term results of transcatheter chemoembolisation (TACE) with doxorubicin loaded DC Bead in the treatment of HCC not eligible to curative treatments K. Malagari, E. Emmanouil, V. Vergadis, D. Letsou, D. Antonopoulos, E. Alexopoulou, K. Chatzimichail, A. Gouliamos, D. Kelekis; Athens/GR ([email protected])

B-332

10:39

Transarterial chemoembolization in down-staging program for hepatocellular carcinoma prior to liver transplantation: The Bologna workin-progress experience A. Cappelli, A. Bazzocchi, E. Giampalma, R. Golﬁeri; Bologna/IT ([email protected]) Purpose: To assess the efﬁcacy and the safety of transarterial chemoembolization (TACE) on long-term survival and tumor recurrence rate in down-staging patients with hepatocellular carcinoma (HCC) listed for liver transplantation (LT): a new allocation policy still under evaluation. Methods and Materials: We prospectively analyzed with an intention-to-treat principle the outcome of 173 patients listed for LT with a diagnosis of HCC. HCCs were divided into three groups: single nodule 3 cm (T1, 38 patients), single nodule b5 cm or multiple nodules b3 with a diameter b3 cm (T2, 94 patients) meeting the conventional Milan criteria, and the down-stage group (T3, 41 patients): single HCC b6 cm or multiple nodules b6 with a total tumor diameter b12 cm. TACE was performed in a total of 114 patients (65.9%): 15 in T1 (39.5%), 61 in T2 (64.9%), 38 in T3 (92.7%). Results: The rate of LT was equally distributed among the three groups (47.4% T1, 55.3% T2, and 56.1% T3) and after a median follow-up of 21 months, the overall tumor recurrence rate was 12% and the overall patient survival was 83%. The HCC recurrence rate was comparable among the three groups (7.1% T1, 14.3% T2, and 13% T3) and the presence of HCC recurrence did not affect signiﬁcantly the patients’ survival. Conclusion: The pre-operative tumor stage did not affect the patients’ survival and in particular the down-stage group had comparable outcome than the other groups. TACE is effective and safe as a down-staging procedure to extend pre-operative selective criteria for HCC patients waiting for LT.

10:48

Regional 90Yttrium microsphere treatment of chemotherapy-refractory colorectal cancer liver metastases T.F. Jakobs, R.-T. Hoffmann, K. Dehm, C.G. Trumm, H.-J. Stemmler, K. Tatsch, T. Helmberger, M.F. Reiser; Munich/DE ([email protected]) Purpose: To present data of radioembolization treatment with 90Yttrium resinmicrospheres in patients with colorectal cancer liver metastases for who standard of care therapies had failed. Methods and Materials: Forty-one patients underwent radioembolization via intra-arterial infusion in a single session, whole-liver treatment. All patients had received prior polychemotherapy. Imaging and laboratory follow-up is available for 36 patients. After treatment, we assessed tumor response by CT/MRI using RECIST-criteria, laboratory and clincial toxicities, and survival. Results: The mean age of patients was 61 years with a mean activity of 1.9 GBq of 90Yttrium microspheres delivered. Imaging follow-up at median 2.9 months after treatment demonstrated partial response, stable and progressive disease in 17, 61 and 9.8%, respectively. A mean and median decrease of tumor marker levels of 18.8 and 27.9% was noted. Median follow-up time was 7.9 months. Median overall survival was 10.5 months with superior survival for patients with low tumor involvement (11.7 months), tumor marker response (19.1 months) and imaging response (29.3 months). Besides one treatment-associated cholecystitis, no severe toxicities were observed. Conclusion: Radioembolization can be performed with an acceptable toxicity proﬁle in patients with colorectal cancer liver metastases who have failed standard-of-care polychemotherapy. Anti-tumoral effect is supported by decrease in tumor size and tumor marker levels. Further investigation is warranted to prove survival beneﬁt.

B-334

10:57

Development of diaphragmatic weakness after transcatheter arterial chemoemboliztion of right inferior phrenic artery: Its frequency and determinant factors D. Lee, J. Chung, H. Jae, W. Lee, J. Park; Seoul/KR ([email protected]) Purpose: To assess the frequency of diaphragmatic weakness after TACE of RIPA in patients with HCC and to ﬁnd out its determinant factors. Methods and Materials: Sixty patients with previous TACE of RIPA who had undergone follow-up right inferior phrenic artery (RIPA) angiography and ﬂuoroscopic diaphragmatic movement assessment were included in this study. The extent of the embolization (selective embolization of the anterior/posterior branch or both branches), use of gelatin sponge pledgets, and tumor size supplied by RIPA were recorded. On follow-up RIPA angiography, the occlusive changes of RIPA have been assessed. Diaphragmatic weakness was determined by the paradoxical or decreased movement on ﬂuoroscopy. Fisher’s exact test was used to correlate development of diaphragmatic weakness and potential determinant factors. Results: Diaphragmatic weakness after TACE of RIPA was developed in 11 patients (11/60, 18.3%). All patients with diaphragmatic weakness after TACE had undergone TACE of both branches of RIPA. On follow-up RIPA angiography, 27 patients showed occlusive change of both branches of RIPA, and among these 27 patients, 10 patients (10/27, 37%) developed diaphragmatic weakness. TACE of both branches of RIPA (p-value: 0.0046) and occlusive change of both branches of RIPA (p-value: 0.0014) on follow-up angiography signiﬁcantly affected the development of diaphragmatic weakness after TACE of RIPA. The tumor size (p-value: 0.3041) and use of gelatin sponge pledgets (p-value: 0.1546) were not signiﬁcant determinant factors. Conclusion: When TACE of RIPA is nonselectively performed with permanent occlusive changes, diaphragmatic weakness could develop as an ischemic complication.

B-335

11:06

Addition of pirarubicin and amiodarone in transarterial chemoembolization of unresectable hepatocellular carcinoma: A pilot study B. Guiu1, R. Loffroy1, S. Guiu1, C. Colin2, C. Deville1, F. Bonnetain1, J.-L. Jouve1, L. Bedenne1, B. Chauffert1, J.-P. Cercueil1, D. Krausé1; 1Dijon/FR, 2Lyon/FR ([email protected]) Purpose: To assess overall survival and prognostic factors after transarterial chemoembolisation (TACE) for unresectable hepatocellular carcinoma (HCC) with the combined use of pirarubicin, gelfoam, lipiodol and amiodarone. Methods and Materials: In June 2000-November 2005, 43 patients (mean age: 65 y, 47-86 y; 40/43 cirrhotic patients) underwent TACE for unresectable HCC (23/43 multifocal) with pirarubicin, gelfoam, lipiodol and amiodarone. Median tumour size was 8 cm (1-20 cm). CLIP-score=0 in 7/40, 1 in 16/40. The regimen of treatment

A

B

C

D

E

F

G

215

Saturday

Purpose: To examine the results of segmental transcatheter arterial chemoembolization (TACE) with doxorubicin-loaded DC Bead in the treatment of hepatocellular carcinoma (HCC) in non-surgical candidates. Methods and Materials: Seventy one patients (60 men; 11 women; mean age 63; range, 46-71 years) with documented HCC of 3-10 cm in diameter (mean 6.2) were enrolled prospectively in the study. All patients had Child A or B cirrhosisrelated HCC. Results: Complete response according to EASL on an intention-to-treat basis was seen in 11 patients who developed complete necrosis (15.5%). Objective response ranged from 66.2% to 85.5% across the four treatments. Survival at 12 months was 97.05%. Sustained complete response was observed in 11 (16.1%), and objective response in 49 (72%). Sustained partial response was seen in 49 patients (72.05%). Survival at 18 months was 94.1%. At 24 months survival was 91.1%. Sustained objective response was seen in 45 patients (66.2%) while sustained complete response was 16.1% (11/68). At 30 months survival was 88.2%. AFP levels decreased signiﬁcantly in measurements one month post each procedure (p 0.001). Bilirubin, G-GT, aspartate aminotransferase (AST), alanine aminotransferase (ALT), and alkaline phosphatase (ALP) showed only transient increases observed during the study period. Severe procedure-related complications were seen in 4.2% (cholecystitis: n=1; liver abscess: n=1; pleural effusion n=1). Post Embolization Syndrome (PES) was observed in all patients. Conclusion: DC Bead is an effective and safe procedure in the treatment of HCC patients with high rates of response and high rates of long term survival.

B-333

Scientific Sessions was one session of TACE repeated every two months depending on tolerance and efﬁcacy. Before treatment and one month after each session, CT scans were performed to assess response (RECIST) and intensity of lipiodol uptake. Overall survival was evaluated according to Kaplan-Meier. Univariate analysis was performed and multivariate Cox models were generated by combining two variables: CLIP-score and lipiodol capture. Results: Mean number of TACE sessions was 3.5 (1-11). Four deaths were related to treatment (two severe sepsis, two digestive hemorrhages). Overall survival rates were 86% (71-93%), 68% (51-80%), 55% (38-70%), 27% (10-48%) at 6, 12, 24, 48 months, respectively. Survival time was signiﬁcantly reduced for patients with a CLIP-score of 2 (p 0.003) and for a lipiodol uptake 25% (p 0.001). In multivariate analysis, these factors were again statistically signiﬁcant. Conclusion: In this series, survival rates are high compared to previously reported studies and despite the large size of tumours. Pirarubicin is more liposoluble than doxorubicin which increases its cellular penetration. Moreover, amiodarone decreases the resistance of tumoral cells to anthracyclins and increases stability of lipiodol-anthracyclin emulsion. We are evaluating this combination in a prospective, randomized and controlled trial.

B-336

11:15

Post chemoembolization treatment of liver metastases from neuroendocrine tumors, survival rate and local control following Nobel protocol treatment using mitomycin C and gemcitabin as therapeutic agents N.A.N. Mohammed, N.N.N. Naguib, K. Eichler, T. Lehnert, T.J. Vogl; Frankfurt a. Main/DE ([email protected]) Purpose: To assess the effectiveness of trans-hepatic arterial chemoembolization using a combined protocol of mitomycin C and gemcitabin on the local tumor control, symptom alleviation and survival rate in patients with liver metastases from neuroendocrine tumors (NET). Methods and Materials: 44 patients (age range 37-77 years) with liver metastases from NET underwent repetitive selective trans-hepatic artery chemoembolization using mitomycin C and gemcitabin chemotherapeutics, in the time period between November 1999 and March 2007, with 4 weeks interval between embolization sessions. Monthly follow-up by CT and MRI scanning and measurement of tumor markers for the evaluation of local tumor control. Results: 244 chemoembolizations were performed with a mean of 5.5 sessions per patient. 25 patients presented multiple metastases, 13 had one metastasis, one had two metastases, and ﬁve had 3-4 metastases. The local chemotherapy protocol consisted of only mitomycin C (n=17), mitomycin C with gemcitabin (n=27). Local tumor control evaluation according to RECIST criteria were as follows: a partial response 13.64%, stable disease 63.64% and progressive disease 22.72%. The 1-year survival rate was 88%, the 2-year survival rate was 74%. The median survival time from the date of diagnosis was 88 months (according to KaplanMaier), after ﬁrst chemoembolization 40.7 months, for the palliative group was 36.9 months, for the symptomatic group was 16 months and for the neoadjuvant group was 46.6 months. Conclusion: Transarterial hepatic chemoembolization using mitomycin C and gemcitabin is an effective tool for local metastases control, symptom relief and survival rate improvement in hepatic metastases from neuroendocrine tumors.

B-337

11:24

Regional transarterial chemoperfusion with combined gemcitabine and mitomycin C in unresectable liver metastases: Tumor response and survival T.J. Vogl1, S. Zangos1, U. Jacob2, R. Bauer1; 1Frankfurt a. Main/DE, 2Bad Heilbrunn/DE ([email protected]) Purpose: To evaluate regional transarterial chemoperfusion of unresectable liver metastases concerning tumor response and survival. Methods and Materials: Between 2002 and 2006, 56 patients were treated with a mean of ﬁve sessions per patient in 4-week intervals. The chemotherapy consisted of combined mitomycin (8.5 mg/m²) and gemcitabine (1,000 mg/m²), administered into the common hepatic artery within one hour by a perfusor. After the session tumor size was measured by MRI and classiﬁed according to the RECIST criteria. Primary tumors were colorectal carcinoma (n=12), breast carcinoma (n=12), cholangiocellular carcinoma (n=10), and others (n=22). Eight patients presented with one solitary lesion, four with two lesions, 17 with 3-5 lesions and 27 with more than ﬁve lesions in the liver. Results: Treatment was well tolerated without any relevant complications and sideeffects. Three patients (5.4%) responded with complete response (CR), nine (16.1%) with partial response (PR), 27 (48.2%) with stable disease (SD) and 17 (30.4%) with progressive disease (PD). Patients with metastases of breast cancer had the

216

A

B

C

D

E

F

G

most favorable outcome with CR (n=3), PR (n=3), SD (n=6), and PD (n=0). Median survival after the ﬁrst local chemoperfusion was 11 months for all patients, 9.33 months for colorectal carcinoma and 12 months for breast carcinoma. Responders (CR+PR) had a signiﬁcantly longer survival time than nonresponders (PD): 13.1 vs 6.8889 months and p=0.00089. Conclusion: Transarterial chemoperfusion is a minimal-invasive outpatient treatment for unresectable liver metastases. The control of tumor size is an important aspect in symptomatic treatment of tumor-associated complaints.

B-338

11:33

Prospective study on the use of transarterial chemoembolization (TACE) in liver metastases of colorectal carcinoma T.J. Vogl1, T. Gruber1, S. Zangos1, J.O. Balzer2; 1Frankfurt a. Main/DE, 2Mainz/DE ([email protected]) Purpose: To evaluate the efﬁcacy of chemoembolization (TACE) in the treatment of liver metastases in colorectal cancer patients regarding local control and survival. Methods and Materials: 207 patients (mean: 68.8 years; range: 39.4-83.5 years) with liver metastases of colorectal cancer were treated with repeated TACE in 4-week intervals. 1,307 chemoembolizations were performed (mean: 6.3 sessions/ patient). 158 patients were treated palliatively, 35 symptomatically and 14 neoadjuvantly. The chemotherapy consisted of Mitomycin C with/without Gemcitabine; embolization was performed with Lipiodol and starch microspheres for vessel occlusion. Tumor response was evaluated by MRI. The change in size was calculated and the response evaluated according to the RECIST criteria. Survival rates from the ﬁrst diagnosis and the ﬁrst TACE session were both calculated according to the Kaplan-Meier method to obtain the median survival. Results: 70% of patients showed multiple metastases, 6% had one metastasis, 5.8% had two metastases and 18.2% had 3-4 metastases. Lesion size and number before, during and after treatment were assessed to deduce the morphological response. Local control results according to the RECIST criteria were: partial response 12% of patients, stable disease 51% and progressive disease 37%. Oneyear survival after TACE was 62%; 2-year survival 38%. Median survival from the date of diagnosis of metastases was 3.4 years (according to Kaplan-Meier), from the start of TACE 1.34 years (palliative group: 1.4 years; symptomatic group: 0.8 years; and neoadjuvant group: 1.5 years). Conclusion: TACE is an effective minimal-invasive therapy for neoadjuvant, symptomatic or palliative treatment of liver metastases in colorectal cancer patients.

B-339

11:42

Transarterial chemoembolization in hepatocellular carcinoma: Role of diffusion weighted imaging and subtracted imaging in necrosis assessment L. Mannelli1, S. Kim1, C. Hajdu1, M. Losada1, T. Clarke1, S. Colagrande2, B. Taouli1; 1New York, NY/US, 2Florence/IT ([email protected]) Purpose: To compare diffusion-weighted MRI (DWI) to contrast-enhanced (CE) MRI with image subtraction for assessment of hepatocellular carcinoma (HCC) response to transarterial chemoembolization (TACE), using liver explant as the reference. Methods and Materials: 18 patients (16 men, 2 women, mean age 58 years) with cirrhosis and HCC who underwent TACE and MRI within three months of liver transplant were evaluated. MRI included pre-contrast T1, T2, dynamic post-contrast 3D GRE T1 (early-arterial, arterial, portal venous and equilibrium phases) and breath-hold single-shot EPI DWI (using b-values of 0-50-500 s/mm²). Two experienced observers measured apparent diffusion coefﬁcient (ADC) values in HCCs and surrounding liver and evaluated % of HCC necrosis on subtracted dynamic post-contrast images. An experienced hepatopathologist evaluated the % of tumor necrosis on explants. ADC and % necrosis from image subtraction were correlated to the % of necrosis from pathology. Results: 20 HCCs (mean size 2.6p0.9 cm, range 1-5.1 cm) were evaluated. There were signiﬁcant differences between ADCs of viable tumor vs necrotic portions (1.37p0.38 vs 1.87p0.56x10-3 mm2/sec, p 0.01). There was a moderate correlation between ADC and % of necrosis at pathology (r=0.64, p 0.003), and strong correlation between % of necrosis at portal phase subtraction and % of necrosis at pathology (r=0.88, p 0.0001). Conclusion: ADC can be used to discriminate between viable and necrotic tumor portions; however, it was outperformed by CE-MRI with image subtraction for the diagnosis of HCC necrosis after TACE.

Scientific Sessions B-340

11:51

Pneumatically driven robotic system for MR-guided interventions: First clinical experiences in Vienna J. Kettenbach, K. Kubin, A. Stadler, J. Lammer; Vienna/AT ([email protected])

14:00 - 15:30

Room A

Musculoskeletal

SS 710 Arthritis Moderators: W.C.G. Peh; Singapore/SG P. Peloschek; Vienna/AT

B-341

14:00

The impact of age, sex, body mass index and handedness on ﬁnger joint space widths measured by computer-aided joint space analysis (CAJSA) in patients with rheumatoid arthritis (RA) A. Pfeil, J. Böttcher, M.L. Schäfer, A. Hansch, H.-J. Mentzel, A. Petrovitch, W.A. Kaiser; Jena/DE ([email protected]) Purpose: To evaluate the associations between sex, age, body mass index (BMI) and handedness regarding the radiogeometric detectable joint space distances (JSD) of the metacarpal-phalangeal (MCP), proximal-interphalangeal (PIP) and distal-interphalangeal articulations (DIP) of the hand in patients suffering from RA. The joint space widths were measured by a new available program named CAJSA. Methods and Material: 128 patients (94 women, 34 men) with RA underwent computerized semi-automated joint space analysis of joint space widths at the JSD-MCP I-V, JSD-PIP II-V and JSD-DIP II-V based on hand radiographs (Radiogrammetry Kit, Version 1.3.6; Sectra; Sweden). The JSD (total) of each articulation was expressed in mm. The patient cohort was differentiated for gender, age, handedness and BMI (BMI 20; BMI 20-25, BMI 25). Results: JSD revealed a signiﬁcant age-related narrowing of 24.8% (MCP), 22.6% (PIP) and 28.7% (DIP) between the ages of 20 and 79. Additionally, males showed a signiﬁcantly wider JSD compared to the female cohort for all age groups (MCP: 10.4%; PIP: 11.7%; DIP: 16.0%). All JSD presented a wider JSD of the left hand (MCP: 3.0%; PIP: 2.1%; DIP: 2.4%) compared to the right side. The JSD-MCP (5.2%) demonstrated signiﬁcant differences regarding the BMI groups. Conclusion: Age and gender seems to affect the width of JSD for peripheral ﬁnger joints as measured by computer-aided joint space analysis. In contrast to JSD-MCP, an effect of the body mass index on measurements of JSD-PIP and JSD-DIP could not be observed. These inﬂuences must be differentiated from disease-related alterations caused by RA.

14:09

Phase-contrast radiography in a clinical setting, using high-resolution digital imaging system with a small-focus X-ray tube instead of synchrotron radiation: II. Quantitative evaluation of bone trabeculae area on early-stage rheumatoid arthritis J. Tanaka1, T. Mimura1, H. Oda1, C. Honda2, M. Tazaki2, Y. Shinden2, T. Unezawa1, N. Funakoshi1, H. Kawasaki1, Y. Wada1; 1Iruma-gun/JP, 2Hachioji/JP ([email protected]) Purpose: To evaluate the clinical usefulness of phase-contrast radiography (PCR) in a clinical setting for the detection of early-stage rheumatoid arthritis (RA). Methods and Materials: The system produces an edge-enhancement effect and therefore should improve the sharpness of X-ray images. The system consists of a tungsten-anode X-ray tube with a 123 μ focal spot and a computed radiography (CR) plate (sampling pitch 43.75 μm): the distances between the focal spot and object, and between the object and CR plate are 0.65 and 0.49 m, respectively, and therefore images are magniﬁed. The distance between the object and CR plate is required to capture the edge-enhancement effect of narrow width, due to a phase-shift phenomenon (refraction). For this clinical trial, PCR and conventional CR (CCR) were performed on the hands of 13 patients with stage-1 and -2 RA, and in 14 healthy subjects. X-ray image-signal values were collected directly from the CR unit to measure the area of bone trabeculae in a ﬁxed area in the hand. Results: The average area of bone trabeculae was 27.3% for patients and 33.0% for healthy subjects (p 0.05). These values were not signiﬁcantly different from those obtained by CCR. Conclusion: PCR images of bones are sharp enough to quantitatively discriminate the subtle decrease in bone trabeculae in RA patients from those in healthy subjects, which was not possible by CCR. This new technique may be beneﬁcial in the diagnosis of early-stage RA.

B-343

14:18

3 T-MRI compared to multidetector-row CT in the evaluation of erosions of both hands in patients with suspected rheumatoid arthritis C.R. Krestan, S. Nemec, T. Kapral, S. Mehrain, C. Czerny; Vienna/AT ([email protected]) Purpose: The purpose of our study was to compare 3 T-MRI with multidetector-row computed tomography (MDCT) for the detection of bone erosions of both hands in patients with early rheumatoid arthritis using MDCT as the gold standard. Methods and Materials: Thirteen patients were referred from the department of rheumatology (10 female, three male; average age: 51 years) for the detection of bone erosions. MDCT studies were performed on a Philips Brilliance 64 MDCT (Philips Medical Systems, Best, The Netherlands) with the following parameters 140 kV, 75 mAs, matrix 768x768; collimation 20x0.625 mm. Axial and coronal MPRs (1.5 mm) were calculated using a bone algorithm (3000/600 HU). Dedicated MRI exams were performed on a 3.0 T MRI unit (Siemens, Magnetom Trio Tim syngo) using a spine coil combined with a body coil. Coronal PD-TSE with fat suppression, coronal T1-TSE, coronal T1-TSE with fat suppression after iv. injection of 0.2 ml/ kg Gd and axial T2-TSE were measured. MDCT and MRI reconstructions were analyzed independently by two musculoskeletal radiologists blinded to patient data in a consensus reading using the EULAR-OMERACT-Ramris criteria. Results: The total number of erosions detected with MDCT was 76. Thirty-two of these lesions were diagnosed with MRI (42% sensitivity). In 483 locations MRI and MDCT yielded negative results, in 13 regions MRI was positive whereas MDCT was negative (97% speciﬁcity). Conclusion: MRI has a low sensitivity in the detection of bone erosions compared to MDCT in patients with rheumatoid arthritis and cannot be recommended for the diagnosis of osseous involvement of the joints of the hand.

B-344

14:27

Intra-articular distribution pattern after ultrasound guided injections in wrist joints of patients with rheumatoid arthritis M. Boesen1, K.E. Jensen2, S. Torp-Pedersen1, M.A. Cimmino3, B. Danneskiold-Samsøe1, H. Bliddal1; 1Frederiksberg, Copenhagen/DK, 2 Copenhagen/DK, 3Genoa/IT ([email protected]) Purpose: To investigate the distribution of an ultrasound-guided intra-articular (IA) injection in the wrist joint of patients with rheumatoid arthritis (RA). Methods: An ultrasound-guided IA drug injection into the wrist joint was performed in 17 patients with 1 ml methylprednisolone (40 mg/ml), 0.5 ml Lidocaine® (5 mg/ml) and 0.15 ml Gadolinium (Omniscan 0.5 mmol/ml). The drug solution was placed in the the wrist between the distal radius and the lunate bone. Coronal and axial MRI sequences were performed after the injection to visualize the distribution. Carpal

A

B

C

D

E

F

G

217

Saturday

Purpose: To evaluate the feasibility of robot-assisted MR-guided percutaneous interventions within the chest, abdomen and retroperitoneum. Methods and Materials: In 10 patients either a biopsy (n=8), a pleural drainage (n=1) or an ethanol instillation (n=1) was performed using a pneumatically driven robotic system (Innomotion, Innomedic GmbH, Herxheim, Germany) and a closedbore 1.5 T-MR scanner (Avanto, Siemens, Erlangen, Germany) for image-guidance. The robotic system had ﬁve degrees of freedom and two manual adjustments for prepositioning. Planning of the skin entry and the target point was performed on a standard PC, receiving DICOM images from the MR scanner via Ethernet link. Registration of the robot to the MR scanner was performed using four spherical Gadolinium-doped markers mounted at the Application Module (AMO) and a 3-D Gradient Echo sequence (FA 50, TR 13, TE 6). For intervention, puncture needles (diameter 15-18G, length 111-150 mm) with a premounted needle holder were attached to the AMO and inserted manually by the physician according to the pre-calculated insertion depth. MR imaging conﬁrmed the appropriate position. For biopsy, 18G-biopsy gun systems were used. Results: All MR-guided interventions were successful without clinically relevant complications. The median overall procedure time was 69 (range 43-140) sec and the median insertion length of the puncture needles was 6.8 (range 4.5-9.1) cm. All biopsy samples were conclusive; both treated tumors showed complete necrosis and the pleural empyema was evacuated successfully. Conclusion: Robot-assisted MR-guided percutaneous interventions were feasible and provide promising features for complex interventions.

B-342

Scientific Sessions distribution (radio-carpal, inter-carpal, and carpometacarpal) and radio-ulnar distribution were recorded. Full distribution in one compartment was given the value 1, partial distribution 0.5 and no distribution 0. A sum of the total distribution for all four compartments was calculated and correlated with the clinical parameters and the MRI OMERACT scores. Results: No uniform pattern was seen in the distribution of the contrast. Only two patients had full contrast distribution to all four compartments, and the mean distribution count for all patients was 2.4 (range=0.5-4). The distribution count correlated with the MRI OMERACT synovitis score (r=0.60, p=0.014), but not with the erosions, bonemarrow oedema scores or any clinical parameters. Conclusion: The MRI distribution of contrast showed patient-speciﬁc and random patterns after IA injections in RA wrist joints. The degree of distribution increased with the MRI-synovitis score, while no association was found with the erosion- and bonemarrow oedema scores. This indicates that a single injection into a standard injection site in the wrist cannot be assumed to distribute - and treat - the whole joint.

B-345

14:36

Value of contrast enhanced grey-scale ultrasound: Assessment of joint vascularity in therapeutic follow-up of rheumatoid arthritis T. De Zordo, D. Junker, P. Soegner, G. Feuchtner, J. Gruber, A. Klauser; Innsbruck/AT ([email protected]) Purpose: To assess the value of contrast enhanced grey-scale ultrasound (CEUS) in detection of vascularity in joints of patients with rheumatoid arthritis (RA) before and after intra-articular corticosteroid injection. Methods and Materials: We assessed 35 clinical active joints in 35 consecutive patients (14 men, 21women; mean age 45p12 years) with clinical diagnosis of RA. CEUS, using a low mechanical index US technique, was performed by a bolus administration of a contrast agent (SonoVue®, Bracco, Milan, Italy). Quantitative estimation of joint vascularity was performed by calculation of time-intensity curves. Follow-up examinations were carried out three weeks after treatment. Results: High peak enhancement was found before therapy in 35/35 joints, whereas after therapy a decreased enhancement with low plateau enhancement characteristics could be seen in 9/35 joints and no enhancement at all in 26/35 joints. Enhancement of calculated time-intensity curves differed signiﬁcantly before and after therapy follow-up (p 0.001). Conclusion: CEUS is sensitive to the presence of joint vascularity decrease in therapeutic follow-up of RA joints, and therefore useful for determination of vascularity at the microvascular level. This technique might have further potential for sensitive monitoring of therapeutic response in patients undergoing systemic antiangiogenetic therapy.

B-346

14:45

Quantiﬁcation of perfusion and endothelial permeability in inﬂammatory joint diseases: First results S. Weckbach, S. Sourbron, M. Notohamiprodjo, J.G. Raya, A. Horng, M.F. Reiser, C. Glaser; Munich/DE ([email protected]) Purpose: Two-compartment modelling of bolus-tracking MRI at high temporal resolution may produce quantitative measures of both perfusion and permeability. The purpose of this study is to evaluate this approach in inﬂammatory joint diseases. Methods and Material: Dynamic contrast enhanced MRI of the hands of ﬁve patients with rheumatoid arthritis were acquired at 3 T (TWIST, 30 partitions of 1.5x0.5x0.5 mm3, Tacq per data set 2 sec) over 5 min. Arterial input functions were measured in the radial artery and four ROIs were set in muscle and inﬂammatory tissue. Pixel- and ROI enhancement curves were ﬁtted to a two-compartment model producing plasma ﬂow and plasma volume and extraction ﬂow. Results were compared to the plasma volume and Ktrans parameter produced by the conventional Tofts model. Results: The two-compartment model accurately ﬁts the data of both muscle and inﬂammation with parameters in physiological range. All three parameters are substantially higher in inﬂammatory tissue than in muscle. Pixel analysis revealed inﬂammatory ROIs with different characteristics (high perfusion and low permeability, or vice versa). Ktrans agrees very well with extraction ﬂow, but the plasma volume produced by the Tofts model is either over- or underestimated. Conclusion: A two-compartment model applied to dynamic MR data with high temporal resolution allows for separate quantiﬁcation of perfusion and permeability in inﬂammatory joint disease. Compared to the existing approach, the additional perfusion parameter and possible improvement in accuracy might improve sensitivity and speciﬁcity in future therapy control studies.

218

A

B

C

D

E

F

G

B-347

14:54

MR characterisation of synovial ﬂuid ex vivo: Is it possible to differentiate between septic and non-septic arthritis? E. Wiener, M. Zanetti, J. Hodler, C.W.A. Pﬁrrmann; Zurich/CH ([email protected]) Purpose: To assess lactate concentration and total protein content in synovial ﬂuid samples using proton magnetic resonance spectroscopy (HMRS) and T2 mapping to differentiate between septic and non-septic arthritis. Methods and Materials: Samples (n=30) of joint aspirations in patients with acute arthritis were analyzed with single voxel HMRS to assess lactate concentration and with multi spin echo measurements to assess T2 relaxation properties. All samples underwent microbiological testing and routine laboratory analysis to quantify lactate concentration and total protein content. Findings between culture positive septic arthritis and non-septic arthritis were compared using the Mann-Whitney U test. The spectroscopy data and T2 relaxation time measurements were correlated with laboratory conﬁrmed data. Results: Specimens from patients with culture positive septic arthritis (n=10) demonstrated elevated concentrations of synovial ﬂuid lactate 11.4p4.0 mmol/L (p 0.001) and total protein content 51.8p10.7 g/L (p 0.01) in comparison to samples from non-septic arthritis (n=20) with 5.2p1.1 mmol/L and 40.4p6.9 g/L. The apparent lactate concentration and the T2 relaxation time showed a moderate correlation to the laboratory conﬁrmed lactate concentration (r2=0.71) and total protein content (r2=0.73). Markedly raised apparent lactate concentrations ( 6 mmol/L) in combination with low T2 values ( 550 ms) of synovial ﬂuids are sensitive MR parameters to identify septic arthritis. Conclusion: Measurements of apparent lactate concentration with HMRS in combination with estimation of protein content using T2 quantiﬁcation may be of value to differentiate septic arthritis from non-septic arthritis.

B-348

15:03

Arthritic changes in patients with ankylosing spondylitis: Scoring at whole-body thin-section low-dose MDCT. Initial experience M.C. Freund, F. Hartig, B. Glodny, E. Mur, K.M. Unsinn, W.R. Jaschke; Innsbruck/AT Purpose: To assess arthritic changes with MDCT by using radiography-like volumerendered images generated from identical MDCT-acquired projection data set as reference standard. To estimate the effective dose from recorded dose-length product of MDCT. Methods and Materials: All patients had ankylosing spondylitis (AS) according to modiﬁed New York criteria. Whole-body low-dose 16-detector row spiral CT (MDCT) was performed in 32 patients (22-66 years) in 2 clusters (neck, chest/abdomen) without contrast material, 0.625 mm slice thickness, 50-75 mAs. Dose-length product (DLP) was recorded. Two sets of reconstructions were performed: a) axial and multiplanar reformated images (Group 1) and b) volume-rendered images of spine, sternum as surrogate imaging modality for computed radiographs (Group 2). Two radiologists arrived at consensus opinion regarding score of arthritic changes (0-4) analogous to Bath Anyklosing Spondylitis Radiology Index scoring system. Results: Following results were obtained (score - number of each joint/patient) (Group 1): intervertebral symphysis [disk] [C2-S1] (3.42-23), manubriosternal (2.8-1) and pubic symphysis (1.4-1); facet (1.9-46), costovertebral (1.6-24), temporomandibular (0.8-2), sternoclavicular (1.6-2), hip (1.3-2) joint(s). The score (meanpSD) of spinal changes between Group 1 (3.42p1.3) and Group 2 (2.32p1.22) was signiﬁcantly different (p 0.05). Conversion of recorded DLP 421pM'YsCM resulted in effective dose of 6.7p1.7 (mSv) for whole-body MDCT. Conclusion: Whole-body low-dose MDCT examination facilitates a comprehensive evaluation of various joints of axial and proximal appendicular skeleton in patients with AS.

B-349

15:12

Diagnostic imaging of gout: Comparison of high resolution US versus conventional X-ray T. Rettenbacher, S. Ennemoser, H. Ulmer, F. Hartig, M. Herold, W.R. Jaschke; Innsbruck/AT ([email protected]) Purpose: The aim was to compare X-ray as the established imaging method and ultrasound (US) in diagnosing gout. Methods and Materials: In a prospective study, 105 consecutive patients with clinical suspicion of gout underwent conventional X-ray und high resolution US in order to help in arriving at a deﬁnite diagnosis. X-ray ﬁndings suggestive of gout included soft tissue opaciﬁcations with densities between soft tissue and bone, articular and periarticular bone erosions, and osteophytes at the margins of

Scientific Sessions opaciﬁcations or erosions. US ﬁndings suggestive of gout included bright stippled foci and hyperechoic soft tissue areas. All images from each method applied were independently reviewed by two experienced radiologists and thereafter a consensus opinion on each imaging sign was obtained. Results: Fifty-ﬁve patients had a deﬁnite diagnosis of gout (102 involved sites), 31 patients were diagnosed as having another disease (59 involved sites), and 19 patients were excluded from the study since a deﬁnite diagnosis could not be established. X-ray suggested gout with a sensitivity of 31% (32/102), speciﬁcity of 93% (55/59), PPV of 89% (32/36), NPV of 44% (55/125), and an accuracy of 63% (102/161) whereas US suggested gout with a sensitivity of 96% (98/102), speciﬁcity of 73% (43/59), PPV of 86% (98/114), NPV of 91% (43/47), and an accuracy of 88% (141/161). US was much more sensitive than conventional X-ray but less speciﬁc. Conclusion: Our data show that US often provided additional diagnostic information in patients with clinical suspicion of gout when laboratory ﬁndings and X-ray results were negative or inconclusive and should therefore be used in these cases.

B-350

15:21

Purpose: To compare signal characteristics of the synovia in knees of asymptomatic volunteers before and after intravenous administration of ultrasmall superparamagnetic iron oxide (USPIO). Methods and Materials: Ten knees of 10 asymptomatic volunteers with a mean age of 34 years (range 22-50) were examined with a 1.5 T MR-system prior to and 36 hours after intravenous administration of USPIO. The magnetic resonance (MR) imaging protocol included T1-weighted spin-echo, T2-weighted fast spin-echo, T2*weighted gradient-echo and short inversion time inversion-recovery sequences. Images were analyzed qualitatively and quantitatively. Signal intensities (SI) of synovia and musculature were measured on pre- and post-contrast images. In addition MR studies were reviewed for presence of cartilage defects, osteophytes, subchondral cysts and joint effusion as signs of minor arthrosis, and for meniscal lesions. The Wilcoxon signed rank test was used to compare relative SI before and after USPIO administration (p 0.05). Results: Qualitative image analysis revealed no changes in signal characteristics. The relative SI of the synovia in T1, T2 and T2* after USPIO administration (mean: T1 0.7, T2 1.38, T2* 0.23) showed no signiﬁcant changes compared with relative SI before injection (mean: T1 0.75, T2 1.44, T2* 0.25) (p=0.52, 0.86, 0.60). None of the volunteers showed moderate or severe arthrosis, one showed minimal arthrosis with cartilage defects involving less than half of the cartilage thickness. Conclusion: In asymptomatic volunteers no changes in SI of the synovia can be depicted after USPIO administration. This means that USPIO-enhanced MR-imaging may be used for assessment of knee disorders with increased macrophage activity.

14:00 - 15:30

Room B

Abdominal Viscera (Solid Organs)

SS 701a Magic bullets for hepatic lesions Moderators: B.I. Choi; Seoul/KR L. Curvo-Semedo; Coimbra/PT

B-351

14:00

Detection of liver metastases from neuroendocrine tumors: [68Ga]DOTATOC-PET/CT and/or liver MRI? C. Pfannenberg, P. Aschoff, M.P. Lichy, H. Schlemmer, M. Reimold, C.D. Claussen; Tübingen/DE ([email protected]) Purpose: To compare the diagnostic value of [68Ga]-DOTATOC-PET/CT and MRI in the detection of liver metastases in patients with neuroendocrine tumors (NET). Methods and Materials: Thirty-eight consecutive patients with proven NET underwent a [68Ga]-DOTATOC-PET/CT and liver MRI within a 1-30 day interval. PET/CT consisted of a multi-phase CT followed by a whole-body PET with the somatostatin receptor ligand [68Ga]-DOTATOC. The MRI comprised T2w TSE fs and T1w multiphase contrast-enhanced 3D GRE sequences. Three readers analyzed images of each modality, blindly and independently for lesion number, size and enhancement.

B-352

14:09

Health economic evaluation of three imaging strategies in patients with suspected colorectal liver metastases: Gd-EOB-DTPA (Primovist)enhanced MRI vs. extracellular contrast media-enhanced MRI and 3-phase MDCT in Germany, Italy and Sweden C.J. Zech1, L. Grazioli2, E. Jonas3, M. Ekman3, R. Niebecker4, S. Gschwend4, J. Breuer4, L. Jönsson3, S. Kienbaum4; 1Munich/DE, 2Brescia/IT, 3Stockholm/SE, 4 Berlin/DE ([email protected]) Purpose: To perform an economic evaluation of hepatocyte-speciﬁc Gd-EOBDTPA (Primovist, Bayer Schering Pharma)-enhanced MRI (PV-MRI) compared to extracellular contrast media-enhanced MRI (ECCM-MRI) and 3-phase MDCT (MDCT) as initial diagnostic modalities in the work-up of patients with known or suspected colorectal liver metastases. Methods and Materials: The economic evaluation was performed with a decisiontree model designed to estimate all aggregated costs for diagnostic imaging and modiﬁed surgical procedures depending on the initial diagnostic modality. Probabilities on the need for further imaging and subsequent surgical planning were collected through interviews with 13 pairs of a radiologist and liver surgeon each in Germany, Italy, and Sweden. Results: The results indicate that PV-MRI leads to cost savings by improving preoperative planning and decreasing intra-operative treatment changes. The rate of further imaging needed was 8.6% after initial PV-MRI, 18.5% after ECCM-MRI, and 23.5% after MDCT leading to lower additional cost for diagnostic work-up when PV-MRI was used for initial imaging. Including both diagnosis and intra-operative treatment changes, a strategy starting with PV-MRI is cost-saving compared to ECCM-MRI and only moderately more costly than MDCT for all three countries. In Sweden, the cost with PV-MRI was 785_ vs 843_ for ECCM-MRI and 766_ for MDCT. For Germany and Italy, the results were similar. Conclusion: Even if the cost of initial imaging is higher for PV-MRI than for MDCT and ECCM-MRI, lower costs for additional imaging and changes in pre-operative planning lead to lower total cost for PV-MRI compared to ECCM-MRI and moderately higher cost compared to MDCT.

B-353

14:18

Differentiation of well-differentiated hepatocellular carcinomas from other hepatocellular nodules in cirrhotic liver: Value of SPIO-enhanced MR imaging at 3.0 T H. Park, J. Lee, S. Kim, J. Han, B. Choi; Seoul/KR ([email protected]) Purpose: To determine the diagnostic value of superparamagnetic iron oxide (SPIO)-enhanced MRI for the differentiation of well-differentiated hepatocellular carcinomas (WD-HCC) from dysplastic nodules (DN) and moderately or poorly differentiated HCCs (MD-/PD-HCC). Methods and Materials: A total of 114 patients with 216 histologically conﬁrmed hepatocellular nodules, i.e. 23 DNs, 37 WD-HCCs, and 156 MD-/PD-HCC, who underwent SPIO-enhanced MRI at 3.0-T, were included in this study. MRI included T2-weighted fast-spin echo sequence and T1- and T2*-weighted gradient echo (GRE) sequences before and after administration of ferucarbotran. Two reviewers analyzed MR images in terms of the relative signal intensity (SI) of each nodule as also their enhancement features on SPIO-enhanced T2- and T2*-weighted images. The enhancement features were categorized into six groups: low SI, iso SI, high SI foci on low- or iso-SI background, subtle high SI, strong homogenous high SI, and strong heterogeneous high SI. Results: The most prevalent enhancement features of WD-HCCs on SPIOenhanced T2-weighted images and on T2*-weighted images were low- or iso-SI (46.0%) and high SI foci on low- or iso-SI background (32.4%) or subtle high SI (24.3%), respectively. Alternatively, 22 of 23 DNs (95.7%) showed complete low- or iso SI, and 145 of 156 (94.9%) MD-/PD-HCCs showed strong high SI. When high

A

B

C

D

E

F

G

219

Saturday

USPIO-enhanced magnetic resonance imaging of the knee in asymptomatic volunteers C.S. Reiner1, A.M. Lutz1, F. Tschirch1, J. Froehlich1, B. Bonnemain2, B. Marincek1, D. Weishaupt1; 1Zurich/CH, 2Roissy/FR

The reference standard was given by the consensus reading of all three methods, follow-up and histology, respectively. Results: In 27 of the 38 patients a total of 240 hepatic metastases were detected. The number of metastases detected on MRI, CT and PET was 239, 200 and 166, respectively. In 44% of patients MRI detected more liver metastases than PET and in 33% of patients more lesions than CT. In one patient with two known metastases only PET revealed an additional lesion. PET failed to detect liver metastases in ﬁve patients (18.5%) in whom MRI or CT was positive. The lesion detectability in PET was inﬂuenced by lesion size with a limit of 8-10 mm. Conclusion: MRI depicted the largest number of hepatic metastases and appears to be the method of choice for liver staging before surgery or ablation in patients with NET. [68Ga]-DOTATOC-PET has a high speciﬁcity, but the sensitivity is limited by receptor status and lesion size.

Scientific Sessions SI foci on low- or iso-SI background or subtle high SI on T2*-weighted images was applied, we could identify 56.7% (21/37) of the WD-HCCs but only 4.4% (1/23) of the DNs and 3.2% (5/156) of the MD-/PD-HCCs. Conclusion: WD-HCCs have characteristic enhancement features that differentiate them from DNs and MD-/PD-HCCs on SPIO-enhanced MRI.

B-354

14:27

Comparison of two fat suppression techniques for MRI regarding image quality and conspicuity of liver lesions T.C. Lauenstein1, P. Sharma1, T. Hughes2, K. Heberlein2, D.R. Martin1; 1Atlanta, GA/US, 2 Erlangen/DE ([email protected]) Purpose: To compare two techniques for fat saturation in T2w MRI regarding image quality and conspicuity of liver lesions. Methods and Materials: 20 patients with hepatic hemangiomas or metastases were studied utilizing a 1.5 T magnet (Magnetom Avanto, Siemens). Two different T2w sequences of the abdomen were acquired: (a) an inversion recovery (IR) single shot fast spin echo (SSFSE) sequence and (b) a spectral pre-saturation attenuated inversion-recovery (SPAIR) SSFSE sequence. Parameters for both sequences included TE/TI=96/150 ms, inﬁnite effective TR, acquisition time=13 s. MR images were analyzed regarding (1) degree of fat saturation in the retroperitoneum and mesenteries by measuring SNR; (2) conspicuity of liver lesions by determining CNR of the lesions; (3) delineation of the bowel wall using a threepoint scale (1=poor, 3=good delineation). Results of each feature were compared using a Student’s t-test. Results: T2w SPAIR images were signiﬁcantly superior for all features compared to T2w IR data (p 0.05). Mean SNR of fat tissue in the retroperitoneum and mesenteries for SPAIR amounted to 18.3 and 11.8, respectively (IR: 47.0 and 30.9). Liver lesions were displayed on SPAIR images with a higher CNR (hemangiomas: 166 vs 124 for IR; metastases 76 vs 52 for IR). Delineation of bowel wall margins was better on SPAIR compared to IR images (average ranking 2.8 vs 1.9). Conclusion: T2w SPAIR sequences provide excellent fat suppression, bowel wall delineation and depiction of liver lesions. This technique may help to improve diagnostic accuracy for the assessment of inﬂammatory or tumor disease throughout the abdomen.

B-355

14:36

Gadobenate dimeglumine (Gd-BOPTA) enhanced MR imaging of small (b20-mm) hepatic nodules seen on arterial phase: Usefulness of the hepatobiliary phase A. Guerrisi, D. Marin, C. Catalano, M. Di Martino, M. Baski, G. De Filippis, R. Passariello; Rome/IT ([email protected]) Purpose: To retrospectively assess whether gadobenate dimeglumine (Gd-BOPTA) enhanced MR imaging during the hepatobiliary phase can improve the characterization of small (b20 mm in diameter) hepatic arterial phase-enhancing (HAPE) lesions that are occult during portal and equilibrium phases and at unenhanced T1-and T2-weighted MR images. Methods and Materials: This study was approved by the institutional review board and a waiver for informed consent was obtained. Two-hundred and ﬁfty patients who underwent Gd-BOPTA enhanced MR imaging were evaluated with breath-hold T2-weighted images and volumetric three-dimensional Gd-BOPTA enhanced T1-weighted GRE MR images acquired in the arterial (25 s), portal venous (60 s),equilibrium (180 s),and hepatobiliary phases (180 min). Two readers retrospectively reviewed the MR images in consensus for small HAPE-only nodules. The ﬁnal study group included 40 patients (27 men and 11 women) aged 23-82 years (median age, 64.2 years) with a total of 55 HAPE-only lesions. Qualitative analysis of MR enhancement features during the hepatobiliary phase was related to pathology reports or imaging follow-up (at least 1 year). Sensitivity, speciﬁcity, and positive and negative predictive values with corresponding 95% conﬁdence intervals (CIs) were determined. Results: Of the 55 HAPE-only lesions, 25 (45%) were hepatocellular carcinomas (HCCs).The remaining 30 (55%) lesions were considered deﬁnite pseudolesions. The sensitivity and speciﬁcity of Gd-BOPTA MR imaging during the hepatobiliary phase for depicting HCCs were 52% (13 of 25) and 90% (27 of 30), respectively. The positive and negative predictive values of Gd-BOPTA MR imaging were 81% (13 of 16) and 70% (27 of 39), respectively. Conclusion: Gd-BOPTA-enhanced MR imaging during the hepatobiliary phase has high speciﬁcity, but low sensitivity for the characterization of small HAPE-only lesions in patients with cirrhosis.

220

A

B

C

D

E

F

G

B-356

14:45

Tumor-liver contrast and subjective tumor conspicuity of respiratorytriggered T2-weighted fast spin-echo sequence compared with T2*weighted gradient recalled-echo sequence for ferucarbotran-enhanced magnetic resonance imaging of hepatic malignant tumors K. Ishiyama, M. Hashimoto, T. Sato, J. Watarai; Akita City/JP ([email protected]) Purpose: This retrospective study proposes to compare respiratory-triggered T2-weighted fast spin-echo (RTT2W-FSE) and T2*-weighted gradient recalledecho (T2*W-GRE) images for visualization of malignant hepatic tumors using ferucarbotran-enhanced magnetic resonance (MR) imaging. Methods and Materials: Ferucarbotran-enhanced RTT2W-FSE and breath-hold long-TE 2D-fast spoiled gradient recalled acquisition in the steady state (FSPGR) images as T2*W-GRE were used to evaluate 128 malignant hepatic tumors (77 metastases, 37 hepatocellular carcinomas, 14 other) in the 62 patients. Tumor-toliver contrast (TLC) was quantitatively compared using the paired Student’s t test, and the score of lesion conspicuity was qualitatively compared using Wilcoxon’s signed rank test. Results: The mean TLC of RTT2W-FSE was signiﬁcantly higher than that of FSPGR (2.54p1.42 vs 1.10p0.82) in all malignant tumors. The mean TLC was signiﬁcantly greater for RTT2W-FSE than for long-TE FSPGR regardless of the scan order. A signiﬁcant difference was noted between the group in which RTT2W-FSE was performed prior to FSPGR (34 lesions in 17 patients, 1.06p0.71 vs 2.02p0.90, p 0.001) and the group in which FSPGR was performed prior to RTT2W-FSE (93 lesions in 45 patients, 1.12p0.86 vs 2.70p1.59, p 0.001). The score of lesion conspicuity of RTT2W-FSE was signiﬁcantly higher than that of FSPGR (4.84p0.52 vs 4.52p0.99) in all malignant tumors. Conclusion: For ferucarbotran-enhanced MR imaging, compared to FSPGR images, RTT2W-FSE images provide greater TLC and subjective conspicuity for malignant tumors.

B-357

14:54

A prospective multicenter trial of contrast-enhanced ultrasound for the differential diagnosis of hepatic tumors in Germany: Results on 1349 liver lesions D. Strobel1, K.-P. Seitz2, W. Blank3, A. Schuler4, C. Dietrich5, A. von Herbay6, C. Greis7, H. Strunk8, T. Bernatik1; 1Erlangen/DE, 2Sigmaringen/DE, 3Reutlingen/DE, 4 Geislingen/DE, 5Bad Merengtheim/DE, 6Tübingen/DE, 7Konstanz/DE, 8Bonn/DE Purpose: To evaluate the diagnostic accuracy of contrast-enhanced ultrasound in the differential diagnosis of liver tumors in clinical practice. Methods and Materials: From May 2004 to December 2006, 1349 patients with a hepatic tumor lacking a deﬁnitive diagnosis based on conventional B-scan ultrasound and color Doppler sonography were examined by contrast-enhanced ultrasound using a standardized protocol. Tumor classiﬁcation was assessed due on vascularity pattern and contrast enhancement pattern seen in focal lesions during the arterial, portal and late phase up to ﬁve minutes based on a standardized protocol (pulse/phase inversion imaging, MI 0.4). The diagnosis established after contrast-enhanced ultrasound was compared to histology (75% cases) or to CT or NMR. Results: The ﬁnal diagnosis of liver tumors included 573 benign hepatic tumors (including hemangiomas n=242, focal nodular hyperplasia n=170, adenoma n=19, other benign lesions n=142) and 755 malignant hepatic tumors (including metastases n=383, hepatocellular carcinoma n=279, other malignant lesions n=93). Contrast-enhanced sonography could correctly assess 723/755 malignant lesions (sensitivity 95.8%) and 476/573 benign lesions (speciﬁty 83.1%), whereas in n=100/1349 tumors (7.4%) a deﬁnitive differentiation of benign/malignant tumors was not possible based on contrast-enhanced sonography. The positive predictive value of contrast-enhanced sonography for the diagnosis of a malignant tumor was 88.2% (versus B-scan 57.0%, power-Doppler 56.2%) and the negative predictive value of contrast-enhanced sonography was 93.7% (versus B-scan 43.5%, powerDoppler 42.3%). Conclusion: Contrast-enhanced sonography clearly improves the differential diagnosis of hepatic tumors and is very helpful when conventional B-scan-morphological criteria are missed.

Scientific Sessions B-358

15:03

Hepatic transit time analysis using contrast enhanced ultrasound: Evaluation of quantitative analysis and “visual” blinded reading in patients with liver metastases and healthy volunteers J. Hohmann1, C. Müller2, A. Oldenburg2, J. Skrok2, B. Frericks2, K.-J. Wolf2, T. Albrecht2; 1Basle/CH, 2Berlin/DE ([email protected])

B-359

15:12

Value of contrast-enhanced ultrasound (CEUS) in cancer patients with subcentimetric focal liver lesions (FLLs) detected at multislice CT (MSCT) F. Sandomenico, O. Catalano, A. Nunziata, E. de Lutio di Castelguidone, A. Siani; Naples/IT ([email protected]) Purpose: MSCT carries a high resolution, with a great number of small FLLs found. Nevertheless, these lesions cannot always be characterised with certainty. We prospectively investigated the additional role of CEUS in cancer patients with small, indeterminate liver lesions seen at MSCT. Methods and Materials: In a 1-year period all patients with extrahepatic tumours showing small ( 10 mm), indeterminate (not conﬁdently diagnosed) FLLs at whole-body or at abdominal CT scanning were enrolled. This group underwent immediate unenhanced US to rule out cystic lesions. Whenever US failed to show a cyst, low-mechanical index, SonoVue-enhanced CEUS was carried out. During the arterial phase, CEUS was focused on the area reﬂecting the CT ﬁnding, while the entire organ was explored during portal-sinusoidal phase. Results: There were 132 patients with MSCT evidence of undetermined, subcentrimetric FLLs (206 lesions, 1-8/patient, mean 2). US proved the cystic nature of 138 lesions in 96 patients. In 45 cases US failed to recognise any abnormality or detected a noncystic image and these subjects underwent CEUS. CEUS results were conﬁrmed at further investigation or follow-up for 43/45 patients (three cysts, eight haemangiomas, 47 metastases, two areas of focal steatosis, two eosinophilic necroses, one granuloma, one abscess, one ﬁstula). In 5/43 cases CEUS ﬁndings changed patients’ management. CEUS failed to detect the three lesions in 2/45 patients, one of which subsequently proved metastasis and two benign. Conclusion: In selected cases, CEUS can be useful in demonstrating or excluding metastasis in cancer patients with MSCT evidence of subcentimetric, indeterminate FLLs.

B-360

14:00 - 15:30

Room C

GI Tract

SS 701b

Saturday

Purpose: Evaluation of hepatic transit time analysis with contrast enhanced ultrasound of patients with known liver metastases and healthy volunteers using a quantitative analysis and a blinded “visual approach”. Methods and Materials: 22 patients with colorectal cancer and liver metastases and 22 healthy volunteers were compared using contrast enhanced ultrasound on a Siemens Sonoline Elegra and SonoVue (Bracco, Milano, Italy). Native baseline and post contrast images were acquired over a time period of 50 s. Arrival times for the major vessels were determined and hepatic transit times were calculated using a quantitative PC-based method and a “visual approach” with two blinded readers. Statistical analysis was done for arrival times and transit times between the two major groups and some subgroups (e.g. involved liver segments). Results: Arrival times did not show signiﬁcant differences due to the metastatic disease. Hepatic transit times were shorter in the patient group (p 0.01) for the arteriovenous and the portovenous transit times. Quantitative analysis gave slightly better results than blinded reading. Sensitivities for the detection of metastatic disease were 77-95%, speciﬁcities 68-82%, positive predictive value 73-83% and negative predictive value 78-94%. Blinded reading agreement was better for the patients. Additional signiﬁcant results were found for the number of involved liver segments. No signiﬁcant age dependency was found in the control group. Conclusion: Hepatic transit time analysis with CEUS is a tool, which has the ability to detect changes in hepatic blood supply due to metastatic liver disease for both quantitative analysis and “visual” blinded reading.

for diagnosis. Quantiﬁcation of costs of resources directly related to the procedure was based on average costs provided by our budget unit. Patients were 398, with 213 hemangiomas, 41 focal nodular hyperplasias (FNH), and 154 pseudolesions (focal fatty spares, focal fatty areas). We performed a CEUS examination for each patient and 98 of them underwent a CECT. All the lesions were followed-up. Results: The cost for a single CEUS exam was _97.02. The cost for a single CECT exam was _202.14. To diagnose hemangiomas we saved _1344.89, _5080.62, _9862.58, and _9115.13 per year. To diagnose FNHs we saved _747.16, _747.16, _1344.89 per year. To diagnose pseudolesions we saved _2689.79, _4931.29, _4931.29, _4034.69 per year. Over the years 2002-2005, by the introduction of CEUS, we saved _44979.36 in diagnosing BFLL. Conclusion: This cost analysis study shows that CEUS is more cost-effective than CECT as second-level method after US for the diagnosis of benign focal liver lesions.

Colon and rectum Moderators: G. Maskell; Truro/UK S. Schmidt; Lausanne/CH

B-361

14:00

Is algorithm designed to pair corresponding regions in prone and supine helpful for interpretation in CT colonography? M. Cadi, I. Chartier-Huynh, D. Chen, P. Grenier, O. Lucidarme; Paris/FR ([email protected]) Purpose: To evaluate an algorithm designed to automatically pair polyps between supine and prone positions. Methods and Materials: This algorithm (Viatronix® V3D colon) was evaluated with 10 CTC, including 50 polyps r5 mm conﬁrmed by optical colonoscopy. First, two independent readers (R1 without and R2 with CTC experience) labelled polyps on both supine and prone axial images and paired manually each polyp seen on supine images with polyps seen on prone images. Finally, the algorithm performed an automatic pairing of the labelled polyps. Discrepancies between manual and automatic procedures were noted and times for both procedures were recorded. Results: The 50 polyps were manually paired in an average of 42 s/polyp for R1 and 26 s/polyp for R2. When polyps were located in the rectum, transverse colon and caecum, the mean pairing time was 15 for R1 and 8 for R2 while for those located in the sigmoid, hepatic and splenic ﬂexures, the mean pairing time was 89 for R1 and 50 for R2. The algorithm was able to automatically pair correctly (without manual adjustment) 15 (30%) polyps and with manual adjustment 31 (62%) polyps. The mean time of “algorithm assisted” pairing was 24 s/polyp for R1 and 11 s/ polyp for R2 (polyps located in the rectum, transverse colon and the caecum were instantaneously paired, while mean times for those located in the sigmoid, hepatic and splenic ﬂexures were 15 s/polyp for R1 and 8 s/polyp for R2). Conclusion: The paired region algorithm decreased the interpretation time of CTC for both experienced and non-experienced radiologists.

B-362

14:09

Two different doses of iodinated fecal tagging agent for CT colonography: Evaluation of tagging quality, homogeneity, patient acceptance and diagnostic accuracy M.H. Liedenbaum, C.I.B.F. Gouw, A.H. de Vries, A.F. van Rijn, S. Bipat, P. Fockens, E. Dekker, J. Stoker; Amsterdam/NL ([email protected])

15:21

Contrast enhanced ultrasonography in benign focal liver lesion characterization: Activity-based cost analysis N. Faccioli, M. D’Onofrio, A. Comai, R. Pozzi Mucelli; Verona/IT ([email protected]) Purpose: To perform a cost analysis of contrast-enhanced ultrasonography (CEUS) in the characterization of benign focal liver lesions (BFLL) with indeterminate appearance at ultrasonography (US) in comparison with contrast-enhanced computed tomography (CECT), using ﬁnal diagnosis (CECT or biopsy) as reference standard. Methods and Materials: A decision model of patients with suspicious BFLL at US who afterwards underwent a CEUS examination from 2002 to 2005 was constructed. We analysed cost-effectiveness of CEUS considering if CECT was necessary or not

Purpose: Comparison of two doses of iodinated water-soluble contrast agent as bowel preparation for computed tomography colonography (CTC) in a Fecal Occult Blood Test (FOBT)-positive screening population. Quality of fecal tagging, lesion detection and patient acceptance were evaluated. Methods and Materials: 99 FOBT-positive, consecutive patients underwent CTC and colonoscopy with segmental unblinding. 50 patients (group 1) ingested 7*50 ml of iodinated contrast agent (meglumine ioxithalamate) and 49 patients (group 2) ingested 4*50 ml of this contrast agent, all combined with a low-ﬁber diet. Measurements of residual stool attenuation and homogeneity (standard deviation of attenuation) were performed per colonic segment. Furthermore the amount and tagging quality of residual stool were evaluated. Two reviewers performed lesion (polyp and carcinoma) detection. Per patient sensitivity and speciﬁcity were deter-

A

B

C

D

E

F

G

221

Scientific Sessions mined. Patient acceptance was assessed with questionnaires. Results: Tagging density was 637 and 615 HU and homogeneity 92 and 86 HU for groups 1 and 2, respectively (no signiﬁcant difference). The amount of residual stool and the tagging quality were not signiﬁcantly different. Sensitivity per patient, identifying lesions r10 mm was 90% in group 1 and 96% for both reviewers in group 2. Speciﬁcity per patient, identifying patients without lesions r10 mm, was 93% for reviewer 1 and 100% for reviewer 2 in group 1 and 100% in group 2 for both reviewers. Patient acceptability concerning burden of diarrhea was signiﬁcantly improved for patients in group 2. Conclusion: Preparation with 4*50 mL compared with 7*50 mL meglumine ioxithalamate results in an improved patient acceptability and has a comparable, excellent image quality and diagnostic performance.

B-363

14:18

Quantitative quality assessment of different faecal tagging (FT) CT colonography (CTC) preparations D. Campanella1, L. Morra2, S. Delsanto2, T.M. Gallo1, V. Tartaglia1, D. Regge1; 1 Candiolo/IT, 2Turin/IT ([email protected]) Purpose: To evaluate the quality of three iodinated FT regimens, differing in administration timing and laxative use. Methods and Materials: Three groups of 35 patients underwent CTC on a 16-slice scanner after observing low residue diet and the following regimens: GroupA: three sachets/die polyethyleneglycol ‘3350’ (Movicol, Norgine Limited) two days preceding CTC; 50 ml Gastrograﬁn (Schering) diluted in one liter of water, two hours before examination; GroupB: 2 ml/kg Gastrograﬁn two days preceding CTC; GroupC: 90 ml Fosfosoda-ﬂeet (Bergamon) and 100 ml Gastrograﬁn one day before CTC. Tagged pools were assigned to colonic segments and mean intensity and standard deviation (SD) assessed. The performance of a CAD scheme designed for liquid preparation was evaluated as a quality index. Patient tolerance was evaluated on a 10-point scale. Mann-Whitney U-test was applied to evaluate differences between groups. Results: Mean pool intensity was above 200 HU in more than 90% of the cases for each segment, with the exception of the rectum and sigmoid in GroupA (84%). SD was signiﬁcantly lower in GroupA with respect to GroupB and GroupC (p 0.001). CAD speciﬁcity was higher for GroupC compared to GroupA and GroupB (p=0.02 and p=0.001). Patients in GroupA and GroupB tolerated preparation better than in GroupC (p=0.002 and p=0.01). Conclusion: GroupC yields best results, but is less tolerated. Sameday-fecal tagging, safer because administered under medical supervision, is well tolerated, but may present insufﬁcient tagging in the rectum/sigmoid.

B-364

14:27

Fecal-tagging CT colonography with structure-analysis electronic cleansing for detection of colorectal ﬂat lesions Y. Xu1, W. Cai1, J. Nappi1, M. Matoba2, H. Yoshida1; 1Boston, MA/US, 2Kanagawa/JP ([email protected]) Purpose: To evaluate the feasibility and sensitivity of the 3D-reading of fecaltagging CT colonography (CTC) with a novel structure-analysis electronic cleansing (SAEC) in detecting colorectal ﬂat lesions in comparison with a prospective 3D reading with Viatronix V3D Colon system (V3D) and 2D reading without electronic cleansing (2D). Methods and Materials: Forty CTC cases with ﬂat lesions were retrospectively observed. The subjects from a multicenter clinical trial underwent cathartic bowel preparation with orally administrated barium-based fecal tagging. Sixty-nine ﬂat lesions were conﬁrmed using colonoscopy and histopathology as a reference standard. The results from SAEC reading were compared with those of prospective V3D and 2D readings. Results: Overall detection sensitivity with SAEC was 51% (35/69), which was statistically higher than that of 32% (22/69) and 29% (20/69) with V3D and 2D readings, respectively (p 0.05). The sensitivities in detecting not-on-fold ﬂat lesions were 61% (23/38), 45% (17/38), and 42% (16/38) with SAEC, V3D, and 2D readings, respectively, whereas those of on-fold ﬂat lesions were 39% (12/31), 16% (5/31), and 13% (4/31), respectively (p 0.05). None of the eight ﬂat lesions (2-9 mm) at cecum was detected by any of the three reading methods. Excluding the ﬂat lesions at cecum, the sensitivity with SAEC for detecting ﬂat lesion r4 mm increased to 84% (31/37). Conclusion: The fecal-tagging CTC with structure-analysis electronic cleansing could yield a high sensitivity for detecting ﬂat lesions r4 mm. The not-on-fold ﬂat lesions were detected with higher sensitivity than on-fold ﬂat lesions. Cecum may be a blind segment in the detection of ﬂat lesions.

222

A

B

C

D

E

F

G

B-365

14:36

Pilot study of dynamic PET/CT in the study of rectal cancer E.M. Anderson, D. Morofke, N.P. Hughes, M. Brady, N.M. Mortensen, D. Kerr, K. Bradley, F. Gleeson; Oxford/UK ([email protected]) Purpose: Dynamic PET/CT is a novel method of quantifying tumour metabolic activity. It is a potential means of assessing the biological aggressiveness and chemoradiotherapy sensitivity of tumours. It has not previously been performed in patients with rectal cancer. We are assessing the feasibility of obtaining dynamic PET/CT data in such patients. Methods and Materials: Five of twenty patients have been prospectively recruited and informed consent obtained. All patients had been selected to undergo downsizing chemoradiotherapy before surgery. Dynamic PET/CT was obtained on a GE PET/CT Discovery LS scanner after injection of 380-400 MBcq F-18 ﬂuorodeoxyglucose. The primary tumour was identiﬁed on the CT scan. From the DICOM data of the imaging frames, a time activity curve for the tumour and adjacent soft tissue (gluteal muscles) was analysed. Results: The bladder and vasculature had characteristic rapid decay and onset curves while the soft tissue and tumour did not. Using a linear regression model, the tumour and soft tissue were shown to have mean uptake values ranging from 9.4 to 316.5 Bq/mL/s and 0.6 to 1.4 Bq/mL/s, respectively. Conclusion: Dynamic PET/CT data can be successfully obtained from patients with rectal cancer and may show a difference between tumour and normal tissue. This may allow the prediction of both the response to downsizing chemoradiotherapy and tumour aggressiveness, such as the development of distant metastatic disease.

B-366

14:45

The feasibility study of MSCT perfusion evaluating intratumoral microvessel density of colorectal cancer X. Xu, Y. Liu, K. Chen; Shanghai/CN ([email protected]) Purpose: To evaluate the feasibility of MSCT perfusion evaluating intratumoral microvessel density (MVD) of colorectal cancer. Methods and Materials: Perfusion CT was performed with sixteen-slice MSCT in 40 patients with colorectal cancer (23 men, 17 women; mean age 63.4 years; age range 30-83years). Before operation, contrast medium-enhanced dynamic CT was performed with a static table position for 65 seconds, and the tissue blood ﬂow (BF), blood volume, mean transit time (MTT), and vascular permeability-surface were analyzed. MVD was evaluated. The lesions were classiﬁed to the microvessels mainly with or without cavity. Then the correlations between MVD and perfusion parameters were analyzed. Moreover, according to Dukes’ stages, patients were grouped and compared. Statistical analysis was performed with the correlation analysis, grouped t test, one-way ANOVA and Kruskal-Wallis test. Results: For the tumor, the meanpsd was 61.77p35.06; 77.07p41.81 mL/100 g tissue/min; 5.79p1.72 mL/100 g tissue; 8.33p4.45 seconds; 15.30p4.28 mL/100 g tissue/min for MVD, BF, BV, MTT, and PS, respectively. There is no correlation between MVD and perfusion parameters (p 0.05). There are no signiﬁcant differences among different lesion categories in perfusion (p 0.05). There were signiﬁcant differences in MVD among different stages of colorectal cancer (p 0.05). However, there were no signiﬁcant correlations between MDV and stages of colorectal cancer (p 0.05). There were no signiﬁcant differences among different stages in perfusion (p 0.05). Conclusion: There is no signiﬁcant correlation between MVD and MSCT perfusion in colorectal cancer. It isn’t feasible to use MSCT perfusion evaluating MVD of colorectal cancer.

B-367

14:54

Comparison of accuracy of CT and MRI in predicting circumferential resection margin with histologic correlation in rectal cancer after preoperative chemoradiotherapy S. Lee1, Y. Kim2, K. Lee2; 1Seoul/KR, 2Seongnam-si/KR ([email protected]) Purpose: To compare accuracies of MR and CT in predicting circumferential resection margin (CRM) of rectal cancer after preoperative chemoradiotherapy by using histology as the reference standard. Methods and Materials: 57 patients with rectal cancer, who had been treated with preoperative chemoradiotherapy and operation, were prospectively enrolled in this study. Contrast-enhanced CT and MRI were obtained before and after chemoradiotherapy. Agreement of post-chemoradiotherapy CRM on CT, MR determined by two independent radiologists and histology was assessed with Bland-Altman plot. Patients were divided into three groups by CRM (group 1: b2 mm, group 2: 2 and 10 mm, group 3: r10 mm). Accuracies of groups on CT and MR were assessed by McNemar's test.

Scientific Sessions Results: In 53 patients excluding total regression cases, the 95% limits and mean difference of agreement of CRM on CT and histology, on MR and histology, and on CT and MR were, respectively, -5.5 to 7.7 mm (1.1 mm), -3.7 to 6.2 mm (1.2 mm), -3.2 to 3.4 mm (0.1 mm) in observer1 and -6.1 to 10.4 mm (2.1 mm), -4.1 to 7.3 mm (1.6 mm), -5.7 to 4.6 mm (-0.6 mm) in observer2. Accuracies of groups on CT and MR were, respectively, 61.4, 70.2% in observer1 and 56.1, 68.4% in observer2. MR was signiﬁcantly more accurate than CT in the pooled data of the two observers (p=0.01). Conclusion: The accuracy of MR in predicting groups was signiﬁcantly higher than that of CT. CRM on MR was more correlated with histology than on CT. But CRM between MR and CT there was a little difference.

B-368

15:03

Transrectal ultrasonography and magnetic resonance imaging in the N-staging of rectal cancer S.R. Rafaelsen, T. Sørensen, A. Jakobsen, C. Bisgaard, J. Lindebjerg; Vejle/DK ([email protected])

B-369

15:12

Spectrum of MR-defecography ﬁndings in patients with anismus A.E. Solopova, C.S. Reiner, R. Tutuian, B. Marincek, D. Weishaupt; Zurich/CH ([email protected]) Purpose: To determine the frequency of MR-imaging ﬁndings in patients with anismus and to compare these ﬁndings to a control group with constipation. Methods and Materials: Forty eight patients (34 females, 14 males; mean age, 48 years) with clinically suspected anismus referred for anorectal manometry and MR-defecography between January 2000 and March 2007 were included. According to the ﬁnal diagnosis, based on the Rome II criteria, patients were divided into two groups: patients with anismus (n=18) and constipated patients without anismus (control group, n=30). MR-images were retrospectively reviewed by two independent radiologists with regard to the time of evacuation, number of attempts to evacuate, changes in the anorectal angle on straining, paradoxical sphincter contraction, and the presence of additional pelvic ﬂoor abnormalities. Interobserver variability was analyzed by calculating kappa statistics. Results: In patients with anismus, MR-defecography revealed evacuation inability in nine patients (50%), increased time of evacuation compared to the patients with constipation (191 vs 113 sec, p 0.001), increased number of attempts to evacuate (mean 6.1 vs 3.5, p 0.05), higher frequency of abnormal anorectal angle changes on straining (9 (50%) vs 1 (3.3%)). Paradoxical sphincter contraction was observed solely in 15 (83.3%) patients with anismus. Interobserver agreement was good-excellent (kappa=0.68-1). Conclusion: In using MR-defecography for the detection of impaired evacuation (including the evacuation inability), increased number of attempts to evacuate and the presence of paradoxical sphincter contraction are highly suggestive of anismus.

15:21

MR-defecography: Prospective comparison between two different rectal enema compositions A.E. Solopova1, F.H. Hetzer2, B. Marincek1, D. Weishaupt1; 1Zurich/CH, 2 St. Gallen/CH ([email protected]) Purpose: To compare intraindividually two different rectal enema compositions in MR-defecography (MRD). Methods and Materials: 20 patients had MRD two times using a 0.5T openconﬁguration system in the sitting position. In the ﬁrst imaging session, MRD was performed with a rectal enema consisting of potato starch mixed with gadolinium (PS-group). In the second session, enema consisted of ultrasound gel mixed with gadolinium (US-group). The imaging protocol consisted of midsagittal T1-w GRE MR images obtained at rest, maximal sphincter contraction, straining and during defecation. All images were analyzed quantitatively by measuring contrast-to-noise ratio (CNR) and reviewed by three independent observers with regard to the visibility and the extent of pelvic ﬂoor abnormalities. Results: The CNR values in the PS-group (167.49p44.4) were signiﬁcantly higher than those obtained in the US-group (150.2p37.8; p 0.05). The visibility of the anterior rectoceles and the intussusceptions scored higher in the PS-group in comparison with the US-group (mean visibility scores in PS-group, 2.8p0.42 and 2.6p0.56; mean visibility scores in US-group, 2.3p0.77 and 2.2p0.74). The size and the number of incompletely emptying anterior rectoceles were higher in the PS-group. Conclusion: Ultrasound gel and potato starch provide a good contrast and depiction of relevant pelvic ﬂoor abnormalities. However, visibility and extent of pelvic ﬂoor abnormalities are dependent on the composition of the rectal enema. In particular, the size and the degree of the anterior rectocele evacuation and the intussusception size are often underestimated when using ultrasound gel for rectal enema.

14:00 - 15:30

Room F1

Interventional Radiology

SS 709a Radiofrequency ablation Moderators: T. de Baère; Villejuif/FR D. Kuhelj; Ljubljana/SI

B-371

14:00

Efﬁcacy and safety of radiofrequency thermal ablation in the treatment of toxic and non-toxic thyroid nodules with pressure symptoms in elderly patients S. Spiezia1, A. Faggiano1, F. Milone1, V. Ramundo1, R. Garberoglio2, A.P. Assanti1, M. D’Andrea1, C. Di Somma1, G. Lombardi1, A. Colao1; 1Naples/IT, 2Turin/IT ([email protected]) Purpose: Surgery and radioiodine therapy are the main therapeutic approaches in patients suffering from thyroid nodules (TNs) with pressure symptoms. However, surgical and post-surgical complications and ﬁnancial costs of thyroid surgery and radioiodine therapy are higher in elderly patients. Radiofrequency thermal ablation (RTA) is a consolidated, safe and simple approach for percutaneous treatment of tumor lesions within the liver. The objective of this pilot study is to test RTA for the treatment of compressive TNs in elderly patients. Methods and Materials: Eighty-three elderly patients (aged 65 yrs) with cytologically benign compressive TNs were prospectively enrolled in the study; 56 of them had non-toxic goiter, 27 had toxic or pre-toxic goiter. RTA was performed by using a RITA© Starburst needle inserted under ultrasonographic real time guide. Results: During the 12-month follow-up, all TNs showed a signiﬁcant decrease after RTA. Mean TN volume decreased from 24.7p2.3 to 6.7p1.5 ml (p 0.001) with a mean percent decrease of 78.3p2.1% 12 months after RTA. A signiﬁcant shrinkage was found both in toxic and non-toxic goiter (p 0.01). Compressive symptoms improved in all cases and disappeared in 85%. The treatment was well tolerated by all patients. Conclusion: RTA is a safe, effective and simple procedure for the treatment of symptomatic TNs in elderly patients. Both toxic and non-toxic goiter are effectively managed by RTA. This percutaneous ablative procedure may be proposed as a valid and non-invasive therapeutic alternative to surgery and radioiodine therapy.

A

B

C

D

E

F

G

223

Saturday

Purpose: To evaluate the effect of experience on N-staging of rectal cancer using transrectal ultrasound (TRUS) and magnetic resonance imaging (MRI). Methods and Materials: A total of 228 consecutive patients with biopsy-proven rectal cancer were examined using TRUS and MRI. An experienced gastrointestinal radiologist (R1) or a general radiologist (R2) performed the evaluations. In our study we excluded 84 patients who had had preoperative radiotherapy, as preoperative radiotherapy can cause nodal shrinkage. A complete histopathological examination was used as gold standard. The pathologist examined the specimens without knowledge of the preoperative results. Results: The pathologist described 66% tumours as N0, and 34% as N1-2. The Nstaging using TRUS obtained a sensitivity of 45% for R1 and 23% for R2; however, this difference fell short of statistical signiﬁcance, p=0.2. The N-staging with MRI produced a higher sensitivity for R1 than for R2 (77/50%). We found no statistical differences in accuracy of N-staging in relation to tumour height, nor in TRUS or MRI; although the experienced MRI reader obtained a higher N sensitivity 10-15 cm from the anal verge (82%) than the less experienced reader (50%) (p=0.1). Conclusion: The highest sensitivity was obtained by MRI. Experience seemed to translate into a tendency of better sensitivity. It is suggested that preoperative MRI is supervised by an expert of the colorectal team and participation in training workshops is recommended.

B-370

Scientific Sessions B-372

14:09

Growth rate of new hepatocellular carcinomas after percutaneous radiofrequency ablation: Evaluation with multiphase CT D. Choi; Seoul/KR Purpose: To evaluate the growth rate of new hepatocellular carcinomas (HCCs) that develop after percutaneous radiofrequency ablation (RFA) at serial follow-up CT examination, and to determine an appropriate follow-up interval for imaging. Methods and Materials: Sixty-two new HCCs appearing after percutaneous RFA in 59 patients who underwent follow-up multiphase CT were retrospectively identiﬁed. The volume of the new HCCs at follow-up CT was measured on a PACS monitor by using an area measure tool and summation-of-areas technique. We calculated the tumor volume doubling time (TVDT) and tumor diameter doubling time (TDDT). The growth rate was represented by the TVDT. We also evaluated the relationships between variables and TVDT by using stepwise multiple linear regression analysis. Results: Mean baseline and follow-up tumor volumes were 580 mm3 (range 85-13,861 mm3) and 2,072 mm3 (range 535-35,937 mm3), respectively. Mean baseline and follow-up tumor diameters were 9.9 mm (range 5.5-29.8 mm) and 15.0 mm (range 10.1-40.9 mm), respectively. Mean TVDT and TDDT were 75 days (range 21-209 days) and 219 days (range 57-897 days), respectively. TVDTs of baseline tumors of a diameter of 1 cm or less were signiﬁcantly shorter than those of the larger baseline tumors (mean 55 days versus 88 days, p=0.024). Conclusion: The growth rate of new HCCs after percutaneous RFA was higher than those reported in natural outcome studies with untreated HCCs. The results of our study suggest that a shorter follow-up interval for imaging, 2.5 months (75 days), is appropriate.

B-373

14:18

Outcome of PET/CT assisted RF ablation in HCC nodules and colorectal liver metastases: Mid-term results H. Kuehl, J. Stattaus, S. Hertel, P. Hunold, G. Kaiser, G. Antoch, A. Bockisch, M. Forsting; Essen/DE ([email protected]) Purpose: The aim was to determine mid-term results of RFA in malignant liver lesions when using FDG-PET/CT for tumor evaluation and follow-up. Methods and Materials: Between January 2002 and June 2006, 55 patients (mean age 63 years) with 78 liver lesions (39 colorectal metastases, 39 hepatocellular carcinoma nodules) were treated with RFA. All patients were examined with PET/ CT prior to intervention. The RFA procedure was performed in conscious sedation under CT-guidance. Post-interventional PET/CT was done in PET-positive patients 24 h after the ablation and was repeated at 1, 3, 6 and every 6 months after the intervention. PET-negative patients received contrast-enhanced CT (CE-CT) at the same time points. The rate of local tumor progression (LTP) and survival rates were assessed for the whole patient population. Results: The 78 lesions (mean size 2.3 cm, range 0.8-5 cm) were treated with 101 consecutive ablation procedures resulting in a technical success rate of 96%. The mean time of follow-up was 25p12 months. 35 of 78 tumors (45%) developed LTP. At the end of follow-up, LTP was found in 22 patients (40%), with intra- and extrahepatic recurrence in 11 patients. 22 patients remained free of hepatic tumors. The 1-, 2-, and 3-year survival rates were 85, 74 and 58%, respectively. Tumor entity, lesion size and localization were signiﬁcant risk factors for LTP. Conclusion: CT-guided RFA of malignant liver tumors is effective, but shows a high rate of local tumor progression. PET/CT supports RFA by early identiﬁcation of residual tumor or LTP.

B-374

14:27

Early detection of residual viable tumour after radiofrequency ablation of colorectal liver metastases: Diffusion-weighted MRI versus unenhanced and gadolinium-enhanced MRI S. Zaid, L. Crespi, P. Marelli, T. Ierace, L. Solbiati; Busto Arsizio/IT ([email protected]) Purpose: To evaluate the accuracy of diffusion-weighted MRI (DWMRI) in detecting residual viable tumour following percutaneous radiofrequency ablation (RFA) of colorectal liver metastases. Methods and Materials: Twenty-ﬁve patients with 52 large colorectal liver metastases underwent abdominal MRI at 1.5 T, including conventional (baseline and contrast-enhanced) MRI and DWMRI (b values of 0, 50, 400 and 800 s/mm²) at 24 hours, one, three and six months following RFA. Two observers reviewed prospectively, independently and, respectively, conventional MRI and DWMRI. Fine-needle aspiration biopsy was performed in all pathological imaging ﬁndings. Results: After RFA, 36/52 lesions (69.23%) treated lesions were found to be necrotic

224

A

B

C

D

E

F

G

on DWMRI at 24 hours check. Sixteen of the 52 metastases showed persistent peripheral hyperintense areas in all DW sequences, characterized by low apparent diffusion coefﬁcient (ADC) values (0.92p0.19x10-3 mm²/s), strongly suggestive of residual tumour. At 24 hours, conventional MRI showed pathological ﬁndings in 7/16 lesions and irregular peripheral thickening in two other lesions. Pathological examination, performed on the 18 suspicious lesions, conﬁrmed the presence of residual viable tumour cells in 12 lesions. At 3-month check, hyperintense areas described on 15/16 lesions became more evident on DWMRI. At one month, conventional MRI showed pathological contrast-enhancement in nine lesions, reaching 15 of the 16 lesions seen at DWMRI after three months. These 15 incompletely treated lesions underwent second RFA, eventually successful. Conclusion: DWMRI presents higher sensitivity (100%) and speciﬁcity (97%) than conventional MRI (sensitivity 60% and speciﬁcity 94.5% after one month) in assessing RFA treatment efﬁcacy at early stage.

B-375

14:36

Multipolar radiofrequency ablation (RFA) of liver tumours: Frequency and causes of local recurrences T. Albrecht, B. Frericks, S. Valdeig, J.P. Ritz, K.-J. Wolf; Berlin/DE ([email protected]) Purpose: RFA of liver tumours shows recurrence rates of up to 30%. We assessed the frequency and causes of local recurrences after multipolar RFA. Methods and Materials: Ninety-seven liver tumours (mainly metastases) in 46 patients were treated with multipolar RFA percutaneously under CT (n=35) or US (n=38) guidance or intraoperatively (n=26). Temporary blood ﬂow reduction during ablation using embolisation or Pringle’s manoeuvre was performed in 24 and 19 lesions, respectively. Up to ﬁve RF-probes were used simultaneously. Mean follow-up by 3-6 monthly MRI was 20p12 (max. 44) months. Results: Local recurrence deﬁned as tumour re-growth within or at the margin of the ablation zone occured in 11/97 tumours (11%) and 6/46 patients (13%). Patients with local recurrence had a higher mean number of lesions treated (4 vs 1.4) and had more frequently previous chemotherapy. 7/11 (63%) tumours with local recurrence were ablated under CT guidance, three (27%) under US guidance and one (9%) intraoperatively. Tumours with local recurrence were treated with a lower number of RF-probes and lower energy (69p50 vs 100p64 kJ). Blood ﬂow reduction was less frequently used in these tumours (1/11 vs 41/86). In 8/11 recurring lesions, the initial post RFA safety margin was b5 mm. The tumour size prior to ablation had no effect on local recurrence. Conclusion: Multipolar RFA of the liver has a relatively low local recurrence rate. Local recurrence was less likely in patients with few tumours without previous chemotherapy and in lesions treated with blood ﬂow reduction and a safety margin exceeding 5 mm.

B-376

14:45

Effect of blood perfusion on coagulation zone size during radiofrequency tumor ablation W. Schramm, D. Haemmerich; Charleston, SC/US ([email protected]) Purpose: It is well known that perfusion affects the size and shape of the coagulation zone created during radiofrequency (RF) tumor ablation. The purpose of this study was to quantify the change in coagulation zone dimensions due to different perfusion rates. Methods and Materials: We created computer models that solve the bioheat equation to determine tissue temperature during RF ablation using a cooled needle electrode. We implemented an accurate perfusion model that includes an initial increase of perfusion at small temperature rise, and subsequent cessation of perfusion using an Arrhenius damage model based on prior experimental data. We report long axis, short axis (diameter), and volume of the coagulation zone at baseline perfusion (1000 mL/kg/min), and at rates of +20% and -20% of baseline. Results: The long axis and short axis (diameter) of the coagulation zone were 26.6 and 44.5 mm for baseline perfusion, 25.6 and 43.8 mm for +20% perfusion, and 28.4 and 45.4 mm for -20% perfusion. The short axis varied by ~7% compared to baseline perfusion, the long axis changed less than 2%. The coagulation zone volumes were 17.61, 13.6 and 17.4 cm3 for baseline, +20%, and -20% perfusion, respectively. Conclusion: Perfusion affects coagulation zone size and shape. However, most of the variation affects the short axis (diameter) of the coagulation zone while the long axis shows only minor variation.

Scientific Sessions B-377

14:54

Variability in effect of radiofrequency ablation in different tissues: Ex vivo and in vivo comparative studies Y. Kim, D. Choi, H. Rhim, Y.-S. Kim; Seoul/KR ([email protected])

B-378

15:03

Percutaneous radiofrequency ablation of hepatocellular carcinoma with assistance of artiﬁcial ascites: Safety and technical efﬁcacy in 143 patients. I. Song, H. Rhim, Y.-S. Kim, D. Choi, W. Lee, H. Lim; Seoul/KR ([email protected]) Purpose: To assess the feasibility, safety and efﬁcacy of radiofrequency ablation (RFA) with artiﬁcial ascites for hepatocellular carcinoma (HCC). Methods and Materials: A total of 143 patients with 180 HCCs underwent ultrasonography (US)-guided percutaneous RFA with artiﬁcial ascites. The reasons for using artiﬁcial ascites were to prevent thermal injury of the adjacent organs (diaphragm (n=107), GI tracts (n=33), peritoneum (n=6)) and to improve sonic window for hepatic dome lesion (n=96). We artiﬁcially induced ascites before RFA by dripping 5% D/W solution. Technical success for introducing artiﬁcial ascites was deﬁned if the dangerous organs were separated 5 mm from the liver. We assessed technical feasibility of introducing artiﬁcial ascites and complication related to artiﬁcial ascites. Technical success rate, primary and secondary technique success rates were assessed by regular follow-up CT (mean, 12 months). Results: RFA with artiﬁcial ascites was successfully achieved in 130 (90.9%) of 143 patients with mean 702 ml D/W solution. The reasons for technical failure were perihepatic adhesion due to surgery (n=4) or TACE (n=1), prominent omental fat (n=1), bare area (n=2), unknown (n=5). Except for one patient with hemoperitoneum, there was no major complication related to artiﬁcial ascites. Primary technique effectiveness rate at one month was 85% (122/143). Secondary technique effectiveness rate at 12 months was 76% (109/143). Conclusion: Artiﬁcial ascites is a simple and useful technical tip to minimize the collateral thermal injury, to reduce the peri-procedural pain, and to improve the sonic window.

B-379

15:12

Fluid-modulated bipolar radiofrequency ablation: An ex vivo evaluation study P. Bruners, H. Mueller, R.W. Guenther, T. Schmitz-Rode, A.H. Mahnken; Aachen/DE ([email protected]) Purpose: To evaluate the potential of different ﬂuids to modulate the effect of bipolar RFA in an ex vivo liver model. Methods and Materials: A bipolar RFA-system with an internally cooled needleshaped applicator (20 mm active tip) was used to induce coagulation lesions in ex vivo porcine liver. Prior to the ablation procedure, one ml samples of seven different ﬂuids were injected (0.9% saline, Gadopentetate-dimeglumine, Ioxithalaminacid, 10% hydroxyethyl-starch, 5% glucose, 95% ethanol, distilled water). Each ﬂuid was used for ﬁve RF-ablations. During the procedure applied energy, resistance and time were recorded. Additionally temperature was measured at a ﬁxed distance of 0.5 and 1.0 cm from the RF-probe. Generator output was set to 20 W as recommended by the vendor. Five ablation procedures without ﬂuid injection served as reference standard. The lesion diameters were measured and volume and an

B-380

15:21

Colorectal liver metastases treated with radiofrequency ablation: Evaluation of FDG-PET/CT for follow-up H. Kuehl, G. Antoch, S. Rosenbaum-Krumme, H. Stergar, G. Kaiser, M. Forsting, A. Bockisch; Essen/DE ([email protected]) Purpose: Colorectal liver metastases were treated with radiofrequency ablation (RFA). FDG-PET/CT is used for follow-up and compared with CT and PET to determine its value in detecting local tumor progression (LTP). Methods and Materials: 22 patients (mean age 63 years) showed PET-positive liver lesions and were treated with CT-guided RFA. Post-interventional PET/CT was performed within 24 h and was repeated after 1, 3, and 6 and then every 6 months. The PET and CT data and the combined PET/CT studies were read by different blinded reader teams. The applicability of the different imaging modalities for the detection of LTP was assessed and compared with the clinical course. Either histological proof by biopsy or resection or the combination of contrast-enhanced imaging after 3, 6 or 12 months with clinical data and tumor markers served as reference standard. Results: The 22 patients had 37 liver lesions (mean size 24 mm) and underwent 51 consecutive RFA sessions. 82 PET/CT examinations were performed during follow-up. The mean follow-up time was 735p350 days. Ten patients died from tumor progression. Compared to reference standard, CT missed 32, PET 18 and PET/ CT 12 LTP. The accuracy and sensitivity for LTP detection were 77 and 64% for CT, 84 and 79% for PET, and 91 and 86% for PET/CT, respectively. PET/CT was signiﬁcantly better than either CT or PET. Conclusion: PET/CT was more sensitive and accurate than either PET or CT when assessing for LTP after RFA of colorectal liver metastases.

14:00 - 15:30

Room G/H

Head and Neck

SS 708 New approaches to examination techniques Moderators: V. Chong; Singapore/SG A. Trojanowska; Lublin/PL

B-381

14:00

Inﬂuence of different mAs-products on image quality in MSCT of the head and neck region U. Baum, K. Anders, C. Ruedel, M. Lell, W.A. Bautz; Erlangen/DE ([email protected]) Purpose: To determine the inﬂuence of different mAs-values on image quality and image noise in multislice CT of the head and neck region. Methods and Materials: Thirty datasets with head and neck malignancies, acquired on a multislice system (4x1 mm, SOMATOM Volume Zoom, Siemens, Forchheim, Germany) were used for image manipulation by means of a dedicated reconstruction software (Syngo Explorer, VAMP GmbH, Möhrendorf). Surplus image noise was added to raw data sets in order to simulate different effective mAs values. Images were reconstructed with a slice width of three mm at 160, 120, 80 and 40 eff.mAs at three predeﬁned anatomical positions (oropharynx, larynx, thoracic inlet). Image noise was measured in all reconstructions, and image quality was rated by two radiologists on a 3-point scale. Results: Mean image noise was not signiﬁcantly different for images reconstructed with 80 eff.mAs, 120 eff.mAs and 160 eff.mAs at the oropharynx, larynx and thoracic inlet. In comparison to the standard reconstruction with 160 eff.mAs no signiﬁcant decrease of image quality was observed for reconstructions with 120 or

A

B

C

D

E

F

G

225

Saturday

Purpose: To determine variability in effect of radiofrequency ablation (RFA) on different tissues in ex vivo and in vivo pigs and to compare the effect between ex vivo and in vivo studies. Methods and Materials: A prototype RFA system with multiplex sequential generating mode was used, which enables to treat three separate tumors at once by switching in order between three electrodes. RFA was applied to three normal tissues (liver, kidney, and thigh muscle) for 10 minutes in ex vivo (n=12) and in vivo (n=17) pigs using a 200 W generator and three 1-cm internally cooled electrodes. Pre-impedance and coagulation diameter were recorded and compared. Friedman’s test and t-test were used to compare RF-induced coagulation areas among three tissues in ex vivo and in vivo studies. Results: RF ablation of kidneys resulted in the smallest coagulation area (0.90p0.40, 0.77p0.40 cm2) compared with that in the other groups in both ex vivo and in vivo studies. The greatest coagulation area (1.75p0.40 cm2) was observed in the muscle in the in vivo study; however, in the ex vivo study, the coagulation area (1.71p0.40 cm2) in the liver was the largest. There is no signiﬁcant correlation between coagulation area and pre-impedance in ex vivo and in vivo studies. Conclusion: Renal RF ablation provided the smallest coagulation area than ablations in the liver and muscle in both in vivo and ex vivo experiments. Pre-impedance as tissue-speciﬁc characteristic did not affect the outcome of RFA in each tissue.

efﬁciency-index (coagulation volume/procedure duration) calculated and compared (ANOVA, Student-Newman-Keuls test). Results: In comparison to the reference standard, ﬂuid injection resulted in an enlargement of mean coagulation volume for all ﬂuids. In comparison to RF-ablation alone (1.42p0.33 cm³), signiﬁcant increase (p 0.001) of lesion size was found using Gadopentetate-dimeglumine, hydroxyethyl-starch, glucose and ethanol. The largest lesions were generated using preinjection of hydroxyethyl-starch. Most energy could be applied after injection of glucose. Comparing the efﬁciency-index, no statistically signiﬁcant differences were found. Conclusion: Besides the modulation of the electrical conductivity, the increase of thermal conductivity also contributes to the enhancement of RFA by the injection of the tested ﬂuids.

Scientific Sessions 80 eff.mAs for the oropharynx and the larynx. However, mean image quality was signiﬁcantly different for images recontructed with 40 and 80 eff.mAs at the larynx and thoracic inlet, respectively. Conclusion: This study demonstrates the feasibility of dose reduction in CT scans of the head and neck region. The collected data may provide a basis for adjustments of mAs values for standard protocols and for the adaptation of automatic dose control systems.

B-382

14:09

Determining optimized scan parameters of multidetector CT of the head and neck: Comparison of different scan delay and total volumes of contrast material Y. Lee, B.-K. Je, S. Yeom, W.-J. Lee, B. Seo, B. Kim; Ansan/KR ([email protected]) Purpose: To determine the optimal scan delay and total volume of contrast material (CM) on multidetector CT (MDCT) of head and neck. Methods and Materials: Eighty patients (M:F=36:44; mean age: 45 years) were randomized into four groups (20 patients per group: group I, 100 ml CM+10 ml saline, 70 sec scan delay; group II, 80 ml CM+30 ml saline, 40 sec; group III, 80 ml CM+30 ml saline, 70 sec delay and group IV, 80 ml CM+30 ml saline 40, 70 and 100 sec). CT imaging was performed using a 64-MDCT with high iodine-concentration CM (370 mg I/ml) administrated to right arm at 2.5 ml/sec. The postcontrast attenuation values were measured in both tonsils, salivary and thyroid glands, muscles, subclavian veins, aortic arch in each patient, and the degree of perivenous artifacts was subjectively assessed. Results: All mean tissue attenuation values in groups II and III were signiﬁcantly lower than in group.i. except muscles (p 0.05), while those of thyroid gland were equivalent in groups I and II (p 0.05). Different scan delays resulted in no signiﬁcant difference in all mean tissue attenuation values, except for thyroid gland and vessels (group IV, p 0.05). Perivenous artifacts were greater at the images obtained with 40 sec than 70 or 100 sec delay. Conclusion: At MDCT of head and neck, decreased tissue enhancement resulting from reduction of total CM volume can be minimized by shortening of scan delay. But perivenous artifact was not reduced on earlier scan delay by saline ﬂush.

B-383

14:18

Dose comparison of dental MSCT and dental cone beam CT E.M. Coppenrath1, F. Draenert1, U. Lechel2, R. Veit2, T. Meindl1, M. Reiser1, U.L. Mueller-Lisse1; 1Munich/DE, 2Neuherberg/DE ([email protected]) Purpose: Three-dimensional imaging of teeth can be performed by a cone beam CT technique as an alternative to MSCT. In order to estimate dose application of both systems, thermoluminescence dosimetry (TLD) was carried out in an anthropomorphic phantom. Methods and Materials: Organ doses and effective dose of a dental investigation were compared in MSCT (Volume Zoom Siemens®) and in cone beam CT (NewTom QR-DVT 9000®) using an Alderson-Rando phantom with TLD elements. Tube voltage of MSCT was 120 kV, tube current time product 70 mAs, rotation time 0.75 s, pitch 0.675, and collimation 1 mm. In cone beam CT collimation (130 mm), voltage (110 kV), and rotation time (75 s) are predetermined. Results: The effective dose of MSCT was 0.33 mSv for women and 0.32 mSv for men. The effective dose of cone beam CT was 0.095 mSv for women and 0.093 mSv for men. The local dose of the mouth was 8-fold higher in MSCT (12.6 mSv) than in cone beam CT (1.5 mSv). Conclusion: Under the aspect of dose application, the dental cone beam CT seems favourable in comparison to MSCT. Both techniques, however, belong to low dose CT investigations. Furthermore, dosimetry by ionisation chamber for estimation of CTDI is an imprecise method for cone beam CT due to its beam geometry. Therefore the time-consuming method of TLD was necessary for dose estimation.

B-384

14:27

3D-CTA in the head and neck using a new software to remove bony structures: Evaluation of the images and clinical role T. Otani, N. Tomura, S. Takahashi, I. Sakuma, K. Yasuda, J. Watarai; Akita/JP ([email protected]) Purpose: To evaluate the quality of images and usefulness of a new software, which has the ability to automatically remove the bony structures (AutoBone Xpress) on the head and neck 3D-CT angiography (CTA). Methods and Materials: Twenty-seven 3D-CTA studies were performed in 26 patients (mean: 68.6 yrs) with carotid or vertebral arterial steno-occlusive dis-

226

A

B

C

D

E

F

G

eases. The scan range included the extracranial vertebral arteries. After AutoBone Xpress automatically removed the bony structures, an aditional post-processing was performed. We measured the time automatically acquired through the AutoBone Xpress and the time for more post-processing. The quality of images in the cervical internal carotid artery (ICA) and vertebral artery (VA) was estimated on a 4-point grade (0-3). Results: The average time through Autobone Xpress was 1 min 04 sec (range: 0 min 25 sec-2 min 00 sec). The average of the additional post-processing time was 8 min 28 sec (range: 2-30 min). The average of grades of quality of the initial image automatically acquired through AutoBone Xpress was 3.1 (ICA: right=3.15, left=3.11) and 1.9 (VA: right=1.85, left=2.07). After post-processing, the average of grades increased in 3.9 (ICA: right=3.88, left=3.92) and 3.5 (VA: right=3.33, left=3.67). In several cases, due to the metallic artifacts, concentration of contrast medium impaired the quality of the images even after post-processing. Conclusion: AutoBone Xpress is useful in 3D-CTA in the head and neck region; it can especially reduce post-processing time. AutoBone Xpress can compose favorable 3D images in head and neck region. Using Autobone Xpress, 3D-CTA cannot be mostly affected by bony structures.

B-385

14:36

Comparison of MRI combined with diffusion-weighted imaging and triphasic dynamic contrast enhanced multislice spiral CT for the diagnosis of parotid tumors H. Yerli1, A.M. Agildere2, E. Aydin2, M. Coskun2, N. Haberal2; 1Izmir/TR, 2Ankara/TR ([email protected]) Purpose: To compare accuracy of MRI combined with diffusion-weighted imaging (DWI) and triphasic dynamic contrast enhanced multislice spiral CT (MSCT) for the diagnosis of parotid tumors. Methods and Materials: Twenty seven patients (mean age, 59) with parotid tumors were evaluated, prospectively. Informed consent was obtained from all patients. After post-contrast whole neck examination at 30 seconds, images only at the level of the parotid tumor were obtained at 90 seconds and at 5 minutes in an MSCT unit. Neck MRI was performed using 1.5 T MRI unit. T1, T2 and fat-suppressed T2-weighted images, DWI with ADC maps and fat-suppressed T1-weighted postcontrast images were obtained. MRI and CT diagnosis was compared with histopathology. Results: The tumors comprised eight Warthin tumors, eight pleomorphic adenomas, six malignant tumors, one basal cell adenoma, two lipoma and two benign cysts. Eighteen patients were assessed with both MRI combined with DWI and triphasic CT methods. One patient, who had cervical arthrosis, and two patients with impaired cooperation were not examined by MRI. Two lipoma and two benign cysts were only examined with plain MRI. Two patients rejected CT examination. T1-weighted postcontrast images did not provide any additional diagnostic information. The accuracy of MRI combined with DWI and triphasic MSCT was 86 and 81%, respectively. Conclusion: Both non-contrast MRI combined with DWI and triphasic dynamic contrast enhanced MSCT may be a valuable modality for the diagnosis of parotid tumors. In the method selection, some factors such as patient cooperation, the presence of ferromagnetic device and cost-effectiveness should be considered.

B-386

14:45

Gadolinium-enhanced T2*-weighted MR imaging of tumors in the extacranial head and neck: Feasibility study and preliminary results using a distributed-parameter tracer kinetic analysis S. Bisdas1, M. Baghi1, I. Burck1, M. Kerl1, J. Wagenblast1, C. Thng2, M. Mack1, T.J. Vogl1, T. Koh2; 1Frankfurt a. Main/DE, 2Singapore/SG ([email protected]) Purpose: To examine the feasibility of ﬁrst-pass dynamic contrast-enhanced (DCE) T2*-weighted MR imaging of tumors in the extracranial head and neck by applying a distributed-parameter (DP) tracer kinetic model in order to quantify the perfusion parameters. Methods and Materials: Eighteen patients with primary malignant and benign tumors in the head and neck underwent DCE-MR studies. Aspin-echo echoplanar-imaging (EPI) MR-sequence was applied for ﬁrst-pass DCE-T2*-weighted imaging. The data were post-processed applying a DP tracer kinetic model which accounts for capillary-tissue exchange. Region of interest analysis was performed in the tumor sites and the adjacent normal tissue. Blood ﬂow (F), intravascular blood volume (v1), extravascular extracellular blood volume (v2), tlag (t0), intravascular mean transit time (t1), permeability (PS), and extraction ratio (E) maps were generated for each patient. Results: All perfusion values in the tumor sites were significantly different (0.000bpb0.01) than those in the normal muscle tissue. The median perfusion values in the tumor tissue were: F=150.5 ml/min/100 g, v1=11.0 ml/100 g, v2=31.5 ml/100 g, t0=4.5 s, t1=8.0 s, PS=96.0 ml/min/100 g, E=32.5. Conclusion: EPI-T2*-weighted DCE-MR in head and neck tumors and quantiﬁca-

Scientific Sessions tion of the perfusion values using DP model physiologic imaging was feasible and the promising initial results have to be validated in future studies.

B-387

14:54

Assessment of nasal mass with diffusion weighted single shot echoplanar MR imaging A.A.A. Abdel Razek, N. Soliman, A. Haroun, N. Nada; Mansoura/EG ([email protected])

B-388

15:03

Clinical evaluation of elastography for the differential diagnosis of thyroid follicular tumors N. Fukunari1, K. Arai1, A. Nakamura1, S. Sakaue1, T. Mitake2; 1Yokohama/JP, 2 Chiba/JP ([email protected]) Purpose: The diagnosis and management of follicular cancer of the thyroid gland remains a controversial topic. The aim of this study was to evaluate the clinical usefulness of elastography imaging for the differential diagnosis of thyroid follicular lesions. Methods and Materials: 88 follicular tumors were examined by elastography. The images of thyroid follicular lesions were recognized as four typical patterns as follows: Pattern 1 - nodule is relatively homogenous and colored with light green. Pattern 2 - the center of nodule is colored with green and its periphery is colored with blue. Pattern 3 - nodule is mixed-colored with light green and red. Pattern 4 - the whole tumor is hard and displayed in blue. Results: All the follicular tumors were surgically and histopathologically diagnosed ﬁnally. In this series, 44 hyperplastic nodules, 28 follicular adenomas and 16 follicular cancers were found. Of the 72 benign thyroid nodules, 60 cases demonstrated Pattern 1, 3 cases Pattern 2 and 9 cases Pattern 3. Of the 16 follicular cancers, on the other hand, 1 case was Pattern 1, 14 cases Pattern 2, and 1 case Pattern 3. Conclusion: In follicular cancers, the major elastography pattern was Pattern 2. Assuming Pattern 2 to be malignant and others to be benign, 69 of 72 benign tumors and 14 of 16 follicular cancers were accurately diagnosed, yielding a sensitivity of 87.5%, a speciﬁcity of 95.3%, and an accuracy of 94.3%. Elastography can provide new useful information for the differential diagnosis of thyroid follicular tumors.

B-389

15:12

Prediction of malignancy in thyroid nodule: Diffusion weighted echoplanar MR imaging A.A.A. Abdel Razek, A. Farouk, A. Tawﬁk, A. Denewer, N. Nada; Mansoura/EG ([email protected]) Purpose: To predict malignancy in thyroid nodule with diffusion weighted echoplanar MR imaging. Methods and Materials: Diffusion weighted MR imaging was conducted on 123 patients (33M, 90 F, mean age 39 ys) with thyroid nodules using a single shot EPI. Diffusion MR imaging was acquired with a b-factor of 0, 250 and 500 sec/mm2. Apparent diffusion coefﬁcient (ADC) value of the thyroid nodule was calculated. Surgey was done for 60 patients and ﬁne needle aspiration cytology (FNAC) was done for 63 patients under ultrasound guidance. Receiver operating curve (ROC) was done with calculation of accuracy, sensitivity, speciﬁcity, positive predictive

B-390

15:21

Radio-guided localization of neck masses: A new approach A. Salmaslioglu, M. Tunaci, G. Acunas, B. Bakir, A. Ucar, K. Guven, I. Adalet, Y. Erbil; Istanbul/TR ([email protected]) Purpose: Neck masses, especially metastatic lymph nodes in previously operated papillary thyroid carcinoma patients and parathyroid adenomas, may be difﬁcult to locate intraoperatively. A new method using preoperative injection of a radiotracer under ultrasonographic guidance is introduced. Methods and Materials: Six patients, four with metastatic papillary thyroid carcinoma veriﬁed by ﬁne-needle aspiration biopsy, and two with parathyroid adenoma veriﬁed by ﬁne-needle aspiration biopsy and PTH measurement of the aspirate have been included in the study. A 20 gauge needle was inserted under ultrasonographic guidance into the center of the lesion and 0.2 ml of 99mTc-labeled rhenium was directly injected. Scintigraphy was performed at 30 min following the injection of radiocolloid to verify the procedure. An intraoperative hand-held gamma-probe (Navigator, Surgical gamma-probe, Auto suture, France) was used to detect the tumor foci. After the lesions were removed, radioactivity was measured in the lesion bed to conﬁrm the success of the dissection. Results: All of the lesions were visible scinitigraphically 30 min following the injection of radiocolloid. Intraoperatively, all lesions were located by the gamma probe and excised. Postoperative pathologic examination veriﬁed the nature of lesions. Conclusion: Ultrasonography-guided intra-lesional radiocolloid injection and intraoperative localization by gamma probe is a safe method which can be used to locate non-palpable neck masses and allow focal surgery to be performed.

14:00 - 15:30

Room I

Physics in Radiology

SS 713 New approaches in MRI Moderators: A. Del Guerra; Pisa/IT O. Dietrich; Munich/DE

B-391

14:00

Prospective navigator-echo-based real-time triggering of fetal head movement for the reduction of artifacts H.M. Bonel, K.A. Frei, M. Meyer-Wittkopf, L. Raio, L. Remonda, R. Wiest; Berne/CH ([email protected]) Purpose: To evaluate the neuroimaging quality and accuracy of prospective real-time navigator-echo acquisition correction versus untriggered intrauterine imaging techniques. Methods and Materials: Twenty women in whom fetal motion artifacts compromised the neuroimaging quality of fetal magnetic resonance imaging (MRI) taken during the 28.7p4 week of pregnancy below diagnostic levels were additionally investigated using a navigator-triggered half-Fourier acquired single-shot turbo-spin echo (HASTE) sequence. Imaging quality was evaluated by two blinded readers applying a visual rating scale from 1 (not diagnostic) to 5 (excellent image quality). Diagnostic criteria included depiction of the germinal matrix, grey and white matter, CSF, brain stem and cerebellum. A quantitative rating scale was generated to evaluate the signal-difference-to-noise ratios (SDNRs) in the white matter and germinal zone. Results: Imaging quality was improved in 18/20 patients using the navigator echo technique (2.4p0.58 vs 3.65p0.73 SD, p 0.01 for all evaluation criteria). In 2/20 patients fetal movement severely impaired image quality in conventional and navigated HASTE. Navigator-echo imaging revealed additional structural brain abnormalities and conﬁrmed diagnosis in 8/20 patients. The overall accuracy improved from 86 to 95%. Average SDNR increased from 7.07p7.27 to 19.83p15.71 (p 0.01).

A

B

C

D

E

F

G

227

Saturday

Purpose: To assess the clinical usefulness of diffusion weighted echoplanar MR imaging in characterization of nasal mass. Methods and Materials: Prospective study was conducted on 55 consecutive patients (34M, 21 F aged 13-64 ys; mean 51 ys) with nasal mass. They underwent diffusion weighted MR imaging using a single shot echo planar imaging with a bfactor of 0, 500 & 1000 sec/mm2. The scanning parameters were: TR=10000 ms, TE=108 ms, NEX=2, bandwidth=125 kHz, slice thickness=3 mm. Apparent diffusion coefﬁcient (ADC) map was reconstructed. The ADC value of the nasal mass was calculated and correlated with surgical ﬁndings or biopsy. Results: Adequate ADC maps were obtained in 51 patients. The mean ADC value of nasal malignant tumors was 1.02p0.13X10-3 mm2/sec and of benign masses was 1.81p0.11X10-3 mm2/sec. There was a statistically signiﬁcant difference in mean ADC values between malignant nasal tumors and benign lesions (p 0.021) and within the malignant tumors (p 0.04). When apparent diffusion coefﬁcient value of 1.37 X10-3 mm2/sec was used as a threshold value for differentiating of malignant tumors from benign lesions, the best result was obtained with an accuracy of 95%, sensitivity 92%, speciﬁcity 96%, positive predictive value 92% and negative predictive value of 96%. Conclusion: Apparent diffusion coefﬁcient value is a new non-invasive imaging parameter that can be used for characterization of nasal masses as it helps in differentiating malignant nasal tumors from benign lesions.

value (PPV) and negative predictive value (NPV). Results: Adequate ADC maps were obtained in 118 patients. Average nodule size was 20p10.9 mm. The mean ADC value of the malignant thyroid nodule (0.98p0.11X10-3 mm2/sec) was signiﬁcantly different (p 0.003) than that of benign thyroid nodules (1.34p0.14X10-3 mm2/sec). Selection of 1.19X10-3 mm2/sec as a threshold value to predict malignancy in thyroid nodule provided accuracy of 91%, sensitivity of 90%, speciﬁcity of 92%, PPV of 92% and NPV of 90%. Conclusion: We concluded that an apparent diffusion coefﬁcient value is a new non invasive accurate parameter for prediction of malignancy in thyroid nodules and differentiation of malignant from benign thyroid nodules.

Scientific Sessions Conclusion: Navigator-echo-based real-time triggering of fetal head movement is a safe and reliable technique that can deliver diagnostic fetal MR image quality despite vigorous fetal movement.

B-392

14:09

A multimodality assessment of global left ventricular function using a moving heart phantom J.M. Groen, M.J.W. Greuter, P.A. van der Vleuten, D.D. Lubbers, M. Oudkerk; Groningen/NL ([email protected]) Purpose: To compare global left ventricular function assessment of Magnetic Resonance Imaging (MRI), 64-slice Multi Detector Computed Tomography (MDCT), Electron Beam Tomography (EBT) and Dual Source Computed Tomography (DSCT). Methods and Materials: Four different imaging modalities were used to scan an anthropomorphic heart phantom with known volumes. Two observers calculated individually the end diastolic and systolic volumes. The calculated ejection fractions and cardiac outputs were compared using Bland-Altman analysis. Results: Results are given with a Bland-Altman (B-A) interval (between square brackets) and an average (between round brackets). An average underestimation of the ejection fraction (EF) was observed for MRI (16%, B-A interval of [-4.35 (-2.48) -0.60]) and EBT (25%, B-A interval of [-5.03 (-3.87) -2.72]). In contrast, MDCT showed an average overestimation of 2.6%, but the variability was very large shown by a fairly wide B-A interval of [-3.40 (0.40) 4.20]. The smallest difference to the reference was found with DSCT. It showed an average underestimation of 0.8% with a B-A interval of [-1.17 (-0.13) 0.91]. Cardiac output (CO) measurements showed similar results. Conclusion: MRI and EBT showed a structural underestimation in global left ventricular function assessment. 64-slice MDCT showed a heart rate dependent underestimation or overestimation of EF and CO output assessment. DSCT-derived global left ventricular function parameters showed the least deviation from known phantom volumes in this experiment.

B-393

14:18

MRI (VIPAR) gel dosimetry: Comparison of two methods for the construction of 2D color radiation dose maps using data obtained directly from a newly designed research PACS system T.G. Maris1, F. Zacharopoulou1, K. Karolemeas2, E. Pappas1, A. Papadakis1, J. Damilakis1; 1Iraklion/GR, 2Athens/GR ([email protected]) Purpose: To compare two analytical methods for the calculation of radiation dose values assessed by means of quantitative MRI (T2-qMRI) utilising image data obtained directly from a research PACS system. Methods and Materials: Three batches of N-Vinyl-Pirrolidone-Argon (VIPAR) gel dosimeters were irradiated from 10 to 50 Gy using a 6 MV X-ray linear accelerator, each having different gelatine concentrations to provide a wide T2 measurement range. T2 values were measured with a 1.5 T clinical MRI system and a 2D MultiSlice-Multi-Echo (32 echoes) Spin Echo (MSMESE) PHAPS train sequence with parameters: (TR/TE1/TE32/FA: 8500 ms/25 ms/800 ms/180o). T2 calculated maps and radiation dose color maps were post-processingly reconstructed using (a) a standard and (b) a weighted linear regression ﬁtting model. Image transferring, archiving and post-processing were performed using a newly designed research PACS system. The two analytical methods (a and b) were compared in terms of precision using CV% comparison graphs and Bland-Altman (BA) tests. Results: 1/T2 values were linearly correlated with radiation doses (r=0.98, p 0.0005). Radiation dose values estimated with method (a) showed a varying precision of 2-12% and a signiﬁcant correlation to the measured dose value (BA). Radiation dose values measured with method (b) showed a varying precision of 2-4% and no correlation to the measured dose value (BA). Conclusion: Radiation dose values, as assessed using color T2 parametric maps and weighted regression analysis methods, seem to be a valuable means for an easy and precise evaluation of dose distributions in MRI gel dosimetry.

B-394

14:27

A switchable, low ﬁeld multi-channel broadband MRI system for hyperpolarized gas and dynamic nuclear polarization (DNP) research M. Paley, E. Krjukov, E.H. Whitby, J. Wild; Shefﬁeld/UK ([email protected]) Purpose: To develop a switchable ﬁeld, multi channel, low ﬁeld MRI system for hyperpolarized gas and dynamic nuclear polarization (DNP) research. Methods and Materials: Hyperpolarization either by optical pumping or DNP is generating immense interest in the MR imaging and spectroscopy community due to the massive increase in SNR possible for many nuclear species. Tunnel magnetic resonance experiments in which nuclear or electron Zeeman levels are crossed with the ground state tunnel levels of a rapidly cooled methyl group quantum rotor

228

A

B

C

D

E

F

G

as pioneered by Clough et al. can be used to generate nuclear polarization without microwave irradiation. A rapid switch resistive magnet was interfaced to a four channel broadband digital quadrature spectrometer and experiments performed using thermally polarized proton and hyperpolarized 3-He samples. Results: The low ﬁeld imaging system has successfully acquired proton and hyperpolarized 3-He images using a four channel receiver array. The magnet could rapidly switch up to 0.02 T and acquire data within 5-10 s depending on the magnitude of ﬁeld switch as required for tunnel magnetic resonance experiments. Conclusion: A ﬁeld switch MRI system has been developed for hyperpolarized gas research and to enable tunnel magnetic resonance DNP to be investigated for relatively high rotational barrier methyl groups. The system will allow DNP using methyl group with tunnel frequencies up to 850 KHz for nuclear and 5.6 GHz for electron Zeeman level crossing experiments in conjunction with a variable temperature cryostat.

B-395

14:36

Principle and performance of a time-resolving scan procedure with a zigzag spiral K. Stierstorfer, E. Klotz, K. Schwarz, C. Süß, B. Schmidt, T.G. Flohr; Forchheim/DE ([email protected]) Purpose: To demonstrate the principle and to evaluate the performance of a zigzag spiral scan procedure for time-resolved scanning with a 64-row system. Methods and Materials: Time-resolved scanning of an object larger than the detector coverage can be performed with an intelligent scan procedure with continuous data acquisition and a table moving back and forth. We investigate this method, presenting various strategies of making use of the data. The pitch factor varies periodically and smoothly. Reconstruction is performed with a generalization of the WFBP (Weighted Filtered Back Projection) method. We evaluated the technique with patient scans. Results: Scan techniques covering two or three times the detector coverage are employed. The generalized WFBP reconstruction yields excellent results even for strongly varying pitch. To avoid quickscan artifacts, full 2pi reconstructions were performed for each image, so the time resolution is equal to the rotation time. Sampling in the time domain varies with the position in the scan range: for central positions, time sampling is 1.2 s; far away from the central plane, it is reduced to 2.4 s. Conclusion: Zigzag spiral scanning is a viable way of extending the coverage of a CT system for time-resolving tasks like perfusion scanning well beyond the detector coverage. Time domain sampling is limited, but in a meaningful range. Any drawback from this may easily be compensated with the beneﬁt of avoiding the physical drawbacks (e.g. scatter) and increased cost of using detectors with increased coverage.

B-396

14:45

Determination of relaxivities of nanosomes doped with monodisperse superparamagnetic nanoparticles of various core sizes M.G. Kaul, O.T. Bruns, U.I. Tromsdorf, N. Bigall, H. Weller, U. Beisigel, G. Adam; Hamburg/DE ([email protected]) Purpose: To determine r1, r2 and r2* of nanosomes doped with monodisperse superparamagnetic nanoparticles (SPN) of various core sizes with a clinical 3 T MR-scanner. Methods and Materials: SPN were synthesized with core sizes of 4.5, 6.0, and 7.5 nm. The sizes of the MnFe2O4 cores were validated by transmission-electronmicroscopy. Nanosomes, which are triacylglycerid-rich-lipoprotein-like micelles doped with SPN, were prepared out of the lipid moiety of human lipoproteins and have a diameter of 100-450 nm. Besides, the same SPN were coated with poly (maleic-anhydride-alt-1-tetradecene). Each of the six samples was diluted in ﬁve concentration levels. The iron concentrations were estimated by atomicabsorption-spectroscopy. Relaxation time measurements were acquired by LookLocker-, multispin-echo- and multigradient-echo sequences with a transmitreceive-head-coil in a clinical 3 Tesla MR-scanner. T1, T2 and T2*-maps were calculated and the relaxivities ri (i=1, 2, 2*) were analysed by linear regression of Ri=1/Ti= ri*concentration (iron)+1/Ti,disolver. Results: With increasing core size of the SPN the relaxivities of the doped nanosomes r2 [m M-1 s-1] increased (124, 144, 219)while r2* [m M-1 s-1] decreased (1056, 974, 693). The longitudinal relaxivities r1 [m M-1 s-1] show no drastic change (0.2, 0.2, 0.3). The signiﬁcant differences of r2* and r2 for the doped nanosomes indicate that the system is ruled by the static dephasing regime. In contrast, the “free-SPN” are ruled by the motional narrowing regime. Here, relaxivities were measured to be r2: 42, 63, 86 and r2*: 50, 77, 103 showing lower relaxivities and smaller differences between r2 and r2*. Conclusion: Nanosomes doped with superparamagnetic nanoparticles might be used as high T2*-relaxivity contrast agents.

Scientific Sessions B-397

14:54

A novel magnetic ﬁeld inhomogeneity correction for ultra-high ﬁeld MR: A demonstration on phase contrast MRA of the brain vasculature S. Vitabile1, S. Sammet2, M.V. Knopp2; 1Palermo/IT, 2Columbus, OH/US ([email protected])

B-398

15:03

Interactive scan-control with online reconstruction improves axial breathhold imaging for continuously moving table MRI T. Baumann, U. Ludwig, G. Pache, O. Schaefer, M. Langer; Freiburg/DE ([email protected]) Purpose: To report on interactive scan-control with online reconstruction for thoracoabdominal imaging with an axial, breath-hold, contrast-enhanced, T1-weighted, fat-saturated gradient-echo sequence (T1-FLASH) acquired during continuous table movement with sliding multislice (SMS). Methods and Materials: A seamless stack of axial T1-FLASH images were acquired with SMS from the lung apex to the pelvic ﬂoor in 50 patients during continuous table movement. Patients were instructed to hold their breath from the beginning of the scan for as long as possible and continue normal breathing afterwards. If breathing motion artifacts became visible on the online reconstruction images, the scan was stopped and another overlapping scan was applied. This procedure was repeated until the pelvic ﬂoor was reached. Image analysis focused on the number of slices being affected by motion artifacts and artifacts remaining in a combination of all segments. Artifact burden was assessed on a 4-point scale (0=”no artifacts” to 3=”severe artifacts”). Results: Two to four segments were necessary for complete thoraco-abdominal coverage (median, 3; mean, 2.6) with an average number of 47 slices per segment and a total 106 non-overlapping slices. Combining segments with a median overlap of 11 slices reduced the number of slices affected by artifacts from 17.8 to 4.5 for the combined examination. On a 4-point scale, the mean artifact rating for the composed images was 0.6 with a median of 0. Conclusion: Online image reconstruction is a valuable tool to avoid breathing motion artifacts during a continuously moving table acquisition with sliding multislice.

B-399

15:12

Scan time reduction in diffusion-weighted imaging of the upper abdomen: Use of tracking only navigator echo (TRON) T. Takahara1, T. Ogino2, M. van Leeuwen1, J. de Jong1, T. Kwee1, P. Luijten1, W.P.H. Mali1; 1Utrecht/NL, 2Tokyo/JP ([email protected]) Purpose: Diffusion weighted Whole-body Imaging with Background body signal Suppression (DWIBS) technique showed that it was possible to obtain clinically useful images even under free breathing. Respiratory triggering (RT), however, is still necessary to minimize image blurring in the upper abdomen. A major disadvantage of RT is considerably prolonged acquisition time, especially in case of irregular respiration. The recently developed Tracking Only Navigator (TRON) technique theoretically offers acceptable quality of respiratory gating within reasonable scanning time, and may be an excellent alternative to RT.

B-400

15:21

Dynamic MRI evaluation of obstructive sleep apnea (OSA) in patients treated with continuous positive airway pressure (CPAP): Early experiences G. Pelle, V. Panebianco, R. Roma, V. Vergari, E. Santucci, R. Passariello; Rome/IT ([email protected]) Purpose: To investigate the efﬁcacy of dynamic MRI in the evaluation of patients affected by sleep apnea (OSA), before and after application of Continuous Positive Airway Pressure (CPAP). To describe the principal morphologic features of pharynx in these patients. To demonstrate the morphofunctional changes after CPAP application. Methods and Materials: Fifteen patients affected by OSA and ﬁve healthy volunteers were enrolled in the study (20 subjects) and submitted to dynamic MRI examination. The protocol included morphologic MR sequences (HASTE and TIRMT2 weighted, TFL-T1 sequences). In all patients the cross sectional area, the shape and the anterior-posterior maximum diameter of the pharynx were evaluated; the latter were obtained in three different regions (base of the uvula, ﬁve mm above the epiglottis and at a level between them). In the OSAS group all the measurements were performed before and after the application of CPAP. Results: The OSAS group demonstrated cross-sectional diameters lower than the control group. In this control group the dominant shape was elliptical, with the largest diameter oriented transversally (L-L). After the application of CPAP a considerable increment in the pharynx diameter was demonstrated in all patients (100%); a variation of the pharynx shape was also noted. Conclusion: Dynamic MRI is a promising and feasible modality for the morphologic evaluation of patients affected by OSAS and in the assessment of the therapy effects obtained with CPAP.

14:00 - 15:30

Room K

Pediatric

SS 712 Fetal imaging Moderators: C. Balassy; Vienna/AT V. Donoghue; Dublin/IE

B-401

14:00

Logistic regression analysis based on MRI lung volume measurement in fetuses with CDH before and after 30 weeks’ gestation: Prediction of mortality and the need for ECMO therapy K.A. Büsing, K. Kilian, T. Schaible, D. Dinter, W. Neff; Mannheim/DE ([email protected]) Purpose: To predict mortality and the need for extracorporeal membrane oxygenation (ECMO) therapy in fetuses with congenital diaphragmatic hernia (CDH) before and after 30 weeks’ gestation using logistic regression analysis based on MR fetal lung volume (FLV) measurement. Methods and Materials: FLV was measured by MRI in 95 fetuses with CDH between 22 and 39 weeks’ gestation using multiplanar T2-weighted half-Fourier acquired single-shot turbo spin-echo (HASTE) imaging. Based on logistic regression analysis, mortality and the need for ECMO therapy were calculated for fetuses before and after 30 weeks’ gestation, respectively. Results: Overall, higher FLV was associated with improved survival and decreasing probability of ECMO therapy (p=0.0004 and p=0.008). Survival at discharge was 0.3% in children with a FLV of 5 ml as compared to 70.8% in patients with a FLV of 25 ml. The corresponding need for ECMO therapy was 56% in children with a FLV

A

B

C

D

E

F

G

229

Saturday

Purpose: Phase-contrast MRA of the brain is a very promising high image quality approach at ultra-high ﬁeld (7T) MR imaging. This study introduces and assesses a correction methodology to enhance the quality for 3D blood vessel reconstruction at high ﬁeld that insures outstanding quality and vascular depiction. Methods and Materials: Original PC-MRA images are affected by static and radiofrequency magnetic ﬁeld inhomogeneity. A Guillemaud based ﬁlter, i.e. a combination of homomorphic ﬁltering and a normalized convolution, was used for inhomogeneity estimation and, consequently, for PC-MRA image correction and enhancement. In this initial feasibility study, we used healthy volunteer acquisitions on a Philips 7T Achieva system. MRA imaging was performed with the following key settings: 2D multi-slice; 60 slices; slice thickness=1.5 mm; repetition time [TR, ms]=6.90; echo time [TE, ms]=3.48; ﬂip angle=10. The Maximum Intensity Projection (MIP) algorithm was used to assess visualization and depiction of the intracranial vasculature. Results: Considerable improvements in vessel depiction are leading to outstanding, unachieved quality in the 3D MIP reconstructed images. As a quantitative measure, we determined the average number of pixels having the maximum intensity value (i.e. the number of pixels representing the reconstructed blood vessels) before and after the correction ﬁlter application for the transverse MIP reconstructed images: 4,297 pixels (original images), 5,896 pixels (enhanced images). Conclusion: We have introduced and assessed a novel method for inhomogeneity estimation and demonstrated its beneﬁt to bring PC-MRA to outstanding image quality. With these improvements, ultra-high ﬁeld MRA appears to be an outstanding clinical application.

Methods and Materials: Ten volunteers underwent MR imaging of the upper abdomen at 1.5 T. In each volunteer, DWI was performed (1) under FB, (2) with RT, and (3) using TRON. Each scan was done twice with different maximum b-values of 50 and 500 s/mm2. For each type of scan, actual scan time was measured, and image quality was graded by two radiologists independently, using a ﬁve-point scale (1=worst, 5=best). Results: Actual average scanning times under FB, with RT, and using TRON, were 99, 305, and 116 s, respectively. Relative scan time compared to FB was 3.08 for RT and 1.17 for TRON. Average image quality was 3.6 for FB, 3.7 for RT, and 4.6 for TRON. Conclusion: TRON offers better image quality within a much shorter time than RT in diffusion weighed imaging of the upper abdomen.

Scientific Sessions of 5 ml and 8.7% in patients with a FLV of 40 ml, respectively. Prognostic power was signiﬁcantly lower before 30 weeks’ gestation. Conclusion: Beyond 30 weeks’ gestation, logistic regression analysis based on MR FLV measurement is a feasible and highly useful method to individually calculate survival rates and estimate the need for ECMO therapy. Prior to 30 weeks’ gestation, the method is not reliable and the FLV measurement should be repeated, particularly in fetuses with small lung volumes, before deciding on therapeutic options.

B-402

14:09

MR relative fetal lung volume in congenital diaphragmatic hernia: An evaluation of seven formulas for predicting survival and the need for extracorporeal membrane oxygenation (ECMO) therapy K.A. Büsing, K. Kilian, T. Schaible, C. Endler, R. Schaffelder, W. Neff; Mannheim/DE ([email protected]) Purpose: Evaluation of seven formulas to calculate the relative fetal lung volume (rFLV) based on prenatal MR planimetry and biometric parameters to predict survival and the need for extracorporeal membrane oxygenation (ECMO) therapy in patients with congenital diaphragmatic hernia (CDH) and a comparison of the results to the prognostic accuracy of the absolute FLV (aFLV) measured at MR planimetry. Methods and Materials: The rFLV was expressed as a percentage of the predicted lung volume calculated with biometric parameters according to seven different formulas previously described in the literature. In total, 68 children with CDH were assessed by MRI FLV measurement within 23-39 weeks’ gestation. Applying the area under the curve (AUC), the various rFLVs were investigated for their prognostic accuracy to predict patients’ survival and the need for neonatal ECMO therapy. Results: All rFLVs and the aFLV revealed a highly signiﬁcant difference in mean lung volume between survivors and nonsurvivors (p=0.001-0.0003) and measurement accuracy was excellent for each method (AUC 0.800-0.900). For predicting neonatal ECMO requirement, differences in FLVs were smaller but still signiﬁcant (0.050-0.009) and measurement accuracy was acceptable throughout (AUC 0.653-0.739). Conclusion: The aFLV measured at MR planimetry and the various rFLVs are each highly valuable in predicting survival in CDH patients and may be helpful in parental counseling. For predicting whether neonatal ECMO therapy is required, the prognostic strength of the aFLV and the rFLVs is weaker but still acceptable.

B-403

14:18

Assessment of various applications of fetal cine MRI in prenatal evaluation T. Feygin, A. Vossough, L.T. Bilaniuk, L.J. States, M. Epelman, R.A. Zimmerman; Philadelphia, PA/US ([email protected]) Purpose: To determine the utility of real-time cine Magnetic Resonance Imaging (MRI) in conjunction with routine anatomic fetal MRI in the evaluation of functional impairment of various fetal organ systems. Methods and Materials: Real-time cine fetal MRI evaluations were performed in 200 patients in addition to routine fetal MR assessment of various organ systems. All patients were referred based on abnormalities identiﬁed or questioned by prenatal sonography. TrueFISP cine MRI was utilized for assessment of the target fetal organ. The fetuses ranged from 19 to 32 weeks gestational age. Results: Fetal swallowing was evaluated in patients with CNS abnormalities and neck masses, and the various phases of swallowing and their associated functional abnormalities were discernable. CSF ﬂow could be seen within the central nervous system, with demonstration of impaired or accentuated ﬂow in fetuses with Chiari, myelomeningocele, Dandy-Walker, and aqueductal stenosis. In cases of fetal lung masses, functional assessment of diaphragmatic excursion was performed. In select cases, movement of the cardiac chambers and cardiac valves could be assessed. Urinary ﬂow was evaluated in patients suspected of upper or lower urinary tract obstruction. Fetal limb movements were also identiﬁed in most fetuses as a random ﬁnding. Conclusion: Real-time cine MRI is a valuable tool in prenatal assessment, with expanding potential applications in the functional evaluation of multiple fetal organ systems. The additional information can be used for planning prenatal and postnatal surgery and intervention.

230

A

B

C

D

E

F

G

B-404

14:27

In utero magnetic resonance imaging and spectroscopy in women with diabetes complicating pregnancy E.H. Whitby, M. Paley, A. Chetwood, T. Farrell, R. Fraser; Shefﬁeld/UK (e.whitby@shefﬁeld.ac.uk) Purpose: To develop a non-invasive technique to detect and quantify lactate in amniotic ﬂuid and fetal brain in pregnancies complicated by maternal diabetes as a marker of fetal hypoxia and distress. Methods and Materials: Eight pregnant women with diabetes and a planned caesarean section underwent an MR scan during the 24 hours prior to birth. Lactate levels in amniotic ﬂuid and fetal brain were assessed using PRESS MR spectroscopy (MRS) at 1.5 T using a 1.5x1.5 cm voxel. Amniotic ﬂuid, collected at delivery, underwent further MRS in vitro using the same size voxel and imaging parameters. Additional chemical analysis of lactate levels was performed. Reference values for MRS were obtained using standard lactate dilutions. Results: MRS detected lactate in amniotic ﬂuid both in utero and in vitro, although no lactate was detected in fetal brain. MR spectroscopy was able to quantify the level of lactate present within lactate solutions of varying concentrations. This closely correlated with the actual concentration (R2=0.96). MRS detected lactate levels in vivo and in vitro. Conclusion: It is possible to identify lactate in amniotic ﬂuid on in utero MRS. MRS appears to be able to quantify the amount of lactate present. If further developed this may enable the non-invasive detection of fetal hypoxia to guide patient management in pregnancies at increased risk of perinatal mortality.

B-405

14:36

Magnetic resonance imaging of the fetal posterior fossa T. Lavanya, R. Batty, P.D. Grifﬁths, J. Eames, M. Paley, E.H. Whitby; Shefﬁeld/UK ([email protected]) Purpose: Magnetic resonance (MR) imaging of the fetus is a rapidly expanding clinical and research tool. The posterior fossa of the fetus has been recently studied to assess the normal changes with gestational age. In this study we assessed the changes in normal fetuses compared to fetuses with spina biﬁda or Dandy Walker. Methods and Materials: Images were obtained using (T2 weighted) single-shot fast spin-echo. Images of fetuses with spina biﬁda (n=40) or Dandy Walker variant (n=13) and normal fetuses (n=63) were retrospectively analysed using proprietary software. Measurements were taken of the trans-cerebellar diameter (TCD), the craniocaudal (CCD), anteroposterior (AP) and midline sagittal surface area (CSA), of the cerebellar vermis. Measurements of the midline sagittal surface area of an area deﬁned as the ‘bony posterior fossa’ (PF) were also obtained. Results: The TCD, CCD, AP, CSA and PF increased with increasing gestational age in all groups. The increasing TCD, CCD, CSA of fetuses with spina biﬁda was normal but the PA increased at a slower rate and did not reach the same size as the normal fetuses when associated with a Chiari deformity. The changes were independent of the size of the spinal defect. The fetuses with Dandy Walker variant also had an increase in size of their cerebellar measurements with gestational age but the vermian measurements were small and did not increase much with gestational age. The PA was larger than the normal fetuses at all gestational ages. Conclusion: Posterior fossa analysis may aid diagnosis and possibly prognosis in the future.

B-406

14:45

In utero tractography of fetal brain pathologies G.J. Kasprian, P.C. Brugger, E. Krampl, C. Lindner, F. Stuhr, D. Prayer; Vienna/AT ([email protected]) Purpose: Diffusion tensor imaging (DTI) is increasingly used to noninvasively investigate human brain development. So far this technique has been limited to postnatal neuroimaging. This study aims to demonstrate normal and abnormal connectivity in cases of pathological fetal brain development using DTI and tractography in utero and in vivo. Methods and Materials: Six unsedated fetuses (aged 19-37 gestational weeks) with brain pathologies suspected by ultrasound underwent fetal MR imaging. Axial DTI sequences (32 diffusion encoding directions, voxel size 1.8x1.8x5 mm, acquisition time 1-2 minutes) and corresponding T2-weighted and T1-weighted sequences in three orthogonal planes were acquired. Multiple regions of interest were deﬁned to calculate right and left corticospinal (CST) and callosal trajectories. Results: In each case at least one trajectory could be successfully depicted. Compared to age matched controls the CST appeared unremarkable in both cases of callosal agenesis (CCA). In one fetus with CCA Probst’s bundle could be identiﬁed.

Scientific Sessions In cases of hydrocephalus (n=2) the CST was located within the thinned intermediate zone. One case with in utero infarction of the middle cerebral artery showed ipsilateral abnormal fractional anisotropy (FA) and ADC values of the brainstem without any trajectories descending or ascending from/to the lesion. Despite normal CST morphology, FA measurements along the trajectory were signiﬁcantly reduced in one fetus with signs of intrauterine death. Conclusion: As fetal brain pathologies can successfully be studied by DTI, a new and powerful tool in the fetal neurologic assessment is added to the in utero neurologic examination of the living human fetus.

B-407

14:54

Prerequisites for in utero tractography of the fetal brain G.J. Kasprian, P.C. Brugger, M. Weber, F. Stuhr, C. Lindner, D. Prayer; Vienna/AT ([email protected])

B-408

15:03

Maturational changes of the olfactory bulbs on MRI F. Floemer, O. Magerkurth, C. Jauckus, J.F.L. Schneider; Basle/CH ([email protected]) Purpose: To describe maturational changes of the olfactory bulb with MR. Methods and Materials: Thin high-resolution coronal T2-spin echo images of the frontobasis were acquired in 119 patients from neonate to adolescent. Olfactory bulbs (OB) were identiﬁed and categorized according to their shape and signal intensity. Results: In the neonatal period, OB shapes were from round to oval with a continuous external layer of neurons and a prominent central area of unmyelinated white matter (median age 15 days, range 1-168 days). In early childhood, OB were U-shaped with thinning and concave deformation of the superior layer. Central white matter was still T2-hyperintense compared to the circumferential neuronal layer (median age 287 days, range 4-643 days). Adult-type OB were small, round or slightly J-shaped with a slightly more prominent lateral neuronal layer. Central white matter could not be identiﬁed any more (median age 1872 days, range 107-7176 days). Conclusion: Maturational changes occur within OB. Reorganization of peripheral neuronal layers and myelination of central white matter can be demonstrated. Adulttype aspect of OB may begin within the ﬁrst year of life and is demonstrated in all children at the end of the second year, paralleling cerebral maturational changes as depicted by conventional MRI.

B-409

15:12

Inner ear malformations in the 22q11 deletion syndrome C. Cellier, J.P. Marie, V. Drouin-Garraud, F. Callonnec, D. Dehesdin, P. Le Dosseur; Rouen/FR ([email protected]) Purpose: Otologic manifestations are frequent in the 22q11 deletion syndrome, especially chronic otitis with subsequent conductive hearing loss. Sensorineural hearing loss (SNHL) seems to be more frequent (about 10%) than in the general population. In this study, we aimed to assess the prevalence of inner ear malformations in that population. Methods and Materials: From 2003 to 2006, 20 patients with proven 22q11 dele-

B-410

15:21

Magnetic resonance imaging of the developing efferent lacrimal pathways and pathogenesis of congenital dacryocystoceles P.C. Brugger, D. Prayer; Vienna/AT ([email protected]) Purpose: To study the prenatal development of the efferent lacrimal pathways with magnetic resonance imaging (MRI) with special reference to the pathogenesis of congenital dacryocystoceles (CDCC). Methods and Materials: A total of 544 MRI studies of singleton foetuses (19-40 gestational weeks (GW)) imaged on a 1.5 T system were retrospectively evaluated. Protocols included at least two high resolution T2-weighted sequences in orthogonal planes of the foetal head. Visibility of the nasolacrimal duct and lacrimal sac was assessed and the maximum transverse diameters were recorded. Results: At least one ﬂuid-ﬁlled lacrimal sac/nasolacrimal duct could be demonstrated in 32.9% of foetuses, either uni- (50.6%) or bilaterally (49.4%). Visualization was age-dependent: not seen before 24 GW, the frequency increased from 24 to 27 GW and peaked between 31 and 33 GW (54-55%). In third trimester, foetus visibility was 51%. Dimensions of the lacrimal sac varied between 1.9 and 4 mm in 87% of foetuses. CDDC (diameter 4 mm) were present in 4.5% of third trimester foetuses. Conclusion: Depending on GW and their size, the efferent lacrimal pathways can be demonstrated with increasing frequency as ﬂuid-ﬁlled structures from 24 GW onwards. This should be considered a normal ﬁnding in third trimester foetuses. Their visualization depends on amniotic ﬂuid entering through the lacrimal puncta. The present data are in accordance with the developmental steps taking place in the prenatal development of the efferent lacrimal pathways and suggest that the development of CDCC relates to variations in the way the lacrimal canaliculi enter the lacrimal sac.

14:00 - 15:30

Room L/M

Vascular

SS 715 MR angiography Moderators: I. Bargellini; Pisa/IT S.G. Ruehm; Los Angeles, CA/US

B-411

14:00

High spatial and temporal resolution MRA of the entire peripheral vascular system using a new 3D time-resolved MRA technique (TWIST) F.M. Vogt1, S. Maderwald1, H.H. Quick1, R. Kroeker2, M. Schmidt2, J. Barkhausen1; 1 Essen/DE, 2Erlangen/DE (ﬂ[email protected]) Purpose: To develop a triple injection scan protocol for dynamic, high-resolution MRA of the entire peripheral vascular system applying a new time-resolved 3D MRA sequence (TWIST). Methods and Materials: Ten patients with PAD underwent dynamic 3D MRA at three consecutive, slightly overlapping stations. Imaging was performed on a 1.5 T system (Avanto, Siemens). 25 consecutive T1w 3D datasets were acquired in coronal planes following injection of 5 cc Gadovist at 3 cc/sec for each station using the TWIST sequence. Spatial and temporal resolution were adapted for each station: Abdominal/pelvic station: (1.3x1.3x1.3 mm³; 2.3 s); thighs: (1.3x1.3x1.3 mm³; 2.0 s); lower limbs: (1x1x1 mm³; 2.0 s). The MRA data sets were evaluated by two experienced radiologists in consensus and all signiﬁcant stenoses ( 50%) as also vessel occlusions were recorded. DSA served as the standard of reference. Results: No technical or reconstruction problems occurred. A total number of up

A

B

C

D

E

F

G

231

Saturday

Purpose: Successful in utero and in vivo diffusion tensor imaging (DTI) and tractography of the fetal brain seems to depend on various factors. The aim of this study is to assess variables affecting in utero DTI image quality. Methods and Materials: An axial DTI sequence (32 diffusion encoding directions, voxel size 1.8x1.8x5 mm, acquisition time 1-2 minutes) acquired in 40 fetuses (aged 18-37 gestational weeks) was evaluated concerning artefacts, maternal breathing movements (MBM) and fetal head movements (FHM). Furthermore, the mean signal intensity of the fetal brain (SIbr) versus surrounding “noise” (SIn) and the minimum distance between the closest coil element and the fetal brain (coildist) were measured using Image J. Image quality was rated as 100% if both corticospinal tracts, the callosal splenium and genu could be visualized. Results: In 13 of 36 fetuses with their heads engaged in the lesser pelvis, all major trajectories could be visualized. In fetuses with breech presentation 75% (two cases), 50% (one case) and 0% (one case) of tracts were found. FHM were present in 12.5% of fetuses, but only one case with signiﬁcant MBM. The most common artefact was “infolding” of maternal adipose tissues (18.6%). A coildist less than 40 mm was associated with better quality, whereas SIbr and SIn statistically did not inﬂuence trajectory visualization. Conclusion: Cephalic presentation and a minimized distance between the coil and the fetal head are important prerequisites for successful in utero tractography. In this fetal position MBM and FHM are negligible.

tion syndrome underwent a temporal bone CT scan reviewed by two radiologists experienced in ENT imaging. Every patient underwent otolaryngologic examination focused on auditory and respiratory evaluation. Results: Inner ear malformations were demonstrated in 11 patients (55%). Bilateral lateral semicircular canal (SCC) dysplasia were present in all: enlarged and dilated vestibule-lateral SCC in eight (40%) or dilated vestibule fused to lateral SCC in three (15%). Two patients (10%) had audiologic abnormalities (sensorineural and conductive hearing loss). In one girl with chronic otitis and mental retardation, CT revealed a bilateral lateral SCC dysplasia. The diagnosis of 22q11 deletion syndrome suspected on the basis of imaging ﬁndings was subsequently conﬁrmed. Conclusion: Bilateral lateral semicircular canal dysplasia seems to be very frequent in children with the 22q11 deletion syndrome and may be a part of it. Imaging should be routinely performed in those patients.

Scientific Sessions to 2000 images per station were reconstructed in about six minutes. Due to the dynamic data acquisition venous overlay did not hamper the assessment of the arterial system in any station. Based on the evaluable segments (274/290, 16 segments not evaluable on DSA) signiﬁcant stenoses and occlusions were determined with excellent correlation (r2=0.976) using DSA as the standard of reference. Conclusion: The triple TWIST protocol provides perfect arterial opaciﬁcation of all vessels without any timing issues and functional information combined with high resolution morphologic information.

B-412

14:09

How to measure SNR and CNR in modern CE-MRA using image postprocessing: Are SNR and CNR calculations reliable? B. Buerke1, T. Allkemper1, H. Kugel1, C. Bremer1, S. Evers1, H. Kooijmann2, W.L. Heindel1, B. Tombach1; 1Muenster/DE, 2Hamburg/DE ([email protected]) Purpose: To evaluate image quality of CE-MRA, SNR and CNR calculations are established based on signal intensity (SI) and standard deviation (SD) measurements in the vessel, the surrounding tissue and the background noise outside the body. However, modern MR-scanners use software algorithms (e.g. CLEAR=ConstantLEvelAppeaRance, Philips Medical Systems, Netherlands) primarily used in parallel-imaging to increase image quality, which may alter SI measurements. The purpose of this intraindividual comparison study was to evaluate the comparability of established SNR/CNR calculations in MRA using CLEAR software algorithm. Methods and Materials: Supra-aortic CE-MRA (Gadovist 0.1 mmol/kg b.w.;2 ml/s) of 11 patients were performed at 1.5 T (Gyroscan Intera®, Philips Medical Systems, Netherlands) using a standard and a CLEAR-based post processed data set. Qualitative analysis regarding image quality and contrast was performed. For quantitative analysis SI and SD were measured in both data sets for intraindividual comparison of SNR and CNR calculations regarding different reference structures (CCA, ICA, soft tissue, background outside the body, reference standard). Results: The qualitative analysis documented an improved image quality and contrast for CLEAR data sets. SNR and CNR calculations of the CCA and the C1 segment were signiﬁcantly different for both reconstruction algorithms when using the background noise outside the body as reference. SNR and CNR calculations based on soft tissue or a reference standard were comparable. Conclusion: For comparative analysis of CE-MRA data sets, SNR and CNR calculations based on SI measurements of the background noise are not applicable for CLEAR post processed algorithms. However, SI measurements in the periarterial tissue enable comparable quantitative analysis.

B-413

14:18

Diagnostic accuracy of contrast-enhanced MR angiography with gadobenate dimeglumine G.K. Schneider1, C. Ballarati2, L. Grazioli3, R. Manfredi4, S. Thurnher5, M.A. Kirchin6; 1Homburg a.d. Saar/DE, 2Como/IT, 3Brescia/IT, 4Verona/IT, 5 Vienna/AT, 6Milan/IT ([email protected]) Purpose: To determine the diagnostic accuracy of contrast-enhanced MR angiography (CE-MRA) with four doses of gadobenate dimeglumine (Gd-BOPTA) for the detection of clinically signiﬁcant steno-occlusive disease versus digital subtraction angiography (DSA). Methods and Materials: Eighty-four patients with suspected disease of the renal (n=16), pelvic (n=41) or carotid (n=27) arteries underwent CE-MRA (3D-spoiled gradient-echo sequences at 1.5 T) and DSA. CE-MRA was performed with GdBOPTA 0.025, 0.05, 0.1 and 0.2 mmol/kg (21, 25, 17 and 21 patients, respectively). Sensitivity, speciﬁcity, and accuracy versus DSA for detection of signiﬁcant disease ( 50% stenosis/occlusion for renal and pelvic arteries; 70% stenosis/occlusion for carotid arteries) was determined by three independent and fully blinded radiologists. All comparisons were tested statistically (McNemar and Chi-square tests) and reader agreement (kappa) was assessed. Results: Sensitivity values of 44.4-70.4, 55.9-67.6, 84.2% (all readers) and 65.2-69.6% were determined for Gd-BOPTA doses of 0.025, 0.05, 0.1 and 0.2 mmol/kg bodyweight, respectively. Speciﬁcity (78.2-90.3, 83.3-91.3, 96.9-99.2 and 83.8-86.5%) and accuracy (75.5-82.1, 80.4-87.0, 95.2-97.3 and 81.3-82.8%, respectively) values similary favored a Gd-BOPTA dose of 0.1 mmol/kg bodyweight. The greater accurancy of the 0.1 mmol/kg dose was signiﬁcant (p 0.01, all readers) compared to all other dose groups. Agreement between the three readers was good to excellent for all dose groups (kappar0.60). Conclusion: Signiﬁcantly better diagnostic performance on CE-MRA of the renal, pelvic and carotid arteries is achieved with Gd-BOPTA dose of 0.1 mmol/ kg bodyweight.

232

A

B

C

D

E

F

G

B-414

14:27

Intra-individual open-label international multicenter study of magnetic resonance angiography (MRA) with the blood pool contrast agent gadofosveset and conventional extracellular contrast agents with intra-arterial digital subtraction angiography as standard of reference in patients with peripheral artery disease F.M. Vogt1, W.A. Willinek2, M. Campos Coy3, F. Ogresta4, E. Salomonowitz5, J. Sánchez Cortazar6, K.F. Kreitner7, H.J. Stahlke Junior8, T. Voigtländer9, A. Beresñak4, S. Rossi4; 1Essen/DE, 2Bonn/DE, 3Monterrey/MX, 4 Buenos Aires/AR, 5St. Pölten/AT, 6Mexico/MX, 7Mainz/DE, 8Curitiba/BR, 9 Frankfurt/DE (ﬂ[email protected]) Purpose: To demonstrate non-inferiority of Gadofosveset (Vasovist®, VV) enhanced magnetic resonance angiography (VV-MRA) versus contrast-enhanced MRA with extracellular contrast media (ECCM-MRA) for assessment of peripheral artery disease using intra-arterial digital subtraction angiography (DSA) as standard of reference (SOR). Methods and Materials: Prospective intra-individual comparison of diagnostic efﬁcacy between ECCM-and VV-MRA using DSA as SOR; off-site blinded read for MRA (average reader) and DSA (one reader). Of 264 recruited patients, 211 were included in efﬁcacy analysis. Up to 21 vessel segments per patient were evaluated regarding diagnostic efﬁcacy for ﬁrst-pass MRA with ECCM (FP-ECCM) and combined assessment of FP and steady-state VV-enhanced MRA (CA-VV). Primary endpoint (diagnostic accuracy) was deﬁned as agreement for stenosis categories (= 50%, 50%- 100%, and 100%) based on SOR. Non-inferiority of CA-VV and FP-ECCM was concluded if the lower limit of the 95% conﬁdence interval (95% CI) of the difference between these two groups was greater than -10%. Results: Based on 2864 vessel segments with SOR, the diagnostic accuracy of CA-VV was 85% and of FP-ECCM 86% with a 95% CI of the difference of -3% to 1%. Number of non-assessable segments was lower for CA-VV than for FP-ECCM in particular for the lower calf vessels (5-18% versus 9-28%). Improvement of vessel wall delineation with VV was observed in 4-15% for FP and 6-57% for CA. Conclusion: This study successfully proved non-inferiority of diagnostic accuracy of CA-VV MRA versus ﬁrst-pass ECCM-MRA. Advantages in secondary endpoints were found for VV-MRA particularly regarding vessel wall delineation and venous imaging.

B-415

14:36

MR angiography with gadobenate dimeglumine (MultiHance): Results of multi-institutional phase III program G. Pirovano1, M.A. Kirchin2, S. Thurnher3, G. Soulez4, N. Anzalone2, G. Schneider5, R. Iezzi6, G. Bongartz7, P.C. Douek8, S. Miller9; 1Princeton, NJ/US, 2Milan/IT, 3 Vienna/AT, 4Montreal, QC/CA, 5Homburg/Saar/DE, 6Chieti/IT, 7Basle/CH, 8Lyon/FR, 9 Tübingen/DE ([email protected]) Purpose: To assess the accuracy of contrast-enhanced MR angiography (CE-MRA) for detecting signiﬁcant peripheral, renal and carotid artery occlusive disease using DSA as gold standard. Methods and Materials: Patients with peripheral (n=272), carotid (n=238) or renal (n=268) artery disease underwent CE-MRA and DSA within 30 days. CE-MRA was performed at 1.5 T using a 3D-SPGRE sequence after administration of 0.1 mmol/ kg gadobenate dimeglumine (MultiHance). Three independent experienced fully blinded radiologists per study evaluated MRA images while a fourth reader evaluated DSA exams. The CE-MRA technical failure rates were calculated and compared to DSA (chi-square test). Sensitivity, speciﬁcity, accuracy, and inter-reader agreement for detection of signiﬁcant disease (r51% peripheral/renal; r61% carotid) were calculated and compared (McNemar's test). Results: Sensitivity, speciﬁcity and accuracy values of 54-86%, 89-95% and 80-87% and good 3-reader agreement (kappa: 0.64-0.69 and % agreement: 80-85) was demonstrated for CE-MRA across all vascular territories. Overall, 89 mild (~98%) or moderate (~2%) adverse reactions considered possibly related to MultiHance were reported for 7.5% (62/828) of subjects, the most common being injection site warmth. Other safety ﬁndings were unremarkable. Conclusion: CE-MRA with 0.1 mmol/kg MultiHance is safe and highly accurate for routine CE-MRA applications. Results for MultiHance compare favorably with Phase III clinical results for other recently approved MR contrast agents.

Scientific Sessions B-416

14:45

High resolution 4D-MRA of the brain using a centra keyhole approach B. Tombach1, B. Buerke2, T. Allkemper2, T. Niederstadt2, D. Kooijmann3, R. Hoogeveen4, W.L. Heindel2; 1Osnabrück/DE, 2Münster/DE, 3Hamburg/DE, 4 Best/NL ([email protected])

B-417

14:54

Whole-body contrast-enhanced MR angiography with gadobenate dimeglumine: Evaluation of subclinical atherosclerotic burden in patients with steno-occlusive disease A. Napoli, M. Anzidei, B. Cavallo Marincola, C. Zini, P.L. Di Paolo, K. Lanciotti, C. Catalano, R. Passariello; Rome/IT ([email protected]) Purpose: To determine the value of high-resolution WB-MRA and total cardiovascular burden in patients undergoing assessment of supra-aortic or peripheral artery steno-occlusive disease. Methods and Materials: Ninety-four patients with suspected steno-occlusive disease of carotid (n=58) or peripheral (n=36) vessels underwent high-resolution WB-gadobenate dimeglumine-enhanced MRA at 1.5 T using 3D-spoiled GRE sequences, parallel imaging and 2x acceleration factor. Sixty-eight surface coil elements and a four-station protocol were employed. Biphasic injection (10 ml @ 1 ml/s followed by 10 ml @ 0.5 ml/s) of gadobenate dimeglumine was performed. The arterial system was divided into 32 segments for evaluation. The presence of steno-occlusive disease, determined at all segments, was compared with ﬁndings from CT or DSA. Segments were classiﬁed as having clinically signiﬁcant disease (r50% diameter reduction) or no signiﬁcant disease ( 50%). Presence or absence of atherosclerotic lesions was evaluated by two observers in consensus. Results: In 58 patients with carotid atherosclerotic disease, 23 (40%) presented extra-carotid lesions that required further diagnostic work-up. Six (10%) of these 23 patients had signiﬁcant stenosis. In 36 patients with peripheral arterial occlusive disease, subclinical lesions were found in 16 (44%); supra-aortic arteries were the most involved region. Signiﬁcant lumen reduction was found in four (11%) patients, one of whom underwent un-programmed surgery (TEA). Overall, WB-MRA revealed additional atherosclerotic lesions in 39 patients, of which 10 required immediate care. Conclusion: WB-CE MRA with gadobenate dimeglumine may be considered a robust modality for imaging atherosclerosis in patients requiring non-invasive assessment of supra-aortic or peripheral artery steno-occlusive disease. The subclinical detection of total atherosclerotic burden has potential for secondary care in this population.

B-418

15:03

High-resolution peripheral MRA featuring a new blood-pool contrast agent and venous signal suppression due to k-space segmentation S. Kinner1, M.O. Zenge2, F.M. Vogt1, M.E. Ladd1, J. Barkhausen1, H.H. Quick1; 1 Essen/DE, 2Erlangen/DE ([email protected]) Purpose: The purpose of our study was to acquire high-resolution contrastenhanced peripheral MRA using a new blood-pool contrast agent (Vasovist) while suppressing venous signal by using an acquisition scheme with k-space segmentation.

B-419

15:12

Foot angiography in PAOD patients: Blood pool contrast media (Vasovist®) CE-MRA vs standard DSA S. Kos, C. Reisinger, A.L. Jacob, G. Bongartz, D. Bielecen; Basle/CH ([email protected]) Purpose: Evaluate visualization of foot vessels and its stenosis or occlusion in a time-resolved sequence using the intravascular agent Vasovist®, compared with standard DSA in PAOD patients. Methods and Material: We prospectively evaluated 20 patients with peripheral artery occlusive disease (PAOD), Fontaine stage I-II. All patients received standard one-sided antegrade DSA and if necessary PTA with foot vasculature depiction (two projections). High-resolution CE-MRA with the blood-pool contrast agent Vasovist® was performed the next day. We used a 1.5-T whole-body scanner with a time resolved 3D gradient-echo sequence (TR 3.81 ms, TE 1.38 ms) and a dedicated foot coil. Eight data sets with an acquisition time of 25 seconds each were obtained consecutively, with a contrast-media injection rate of 2 ml/s. Qualitative assessment of stenosis visualization and venous contamination were performed in comparison between the two techniques regarding the following criteria: a) diagnostic usability: 1-good, 2-suboptimal and 3-not usable, b) venous contamination: 1- no venous signal, 2-minmal venous signal, 3-strong venous contamination, c) Grading of stenosis: 0-no stenosis, 1-moderate stenosis (1-49%), 2-relevant stenosis (50-99%), 3-occlusion. Results: We demonstrate the ﬁrst comparison between CE-MRA with blood pool contrast media (Vasovist®) and standard DSA regarding the depiction of foot vessels and associated stenosis or occlusion in PAOD patients. We show that CE-MRA using Vasovist® provides very good results (high sensitivity and speciﬁcity values), comparable to the actual gold-standard, the DSA. Conclusion: Vasovist® CE-MRA may well be used for the foot vessel assessment in PAOD patients.

B-420

15:21

MR angiography in rabbits: Comparison of a new blood pool contrast agent with Dotarem® S. Kinner1, S. Maderwald1, N. Parol1, S.L. Göricke1, J. Albert1, C. Corot2, P. Robert2, J. Barkhausen1; 1Essen/DE, 2Aulnay sous Bois/FR ([email protected]) Purpose: MR angiography is routinely performed using extracellular Gd-based compounds, although these agents have several limitations including short intravascular half-life time requiring bolus timing and the risk of nephrogenic systemic ﬁbrosis. Our study aimed to compare a new iron-based blood-pool contrast agent (P904, Guerbet, France) with Dotarem® as non-speciﬁc reference. Methods and Materials: Six rabbits underwent two MRA examinations on different occasions using Dotarem® (0.1 mmol Gd/kg) and P904 (0.05 mmol Fe/kg) in random order. All measurements were performed on a 1.5 T MR (Magnetom Espree®, Siemens, Germany). The rabbits were placed between two phased array coils. 3D-ce-MRA (TR=5.1; TE=1.78; ﬂip angle=30°; slice thickness=0.7 mm; acquisition time=20 sec; FOV=210 mm) of the abdominal aorta, including visceral side branches and iliac arteries were performed prior to the application of the contrast medium, during the bolus phase and every minute up to 10 minutes thereafter. Two experienced radiologists evaluated the images qualitatively by consensus (4=excel-

A

B

C

D

E

F

G

233

Saturday

Purpose: To investigate the feasibility of high resolution 4D CE-MRA of the brain using a combination of parallel imaging (SENSE) and CENTRA keyhole data acquisition compared to non-enhanced techniques (TOF and PCA). Methods and Materials: Using a SENSE-capable head coil a keyhole-technique was evaluated in a total of 34 patients suspicious of intracranial vascular pathologies on a clinical 1.5 T scanner (Intera, Philips, Best, The Netherlands). The intracranial vessels were imaged with 0.1 mmol/kg bw 1 M Gadobutrol (ﬂow rate 2 cc/sec, followed by 20 ml saline ﬂush, ﬁxed delay 10 sec) using four dynamic keyhole scans (TA 7.3 sec) followed by a 40.8 sec acquisition of peripheral k-space lines. Qualitative evaluation was performed by two experienced radiologists in consensus with regard to arterial contrast, venous overlay and overall image quality. Quantitative evaluation included SNR and CNR measurements. Results: The keyhole technique with a temporal resolution of 7.3 sec and a high spatial resolution of 0.39x0.39x0.7 mm (reconstructed voxel size) allowed a discrimination of arterial and venous phases in 30/34 patients and an excellent image contrast (34/34). Compared to non-enhanced techniques the total acquisition time of 63 sec was signiﬁcantly shorter and the diagnostic impact was much higher, for e.g., exclusion of intracranial venous thrombosis. Conclusion: The keyhole technique as a robust diagnostic imaging tool for intracranial CE-MRA allows the discrimination of arterial and venous phases within a total acquisition time of 63 sec and offers the potential to replace TOF- and PC-angiography.

Methods and Materials: Scanning was performed on a 1.5 T Magnetom Avanto (Siemens Medical Solutions, Erlangen). Multi-station peripheral MRA with Vasovist (Bayer-Schering Pharma, Berlin) was performed in eight patients with known PAOD. k-space of the 3D data sets was segmented such that the central parts of k-space were acquired during the ﬁrst-pass of the blood-pool contrast agent, while the peripheral parts of k-space were acquired in the steady-state of the contrast dynamics. First-pass 3D MRA data sets and higher-resolution 3D data sets, combining the ﬁrst-pass and steady-state data, were displayed for direct image comparison.Two blinded radiologists compared according to a three-point scale vessel sharpness (1=blurred, 3=sharp edged), vessel conspicuity and venous contamination in the ﬁrst-pass and fused images. Results: Comparison of vessel sharpness revealed statistically signiﬁcant differences for the ﬁrst-pass data sets and the fused data sets in all three stations (1.8 vs 2.9). Vessels were depicted in the fused images over longer parts while veins were almost completely suppressed in both data sets. Conclusion: Peripheral CE-MRA with ﬁrst-pass and steady-state k-space segmentation using a blood-pool contrast agent proved feasible and provided highresolution data with sharp delineation of the arterial vessel tree while suppressing venous overlay. Further developments aim towards combining this technique with data acquisition during continuous table movement.

Scientific Sessions lent up to 1=non-diagnostic). SNR and CNR of the arteries were calculated. For statistical evaluation a Wilcoxon test was used. Results: In the bolus phase the average aortic SNR and CNR were signiﬁcantly higher with Dotarem® than with P904 (p=0.03). Ten minutes after injection P904 demonstrated an approximately threefold higher aortic SNR and nearly ﬁvefold higher CNR compared to Dotarem® (p=0.011). Image quality (IQ) was rated excellent for P904 images at all time points whereas with Dotarem® only the bolus phase was rated excellent. Conclusion: P904 as new blood pool agent appears to be well suited for highquality ﬁrst pass and steady-state MRA.

14:00 - 15:30

SS 711 Grey and white matter Moderators: S.P. Morozov; Moscow/RU Z. Patay; Memphis, TN/US

14:00

Voxel based morphometry and voxel based relaxometry in Parkinsonian variant of multiple system atrophy L.C. Tzarouhi, L.G. Astrakas, S. Konitsiotis, P.N. Margariti, S. Tsouli, M.I. Argyropoulou; Ioannina/GR ([email protected]) Purpose: Multiple System Atrophy (MSA) is a sporadic, progressive, adult-onset neurodegenerative disorder characterized by several degrees of parkinsonism, cerebellar ataxia and autonomic failure. Clinically, a parkinsonian variant (MSA-P) and a cerebellar variant (MSA-C) are distinguished. Our purpose was to study atrophy and to investigate brain iron content by using Voxel Based Morphometry (VBM) and Voxel Based Relaxometry (VBR), respectively, in MSA-P. Methods and Materials: We studied 11 MSA-P (9 men, 3 women, age 61.9p11.7 years, range 38-79) and 11 age and sex matched controls. VBM was applied to three-dimensional high resolution T1-weighted images (TR=25 sec, TE=4.6 msec, ﬂip angle=30, slice thickness=0.8 mm, acquisition matrix=256, reconstruction matrix=256) using SPM5. The images were segmented in grey matter, white matter and cerebrospinal ﬂuid compartments, smoothed and transformed to a normalized space. VBR was used to estimate the T2 relaxation rate (R2) derived from multiecho T2-weighted images (TR=2200 sec, TE=32-112 sec, slice thickness=5 mm, gap=0.5, acquisition matrix =156, reconstruction matrix=256). A p value equal to 0.001 uncorrected for multiple comparisons was used. Results: Patients compared to controls showed signiﬁcant decrease of grey matter in the anterior and posterior cerebellar lobes, thalami, caudate nuclei, lentiform nuclei, middle frontal gyri, precentral gyri, inferior parietal lobules and of white matter in small areas of the pons. There was no signiﬁcant difference in the R2 between patients and controls. Conclusion: In MSA-P, grey matter atrophy of the cerebellum, basal ganglia and motor cortex is a main ﬁnding but does not coexist with any changes in brain iron content.

B-422

14:09

Differentiation of idiopathic Parkinson’s disease from multi system atrophy with routine MR imaging and diffusion weighted MR imaging A.A.A. Abdel Razek, A. Elmongy, M. Hazem, W. Gabr; Mansoura/EG ([email protected]) Purpose: To assess the role of routine MR imaging and diffusion weighted MR imaging in differentiation of idiopathic Parkinson’s disease (IPD) from Parkinson’s variant of multisystem atrophy (MSA). Methods and Materials: Prospective study was conducted upon elderly patients with IPD (n=55), MSA (n=15) & control subjects (n=30). Routine T1 and T2 weighted MR imaging of the brain was done. Diffusion MR weighted imaging was done on 1.5 Tesla unit (Symphony, Siemens) using a single shot spin echo type of echo planar imaging. The parameters used were TR/TE: 10000/38 msec, B value=0, 500 & 1000 sec/mm2. The ADC map was reconstructed. The signal intensity on T2 weighted images and ADC values of the substania nigra and putamen were calculated. Results: There was statistically signiﬁcant difference in ADC values between patients with MSA and IPD within putamen (p 0.001) and insigniﬁcant difference within substania nigra (p 0.6). Also, there was statistically signiﬁcant difference in signal intensities between patients with MSA and IPD within putamen (p 0.004)

234

A

B-423

14:18

Changes of white matter in aging and early Alzheimer’s disease detected by diffusion tensor imaging H. Ni1, V. Kavcic2, J. Zhong2, J. Qi1; 1Tianjin/CN, 2Rochester, NY/US ([email protected])

Room N/O

Neuro

B-421

and insigniﬁcant difference within substania nigra (p 0.07). There is signiﬁcant correlation between ADC value and signal intensity of putamen (r=0.89, p 0.002). Conclusion: We concluded that signal intensity on T2 weighted images and apparent diffusion coefﬁcient value of the putamen can be used to differentiate idiopathic Parkinson’s disease from multisystem atrophy.

B

C

D

E

F

G

Purpose: To investigate white matter (WM) changes in both aging and early stage of the Alzheimer’s disease (AD) using diffusion tensor imaging (DTI). Methods and Materials: DTI (diffusion weightings were applied in 20 orientations with b=1000 s/mm2) was performed with 20 young and middle age subjects (YN, mean age=38.1, age range 23-57), 18 elder subjects (ON, mean age=75.6, age 60; mean MMSE=29 with range 27-30) and 14 mildly impaired AD patients (mean age=75.4, age 60; mean MMSE=21 with range 20-24 to ensure AD patients were in early stage). Fractional Anisotropy (FA) and mean diffusivity (D) were computed and compared on nine regions of interest (ROI): Subcortical WM in the prefrontal and parietal-temporal regions; Corpus Callosum (CC): at genu, body and splenium sites; Cingulum and Superior Longitudinal Fasciculus in middle and posterior sites. Results: Compared with YN, ON showed signiﬁcantly increased D in all ROIs ( 10%, p 0.0005), while FA signiﬁcantly decreased only in genu and body of CC ( 8%, p 0.004). AD patients, as compared to ON, showed signiﬁcant decrease in FA in the parietal-temporal subcortical WM and splenium ( 6%, p 0.04), and signiﬁcant increase in D in splenium (5%, p=0.05). Conclusion: DTI showed greater WM changes related to aging than AD. Results also indicated that with age there was a global and widespread increase in diffusivity, while decrease in anisotropy was observed only at the anterior and middle portions of the CC. AD-related WM changes at the posterior sites may reﬂect the early spread of AD neuropathology.

B-424

14:27

The potential role of diffusion tensor MR imaging in assessing the differential diagnosis between mild cognitive impairment and Alzheimer´s disease D.B. Parente, L.C.H. da Cruz Jr, E.L. Gasparetto, R.C. Domingues, T. Doring, R.S. Borges, R.C. Domingues; Rio de Janeiro/BR ([email protected]) Purpose: To evaluate the fractional anisotropy (FA) values of several white matter tracts, aiming to differentiate the normal population from mild cognitive impairment (MCI) and Alzheimer’s disease (AD) patients. Methods and Materials: Seventy-nine patients with memory impairment (MCI, possible AD, and probable AD) and 16 age- and gender-matched controls were studied. Magnetic resonance imaging studies were performed on a 1.5 T scanner. Conventional images and diffusion tensor imaging (DTI) were obtained in all participants. The DTI data were post-processed and based on FA maps; FA values were measured in ROIs positioned in the hippocampal formations, superior longitudinal fascicles, posterior cingulate gyri, and the splenium of the corpus callosum. Kruskal-Wallis and Steel-type multiple comparisons non-parametric tests were applied for the statistical analysis. Relations between variables were assessed with non-parametric Spearman and Pearson correlations. Results: The FA values for the splenium of the corpus callosum, bilateral posterior cingulate gyri, and bilateral superior longitudinal fasciculi of MCI and probable AD patients were signiﬁcantly lower than controls (p=0.01 and 0.03). No signiﬁcant difference was noted in FA values when comparing MCI versus possible AD and probable AD, and probable AD versus possible AD. No differences in FA values were seen in hippocampal formations among patients from all groups. Conclusion: In conclusion, the anatomical regions ﬁrst involved in patients with cognitive impairment demonstrated reduced FA values compared with controls. However, the FA values did not allow the differentiation between MCI, possible and probable AD.

Scientific Sessions B-425

14:36

Proton MR spectroscopy and diffusion tensor imaging in the posterior cingulated regions of patients with Alzheimer’s disease: A preliminary study B. Ding, H.-W. Ling, H. Zhang, W.-M. Chai, K.-M. Chen; Shanghai/CN

B-426

14:45

The role of diffusion tensor imaging (DTI) in the detection of white matter lesions in patients with mild cognitive impairment D. Sali1, D.A. Verganelakis2, J. Papatriantafyllou1, C. Karageorghiou1, T. Thomaides1, E. Kapsalaki3, G. Hadjigeorghiou3, A. Papadimitriou3; 1Athens/GR, 2 Halandri/GR, 3Larissa/GR ([email protected]) Purpose: Numerous neuro-imaging studies have been employed in the evaluation of patients with mild cognitive impairment. Diffusion Tensor Imaging (DTI) is a non invasive MRI technique used to visualize white matter tracts. Quantitative measurements of Fractional Anisotropy (FA) and Apparent Diffusion Coefﬁcient (ADC) can also be measured using the above technique. The purpose of our study is to examine the role of DTI in the detection of white matter lesions in patients with Mild Cognitive Impairment (MCI). Methods and Materials: In our prospective study, eight patients with MCI, (mean age 68 yrs) and 10 normal volunteers (mean age 72 yrs) underwent brain MRI including DTI technique. The diagnosis of MCI in our patients was based on Petersen’s criteria (1999). All patients and volunteers also underwent neuropsychological tests (MMSE, 9w-California Verbal Learning Test, Modiﬁed Rey). MRIs were performed on a 1.5 T MR scanner. Fractional anisotropy and diffusion of the white matter tracts were measured in all patients and volunteers at standard anatomical areas (genu and splenium of the corpus callosum, posterior cingulum, superior longitudinal fasciculus) in both hemispheres. Results: FA was smaller at the genu in patients compared to volunteers (p=0.019) and ADC was larger (p=0.028). At the right cingulum ADC was larger in patients than in volunteers (p=0.019) as well as at the left, but not in a statistically signiﬁcant manner (p=0.056). Conclusion: Our initial results suggest that FA and ADC can detect white matter changes and could be a useful tool in the diagnosis of patients with MCI.

B-427

14:54

Brain lesions in septic shock: A magnetic resonance imaging (MRI) study D. Mompoint, E. Kocheida, S. Mokhtari, M. Louvet, T. Sharshar, C.A. Vallée, R.Y. Carlier; Garches/FR ([email protected]) Purpose: Septic shock related mortality and morbidity are partly ascribable to brain dysfunction whose clinical diagnosis can be difﬁcult and mechanisms still remain hypothetical. This study assesses the usefulness of brain MRI in septic shock. Methods and Materials: Brain MRI was requested in septic shock patients who required a body MRI or developed brain failure. Brain MRI included gradient echo T1 weighted, ﬂuid-attenuated inversion recovery (FLAIR), T2 weighted and diffusion weighted images (DWI) with Apparent Diffusion Coefﬁcient (ADC). Patients with contraindications for MRI were excluded. Results: Fifteen consecutive mechanically ventilated patients (women: 7, age: 68p13 years) underwent brain MRI, notably for neurological reasons in nine (66%).

B-428

15:03

Posterior reversible encephalopathy syndrome: A preliminary study of the relationships between its cause, radiologic feature and clinical outcome B. Gao; Yantai/CN ([email protected]) Purpose: To investigate the effects of of clinicoradiologic factors on the reversibility of posterior reversible encephalopathy syndrome (PRES). Methods and Materials: The clinical and radiologic data were analyzed retrospectively in 25 patients, who were divided into ﬁve groups according to their clinical causes: eclampsia group (13 patients), hypertension group (4), uremia group (3), vasculitis group (3), neurotoxity group (2) induced by cyclosporine A (CsA). Conventional MRI studies were performed in all 25 cases, of which diffusion-weighted imaging (DWI) studies were appreciated in 14 cases. Results: Differences between eclampsia group, hypertension group and other three groups were statistically signiﬁcant (p 0.01). The lesions located in cortical and subcortical white matter had more reversibility than those in deep structures, such as deep white matter, brain stem, basal ganglia regions and corpus callosum (p 0.01). The reversibility between the lesions with hyperintensity on DWI and those with iso-intensity or hypo-intensity on DWI had no signiﬁcant difference (x2=4.289, p 0.05). However, the reversibility between diffusion restricted lesions and unrestricted lesions on ADC maps had markedly signiﬁcant difference in either hyperintense or iso-, hypo-intense lesions on DWI (x2=27.50 and 14.13, respectively, p 0.01). Conclusion: The reversibility of PRES is not only related to the anatomical locations of the lesions but also to the clinical causes. DWI together with ADC maps would reliably estimate the prognosis.

B-429

15:12

Quantiﬁcation of intracerebral GABA by 1H-MRS: How reproducible are results? W. Bogner1, S. Gruber1, A. Stadlbauer2, M. Doelken2, S. Trattnig1, A. Doerﬂer2, H. Stefan2, T. Hammen2; 1Vienna/AT, 2Erlangen/DE ([email protected]) Purpose: Gamma-aminobutyric-acid (GABA) is an important inhibitatory neurotransmitter in human brain with anticonvulsive character. Because of an increased interest in GABA metabolism we tested the reliability of special editing sequences which allow non-invasive intracerebral measurements of this neurotransmitter in healthy adults. Methods and Materials: We determined the reproducibility of results obtained with a J-coupling-based editing sequence using MEGA suppression on a 3 TTIM-Trio (Siemens) by placing single-voxel-spectroscopy (SVS) ROIs in left/right occipital lobe of subjects. Intra- and inter-subject reproducibility was evaluated. Nine volunteers were measured in total, one proband repeatedly on several days. Side differences between left and right hemispheres were investigated. GABA/Cr and GABA/H2O-ratios were compared. Furthermore a detailed comparison of two different post-processing routines for the evaluation was accomplished. Integration of signals (GABA/Cr/H2O) was done using MestreC, compared to time-domain line-ﬁtting-routine (AMARES-jMRUI). Mean value, standard deviation (SD) and coefﬁcient of variation (CV) were calculated among groups. Results: Using data from all subjects lower CVs can be observed for line-ﬁtting compared to integration. GABA/Cr-ratio performs better than GABA/H2O or GABA alone (GABA/Cr: 13.6% and 21.55%; GABA/H2O: 14.8% and 21.68%; GABA: 20.62% and 26.42%). Single subject 4-day measurements showed slightly better intra- than inter-subject reproducibility (GABA/Cr 12.23%). No left/right difference could be found (GABA/Cr: 0.372p0.058 and 0.366p0.042). Conclusion: With a CV of ~13% for GABA/Cr, the technique turns out to be a precise and promising tool which easily detects changes of up to +300% found during antiepileptic drug intake. Due to minor liquor contamination GABA/Cr-ratios are generally more reliable than GABA/H2O and GABA alone. Automated line-ﬁtting performed better than integration.

A

B

C

D

E

F

G

235

Saturday

Purpose: To compare 1H-MRS and DTI ﬁndings in the posterior cingulated regions in patients with Alzheimer’s disease (AD) and normal elderly controls. Methods and Materials: 15 mild AD, 20 moderate-severe AD and 20 normal aging subjects were recruited. Single voxel MRS and DTI were performed on a 1.5 T MRI scanner. A region of interest was positioned in the posterior part of the cingulate. Metabolite ratios were estimated, including the ratios of NAA/Cr, Cho/Cr, mI/Cr. DTI was performed using a single shot EPI sequence, with 25 directions. Mean spectroscopy data and DTI values (fractional anisotropy, FA and mean diffusivity, MD) for each group were analysed with nonparametric test. Results: Mild AD group showed a signiﬁcantly lower FA value in the left posterior cingulum bundle than normal aging group (p 0.05) whereas moderate-severe group versus mild AD group revealed signiﬁcantly elevated MD value and a decrease in FA value in the right posterior cingulate (p 0.01). Obvious increasing mI/Cr ratio was found in mild AD. NAA/Cr ratio showed gradual decrease in AD groups. Partial correlations analysis revealed a positive correlation between mI/Cr ratio and left posterior cingulated FA value in mild AD group (r=0.553, p=0.040) and negative correlation between NAA/Cr and MD value in the right posterior cingulated region (r=-0.600, p=0.007). Conclusion: These ﬁndings suggested that there were different regional and temporal patterns in different courses of AD disease, resulting from axonal loss or gliosis. Combining MRS with DTI alternations could better explain the pathological changes in AD progression.

Neuroimaging showed multiple strokes in two patients. In one patient with disseminated intravascular coagulopathy (DIVC), multiple foci of low signal were visible on T2 images. In four patients, MR displayed small centrum semi-oval white matter (WM) lesions (1) predominating around Virchow-Robin spaces (2) multiple to diffuse and (3) with a bright signal on FLAIR images and, for the large ones, DWI hypersignal and increased ADC. In one patient, a post-mortem study disclosed diffuse vasogenic oedema. WM lesions increased in the course of septic shock and could last one month. Among patients with abnormal MRI, ﬁve died (strokes: 2, DIVC 1, WM lesions: 2) and three had a neurological disability at ICU discharge. Conclusion: MRI detects brain damage during septic shock. Visible brain lesions consist of previously described strokes and multiple haemorrhages associated with DIVC and also with non previously reported leucoencephalopathy.

Scientific Sessions B-430

B-432

15:21

Brain metabolic changes in hypertensive patients: An MR spectroscopy study S. Torres del Río, J.M. Garcia Santos, L.J. Fuentes, C. Antúnez, J.B. Vidal, G. Ortega, M.M. Antequera, J.G. Sevilla; Murcia/ES ([email protected]) Purpose: To evaluate brain metabolite changes related to arterial hypertension (HT). Methods and Materials: We evaluated 56 elderly hypertensive patients without other related abnormalities. 26 had controlled HT (P1), 16 received more than three drugs to control HT (P2), and 14 had uncontrolled HT despite pharmacological treatment (P3). 20 age-matched normal subjects were the control group. Subjects underwent four single-voxel MR spectroscopy acquisitions (PRESS TR/ TE 1500/35 ms) in the posterior cingulate gyrus (PCG), visual cortex (VIC), left frontal white matter (FWM) and left hippocampus (HIP). N-Acetyl-Aspartate (NAA), Creatine (Cr), Choline (Cho) and Myoinositol (mI) were assessed by means of metabolite ratios: NAA/Cr, NAA/Cho, NAA/mI, mI/Cr, Cr/Cho, H2O/NAA, H2O/Cr, H2O/Cho, and H2O/mI. Inter-group differences of metabolite ratios were evaluated with the ANOVA test and Bonferroni correction. Correlations between age and metabolite ratios were assessed with the non-parametric Spearman test. Metabolite dependence on sex and disease duration was also evaluated. Results: Metabolite ratios showed no sex and disease duration dependence. There were no metabolite differences between groups in the PCG, VIC and HIP. H2O/Cho ratio was different between groups in the FWM (p=0.012), and NAA/Cho showed a strong trend to a signiﬁcant difference (p 0.07). In P1, H2O/Cho ratio was higher than in P3 patients (p=0.025) and showed a strong trend to difference with P2 subjects (p=0.061). Unlike other frontal lobe ratios, NAA/Cho, Cho/Cr and H2O/Cho did not show age correlations. Conclusion: In HT patients, disease severity can be associated with changes in choline metabolism in frontal lobes.

14:00 - 15:30

Room R1

Chest

SS 704 Pulmonary embolism Moderators: A.P. Parkar; Bergen/NO E.J. Stern; Seattle, WA/US

B-431

14:00

The use of radio protective bismuth breast shields does not impair image quality in CT pulmonary angiogram examinations in adult females C.P.J. Murray; Perth/AU ([email protected]) Purpose: To determine the effect of radioprotective breast shields on the image quality of multidetector CTPA examinations in adult females. Method and Materials: 4-ply bismuth radioprotective breast shields (AttenuradTM) were applied to the anterior chest wall of 20 consecutive adult females undergoing CTPA. All patients underwent a standard protocol 64-slice CTPA. Differences in the image quality of the lungs and pulmonary arteries inside or outside the ﬁeld of the shields were assessed subjectively by two thoracic radiologists independently. Image noise inside and outside the ﬁeld of shielding was measured objectively by comparing the standard deviation in Hounsﬁeld units in the lungs, pulmonary arteries, and breast tissue using the paired t-test. The dose reduction of the breast shields was calculated using thermoluminescent detectors on an anthropomorphic phantom. Results: Beam attenuation artefact from the shields was identiﬁed in the breast tissue on all examinations, conﬁrmed by a statistically signiﬁcant increase in noise in the breast tissue in the ﬁeld versus outside the shield ﬁeld (p=0.0001). None of the CTPA examinations showed a subjective reduction in image quality. No statistically signiﬁcant increase in noise was found in the lung tissue or pulmonary arteries in the ﬁeld of shielding. The reduction in radiation dose to the breast tissue from shielding was 43%. Conclusion: Bismuth breast shields reduce radiation dose to the breast of adult females during CTPA without any signiﬁcant subjective or objective reduction in image quality. The use of bismuth breast shields for CTPA in adult females is recommended.

236

A

B

C

D

E

F

G

14:09

Comparative study on scintigraphic algorithms for the diagnosis of acute pulmonary thromboembolism based on computed tomographic pulmonary angiography: Multi-centre study for IAEA coordinated research project N. Watanabe1, N. Kücük2, J. Fettich3, P. Choudhury4, F. Mut5, O. Kraft6, K. Endo7, S. Sharma8, M. Dondi1; 1Vienna/AT, 2Ankara/TR, 3Ljubljana/SI, 4Delhi/IN, 5 Montevideo/UY, 6Ostrava/CZ, 7Maebashi/JP, 8New Delhi/IN ([email protected]) Purpose: The aim of this study was to compare two scintigraphic algorithms, namely the Prospective Investigative Study of Acute Pulmonary Embolism Diagnosis (PISAPED) and the Prospective Investigation of Pulmonary Embolism Diagnosis (PIOPED) study in patients with clinical suspicion of acute pulmonary thromboembolism (PE) based on computed tomographic pulmonary angiography (CTPA) interpretation. Methods and Materials: A total of 181 patients with clinical suspicion of acute PE was examined with ECG, blood chemistry, chest X-ray, pulmonary ventilation/ perfusion scintigraphy and spiral CTPA. The follow-up time was 6-12 months. The scintigraphic and CTPA interpretation was reviewed by different specialists, without being aware of each others’ results. Results: Positive CTPA was shown in 110 patients with suspected PE and 109 of the 181 were clinically diagnosed with PE. Based on CTPA interpretation, the sensitivity and speciﬁcity of PISAPED are 78.2% (86/110) and 69.0% (49/71), respectively, while that of PIOPED are 63.6% (70/110) and 45.1% (32/71). Conclusion: The diagnostic ability of PISAPED algorithm is higher than that of PIOPED algorithm. Pulmonary perfusion scintigraphy without pulmonary ventilation scintigraphy, linked to the diagnostic algorithm of PISAPED, may be applicable for the diagnosis of acute PE in case CTPA is not available.

B-433

14:18

Computer-aided detection of acute pulmonary embolism with dual-source CT: Impact of the scanning conditions and overall image quality in the detection of peripheral clots M. Dewailly1, M. Rémy-Jardin1, J.-B. Faivre1, A.-M. Bakai2, V. Pansini1, J. Rémy1; 1 Lille/FR, 2Forchheim/DE ([email protected]) Purpose: To evaluate the performance of a computer-aided detection (CAD) system for diagnosing peripheral acute pulmonary embolism (PE) with dual-source CT. Methods and Materials: Two radiologists investigated the accuracy of a software aimed at detecting peripheral clots (PECAD, version 7, Siemens) by applying this tool for the analysis of the pulmonary arterial bed of 67 CT angiograms obtained with dual-source CT (Deﬁnition, Siemens) using similar injection protocols. The study population included: (1) various scanning conditions (nongated (n=20), ECG-gated (n=33) and dual-energy CT (n=14) angiograms); (2) various image qualities (IQ): Group 1 (excellent IQ; n=42), and Group 2 (lower IQ due to lower levels of arterial opaciﬁcation and/or presence of noise; n=25). Results: The software detected 61 out the 75 peripheral clots present in the 20 patients (60 segmental and 15 subsegmental clots). The sensitivity of the CAD tool was 81.3% for detecting peripheral clots, 78.3% at the segmental and 93.3% at the subsegmental levels. A mean number of 6.4 false-positives was found per patient (total: 423 false-positives), mainly linked to the presence of hilar lymph nodes (n=133; 31%) and perivascular airspace consolidation (n=65; 15%). No signiﬁcant difference was found in the sensitivity of the CAD software between Group 1 and Group 2 (p=0.83). The sensitivity for the CAD tool was 84% for ECG-gated scans, 82% for dual-energy scans and 76% for nongated scans (p=0.86). Conclusion: The evaluated CAD software is a sensitive tool in detection of peripheral PE, not inﬂuenced by the scanning conditions or the overall image quality.

B-434

14:27

Incidence and radiological characteristics of falsely highlighted lesions using CAD to detect lung embolism A. Malich1, D. Hentrich1, A. Ulrich1, P. Hannemann2, M. Ibe1, S. Mikulik1, A. Kott1, A. Hansch3; 1Nordhausen/DE, 2Neustadt/DE, 3Jena/DE ([email protected]) Purpose: To characterize false positive markers set by CAD on thoracic CT images obtained to verify PE. To compare CAD-detected number of emboli vs radiologists. To verify affection of number of CAD-detected emboli and severity of PE. Methods and Materials: 160 CT examinations performed to verify suspected PE were analyzed using ImageCheckerLung (V 3.1, Hologic). Two radiologists in consensus analyzed images without CAD-results. Highlighted lesions were analyzed radiologically in order to characterize false markers and to prove feasibility of CADapplication. D-Dimer was obtained and matched to the number of detected emboli

Scientific Sessions by CAD using group1: 0 marks; group2: 1-3 marks; group3: 4-6 marks; group4: 7-9 marks; group5: 9 CAD-marks. Results: 609 CAD markings were set (3.8/case). Radiologists marked 227 emboli. In total in four cases lymphatic tissue, in 136 cases perivascular tissue, in 11 cases vessels and in 24 cases lung scars/ﬁbrosis were marked falsely by CAD. CAD highlighted peripheral PE not detected by the radiologists in 28 cases and did not highlight two large central PE. It remains difﬁcult to judge whether CADmarker in the periphery close to the pleura highlights a thrombosed pulmonary vessel. 3D-analysis and enlargement are of use to clarify these markers. D-Dimer increases with increasing number of CAD-marks (group 1-5: 0.568; 0.783; 0.920; 1.580; 1.230, respectively). Conclusion: CAD allows a better analysis of peripheral PE on lung-CT. Majority of false positive markers highlight perivascular soft tissue. There is an association of the total number of CAD-markers/case with D-Dimer (and thus severity of the disease).

B-435

14:36

Purpose: To assess the incremental value of pelvic vein evaluation with Computed Tomographic Venography (MDCT-V) for the diagnosis of thromboembolic disease in patients suspected of Pulmonary Embolism (PE). Methods and Materials: We retrospectively reviewed CT images of 440 patients (228 women and 212 men; mean age 62p18 years) who underwent combined Multidetector Computed Tomographic Pulmonary Angiography (MDCT-PA) and MDCT-V for suspected PE. Frequency of PE, deep venous thrombosis (DVT) and venous thromboembolism (VTE), and the incremental value of MDCT-V were recorded; the distribution of thrombi and the frequency of isolated clots in the inferior vena cava or iliac veins were also assessed for each patient. Results: PE was detected in 146 (33%) of 440 patients, DVT in 147 (33.4%) and VTE in 196 (44%). MDCT-V showed thrombi in 50 patients who did not have signs of PE at MDCT-PA. Thus, the addition of MDCT-V resulted in a 25% increase in thromboembolic disease detection (99% conﬁdence interval: 22%, 28%). Among 147 patients with DVT, only four had thrombi in the iliac and/or in the inferior vena cava without femoral or popliteal vein thrombosis; however, all had PE detected on MDCT-PA. Thus, the incremental diagnostic value of MDCT pelvic vein evaluation was 2.7% in the diagnosis of DVT and 0% in the diagnosis of VTE. Conclusion: In patients with suspected PE, pelvic radiation at MDCT-V can be reduced by omitting pelvic veins without decreasing thromboembolic disease detection rate.

B-436

14:45

Unsuspected pulmonary emboli in oncology inpatients detected on chest multidetector computed tomography (MDCT): Inﬂuence of clinical data on disease prevalence L. Menchini, A. Larici, L. Calandriello, F. Maggi, M. Torge, L. Bonomo; Rome/IT ([email protected]) Purpose: To evaluate if the prevalence of unsuspected Pulmonary Emboli (PE) in oncology hospitalized patients may be inﬂuenced by clinical data. Methods and Materials: From May to November 2005, 1221 oncology inpatients underwent 16-row MDCT examinations of the chest for indications other than pulmonary emboli. All images were retrospectively reviewed by two chest radiologists in consensus to look for PE. Clinical data and symptoms at the time of the examinations were retrospectively retrieved from our database. Patients with signs and symptoms suggestive of PE (dyspnoea, chest pain) were ruled out from the evaluation. 841 out of 1221 oncology inpatients were therefore included. Information about follow-up of the PE patients were obtained from clinical charts, also looking for other risk factors. Results: Nineteen out of 841 oncology inpatients had unsuspected PE (prevalence: 2.2%). All PE ﬁndings were correctly identiﬁed at the time of the report and all patients underwent anticoagulant therapy. Two out of 19 (10.5%) patients also showed hypercoagulability states. During 6-month follow-up no patient died from PE and no recurrences of PE or complications from anticoagulant therapy were observed. In the patients with coagulopathy physicians changed the duration of the therapy on the basis of the incidental ﬁnding. Conclusion: The prevalence of incidental PE in oncology inpatients is lower than that reported in literature when clinical data are used in the selection of the population. The assessment of risk factors other than neoplasm could be important in order to select the appropriate therapy and prevent further PE episodes.

14:54

Assessment of pulmonary perfusion defect with dual energy CT in acute pulmonary embolism and correlation with CT angiographic obstruction score and RV/LV diameter ratio J. Seo, Y. Jang, E. Chae, J. Song; Seoul/KR ([email protected]) Purpose: To assess severity of perfusion defect (P_score) with dual energy CT (DECT) in acute pulmonary embolism and to correlate it with CT angiographic obstruction score (CTA_score) and RV/LV diameter ratio (RVD). Methods and Materials: Thirty patients (13 men, 17 women; mean age, 55 years) with acute pulmonary embolism underwent contrast enhanced DECT scans. CT scan was done with following protocols: dual energy mode; tube voltages, 140 and 80 kVp; ratio of tube currents, 1:4.25; collimation, 1.2 mm; gantry rotation time, 0.33 s; injection of 100 ml of contrast agent (370 mg I/ml) at a rate of 3.5 ml/s; scan delay time, 20 s. Both average weighted images and contrast agent images were generated, which were evaluated separately. Two readers independently assessed CTA_score and RVD at average weighted image. On contrast agent images, the P_score was assessed on segmental basis with 3-grade scoring system (0: normal perfusion, 1: decreased perfusion, 2: no perfusion). The P_score was calculated by dividing summed score by maximal total score. Correlations among P_score, CTA_score, and RVD were tested with Pearson correlation. Results: The P_score (21p14%) and the CTA_score (28p20%) were well correlated (r=0.81, p 0.001) with an excellent concordance between readers for both P_score (r=0.89, p 0.001) and the CTA_score (r=0.93, p 0.001). Both indices showed good correlation with RVD (r=0.69, p 0.001 for P_score, and r=0.66, p 0.001 for CTA_score, respectively). Conclusion: Simultaneous evaluation of arterial occlusion and perfusion defect is possible with contrast enhanced DECT in acute pulmonary embolism. The P_score was well correlated with CTA_score and RVD.

B-438

15:03

Lung perfusion with dual-energy multidetector-row CT (MDCT): Application to the management of acute pulmonary embolism F. Pontana1, J.-B. Faivre1, M. Rémy-Jardin1, T.G. Flohr2, V. Pansini1, J. Rémy1; 1 Lille/FR, 2Forchheim/DE ([email protected]) Purpose: To investigate the accuracy of dual-energy CT in the depiction of perfusion defects in patients with acute pulmonary embolism (PE). Methods and Materials: 117 consecutive patients with the clinical suspicion of acute PE underwent a dual-energy MDCT angiogram of the chest (Deﬁnition, Siemens Medical Systems, Germany) using the following acquisition parameters (tube A: 140 kV; 50 mAs, tube B: 80 kV; 330 mAs) with a standard injection protocol. Two radiologists evaluated by consensus the presence of: (a) endoluminal clots on transverse “diagnostic” scans (1-mm thick averaged images from tubes A and B), and (b) lung perfusion scans. Results: Seventeen patients showed CT features of acute PE with depiction of 75 clots within lobar (n=15), segmental (n=43) and subsegmental (n=17) pulmonary arteries. A total of 17 clots were identiﬁed as complete ﬁlling defects (i.e., obstructive clots), located within segmental (12/17) and subsegmental (5/17) arteries. Fourteen of the 17 obstructive clots were seen with the concurrent presence of perfusion defects whereas cardiac-motion and/or contrast-induced artefacts within lung segments in close contact to the upper mediastinum and the cardiac cavities precluded conﬁdent recognition of perfusion abnormalities. Four subsegmental perfusion defects were depicted without visualization of endoluminal thrombi within the corresponding arteries. Perfusion defects were identiﬁed beyond ﬁve nonobstructive clots. The mean DLP value for dual-energy CT angiography was 275p37 mGy cm (mean effective dose: 4.67p0.63 mSv). Conclusion: Simultaneous information on the presence of endoluminal thrombus and lung perfusion impairment can be obtained with dual-energy CT.

B-439

15:12

Dual-energy CTA of pulmonary embolism: Initial results C. Fink1, T.R. Johnson2, K. Nikolaou2, B. Krauss3, M. Sedlmair2, M. Reiser2, C. Becker2; 1Mannheim/DE, 2Munich/DE, 3Forchheim/DE (christian.ﬁ[email protected]) Purpose: To evaluate dual-energy CTA for lung perfusion imaging in suspected pulmonary embolism (PE). Methods and Materials: 24 patients with suspected PE were examined with a single-acquisition, dual-energy CT angiography protocol (140 kV, 20-97 mA; 80 kV, 133-350 mA) on a dual-source CT (Somatom Deﬁnition, Siemens Medical Solutions, Forchheim, Germany). Voxels containing iodine and air were color-coded by a dedicated software. Lung perfusion was assessed by two radiologists. Perfusion

A

B

C

D

E

F

G

237

Saturday

MDCT venography in patients with suspected pulmonary embolism: Diagnostic impact of pelvic veins evaluation in thromboembolic disease detection B. Feragalli, B. Seccia, V. Di Mizio, P. D’Ambrosio, R. Patea, M.L. Storto; Chieti/IT ([email protected])

B-437

Scientific Sessions defects were classiﬁed as being consistent or non-consistent with PE (e.g. caused by artefacts). Image quality of the perfusion maps and CTA were rated using a 5-point score (1:excellent, 5:poor). Peak attenuation of the pulmonary arteries, lung parenchyma, and perfusion defects was assessed. Finally, the CTA data alone was assessed for PE by an independent third reader. Results: Perfusion defects consistent with PE were identiﬁed in four patients. In all these cases PE was conﬁrmed by the independent reader. Perfusion defects rated as non-consistent with PE were most frequently caused by streak artefacts from dense contrast material in the great mediastinal vessels or heart. Other reasons included atelectasis or pulmonary edema. The median score for the image quality of both the perfusion maps and CTA was 2. The median attenuation of pulmonary arteries, normal lung parenchyma, and perfusion defects was 453 HU, 56 HU, and 18 HU, respectively. Conclusion: Dual-energy CTA of pulmonary embolism is feasible and allows the assessment of perfusion defects caused by PE. Further optimization of the injection protocol is required to reduce artefacts from dense contrast material.

B-442

B-440

15:21

Dual source CT for chest pain assessment T.R. Johnson, C.R. Becker, A. Becker, C. Rist, M.F. Reiser, K. Nikolaou; Munich/DE ([email protected]) Purpose: CT angiography protocols offering a simultaneous evaluation of pulmonary embolism, coronary stenoses and aortic disease are gaining attractiveness with recent CT technology. The aim of this study was to assess the diagnostic accuracy of a speciﬁc Dual Source CT protocol for chest pain assessment. Methods and Materials: 109 patients with acute chest pain were scanned on a Somatom Deﬁnition (Siemens, Forchheim, Germany) using a body-weight adapted contrast material injection regimen. Two readers evaluated the images for initial diagnosis. A third reader correlated the coronary ﬁndings to invasive coronary angiography in 29 patients, performed dose calculations and reviewed the ﬁles of patients with negative exams for further diagnoses. Results: Heart rates ranged from 58 to 124 bpm with a median of 72. The cause of chest pain was identiﬁed in 83 patients. Regarding a one-month follow-up as standard of reference, the overall sensitivity for the identiﬁcation of the cause of chest pain was 98%. Among the most frequent diagnoses were coronary stenoses, valvular and myocardial disease, pulmonary embolism, aortic aneurysm and dissection. Correlation to invasive coronary angiography showed a high diagnostic accuracy in the patient-based evaluation with no non-diagnostic exams and 100% negative predictive value. In the segment based analysis, sensitivity, speciﬁcity, positive and negative predictive values amounted to 100, 99, 79 and 100%, respectively. Conclusion: Dual Source CT offers a very high diagnostic accuracy in chest pain assessment, even in high heart rates. ECG gated DSCT angiography can be a helpful tool in the triage of chest pain patients.

14:00 - 15:30

Room R2

SS 703 CT/MR in cardiovascular risk assessment Moderators: A. Bücker; Homburg a.d. Saar/DE R. Vliegenthart; Groningen/NL

B-441

14:00

NIMIS-CAD study: Inﬂuence of CAD risk, BMI, and coronary artery calcium deposits over the negative predictive value of MDCT in patients scheduled for coronary angiography F. De Cobelli1, R. Marano2, I. Floriani1, T. Nimiscad Study Group1, L. Bonomo2, A. Del Maschio1; 1Milan/IT, 2Rome/IT ([email protected]) Purpose: To investigate whether the NPV of MDCT in evaluation of coronary stenosis is affected by pre-test CAD-risk, BMI, and CACS. Methods and Materials: Twenty participating sites prospectively enrolled 367 patients scheduled for X-ray Coronary Angiography (CA) from July 2004 to June 2006 with a sample size (350) according to the targeted number. All patients had a non-contrast prospective-ECG-gated MDCT to evaluate the Agatston calciumscore (ACS); MDCT ﬁndings obtained during iv administration of cm (Iopromide370; Bayer-Schering-Pharma AG, Berlin, Germany) with 16- and 64-MDCT of different manufacturers have been compared with quantitative-CA. Patients were divided into three categories by the pre-test CAD-risk (low, intermediate, high) according

A

B

C

D

E

F

G

14:09

Coronary calcium scoring in cardiac asymptomatic patients with peripheral arterial disease: Is it useful? A. Rutten1, A.M. de Vos1, H. van der Zaag2, F. Zijlstra2, B. Rensing3, A. Mosterd4, W.P.T.M. Mali1, M. Prokop1, M.L. Bots1, M. Oudkerk2; 1Utrecht/NL, 2Groningen/NL, 3 Nieuwegein/NL, 4Amersfoort/NL ([email protected]) Purpose: We determined whether coronary calciﬁcation and cardiovascular risk factors can predict the occurrence of a signiﬁcant left main (LM) (or equivalent) coronary artery stenosis, eligible for surgery, in peripheral arterial disease (PAD) patients without cardiac symptoms but at a high risk for cardiac morbidity and mortality. Methods and Materials: Risk factor assessment, non-contrast-enhanced computed tomography (CT) scanning of the heart for calcium scoring and contrast-enhanced coronary CT angiography (cCTA) to determine the presence of an LM stenosis were performed in 102 cardiac asymptomatic PAD patients. If positive, invasive coronary angiography was performed and if indicated, surgery followed. Relations between risk factors and Agatston calcium score were determined. We used logistic regression modeling to evaluate the role of risk factors and Agatston score for the prediction of the presence of a signiﬁcant LM stenosis, eligible for surgery. Results: Median Agatston score was 259 (range 0-5135). Only an increase in age was signiﬁcantly related to Agatston score (p 0.001). Thrteen of 102 patients had a signiﬁcant LM stenosis on cCTA. Seven of these 13 patients underwent surgery. Diabetes and Agatston score 400 showed an 8.4 and 10.8-fold increased risk, respectively, for signiﬁcant LM stenosis, eligible for surgery. These two variables were included in a multivariable logistic regression-model for the presence of a signiﬁcant LM stenosis, eligible for surgery (ROC-area 0.82; 95% CI 0.64-1.0). Conclusion: Calcium scores and diabetic status may aid in predicting the occurrence of signiﬁcant LM stenosis, eligible for surgery, in cardiac asymptomatic PAD patients, despite their high cardiovascular risk status.

B-443

Cardiac

238

to the Guideline of National Cholesterol Education Program, by the BMI (normal: 25 kg/m²; overweight: 25-29 kg/m²; obese: 30 kg/m²), and by the ACS (0-100; 101-400; 400). Results: Among the 327 patients (93%) eligible for per-patient analysis, 115 (35%) were at low-risk, 156 (48%) at intermediate-risk, and 56 (17%) at high-risk for CAD. Even if not statistically signiﬁcant, there was evidence in favour of a trend of decreasing NPV as risk increase (low-risk: 93%; intermediate-risk: 92%; highrisk: 79%; p=0.07). According to the BMI, 102 patients (31.2%) were normal, 181 (55.3%) overweight, and 44 (13.4%) obese. NPV was not signiﬁcantly affected by BMI (p=0.59). According to the CACS, 151 patients (46.2%) had low-ACS, 86 (26.3%) had intermediate-ACS, while 90 patients (27.5%) had ACS 400, with a signiﬁcantly higher NPV in patients with lower score (0-100: 94%, 101-400: 86%, 400: 62%; p=0.003). Conclusion: MDCT is able to exclude signiﬁcant stenosis in patients at low/ intermediate-risk for CAD and in patients with low amount of calciﬁcations.

14:18

Predictive value of stenosis severity and atherosclerotic plaque burden at coronary CT angiography for major adverse cardiac events (MACE): A population based outcome study H.-R. Kim1, C. Thilo1, A. Flauaus1, B. Ruzsics1, H. Lee1, J.A. Abro1, S.A. Nguyen1, T.J. Vogl2, P. Costello1, U.J. Schoepf1; 1Charleston, SC/US, 2Frankfurt a. Main/DE ([email protected]) Purpose: To evaluate the predictive value of stenosis severity and the calciﬁed and non-claciﬁed plaque burden at coronary CTA (cCTA) for cardiac events. Methods and Materials: The outcomes of 250 consecutive patients who had undergone 64-slice cCTA for suspected coronary artery disease were determined using medical record review and phone calls to patients and referring physicians. Major adverse cardiac events (MACE; myocardial infarction, urgent revascularization, cardiac death) within a follow-up period of 12 months were recorded. Patient outcome was correlated with the presence and type (non-calciﬁed, calciﬁed, mixed) of atherosclerotic plaque burden and the presence of stenosis 50% at cCTA using regression analysis. Results: Twenty-three patients were lost to follow-up. Of the 227 remaining patients (111 male, 116 female), 86 patients did not show any atherosclerosis at cCTA. Of these, none had a MACE during follow-up (100% NPV). 141 patients had atherosclerosis at cCTA (eight purely non-calciﬁed, 52 purely calciﬁed, 81 mixed). 44/141 patients had stenosis 50%. 51 patients had a MACE during 12-month follow-up. MACE rates were not signiﬁcantly different between patients with different plaque types (p 0.05). The presence of stenosis 50% was signiﬁcantly (p=0.003) associated with MACE during follow-up.

Scientific Sessions Conclusion: The absence of atherosclerosis at coronary CTA reliably excludes MACE during follow-up. Among patients with atherosclerosis, presence of stenosis 50% is the strongest predictor of MACE during follow-up.

B-444

14:27

The relationship between coronary artery atherosclerosis evaluated by MDCT and carotid plaque by carotid ultrasound J.I. Jung, S.K. Yoon, Y.S. Choi, H.J. Yoon, S.T. Hahn; Seoul/KR ([email protected])

B-445

14:36

Plaque burden assessment with 64-slice computed tomography coronary angiography in diabetic patients compared with nondiabetic patients F. Cademartiri1, S. Seitun2, M. Romano2, G. Messalli3, S. Tresoldi4, E. Maffei1, A. Palumbo1, M. Fusaro1, N.R. Mollet5; 1Parma/IT, 2Genoa/IT, 3Naples/IT, 4Milan/IT, 5 Rotterdam/NL (ﬁ[email protected]) Purpose: We sought to evaluate plaque burden with Multislice Computed Tomography Coronary Angiography (MSCT-CA) in diabetic subjects compared with nondiabetics. Methods and Materials: One hundred diabetic patients (mean age: 66p9 years; males: 66) and 300 nondiabetic subjects (mean age: 63p11 years; males: 200) underwent contrast enhanced (Iomeprol 400, Bracco) MSCT-CA (Sensation 64, Siemens) with the aim to rule out CAD. Sixty-eight (68%) diabetics and 200 (67%) nondiabetics were without history of CAD, while the remaining were previously revascularized and/or had history of myocardial infarction. We evaluate the prevalence of obstructive disease (luminal narrowing 50%), nonobstructive disease ( 50%) and normal coronary arteries. The mean number of diseased segments was determined per patient and each diseased segment was classiﬁed as showing obstructive lesion or not. Calcium scoring was assessed. Among patients with nonobstructive disease, plaque type (calciﬁed, mixed, soft) was determined. Results: Diabetics showed a more diseased coronary segments (5.8p4.8 vs 4.4p4.4; p=0.007), either nonobstructive and obstructive; a higher prevalence of obstructive disease (52 vs 37%; p=0.01), a lesser prevalence of normal coronary arteries (14 vs 26%; p=0.02) and a higher total Agaston calcium score (472 p694 vs 196p406; p 0.0001). Among patients with nonobstructive disease, no differences were observed in plaque type between diabetics and nondiabetics. Conclusion: Diabetic patients had a higher extent of plaque burden compared with nondiabetic subjects. MDCT-CA may be used as a useful diagnostis tool for risk stratiﬁcation.

14:45

Prognostic value of 64-slice coronary angiography in patients with diabetes mellitus with known or suspected coronary artery disease compared with nondiabetic population F. Cademartiri1, S. Seitun2, M. Romano2, S. Tresoldi3, G. Messalli4, E. Maffei1, A. Palumbo1, M. Fusaro1, N.R. Mollet5; 1Parma/IT, 2Genoa/IT, 3Milan/IT, 4Naples/IT, 5 Rotterdam/NL (ﬁ[email protected]) Purpose: To assess the prognostic value of Multislice Computed Tomography Coronary Angiography (MSCT-CA) in a population with diabetes with known or suspected coronary artery disease (CAD) compared with non diabetic population. Methods and Materials: We followed up 108 patients with type-2 diabetes (group I) and 303 nondiabetic patients (group II), who consecutively underwent contrast enhanced (Iomeprol 400, Bracco) MSCT-CA (Sensation 64, Siemens) with the aim to rule out CAD. End-points were total cardiac event rate and hard cardiac event rate. Cox proportional-hazard models, with and without adjustment for risk factors, were developed to predict outcome. Results: At mean follow-up of 20p3.5 months, univariate analysis of survival showed statistically signiﬁcant differences in the comparison between groups for both total (36.4 vs 15.2%, pb0.0001) and hard (10.6 vs 2%; p=0.001) event rate. Total cardiac event rate resulted in signiﬁcant association with obstructive disease (luminal narrowing 50%) in both groups (p 0.0001). In multivariate analysis of all patients (group I+group II, n=411), independent predictors of events were smoking, DM, numbers of coronary segments with nonobstructive disease (luminal narrowing 50%) and obstructive disease. In multivariate analysis, obstructive CAD ( 50%) was the strongest independent predictor of events in both groups (p 0.0001). In group I, an adjunctive signiﬁcant predictor of events was the number of coronary segments with nonobstructive disease (p=0.011). Conclusion: MSCT-CA is a useful method for further risk stratiﬁcation and, together with the diagnostic accuracy, provides independent prognostic information.

B-447

14:54

Can MDCT be useful in diagnosing CHD in patients with low pre-test probability of disease? F. Cademartiri1, S. Tresoldi2, G. Messalli3, M. Romano4, S. Seitun4, E. Maffei1, A. Palumbo1, M. Fusaro1, N.R. Mollet5; 1Parma/IT, 2Milan/IT, 3Naples/IT, 4Genoa/IT, 5 Rotterdam/NL (ﬁ[email protected]) Purpose: To assess the diagnostic value of Multi-Detector Computed Tomography (MDCT) in the evaluation of patients suspected for coronary heart disease (CHD) and stratiﬁed by Framingham Risk Score (FRS) and Agatston Calcium Score (ACS). Methods and Materials: We studied 586 patients (346 males, 240 females; mean age 61p13 years) suspected for CHD on the basis of symptoms or provocative tests, who underwent contrast enhanced (Iomeprol 400, Bracco) 64-slice MDCT (Sensation 64, Siemens) coronary angiography. For each patient we evaluated pre-test risk for CHD using the FRS ( 5%=low; 5-20%=moderate; 20%=high), ACS, and CHD using MDCT (sound coronary arteries, non-signiﬁcant disease, signiﬁcant disease). We then correlated MDCT ﬁndings with the results of FRS and ACS. Results: Sixty-eight (11.6%) patients were considered at low risk, 384 (65.4%) at moderate risk, 135 (23%) at high risk, with FRS; ACS was 0-10 in 322 (54.9%) patients, 11-100 in 93 (15.8%) patients, 101-400 in 74 (12.6%) patients, 401-1000 in 47 (8%), 1000 in 51 (8.7%) patients. With MDCT 254 (43.3%) patients had sound coronary arteries, 229 (39%) non-signiﬁcant disease, 104 (17.7%) signiﬁcant disease. All patients considered at low risk had ACS 100. Considering patients at high pre-test risk, 44 (32.6%) had ACS 400: 25 had signiﬁcant disease at MDCT, 19 non-signiﬁcant disease. Twenty-one (6%) out of 351 patients considered at low or moderate risk and with ACS 100 were diagnosed at MDCT with signiﬁcant disease. Conclusion: MDCT can detect signiﬁcant disease in patients at low FRS and ACS.

B-448

15:03

Role of myocardial MR stress perfusion imaging in risk stratiﬁcation of patients with intermediate grade coronary artery stenoses J.S. Döring, C. Dösch, A.H. Seeger, B. Klumpp, U. Kramer, M. Fenchel, A.E. May, C.D. Claussen, S. Miller; Tübingen/DE ([email protected]) Purpose: In patients with suspected intermediate grade coronary artery stenoses, we investigated the value of myocardial MR stress perfusion imaging (MRMPI) for prediction of cardiac adverse events and/or identiﬁcation of patients at risk who may proﬁt from a close follow-up and early revascularization. Methods and Materials: At 1.5 T, 68 consecutive patients (57 men, 11 women, mean age 64.2 years, range 43-79 years) with CAD and coronary stenoses of

A

B

C

D

E

F

G

239

Saturday

Purpose: The aim of this study was to correlate the ﬁndings of coronary artery atherosclerosis (CAA) evaluated by multi-detector CT (MDCT) and carotid plaque evaluated by carotid ultrasound (CU). Methods and Materials: The study group consisted of 126 consecutive patients (76 men, 50 women; mean age 61p10 years) who underwent both ECG-gated 64-slice MDCT coronary angiography and carotid ultrasound within a 3-month period. We evaluated calcium score, presence of CAA, and presence of atherosclerosis at the left main coronary artery with 64-slice MDCT. The severity of CAA was determined by the number of stenotic vessels deﬁned as 50% area stenosis among the three coronary arteries. With US we evaluated carotid intima thickness, initima-media thickness (IMT), carotid artery dimension, carotid ﬂow velocity, and carotid plaque area on both sides of vessel (near wall vs far wall). MDCT ﬁndings were compared with CU parameters. Results: 1. Mean IMT signiﬁcantly correlated with the severity of CAA on MDCT (p 0.05). 2. Plaque size at the near wall of the right carotid artery signiﬁcantly correlated with severity of CAA on MDCT (p 0.05). 3. Mean carotid ﬂow velocity was associated with severity of CAA on MDCT (p 0.05). 4. Maximum/mean IMT correlated with calcium score on MDCT (p 0.01). Conclusion: Calcium score and severity of CAA by MDCT correlated well with carotid IMT and plaque size at the near wall of carotid arteries. Therefore MDCT evaluation of coronary artery atherosclerosis might be useful in predicting the severity of systemic atherosclerosis.

B-446

Scientific Sessions intermediate angiographic severity (r50% and b75%) were included in the study. MRMPI was obtained under adenosine stress and resting conditions using a turboFLASH 2D sequence. A dose of 0.1 mmol Gd-DTPA was injected at both times. Perfusion deﬁcits (PD) in myocardial segments were assigned to the coronary artery territories according to the AHA segmentation model. All patients were followed up personally or by telephone. Results: In 35/68 (51%) patients a PD was detected by MRMPI, while 33/68 (49%) patients had no PD. All patients were initially treated conservatively. During the follow-up period of 15p5 months, 15 patients presented with clinical aggravation of symptoms requiring repeat coronary angiography followed by interventional therapy (stenting, n=8) or bypass surgery (n=7). PD was observed in 13/15 cases so that 37% (13/35) of patients with a PD required revascularization within the follow-up period. In contrast, only two out of 33 patients (6%, p 0.05) without a PD developed disease progression. Conclusion: In patients with CAD and intermediate grade coronary stenoses, MRMPI is helpful in predicting adverse cardiac events and to identify patients who beneﬁt from early coronary revascularization.

B-449

Non-calciﬁed atherosclerotic plaque burden at coronary CTA: A better predictor of ischemia at myocardial perfusion imaging than calcium score and stenosis severity R.W. Bauer1, C. Thilo2, S.A. Chiaramida2, T.J. Vogl1, P. Costello2, U.J. Schoepf2; 1 Frankfurt a. Main/DE, 2Charleston, SC/US ([email protected]) Purpose: To examine the relationship of calciﬁed and non-calciﬁed plaque (NCP) burden and stenosis severity determined by coronary CT angiography (cCTA) with ischemia at myocardial perfusion imaging (MPI). Methods and Materials: 75 patients with suspected or known coronary artery disease underwent ECG-gated 64-slice cCTA and stress/rest SPECT MPI for clinical reasons. Calcium scoring was performed; cCTAs were analyzed for coronary artery stenosis and NCP or mixed plaque on a per-patient and per-vessel basis. An automated plaque analysis tool was used to calculate the volume of NCP components. MPI studies were analyzed for perfusion defects (PD) and their respective coronary territories. Results: 225 calciﬁed, 60 mixed, and 58 NCP were detected. 46 stenoses were rated 50% with 18 70%. 39/298 vessels in 26/75 patients showed PD (19 reversible, 20 ﬁxed) in their respective territories. There was no signiﬁcant difference in Agatston (118.0p267.0 vs 51.7p156.8; p=0.22) or Calcium Volume Score (96.2p210.0 vs 43.4p125.6; p=0.21) between vessels with and without PD. Sensitivity, speciﬁcity, PPV, and NPV of 50% ( 70%) stenosis for predicting PD were 18% (28%), 98% (91%), 54% (31%), and 89% (89%). There was a signiﬁcant difference in NCP volume (45.3p76.5 mm3 vs 19.7p58.4 mm3; p=0.028) between vessels with and without PD. Vessels with a total NCP volume of 100 mm3 had signiﬁcantly more PD (34.8%) than vessels without NCP (10.3%; p=0.005) or 1-100 mm3 (11.4%; p=0.026). Conclusion: Non-calciﬁed plaque burden as a measure of active coronary atherosclerosis is a better predictor of myocardial ischemia on MPI than calcium score or stenosis severity at cCTA.

15:21

Diagnostic accuracy of 64-slice computed tomography coronary angiography in patients with different pre-test-probabilities and prevalences of coronary artery disease L. Husmann, I. Valenta, P.T. Siegrist, O. Gaemperli, T. Schepis, H. Scheffel, P. Stolzmann, S. Leschka, H. Alkadhi, P.A. Kaufmann; Zurich/CH ([email protected]) Purpose: The aim of the study was to compare the diagnostic accuracy of 64-slice computed tomography coronary angiography (CTCA) in groups of patients with low, intermediate, and high pre-test-probabilities for coronary artery disease (CAD). Methods and Materials: Eighty-eight consecutive patients with suspected CAD (40 women, mean age 64.3p9.4 years, age range 39-82 years) underwent CTCA, calcium-scoring and invasive coronary angiography (ICA), and were grouped according to their Framingham 10-year-risk for hard coronary events into: low ( 10%), intermediate (10-20%), and high ( 20%)risk. On an intention-to-diagnose-basis, no coronary segment was excluded and non-evaluative segments were rated false positive. Results: Per-patient sensitivity, speciﬁcity, negative predictive value (NPV), and positive predictive value (PPV) were 90.0, 79.2, 95.0, and 64.3% in the subgroup with low pre-test-probability (n=34), 87.5, 92.3, 85.7, and 93.3% in the subgroup with intermediate pre-test-probability (n=29), and 100, 75.0, 100, and 89.5% in the subgroup with high pre-test-probability (n=25). Mean calcium-score-units were 90,

240

A

B

14:00 - 15:30

C

D

E

F

G

Room R3

Interventional Radiology

SS 709b Musculoskeletal intervention Moderators: M.B. Matson; London/UK P. Reimer; Karlsruhe/DE

B-451

15:12

B-450

220, and 312, and the prevalence of CAD was 29.4, 55.2, and 68.0% in subgroups with low, intermediate, and high pre-test-probabilities, respectively. Conclusion: The NPV of CTCA is invariably high in patients with different pretest-probabilities. The PPV increases with increasing prevalence of CAD, while the speciﬁcity decreases due to a higher calcium load.

14:00

Percutaneous ﬁxation of unstable sacroiliac fractures in trauma patients using CT guidance M.G. Mack, B. Maier, J. Frank, K. Eichler, A. Thalhammer, F. Walcher, I. Marzi, T.J. Vogl; Frankfurt a. Main/DE ([email protected]) Purpose: To evaluate feasibility and accuracy of CT percutaneous ﬁxation of sacroiliac fractures in trauma patients. Methods and Materials: 22 patients (nine female, 13 male, mean age 44.6 years, range 16-70 years) with unstable traumatic pelvis fractures had been advised to minimize complications and avoid further injury. Patients were positioned in supine position under general anesthesia on the CT table. Treatment planning was with a MSCT scan with 4x2.5 mm collimation (120 KV, 80 mAs): Afterwards, skin entrance points were marked. Guide pins (diameter 2.5 mm) were placed under CT guidance. Following conﬁrmation of depth, position, and angulation of guide pin with CT images, a 7-mm cannulated screw of appropriate length was placed through a small skin incision. Results: It was possible to place all planned guide pins under CT-guidance without complications. A total of 55 guide pins were inserted as follows: 25 guide pins were placed on the right side, 30 were placed on the left side. Seven patients got percutaneous ﬁxation on both sides. All guide pins could be placed correctly in the ﬁrst and second sacral segments through the sacroiliac joint. All cannulated screws could be placed via guide pins without complications. Postoperative CT-scan veriﬁed the correct positioning of the screws without perforation in anterior or posterior direction and without affecting the foramen of the nerve roots. Conclusion: CT-guided ﬁxation of unstable pelvic fractures is a safe and feasible method to minimize complications and to avoid further injury.

B-452

14:09

Final results of combined intradiscal and periganglionic injection of medical ozone and periganglionic administration of steroids and anesthetic for the treatment of lumbar disk herniation: Effects on disk size and lumbar radiculopathy T. Lehnert1, F. Berg-Schlosser1, J.O. Balzer2, V. Jacobi1, M.G. Mack1, T.J. Vogl1; 1 Frankfurt a. Main/DE, 2Mainz/DE ([email protected]) Purpose: To evaluate the therapeutic beneﬁt and morphologic changes in herniated lumbar disk after CT-guided intradiscal and periganglionic ozone-oxygen injection combined with a periganglionic administration of steroids and anesthetic. Methods and Materials: 110 patients with lumbar radiculopathy received an intradiscal (3 mL) and periganglionic (7 mL) injection of an ozone-oxygen mixture (ratio 3:97), followed by a periganglionic injection of corticosteroid (1 mL of Celestan®Depot) and anesthetic (2 mL of Carbostesin® 0.25%) in the same session. Under CT guidance, intradiscal and periganglionic injection was administered by means of an extraspinal lateral approach, using a 22-gauge 17.8-cm spinal needle. Six months after treatment, clinical outcome was assessed by applying the modiﬁed MacNab method. The effects on disk matrix and disk volume were evaluated by MRI. Results: Treatment was successful in 84 patients (76.4%). In the remaining 26 patients (23.6%), treatment was considered to have failed. Among the patients whose treatment was a success, outcome was excellent in 49 patients (58.3%) and good in 35 patients (41.7%). Among the patients whose treatment was a failure, this was poor in 21 patients (80.8%) and poor with recourse to surgery in ﬁve patients (19.2%). Six months after treatment, in patients with excellent outcome, mean disk volume reduction was 15.3%, in patients with good outcome 8.8%, and in patients with poor outcome 3.7%.

Scientific Sessions Conclusion: This treatment is a therapy option for lumbar disk herniation that has failed to respond to conservative management and affects both the mechanical and the inﬂammatory components of pain.

B-453

14:18

Bipolar percutaneous radiofrequency ablation (RF) of osteoidosteomas M.G. Mack, K. Eichler, S. Zangos, T. Lehnert, D. Proschek, K. Hochmuth, T.J. Vogl; Frankfurt a. Main/DE ([email protected])

B-454

14:27

Initial experience of concurrent image-guided intramuscular botulinum toxin injection in patients with idiopathic cervical dystonia Y. Yoon, I. Lee; Seoul/KR ([email protected]) Purpose: To present the initial experience of concurrent EMG-, and image-guided intramuscular Botulinum toxin (BTX) injection in patients with idiopathic cervical dystonia (ICD). Methods and Materials: Twelve patients with ICD underwent EMG-guided or blind intramuscular BTX with high standardized uptake value (SUV) on PET-CT scan between October 2005 and January 2007. Six patients with ICD who showed high SUV in deeply located cervical muscles or focal high SUV in muscles underwent image-guided BTX injection. The clinical outcomes were measured by Tsui rating scale and the Toronto Western Spasmodic Torticollis Rating Scale (TWSTRS). Results: Twenty-two BTX injections into 14 muscles of 6 patients were performed under image guidance. Total 13 sessions of BTX injection (6 CT guided and 7 US guided) were performed in 6 patients and technical success was 100%. All six patients showed signiﬁcant reduction of Tsui and TWSTRS scores. There were two minor complications (dysphasia and hematoma) related to the image-guided BTX injection, which subsided soon without management. Conclusion: The image-guided BTX injection can be a useful treatment modality in patients with ICD when the target muscle locates deeply, or a focal area of muscle should be injected.

B-455

14:36

Osteoid osteoma of the spine: CT-guided radiofrequency ablation J. Martel, Á. Bueno, E.J. Ortíz; Alcorcón/ES ([email protected]) Purpose: Our aim is to present our experience with percutaneous CT-guided radiofrequency ablation of spinal osteoid osteoma as an alternative to surgical resection. Methods and Materials: We report 12 cases of spinal tumors adjacent to the neural elements. Ten of these cases were treated with CT-guided radiofrequency coagulation. The procedures were performed on an outpatient basis. One osteoid osteoma was localized in the cervical spine, four in the thoracic spine and the remaining ﬁve in the lumbar area. The pedicle was involved in three cases, the pars interarticularis and superior facets in four cases and the laminae in the other three cases. CT has been the standard imaging technique to evaluate the location and the size of the nidus of the osteoid osteoma, and to establish if the cortex was intact between the nidus and the spinal canal. Results: Pain disappeared post-RF ablation in seven patients within 10-15 days.

B-456

14:45

Interventional radiology with transoral approach on the 2° cervical vertebra (C2): Result and complications in different pathologies on 15 patients G.C. Anselmetti, A. Manca, S. Cirillo, G. Chiara, F. Russo, D. Regge; Candiolo/IT ([email protected]) Purpose: To assess safety, effectiveness, results and complications of the transoral approach in spine intervention of C2. Methods and Materials: Since 2003, 15 patients (31-88 years, mean 65) underwent spine intervention on C2 with transoral approach. Vertebroplasty (PV) was performed under ﬂuoroscopic guidance in 14 patients (eight metastases, one myeloma, one symptomatic angioma, four traumatic fractures) injecting 1-4 ml of polymethylmethacrylate through a 15- (nine cases) or 13- (ﬁve cases) gauge needle while radiofrequency thermoablation (RFA) of one osteoid osteoma was executed under combined CT and ﬂuoroscopic guidance. Procedures were performed in general anesthesia. All patients treated with PV had pain; nine (ﬁve malignant, four traumatic) were obliged to wear Halo-jacket. Only after conservative therapies failed and experienced neurosurgeons contraindicated surgery was interventional radiology performed. Results: All procedures were completed without complications. Patients with traumatic fracture of C2 referred signiﬁcant pain remission, evaluated with Visual Analog Scale (VAS from 6.7p0.6 to 0.7p0.6) before and after procedure (follow-up from 48 to 15 months); in one patient 1-year follow-up CT revealed asymptomatic pseudoarthrosis. Same clinical result was obtained in malignancy (VAS from 7.8p2.4 to 1.4p1.7); one patient required surgical posterior consolidation after 3 months. Patients could dismiss Halo. Patient with angioma had complete pain remission (VAS from 8 to 0). Patient with osteoid osteoma had partial pain remission (VAS from 9 to 3; follow-up 26 months) but could suspend painkillers. Conclusion: Transoral interventional radiology on C2 is a safe and effective procedure; it should be considered when surgery is contraindicated.

B-457

14:54

Pedicle screw placement in trauma patients with unstable fractures of the thoracic spine using CT assistance M.G. Mack, B. Maier, C. Ploss, K. Eichler, J. Frank, I. Marzi, T.J. Vogl; Frankfurt a. Main/DE ([email protected]) Purpose: To evaluate feasibility and accuracy of CT-assisted placement of pediclescrews in patients with fractures of thoracic spine. Methods and Materials: 50 patients (14 female, 36 male, mean age 46.39 years, range 3.3-80.9) with traumatic fractures of thoracic-spine were treated. Patients were positioned in prone position under general anesthesia and treatment planning was performed with a MSCT scan with 4x2.5 mm collimation (120 KV, 40 mAs): After marking skin entrance points, guide-wires (diameter 1.6 mm) were placed under CT guidance through pedicles using a percutaneous approach. Positioning of all guide-wires was veriﬁed by a MSCT scan. These images were used for measuring the length of the screws. In the operation room screws and internal ﬁxation were placed via inserted guide-wires. Results: It was possible to place all planned guide-wires under CT-guidance without complications. 425 guide-wires were inserted as follows: 18 guide-wires were placed in C4-C7, 119 were placed in T1-T4, 194 were placed in T5-T8, and 88 were inserted in T9-L1. All guide-wires could be placed through the centre of the pedicle into vertebal body without perforation into spinal canal or paravertebral soft tissues. Intraoperatively all pedicle screws could be placed via inserted guide-wires. Mean inner diameter of pedicles was 3p1 mm, mean outer diameter of pedicles was 5p1 mm, representing typical appearance of small thoracic pedicles. Conclusion: CT guidance allows accurate placement of the guide-wire, which allows a safe placement of pedicle-screws and internal ﬁxation system and reduces the risk of wrong placements of pedicle-screws signiﬁcantly.

A

B

C

D

E

F

G

241

Saturday

Purpose: To analyze the results of bipolar RF-ablation of osteoidosteomas. Methods and Materials: Twenty patients (14 male and six female; mean age, 27.7, range, 9-60) with clinically and radiologically suspected osteoidosteoma were treated with CT-guided RF-ablation. The procedure was performed in seven patients under general anesthesia and in 13 patients with deep analgo-sedation. After localization of the nidus, an eccentric drill bit (Bonopty) with a diameter of 1.7 mm was used to create a hole of about 2.3 mm and a penetration cannula with a diameter of 2.1 mm was inserted. A bipolar 18-gauge RF probe (Celon ProSurge micro) with a 9 mm active tip was inserted through the penetration cannula and connected to the RF generator. Energy application was started at 3 W and subsequently increased to a maximum of 5 W. The procedure was terminated if a resistance of 900 Ohm was reached. 3-4 treatment circles were performed. Saline solution was injected through the penetration cannula. Results: The osteoidosteomas were localized as follows: femur n=6, tibia n=6, humerus n=3, spine n=3, acetabulum n=2. The mean duration of energy deposition was 8.1p2.5 min (minimum 2.6 min, maximum 11.75 min), with a mean energy application of 2.3p2.0 kJ (range, 0.3-6.24 kJ). All 20 patients remained free of symptoms during follow-up. The follow-up time was 31.5p24.2 months (median follow-up time 25.5 months, range, 3.8-85 months). No complications occurred. Patients had no restrictions of normal activity within 24 hours. Conclusion: Bipolar RF ablation is an efﬁcient and safe treatment of osteoidosteoma.

One patient reported pain in the course of the ﬁrst month after the procedure. Recurrence of the pain was observed in two patients in the whole study group. The RF ablation was repeated in these cases, achieving successful results in both patients. Our primary and secondary clinical success rates were 80% and 100%, respectively. Conclusion: CT-guided percutaneous radiofrequency ablation of vertebral osteoid osteoma is a safe and effective method that should be considered the procedure of choice for most of these cases. The technique may be contraindicated when no intact cortex is evident between the nidus and the neural structures.

Scientific Sessions B-458

15:03

Long-term follow-up of arterial embolization of aneurismal bone cysts: A single center experience M. Cellerini, S. Mangiaﬁco, G. Roselli, G. Beltrami, D. Campanacci, R. Capanna; Florence/IT ([email protected]) Purpose: We report our experience with long-term follow-up of therapeutic arterial embolization of Aneurismal Bone Cyst (ABC) in a series of consecutive patients. Methods and Materials: Between January 1994 and June 2005, 13 patients with ABC (female: 4, male: 9, age range: 7-35 aa) underwent arterial embolization for a total of 18 procedures. Deﬁnitive diagnosis was established with biopsy. An acrylic glue (Histoacril: #9, Glubran2: #4) was employed in 13 procedures and Onyx in four procedures. In 10 procedures, 150-250 micron particles (PVA contour) were jointly employed. The follow-up program consisted of clinical and radiological examinations 3, 6, 12 and 18 months after the procedure and a yearly CT/MRI examination for the ﬁrst two years. Results: Complete embolization (occlusion of 80% of the pathological vessels) was obtained in nine patients. In all patients, symptoms regressed completely after embolization and they recurred in only ﬁve patients 6-12 months after a complete (n°=1) or partial (n°=4) embolization. All were referred for a second embolization followed by surgical excision in two. The mean follow-up period was 6.7 years. Two patients were lost to follow-up. Total and partial regression rates were 63% (7/11) and 36% (4/11), respectively. Conclusion: Our long-term results indicate that arterial embolization of ABC should be considered a valid therapeutic alternative especially in young patients with a difﬁcult surgical approach or cases of postsurgical recurrence.

B-459

15:12

Percutaneous echo-guided treatment of calciﬁed tendonitis of the shoulder: Clinical response and radiological ﬁndings G. De Conti, U. Marchioro, A. Dorigo, F. Angelini; Padova/IT ([email protected]) Purpose: Clinical response evaluation to echo-guided percutaneous needle aspiration in shoulder calciﬁed tendonitis and radiological ﬁndings after treatment. Methods and Materials: Sixty-two shoulders of 62 consecutive symptomatic patients (38 women, 24 men; mean age: 46.8 years) were treated percutaneously by using two 16-gauge needles and US guidance. Patients were prospectively evaluated by using Shoulder Pain and Disability Index. Radiographs, US examination and magnetic resonance images were obtained immediately before the treatment, and at two and six months. Results: In 45 patients (72.6%) signiﬁcant improvement was seen in shoulder pain and disability index, from 75.4 points (SD 12.88) preoperatively to 17.65 points (SD 9.43) at two months (p 0.001). Of those, 16 (35.6%) demonstrated not complete resolution of calciﬁcations; 11 (23.9%) revealed MR signs of impingement syndrome. Seventeen still-symptomatic patients (27.4%) required retreatment for subacromial bursitis; of those, 10 (58.8%) demonstrated not complete resolution of calciﬁcations and 12 (70.6%) showed impingement syndrome at MRI. At six months a signiﬁcant clinical improvement was maintained in all patients from 71.1 points (SD 20.9) to 23.5 points (SD 13.8) (p 0.001). MR examination demonstrated integrity of treated tendons in overall 100% of cases. Conclusion: Percutaneous echo-guided treatment of calciﬁed tendonitis of the shoulder proved to be effective, despite persistence of calciﬁc remnants in some patients. Retreatment is needed when subacromial bursitis occurs; this complication is more frequent when subacromial impingement exists. In all patients, no tendon damage was seen after treatment.

B-460

15:21

Guided removal of soft tissues foreign bodies: Personal experience L. Callegari, G. Sturniolo, A. Leonardi, D. Mariani, C. Sabato, E. Genovese, C. Fugazzola; Varese/IT ([email protected]) Purpose: Foreign bodies in soft tissues are a frequent cause of medical consultation; they can cause pain and complications. The surgical removal of foreign bodies is not always successful; it can be invasive, expensive and burdened by complications. In our study we evaluated the advantages of US-guided removal of soft tissue foreign bodies versus surgery. Methods and Materials: Twenty-two patients (18 male, four female, age range 1365) came to our attention for clinical suspicion and/or RX evidence of foreign bodies in soft tissues. All patients were primary evaluated with US examination (performed with ATL 5000 and PHILIPS IU-22, with linear high frequency probe) that conﬁrmed the presence and location of 42 foreign bodies. Four patients with organic foreign bodies (one agave fragment, four wood splinters); 14 patients with metallic foreign

242

A

B

C

D

E

F

G

bodies (nine fragments, two nails, two needles, one bullet); four patients with glass foreign bodies (23 fragments). The average dimension of the foreign bodies was from less than 1 mm of some glass fragments to 3 cm of the agave fragment. All the patients underwent US-guided procedure of foreign bodies removal. Results: In 42/42 cases US-guided removal was successful. No complications related to procedure occurred. Thirty cutaneous incisions were performed, in 28/30 cases incision was repaired with Steri-Strip. In all the cases antibiotic therapy was established for seven days after procedures. Conclusion: Compared to surgery, US-guided removal of soft tissue foreign bodies is more successful, is less expensive and invasive with evident advantage. The results of our study allow us to propose the US-guided removal as primary therapeutic option.

Scientific Sessions

A

B

C

D

E

F

G

243

Sunday

Sunday, March 9

Scientific Sessions room A 2nd level

room B 2nd level

room C 2nd level

room E1 entr. level

room E2 entr. level

room F1 entr. level

room F2 entr. level

room G/H lower level

room I lower level

07:00

07:00

07:30

07:30

08:00

08:00

08:30

08:30 CC 916 MRI Neuro: Conquering important territories (p. 60)

09:00

09:30

CC 917 Multidetector CT Made Easy Bones and joints: CT is still alive! (p. 60)

NH 9 New Horizons Session Above 3 T: Clinical tool or toys for the boys? (p. 61)

RC 910 Musculoskeletal Hip (p. 62)

E3 920a Interactive Teaching Session Classifying breast lesions: The BI-RADS™ exercise (p. 63)

E3 920b Interactive Teaching Session Inﬂammatory small bowel diseases (p. 63)

SF 9a Special Focus Session Cardiothoracic CT: One test solves all (p. 63)

SF 9b Special Focus Session Hybrid imaging: PET/CT, SPECT/ CT, PET/MRI …, where next? (p. 64)

RC 913 Physics in Radiology MR-guided cardiovascular interventions (p. 65)

09:00

09:30

10:00

10:00

10:30

10:30 EM 3 ESR meets Israel Views and impressions from radiology in Israel (p. 70)

11:00

11:30

SS 1010 Musculoskeletal Osteoporosis and hip disorders (p. 246)

SS 1001b GI Tract CT colonography (p. 248)

SY 8 GE Healthcare Satellite Symposium (p. 540)

E3 1020 Interactive Teaching Course: Useful Signs in Imaging Signs in chest imaging (p. 71)

SS 1007 Genitourinary Imaging of the urinary tract: Do new techniques give new answers? (p. 250)

SS 1002 Breast MR: New clinical ﬁndings (p. 252)

SS 1005 Computer Applications Radiological reporting and teaching (p. 254)

SS 1013 Physics in Radiology Dose, risk assessment and optimisation (p. 256)

12:00

11:00

11:30

12:00 Honorary Lecture 2 (p. 72)

12:30

12:30 SY 10 GE Healthcare Satellite Symposium (p. 542)

13:00

SY 11 Hitachi Satellite Symposium (p. 542)

SY 12 Bracco Satellite Symposium (p. 542)

13:00

13:30

13:30

14:00

14:00

SY 14 GE Healthcare Satellite Symposium (p. 544)

IIQ Image Interpretation Quiz

14:30

15:00

SY 15 Guerbet Satellite Symposium (p. 544)

SY 16 Hologic Satellite Symposium (p. 544)

14:30

15:00

15:30

15:30

16:00

16:30

EPOS™ - scientiﬁc exhibition

18:00

technical exhibition

17:30

registration

17:00

CC 1217 Multidetector CT Made Easy The abdomen: The realm of MDCT (p. 73)

CC 1216 MRI Vascular: From head to toe (p. 72)

SA 12 State-of-the-Art Symposium Coronary CTA comes of age: Technical advances, decision modeling and cost effectiveness (p. 73)

16:00

RC 1210 Musculoskeletal Spine (p. 74)

E3 1220 Interactive Teaching Course: Useful Signs in Imaging Signs in musculoskeletal imaging (p. 75)

RC 1207 Genitourinary Imaging of the adnexal masses (p. 75)

SF 12 Special Focus Session Whole body MR imaging: Valuable screening or life style choice for the worried rich? (p. 76)

RC 1211 Neuro Imaging of vascular brain disease (p. 77)

RC 1213 Physics in Radiology Physics and technology of multi-modality imaging systems (p. 77)

16:30

17:00

17:30

18:00

18:30

18:30

19:00

19:00

244

A

B

C

D

E

F

G

Scientific Sessions room K lower level

room L/M 1st level

room N/O 1st level

room R1 1st level

room R2 1st level

room R3 1st level

room U 2nd level

room X 2nd level

room XI, Z & Y 2nd level

room Z 2nd level

07:00

07:00

07:30

07:30

08:00

08:00

08:30

08:30

09:00

09:30

SF 9c Special Focus Session Imaging your youngest patients: Fetal MRI (p. 66)

RC 914 Radiographers MRI (p. 67)

MC 919 Molecular Imaging Cancer and inﬂammation (p. 67)

RC 903 Cardiac The atherosclerotic plaque (p. 68)

RC 901 Abdominal and Gastrointestinal Small nodules in the cirrhotic liver: What to do? (p. 69)

RC 907 Genitourinary Imaging of the male genitalia (p. 70)

WS 921 Image Guided Breast Biopsy: How to do it

09:00

09:30

10:00

10:00

10:30

10:30

11:00

11:30

SS 1012 Pediatric Chest and abdomen (p. 259)

SS 1014 Radiographers Development of the professional role (p. 261)

SS 1011 Neuro Functional MRI (p. 263)

SY 9 Carestream Health Satellite Symposium (p. 541)

SS 1003 Cardiac Myocardial and pericardial diseases (p. 265)

SS 1009 Interventional Radiology Abdomen and pelvis (p. 267)

WS 1021 Image Guided Breast Biopsy: How to do it

WS 1022 Experience Vascular Procedures Using Simulators

WS 1023 Tips and Tricks in Radiofrequency Ablation

WS 1024 Test-Driving your Workstation

11:30

12:00

12:00

12:30

SY 13 Philips Healthcare Satellite Symposium (p. 543)

13:00

13:00

13:30

13:30

14:00

14:00

SY 17 Toshiba Satellite Symposium (p. 544)

14:30

15:00

SY 18 Philips Healthcare Satellite Symposium (p. 544)

WS 1121 Image Guided Breast Biopsy: How to do it

WS 1122 Experience Vascular Procedures Using Simulators

WS 1123 Tips and Tricks in Radiofrequency Ablation

WS 24C1 Test-Driving your Workstation

15:00 WS 24C2 Test-Driving your Workstation

15:30

16:00

17:00

14:30

15:30

16:00

RC 1212 Pediatric Gastrointestinal tract (p. 78)

RC 1214 Radiographers Professional proﬁle developments: The extended role of the radiographer (p. 79)

RC 1206 Molecular Imaging and Contrast Media Angiogenesis imaging in clinical practice today (p. 79)

RC 1204 Chest Bronchogenic carcinoma: Screening and staging (p. 80)

CC 1218 Breast: From Basics to Advanced Imaging Characterisation of breast lesions by US and MRI (p. 81)

PC 12 Professional Challenges Session Teleradiology: Film reading at the sea side? (p. 81)

16:30

17:00

17:30

17:30

18:00

18:00

18:30

18:30

19:00

19:00

A

B

C

D

E

F

G

245

Sunday

12:30

16:30

11:00

Scientific Sessions 10:30 - 12:00

Room B

Musculoskeletal

SS 1010 Osteoporosis and hip disorders Moderators: J.L. Bloem; Leiden/NL A. Klauser; Innsbruck/AT

B-461

10:30

Diffusion weighted MRI and 1H-MR-spectroscopy at 3 T applied to patients with osteoporosis: Preliminary data supporting a new procedure to evaluate bone fracture risk S. Capuani, M. Rebuzzi, F. Fasano, G.H. Hagberg, B. Maraviglia; Rome/IT ([email protected]) Purpose: Bone Mineral Density (BMD) accounts only for about 60% of the global risk of bone fracture. Recent data suggest that a relevant proportion of women reporting osteoporotic fractures have BMD values in the normal range. Bone marrow contained in spongy bones also contributes in determining bone strength. Aim of this study was to combine diffusion weighted imaging (DWI) and 1H-MRS data to obtain information in vivo on bone strength in humans. Methods and Materials: Ten women with osteoporosis (T-score -2.5 or less; age range: 60-70 years) and ten age-matched healthy women (T-score -1.0 or more) were investigated at 3.0 T. Diffusion-weighted images, using phase diffusion gradients (“b values” 0-8000 s/mm2), and bone marrow 1H-spectra (Single-Voxel Spectroscopy) were collected from calcanei of each subject. 1H-spectra were used to calculate individual percentages of bone marrow fat content (Mfc%) by assessing the methylene-to-water peak area ratios. DWI was used to derive the ADC values from the corresponding calcaneal regions. Statistical group comparisons were performed using Pearson’s correlation coefﬁcient. Results: ADC values in subjects with Mfc% lower than 87 were signiﬁcantly higher in osteoporotics compared to healthy women (ADC values: 69.3p3.7x10-12 and 39.2p3.2x10-12 m2/s, respectively; p=0.0005). In contrast, subjects with Mfc% higher than 87 showed homogeneous ADC values (32.2p7.6x10-12 m2/s) regardless of group belonging to. Conclusion: These preliminary ﬁndings indicate that calcaneal ADC is a sensitive measure to discriminate between normal and osteoporotic women. Moreover, the combination of DWI and 1H-spectroscopy assessment might contribute to a better prediction of bone fracture risk.

B-462

10:39

MDCT evaluation of lumbar pedicles in osteoporotic spines A. Papadakis, A.H. Karantanas, G. Papadokostakis, A. Voloudaki, J. Damilakis, N. Gourtsoyiannis; Iraklion/GR ([email protected]) Purpose: Lumbar operations using pedicular screws require stiff substrate to allow efﬁcient and long lasting ﬁxation. The aim of this study was to assess the physical status of the cortical bone of the lumbar pedicles in osteoporotic female patients using images obtained from routine abdominal MDCT examinations. Methods and Materials: The outer (O) and the inner (I) dimensions of the pedicles of two lumbar vertebrae (L3 and L4) were measured on MDCT (16-row) images of 16 osteoporotic and 16 healthy female individuals. Measurements were performed on the left and right pedicles of each vertebra in the coronal and the sagittal planes. Image thickness was 0.75 mm. The widths of the cortical bone of the pedicle (WCBP), deﬁned as (O-I)/2, and the cortical bone of the pedicle index (CBPI), deﬁned as (O-I)/O, were calculated. Results: Mean WCBP values averaged over the coronal and the sagittal planes of the L3 and L4 vertebrae were 0.17 cm (p0.03) for osteoporotic and 0.28 cm (p0.04) for healthy individuals. The corresponding mean CBPI values for osteoporotic and healthy females were 0.26 (p0.05) and 0.50 (p0.06), respectively. Differences in mean WCBP and CBPI values between the two groups were statistically signiﬁcant (paired t-test, p 0.0001). Conclusion: WCBP and CBPI values of women with osteoporosis were signiﬁcantly lower than those of healthy women. These indices may be used as indicators for the effectiveness of the pedicular screw ﬁxation and establish exclusion criteria in patients referred for lumbar spine stabilizing operations.

246

A

B

C

D

E

F

G

B-463

10:48

Automated trabecular bone structure analysis of the proximal femur: A comparison with BMD in assessment of osteoporosis J.S. Bauer1, T. Baum1, M. Huber2, J. Carballido-Gamio2, D. Müller1, F. Eckstein3, T.M. Link2, E.J. Rummeny1; 1Munich/DE, 2San Francisco, CA/US, 3Salzburg/AT ([email protected]) Purpose: For the assessment of osteoporosis, a highly standardized analysis of the bone is needed. The goal of this in vitro study was to investigate whether trabecular structure parameters, measured with an automated analysis of Multidetector Computed Tomography (MDCT) scans of the proximal femur, were able to improve the prediction of biomechanical bone strength compared to bone mineral density (BMD) as a standard parameter. Methods and Materials: 170 left femur specimens were harvested from formalinﬁxed human cadavers (mean-age 80p10 years). BMD was determined using Dual X-ray Absorptiometry (DXA). High resolution images were acquired using a 16-row MDCT. Regions of interest were automatically determined in the trochanter, the neck and the head of the femur. Nonlinear (NL) and standard morphometric (SM) structure-parameters were calculated in these regions. A 3-point bending test, simulating a fall on the greater trochanter, was used to determine bone strength (MCS) of the proximal femur. Results: DXA and structure parameters both showed signiﬁcant correlations (p 0.01) with MCS. The highest correlations were found in the trochanteric region with both modalities (r up to 0.80 for NL, 0.74 for SM, 0.71 for BMD). Although correlations were higher for nonlinear structure parameters, differences were not signiﬁcant (p 0.05). Conclusion: Feasibility of automated structure analysis using MDCT images of the proximal femur was demonstrated, achieving high correlations between structure measurements and bone strength in particular for a nonlinear structure parameter in a standardized region in the trochanter. A trend to higher correlations was found compared to DXA BMD; yet, differences were not signiﬁcant.

B-464

10:57

Can routine abdominal multi-detector CT diagnose spinal osteoporosis? A. Papadakis, J. Damilakis, K. Perisinakis, A.H. Karantanas, N. Gourtsoyiannis; Iraklion/GR ([email protected]) Purpose: Although quantitative CT (QCT) is a sensitive method for the diagnosis of osteoporosis, its widespread clinical use has been limited due to increased patient radiation burden. The aim of this study was to investigate whether densitometric data acquired from routine abdominal MDCT examinations can be used to differentiate patients with osteoporosis from healthy individuals. Methods and Materials: Twenty-ﬁve female patients (group A) with history of radiotherapy were scanned with both routine abdominal MDCT and standard QCT to generate a MDCT-to-QCT conversion equation. Twenty-one osteoporotic (group B) and 23 healthy females (group C) were also recruited in the study. Patients of groups B and C underwent routine abdominal MDCT examination for various clinical indications. Mean bone mineral density (BMD) values of the patients in groups B and C were calculated based on the MDCT-to-QCT conversion equation derived from patients of group A. Results: Mean BMD in patients of group A was 103.4p32.8 mg/ml with routine abdominal MDCT and 91.0p28.5 mg/ml with QCT. Quantitative CT BMDQCT values can be calculated based on the BMDMDCT values derived from routine abdominal MDCT scans utilizing the equation: BMDQCT=0.78 * BMDMDCT+10.13 (r=0.90, p 0.0001). The calculated QCT densitometric data differentiated adequately osteoporotic from healthy female individuals (area under ROC curve 0.828, p=0.05). Conclusion: Female individuals referred to routine abdominal MDCT may be additionally informed about their bone density status without any excess radiation burden. Routine abdominal MDCT may differentiate osteoporotic from healthy female individuals.

B-465

11:06

Relationship between RANKL/OPG system and bone mineral density in children with inﬂammatory bowel disease: A DXA study C.V. Albanese, M. Paganelli, F. Civitelli, M. Osimani, R. Passariello; Rome/IT ([email protected]) Purpose: Recently, two critical extracellular regulators of osteoclast differentiation and activation have been identiﬁed: receptor activator of nuclear factor (NF-kappaB) ligand (RANKL) and osteoprotegerin (OPG). Disorders of the RANKL/OPG system have been linked to several human diseases including inﬂammatory diseases and bone loss. We evaluated the relationship between RANKL/OPG system and bone mineral density (BMD) in children with Crohn’s disease (CD) and ulcerative colitis (UC).

Scientific Sessions Methods and Materials: Lumbar spine BMD was measured by DXA and bone mineral apparent density (BMAD) was calculated in 39 children (26 CD, 13 UC). Plasma levels of total and soluble RANKL and OPG were measured and correlated with BMAD, body mass index (BMI) and duration of disease. Results: BMAD z-score was -1.3p1.5, with no differences between CD and UC. Plasma levels of RANKL and OPG were 13.9p12.2 ng/ml and 305p108 pg/mL, respectively. No correlation was found between BMAD and OPG or RANKL. A direct relationship was found between the RANKL/OPG ratio and BMAD (r=0.4) only in CD. In IBD RANKL was inversely correlated to BMI z-score (r=-0.42). A signiﬁcant correlation was found between BMI and both RANKL and OPG (r=0.4) in CD but not in UC. Conclusion: The direct correlation between RANKL/OPG ratio and BMAD in CD proves the protective role of OPG and the negative role of RANKL on BMD. The signiﬁcant relationship between both RANKL and OPG and nutrition status seems to be implicated in the pathogenesis of low BMD in pediatric IBD.

B-466

11:15

Trabecular bone architecture of the calcaneus measured using nonlinear structure parameters: Comparison of high resolution MR imaging at 3.0 and 1.5 Tesla using noise simulations J.S. Bauer1, D. Müller1, C. Räth2, R. Monetti2, M. Matsuura1, F. Eckstein3, T.M. Link4, E.J. Rummeny1; 1Munich/DE, 2Garching/DE, 3Salzburg/AT, 4 San Francisco, CA/US ([email protected])

B-467

11:24

Physical load-associated differences in femoral neck MRI texture L.C.V. Harrison, R. Nikander, H. Sievänen, H. Eskola, P. Dastidar, S. Soimakallio; Tampere/FI (lara.harrison@tut.ﬁ) Purpose: To evaluate load-associated differences in femoral neck MRI texture in order to ﬁnd out feasible exercise types that have the potential to strengthen proximal femur and thus partly prevent hip fractures in later life. Methods and Materials: Ten female top-level squash and football players and nine normally exercising referents were included in this preliminary study. Standardized axial T2-weighted gradient echo images were obtained with 1.5 T MRI (TR 40 ms, TE 17 ms, slice thickness 1 mm). Analysis was performed with Mazda texture analysis software (Technical University of Lodz, Poland). The trabecular bone was divided manually into ROIs representing four anatomic quadrants. Histogram, gradient, run-length matrix, co-occurrence matrix, autoregressive model and wavelet texture parameters were used. Classiﬁcation error probability combined with average correlation coefﬁcient was used to ﬁnd out the most signiﬁcant texture features to discriminate and classify the load-inﬂuence on bone. Raw data analysis and principal component analysis were performed. One-nearest-neighbour classiﬁcation was performed to ﬁnd out the ability of analysis subsets to classify tissues of interest. Results: The most discriminative texture parameters and classiﬁcation results varied with anatomical location of ROI. The clearest difference between loading groups was seen in the anterior quadrant of the femoral neck: raw data and principal component analyses both led to 1/19 misclassiﬁcation; medial quadrant 3/19; posterior quadrant 4/19 and lateral quadrant raw data 7/19 and principal component analysis 8/19 misclassiﬁcation.

B-468

11:33

Line-scan diffusion MR imaging in ischemic epiphysis of femoral head: Experimental study X. Li1, J. Qi1, L. Xia1, S. Maier2; 1 Wuhan/CN, 2Boston, MA/US ([email protected]) Purpose: To determine whether diffusion changes with ischemia of increasing duration, and whether line-scan diffusion weighted imaging (LSDWI) can provide temporal information about ischemia of epiphysis. Methods and Materials: Ischemia was surgically induced in one hip of each piglet (n=12) and the other hip served as a normal control. Piglets were imaged at three hours, 72 hours and 1, 3 and 6 weeks after surgery by using LSDWI. Apparent diffusion coefﬁcients (ADC) were calculated. Signiﬁcant differences in ADC between ischemia and control were evaluated by using statistics. At six weeks after surgery, piglets were sacriﬁced for histological study. Results: ADC in ischemic femoral head (1.21p0.39) decreased signiﬁcantly at three hours after surgery (P 0.05), compared to that in control (1.74p0.31), and increased at 72 hours (2.18p0.43 versus 1.70p0.28, P 0.05). ADC was kept increasing until six weeks after surgery (P 0.01). The signiﬁcant difference of ADC between two continuous time points was found (P 0.05), except that between three and six weeks after surgery. Histological ﬁndings revealed abnormal epiphyseal cartilage thickening, cartilaginous islands within ossiﬁed tissues. Growth disturbance was found in form of regions of focal growth plate disruption. Conclusion: ADC remained changed in this model of prolonged ischemia. LSDWI can be useful for evaluating duration and extent of ischemic epiphyseal disruption. This research is supported by NSFC (30370430).

B-469

11:42

An abnormal vessel supply of an affected lower extremity in patients with proximal femoral focal deﬁciency M. Horak, Z. Ryznarova, M. Kaspar, M. Masek, J. Chomiak, M. Frydrychova; Prague/CZ ([email protected]) Purpose: Patients with proximal femoral focal deﬁciency (PFFD) have bone and joint abnormalities but in rare cases can also have vessel abnormalities. An abnormal vessel supply of affected lower extremity has been diagnosed using CT angiography. Methods and Materials: CT angiography has been used to examine 23 patients with PFFD to visualize bone and joint changes on affected lower extremity and to determine the grade of affectation according to Pappas. At the same time projection of vessels in arterial and venous phases has allowed the diagnosis of an abnormal vessel pattern. For better presentation of vessels an MIP and VRT technique has been used. Results: In 2 patients out of 23 patients with PFFD the affected extremity had an abnormal vessel pattern. Arteria iliaca interna and its branch arteria glutea inferior have become the main supplying arteries. The artery continued in thigh in course of ischiadic nerve as ischiadic artery. Arteria iliaca externa have been thin just supplying adductors of thigh. At the same time in one case there has been even an abnormal vein drainage when a deep vein of a thigh accompanied abnormal arterial pattern and collected blood to vena iliaca interna. Conclusion: The examination of patients with proximal PFFD using CT angiography can provide better visualization of bone and joint changes as well as visualization of vessels of affected lower extremity. This new described ﬁnding shows the complexity of PFFD but it can also allow a better planning of a surgical access for a correction treatment.

B-470

11:51

Herniation pits of the femoral neck: A radiographic indicator of femoroacetabular impingement? J.A. Kim, J.H. Choi, J.S. Park, W. Jin, K.N. Ryu; Seoul/KR Purpose: To assess the signiﬁcance of herniation pits at the femoral neck in radiographic diagnosis of femoroacetabular impingement (FAI). Methods and Materials: Of 3014 patients, 103 patients showed herniation pits on the anteroposterior hip radiographs. Two radiologists diagnosed herniation pits based upon the radiographic criteria; location at the anterosuperior femoral neck, close to the physis and in a diameter of 3 mm. Thirty-eight AP hip radiographs without neutral pelvic orientation were excluded for a reliable evaluation. Seventyﬁve hips of 65 patients (bilateral in 10) were suggestive of FAI based on three

A

B

C

D

E

F

G

247

Sunday

Purpose: To investigate the trabecular bone architecture in calcaneus cadaver specimens obtained from elderly human donors with high resolution magnetic resonance imaging (MRI) at 3.0 T and 1.5 T and to compare the structural information revealed at both ﬁeld strengths using noise simulations. Methods and Materials: Sagittal MR images of 42 calcaneus specimens were obtained using high resolution gradient echo sequences (donor age: 81p10 years, male:female=22:20). Noise was added to obtain ﬁve sets of images with decreasing signal to noise ratios (SNR), similar at both ﬁeld strengths. Micro-CT (μCT) images were obtained and bone strength (MCS) was determined in a biomechanical test. Histomorphometric parameters (HM) and scaling indices were calculated in corresponding regions for μCT and all noise levels at 1.5 T and 3.0 T. Results: SNR was 7.3 at 3.0 T and 3.5 at 1.5 T. In the simulated noise levels, SNR decreased from 3.5 to 1.0, similarly for both ﬁeld strengths. Down to a level of SNR=2.3, adding noise did not change the correlations of MR-derived parameters versus μCT parameters. Comparing correlations of MR vs μCT at similar noise levels, 3.0 T consistently revealed higher r values than 1.5 T (e.g. bone fraction at SNR=2.3: r[3.0 T]=0.90; r[1.5 T]=0.81). Conclusion: To obtain valid bone structure parameters by MR, an SNR of 2.3 was sufﬁcient. Thus, the improvement of 3.0 T cannot be explained by higher SNR, but by the pronounced susceptibility that enhances the visualization of thin trabecular structures at similar resolutions. Therefore, the higher signal at 3.0 T should better be used to increase spatial resolution instead of SNR.

Conclusion: MRI texture analysis provides non-ionizing quantitative imaging method for detecting and classifying bone structural differences due to exercise loading.

Scientific Sessions radiographic signs; lateral center edge angle (LCE) 39, acetabular index (AI)b0, extrusion index (EI) 25%. The diameters of all herniation pits were measured. Patients radiographically suggestive of FAI were correlated with their symptoms. Results: The mean values of the LCE, AI, EI of all hips were 31.0° (range, 22.141.8), 8.2° (range, -0.3-15), 13.3 (range, 3.1-23.5), respectively. Hips with positive radiographic signs of FAI were 4/75 (5.3%) with LCE, 2/75 (3.7%) with AI, 75/75 (100%) with EI. Only one out of 75 hips showed all of the signs. The mean EI of hips with at least one positive sign (ﬁve hips) was 7.3. The average diameter of herniation pits of all hips and of hips with at least one positive sign (ﬁve hips) were 8.9 mm (range, 4-15) and 8.4 mm each. Also, among those ﬁve hips, two hips of one patient had symptoms implying FAI clinically. Conclusion: The very low frequency of positive radiographic signs indicating FAI among the patients with herniation pits suggests that herniation pits have very limited signiﬁcance in the diagnosis of FAI.

10:30 - 12:00

Room C

GI Tract

SS 1001b CT colonography Moderators: E. Neri; Pisa/IT S. Somers; Dundas, ON/CA

B-471

B-473

10:30

CT colonography with limited bowel preparation: Diagnostic performance in a FOBT positive screening population M.H. Liedenbaum1, A.F. van Rijn1, A.H. de Vries1, H.M. Dekker2, P. Fockens1, E. Dekker1, J. Stoker1; 1Amsterdam/NL, 2Nijmegen/NL ([email protected]) Purpose: To determine whether CT-colonography (CTC) is an accurate method for detection of lesions (carcinomas and polyps) r10 mm and r6 mm in persons with a positive Fecal Occult Blood Test (FOBT) in a population-screening setting. Methods and Materials: 150 consecutive FOBT positive individuals (32 guiac FOBT (Hemoccult), 118 immunochemical FOBT (OC-Sensor)) underwent a CTC with limited bowel preparation, which was read by two independent observers. Reference standard was colonoscopy with segmental unblinding. Per polyp sensitivity and per patient sensitivity and speciﬁcity were calculated for both observers (‘double reading’) with two cut-off points: lesions r10 mm (category 1) and lesions r6 mm (category 2). Results: In total 7% of FOBT positive patients had a carcinoma; all identiﬁed at CTC, with one false positive CTC ﬁnding. Of FOBT positives 50% had a category 1 lesion and 70% a category 2 lesion. Per polyp sensitivity for CTC was 94% in category 1 and 87% in category 2. Per patient sensitivity for CTC was 96 and 93% in categories 1 and 2, respectively (OC-Sensor 96 and 94%, Hemoccult 95 and 92%, respectively). The per patient speciﬁcities were 92 and 77% for categories 1 and 2, respectively (OC-Sensor 91 and 82%, Hemoccult 100 and 50%, respectively). Conclusion: CTC with limited bowel preparation is an accurate technique for carcinoma and polyp detection in a FOBT positive population. CTC as triage technique can reduce the number of colonoscopies, although number is inﬂuenced by true positive FOBT rate.

B-472

10:39

Prospective comparison of colonoscopy, CT colonography, and stool tests in an average risk population: Results from the Munich colorectal cancer prevention trial A. Graser, P. Stieber, D. Nagel, C.R. Becker, H. Kramer, S. Geisbuesch, C. Schaefer, H. Diepolder, A. Lottes, A. Wagner, M.F. Reiser, B. Goeke, F.T. Kolligs; Munich/DE ([email protected]) Purpose: To assess the sensitivity and speciﬁcity of colonoscopy (OC), CT colonography (CTC), sigmoidoscopy (SC), fecal occult blood testing (FOBT), and immunochemical fecal hemoglobin testing (imHb) in colorectal cancer screening in asymptomatic adults. Methods and Materials: 307 average risk adults (135 women, 172 men; 60.3p6.7 years) underwent same day 64-MDCTC and OC using segmental unblinding. All CTC scans were read by experienced readers using a primary 3D approach. Polyps located in the sigmoid colon and rectum were included for analysis of SC. Per-patient and per-polyp sensitivities and speciﬁcies were calculated for CTC, SC and OC; for FOBT and imHb, patient-based sensitivities and speciﬁcities were

248

A

B

C

D

calculated. Furthermore, we calculated sensitivities for different combinations of screening tests. Results: 511 polyps were detected in the screening population, out of which 222 were adenomas. 27 patients had at least one adenoma 9 mm (total number of adenomas 9 mm=36; total number of polyps 9 mm=40 in 29 patients), 48 patients had at least one adenoma 5 mm (total number of adenomas 5 mm=79; total number of polyps 5 mm=100 in 61 patients). Per-patient sensitivities of CTC, OC, SC, FOBT and imHb were 92.6% (25/27), 100% (27/27), 66.7% (18/27), 21.7% (5/23), and 30.4% (7/23) for adenomas 9 mm; 93.7% (45/48), 95.8% (46/48), 64.6% (31/48), 16.7% (8/48), and 37.5% (18/48) for adenomas at least 6 mm in size. Per-polyp sensitivities were 91% (91/100) for CTC and 92% (92/100) for OC for polyps 6 mm, and 95% (38/40) for CTC and 100% (40/40) for OC for polyps 9 mm. For polyps at least 10 mm in size, speciﬁcities were 98.6% for CTC, 67.4% for FOBT, and 85.5% for imHb, while OC and SC by deﬁnition have a 100% speciﬁcity each. For polyps 6 mm and larger, CTC reached a 95.9% speciﬁcity, while FOBT showed a speciﬁcity of 66.0%, and imHb, 88.1%. One carcinoma was detected by FOBT, imHb, CTC and OC, but missed by SC as it was located in the transverse colon. Conclusion: The accuracy of CTC in the detection of polyps 5 mm is comparable to colonoscopy, and CTC has high speciﬁcity. Sigmoidoscopy misses signiﬁcant polyps, and missed one carcinoma. imHb is signiﬁcantly better than FOBT in the detection of adenomas, and shows less false positive results.

E

F

G

10:48

Screening CT colonography in an asymptomatic average-risk Asian population: A two-year experience in a single institution S. An1, K. Lee2, Y. Kim2, S. Park1, H. Kim2, S. Kim2, S. Kim1, N. Kim2; 1Seoul/KR, 2 Seongnam-si/KR Purpose: To evaluate the results of screening CT colonography (CTC) in an asymptomatic average-risk Asian population. Methods and Materials: Our institutional review board approved this retrospective study and waived informed consent. In 2005 and 2006, 1,015 adults (609 males and 406 females; mean age, 51 years) underwent screening CTC using a 16-detector row CT scanner and an automated CO2 delivery system. During the study period the protocols were changed to use less vigorous purgation and lower radiation doses; stool tagging (n=890) and primary three-dimensional interpretation (n=966) were generally used. CTC results were categorized: C0, inadequate; C1, no signiﬁcant polyp; C2, one or two 6-9 mm polyps; C3, polyps r10 mm or rthree 6-9 mm polyps; and C4, mass. Patients with positive CTC results were referred to gastroenterologists for follow-up or management planning. Results: Categories C0-C4 were assigned to 21 (2.1%), 916 (90.2%), 54 (5.3%), 23 (2.3%), and one (0.1%) patients, respectively. Fifty-four patients with C3-C4 (n=21) or C2 (n=33) underwent subsequent optical colonoscopy: complete (n=53) and incomplete (n=1). Per-patient positive predictive values (PPVs) for categories C3-C4 and C2-C4 were 90 (18/20) and 74% (39/53), respectively. Per-polyp PPVs at 10- and 6-mm thresholds were 92 (22/24) and 69% (45/65), respectively. Diagnostic yield for advanced neoplasm was 1.5% (15/1,015). Conclusion: Our results seem comparable to Western experiences, showing the reproducibility of a successful screening CTC program.

B-474

10:57

Evaluation of potential advantages of computer-aided diagnosis (CAD) in reporting CT colonography with a primary 2D approach R. Ferrari, M. Rengo, P. Paolantonio, F. Iafrate, P. Lucchesi, C. De Cecco, A. Laghi; Rome/IT ([email protected]) Purpose: To determine potential advantages of computer-aided diagnosis (CAD) system in the detection of polyps at CT colonography (CTC). Methods and Materials: Three readers, with different rates of experience, (10 years experienced radiologist, two years experienced radiologist and a senior radiology resident) independently evaluated 46 CTC examinations retrospectively chosen as a polyp-enriched population from our database (96 polyps and six cancers). CTC had been performed using 4-row (Volume zoom, Siemens, Erlanger, Germany) and 64-row MDCT (VCT, General Electric Healthcare, Milwaukee, USA). Readings were performed using primary 2D approach with and without CAD software (aw 4.4 software. General Electric healthcare) as a concurrent reader. Results: Same diagnostic sensitivity with and without CAD (96%) was observed for expert radiologist. No statistically signiﬁcant difference (86% without CAD and 90% with CAD software) was observed for intermediate experienced radiologist. Signiﬁcant improvement of sensitivity (63% for 2D approach, 82% for 2D+CAD) was observed in non-expert radiologist. Mean reporting time was 13 min for primary 2D approach and 10 min and 35 sec for 2D+CAD software. CAD revealed 7.3 false positive ﬁndings per patient.

Scientific Sessions Conclusion: The use of CAD software increased sensitivity of CTC to detect colonic polyps only in non-expert radiologist. CAD used as a concurrent reader reduces reporting time.

B-475

11:06

Evaluation of computer-aided detection for CT colonography with multiple bowel preparation regimens J. Näppi1, H. Yoshida1, P. Lefere2, W. Cai1, M.E. Zalis1; 1Boston, MA/US, 2 Roeselare/BE ([email protected]) Purpose: Conventional computer-aided detection (CAD) systems have been designed for CT colonography (CTC) with rigorous laxative bowel cleansing and/or consistent outcomes. We evaluated our previously validated fully automated CAD system in a more realistic clinical setting involving seven different patient regimens ranging from laxative to laxative-free bowel preparations. Methods and Materials: The CAD system was trained to detect polyps 6 mm with 146 CTC cases. The multi-center testing data were independent from the training data and included 150 clinical CTC cases with 24-hour or 48-hour bowel preparations from two laxative (Group 1), two minimally laxative (Group 2), and three laxative-free (Group 3) regimens. There were 50 patients in each group, and all but 25 patients of Group 1 had fecal tagging. The CTC cases were acquired in supine and prone positions with 1-5 mm collimations, 1-5 mm reconstruction intervals, 50-120 mA currents, and 120-140 kVp voltages. Results: The detection sensitivity of CAD was 100% for colonoscopy-conﬁrmed lesions 10 mm in all groups (n=20), with 5.5 FP detections per patient on average. For lesions 8 mm, the detection sensitivity was 92% in Group 1 (n=12), 90% in Group 2 (n=10), 100% in Group 3 (n=7), and 93% in all groups (n=29), with 8.5 FP detections per patient on average. The CAD accuracy did not vary signiﬁcantly (p 0.05) between the groups. Conclusion: The CAD system provides consistently high detection sensitivity with clinically acceptable low false-positive rate for polyps/masses 8 mm regardless of bowel preparation in CTC.

11:15

Evaluation of a prototype colon CAD system on polyps submerged in tagged material S. Lakare, M. Wolf, M. Dundar, L. Bogoni, M. Salganicoff; Malvern, PA/US ([email protected]) Purpose: Evaluate the performance of a prototype Colon CAD system on polyps that are either partially or fully submerged (PoFS) in tagged material in CTC cases using a reduced cathartic preparation with fecal tagging. Methods and Materials: The data consisted of 702 high resolution MDCT scans (Siemens and GE scanners) acquired at supine and prone positions. All these patients had undergone tagged preparation. The data was randomly split into training (170, 25%) and test (532, 75% not used in development) sets. From the test set, all cases with at least one colorectal polyp r6 mm were considered in this study (147). Results: The test set included 134 mid-size (MS: 6-10 mm), 50 large-size (LS: 10-25 mm) polyps; 20 MS and eight LS polyps were PoFS on either patient scan. The sensitivity for PoFS polyps was: 16/20 (80%) MS, 8/8 (100%) LS and 24/28 (85.71%) on clinically signiﬁcant (CS) polyps ( 6 mm); for non-PoFS polyps the sensitivity was: 98/114 (85.96%) MS, 35/42 (83.33%) LS and 85.26% CS polyps. The overall sensitivity was 85.07 MS, 86 LS and 85.33% CS polyps. The average and median FP rates per volume were 3.14 and 2.0, respectively. Conclusion: The performance for the detection of PoFS and non-PoFS polyps was shown to be comparable. Yet, as PoFS polyps are usually more difﬁcult to detect by a human, a CAD system that aids in their detection should prove to be even more valuable to a radiologist and will accommodate a wider set of preparations.

B-477

11:24

CT colonography: Computer-aided detection of morphologically ﬂat early (T1) colonic carcinoma S.A. Taylor1, G. Iinuma2, J. Zhang3, S. Halligan1; 1London/UK, 2Tokyo/JP, 3Beijing/CN ([email protected]) Purpose: To retrospectively evaluate the ability of computer-aided detection (CAD) software to detect morphologically non-polypoidal (ﬂat) early colonic carcinoma using CT colonography. Methods and Materials: Ethical permission was obtained. Twenty-six consecutive early (stage T1) carcinomas endoscopically classiﬁed as non-polypoidal (“ﬂat”), [Paris classiﬁcation stage 0-Il] were collected from a symptomatic population undergoing staging with multi-detector row CT colonography (CTC). The location of

B-478

11:33

Performance analysis of a CAD system on ﬂat, sessile, and pedunculated polyps M. Wolf, S. Lakare, M. Dundar, L. Bogoni, M. Salganicoff; Malvern, PA/US ([email protected]) Purpose: Routine CT colonography examinations include a spectrum of polyp types, each presenting detection challenges to radiologists and CAD systems. We report the performance of a prototype CAD system* with respect to different polyp types. Methods and Materials: The database of 418 high-resolution MDCT datasets was partitioned randomly into training and testing. The entire set contains 174 polyps (6-25 mm in extent). Data, untagged with full cathartic preparation, was obtained from 10 sites in the US and Europe, acquired using different scanner types and manufacturers, with a wide range of acquisition parameters. The CAD system was optimized to provide best sensitivity for polyps between 6 and 25 mm. Datasets deemed suitable for radiologist review were included regardless of distension, artifacts or residual material. Results: The training set contains 208 cases, 99 positive, 65 mid-sized (MS, 69.9 mm), 38 large-sized (LS, 10-24.9 mm) polyps. The independent test set contains 210 cases, 102 positive, 42 MS, 29 LS polyps. Sensitivity obtained for sessile and pedunculated polyps: 100% (25/25) for MS, 100% (16/16) for LS. Flat polyps: 93.8% (15/16) for MS, 92.3% (12/13) for LS. Shallow ﬂat polyps (height 3 mm): 92.9% (13/14) for MS. Sensitivity for all polyp types: 97.6% (MS), 96.5% (LS) with an average of 2.7 and median of 2 false-positives per volume. Conclusion: The prototype system demonstrates good and uniform overall system performance on a database containing the full spectrum of different polyp types, acquisition protocols and scanners. *Prototype not yet commercially available.

B-479

11:42

Comparative performance assessment of CAD and unaided human reading in multicenter clinical trials of CT colonography H. Yoshida1, J. Nappi1, D.C. Rockey2, R. Choi3, M.E. Zalis1, A.K. Singh1; 1 Boston, MA/US, 2Dallas, TX/US, 3Washington D.C., AE/US ([email protected]) Purpose: To retrospectively compare the performance of computer-aided detection (CAD) with that of prospective unaided human reading for detection of adenomas/carcinomas in 1056 CT colonography (CTC) cases obtained from two large multicenter clinical trials. Methods and Materials: The two multicenter clinical trials involved 15 U.S. medical centers. The patients were prepared with rigorous cathartic preparation without (57%) or with (43%) orally administered fecal tagging. The locations of colonoscopyconﬁrmed lesions r6 mm were identiﬁed retrospectively based on colonoscopy and unblinded retrospective CTC reading. Cases were subjected to an in-house CAD system that was trained with an independent set of 121 CTC cases. The retrospective per-polyp detection sensitivity of CAD was compared with that of prospective unaided human CTC reading conducted during the clinical trials. Results: There were 191 retrospectively visible biopsy-proven adenomas/carcinomas r6 mm in 142 patients, among which 107 lesions were r8 mm and 72 lesions were r10 mm. The per-polyp detection sensitivity of unaided human CTC reading was 70, 69, and 69% for adenomas/carcinomas r6 mm, r8 mm, and r10 mm, respectively. The corresponding per-polyp detection sensitivity of CAD was 76, 88, and 90%, respectively, at an average false-positive rate of 4.5 per scan. The difference in sensitivities between CAD and human reading was statistically signiﬁcant.

A

B

C

D

E

F

G

249

Sunday

B-476

each tumour within CTC datasets was noted by three experienced radiologists in consensus aided by the endosocpic report. CAD software (ColonCAD 4.0, Medicsight PLC) was then applied at three settings of sphericity (0, 75[default], and 100]. Computer prompts for each patient were categorized as either true-positive if the mark overlapped tumour boundary, or false-positive if located elsewhere. True-positive CAD prompts were subclassiﬁed as focal (overlaying a focal elevation beyond the contour of the main lesion), or non focal. Results: The 26 cancers were endoscopically classiﬁed as type IIa (n=12) and type IIa+IIc (n=14). The mean size (range) was 27 mm (7-70 mm). CAD detected 22 (84.6%), 19 (73.1%), and 15 (57.7%) of the 26 cancers at ﬁlter settings of 0, 75, and 100, respectively. Most missed cancers were type IIa lateral spreading. The mean total number of false-positive CAD marks per patient at each ﬁlter setting was 35.6, 20.4, and 9.3, respectively, excluding polyps. At all settings, 96.1% of CAD true-positives were classiﬁed as focal. Conclusion: CAD is effective for the detection of morphological non polypoidal cancer, although minimally raised lateral spreading tumors remain problematic.

Scientific Sessions Conclusion: Based on a large cohort of 1056 patients and 191 adenomas/carcinomas r6 mm, CAD has superior detection sensitivity compared to unaided human reading for adenomas/carcinomas in all signiﬁcant size categories with a clinically acceptable FP rate.

B-480

11:51

CT colonography and computer-aided detection: Effect of false positives on reader speciﬁcity and reading efﬁciency in a low prevalence screening population S.A. Taylor1, R. Greenhalgh1, R. Ilangovan2, E. Tam2, V.A. Sahni2, D. Burling2, J. Zhang3, P. Pickhardt4, S. Halligan1; 1London/UK, 2Harrow/UK, 3Beijing/UK, 4 Madison, WI/US ([email protected]) Purpose: To retrospectively evaluate the effect of increasing numbers of computer-aided detection (CAD) generated false positive (FP) detections on reader speciﬁcity and reporting times using CT colonography (CTC) in a low prevalence screening population. Methods and Materials: Four trained readers each read 48 CTC datasets (26 male, 22 female mean age 57) from a screening population (3 containing polyps), without CAD followed by review of the CAD output (as a second read), recording ﬁndings and diagnostic conﬁdence. The 45 designated normal datasets were chosen such that 22 and 23 generated b15 FP and 15 FP CAD prompts, respectively. Sensitivity, speciﬁcity, and ROC curves were calculated with and without CAD assistance. The relationship between CAD FP number and reader conﬁdence, reporting times and correct dataset classiﬁcation was analyzed using linear and logistic regression. Results: Across all readers, CAD resulted in four additional FP detections. Overall reader sensitivity and speciﬁcity (6 mm polyp threshold) pre- and post-CAD assistance (95% CI) were 0.75 (0.43, 0.95) vs 0.83 (0.52, 0.98) and 0.96 (0.91, 0.98) vs 0.93 (0.88, 0.96), respectively. AUC increased from 0.57 (0.34, 0.80) to 0.61 (0.42, 0.80). There was no correlation between increasing CAD FP and reader conﬁdence (p=0.71), or correct study classiﬁcation (p=0.23), but there was a positive correlation with CAD-assisted reading times (0.06 [0.02, 0.10], p=0.002). Conclusion: Increasing numbers of CAD FP do not adversely inﬂuence correct reader study classiﬁcation or diagnostic conﬁdence, although they increase reporting times.

10:30 - 12:00

Room F1

Genitourinary

SS 1007 Imaging of the urinary tract: Do new techniques give new answers? Moderators: J.J. Fütterer; Nijmegen/NL S. Hanna; Cairo/EG

B-481

10:30

ESUR CT urography practice guidelines 2008: The latest A.J. van der Molen1, N.C. Cowan2, U.G. Mueller-Lisse3, C.C.A. Nolte-Ernsting4, S. Takahashi5, R.H. Cohan6, C.T.U. Working Group of the ESUR1; 1Leiden/NL, 2 Oxford/UK, 3Munich/DE, 4Hamburg/DE, 5Nijmegen/NL, 6Ann Arbor, MI/US ([email protected]) Purpose: To develop practice guidelines for the use of CT Urography (CTU). Methods and Materials: Peer-reviewed papers, reviews and abstracts were systematically scrutinized by a group of CT/CTU experts from the European Society of Urogenital Radiology (ESUR). To reach consensus, a summary document was produced and input from clinical colleagues obtained at the ESUR 2006 and ECR 2007 meetings was integrated. Results: Expert-based guidelines on indications and details of the CTU technique were produced, but true evidence-based guidelines could not be derived from the currently available evidence. The use of CTU as a ﬁrst-line test is justiﬁed for patients with macroscopic haematuria, at high risk for urothelial cancer. For other indications CTU can be used for problem-solving if the results of cystoscopy and ultrasound (and/or intravenous urography) are equivocal. A differentiated approach using 1-, 2- or 3-phase protocols is recommended in which the clinical indication and the nature of the patient population determine the number of phases. All protocols employ oral hydration, low-dose furosemide, and thin-slice acquisition and reconstruction parameters. Either a combined nephrographic-excretory phase following

250

A

B

C

D

E

F

G

a split-bolus injection of contrast medium, or separate nephrographic or excretory phases following a single-bolus injection can be used. For benign conditions a reduced dose (CTDI vol 5-6 mGy) and for potential malignant disease a normal dose (CTDI vol 9-12 mGy) is used. A low-dose (CTDI vol 2-3 mGy) unenhanced phase may be added on indication. Conclusion: These expert-based practice guidelines provide recommendations to optimize CTU and to unify radiologists’ approach to CTU.

B-482

10:39

Inﬂuence of bladder distension on opaciﬁcation of urinary collecting system during CT urography B. Brkljacic, J. Curic, M. Vukelic-Markovic, P. Marusic, A. Hrkac-Pustahija; Zagreb/HR ([email protected]) Purpose: To compare degree of opaciﬁcation of renal collecting system and ureters during CT urography between patients in who exam was performed after 20 minutes and after one hour following oral ingestion of 1000 ml of water. Methods and Materials: CTU was performed in 89 patients, and 168 collecting systems and ureters were evaluated. Sixteen-detector-row scanner was used. In 44 patients water was administered 20 minutes before, and in 45 patients one hour before examination. After water ingestion two-phase-split-bolus CTU protocol was applied (50 ml of contrast at 2 ml/sec, followed 10 minutes later with 70 ml at 2 ml/ sec). Unenhanced images were obtained, and combined nephrographic-excretory phase was obtained 100 seconds after start of second injection. 3D reconstructions were used to evaluate opaciﬁcation of collecting system and ureters, which was graded as poor, near-complete, or complete. Results: CTU one hour after water ingestion demonstrated complete opaciﬁcation of calices in 87.5%, renal pelvis in 97.5%, upper ureter in 91.8% and lower ureter in 87.5% of patients. CTU 20 minutes after water ingestion demonstrated complete opaciﬁcation of calices in 79.5%, renal pelvis in 85%, upper ureter in 62.5% and lower ureter in 54.5% of patients. In one-hour-delay group incidence of complete opaciﬁcation was signiﬁcantly higher than in 20-minutes delay group for upper (C2=8.12; p 0.01) and distal ureters (C2=9.40; p 0.01), while difference of opaciﬁcation of calices and pelvis was not signiﬁcant. Conclusion: CTU performed with distended bladder, one hour after oral ingestion of water, enables excellent opaciﬁcation of collecting system, including distal ureters.

B-483

10:48

Contrast-enhanced dual-energy computed tomography for detection of urinary stone disease H. Scheffel, P. Stolzmann, T. Frauenfelder, T. Schertler, L. Desbiolles, S. Leschka, B. Marincek, H. Alkadhi; Zurich/CH ([email protected]) Purpose: The purpose of the study was to evaluate the value of dual-energy contrast-enhanced CT imaging for the detection of urinary stone disease using dual-source CT. Methods and Materials: 40 consecutive patients (mean age 46.6p16.2 years, range 27-85 years) suspicious of suffering from urinary stone disease underwent dualsource CT. A three-phasic CT scan protocol including a standard unenhanced scan, a nephrographic and an excretory phase of contrast enhancement was performed. Nephrographic phase scan was acquired in the dual-energy mode (80 kV/400 mAs and 140 kV/95 mAs) allowing reconstruction of virtual unenhanced images. Number, size, and location of urinary stones were assessed by two blinded readers independently on standard and virtual unenhanced CT images. The anteroposterior abdominal diameters were measured to determine abdominal obesity. Results: Standard unenhanced CT revealed 35 uroliths in 18/40 (45%) patients, virtual unenhanced CT demonstrated 26 uroliths in 15/40 (38%) patients (kappa value 0.89). Three false-negative (FN) and no false-positive (FP) ratings occurred using virtual unenhanced CT. FN ratings occurred solely in obese patients. Sensitivity, speciﬁcity, positive predictive and negative predictive values for virtual unenhanced CT for the diagnosis of urinary stone disease were 83, 100, 100, and 88%, respectively. Conclusion: Virtual unenhanced CT images reconstructed from contrast-enhanced dual-energy CT allow detection of urinary stones with a good sensitivity and excellent speciﬁcity, but sensitivity is decreased in abdominally obese patients.

Scientific Sessions B-484

10:57

Differentiation of uric acid stones using dual-energy computed tomography: An ex-vivo evaluation P. Stolzmann, H. Scheffel, K. Rentsch, T. Schertler, T. Frauenfelder, S. Leschka, T. Sulser, B. Marincek, H. Alkadhi; Zurich/CH ([email protected]) Purpose: The management of urinary calculus disease depends on the presence or absence of uric acid (UA). The purpose of this study was to assess the potential of dual-energy computed tomography (CT) for the differentiation between UA-containing and non-UA-containing urinary stones in an ex-vivo setting. Methods and Materials: Twenty urinary stones of 10 different compositions (eight of pure composition: calcium oxalate, apatite, brushite, struvite, UA, ammonium acid urate, sodium hydrogen urate, or cystine; two of mixed composition: UA-sodium hydrogen urate and oxalate-brushite-struvite) and of two different sizes (smaller and larger than 5 mm) were examined in an ex-vivo work bench model using a dual-source CT scanner in a dual-energy mode (tube voltage 80 kV and 140 kV). The stones were classiﬁed as UA-containing or non-UA-containing. Sensitivity, speciﬁcity, positive predictive and negative predictive values were calculated using the crystallographic stone analysis as the gold standard. Results: Twelve out of 20 stones (60%) were non-UA-containing; and eight (40%) were UA-containing. The software automatically mapped 19/20 (95%) stones. Only one stone (UA, 2 mm) was missed. The software correctly classiﬁed all stones. The sensitivity for the detection of UA-containing stones (7/19) was 100% (95% CI: 59100%), the speciﬁcity was 100% (95% CI: 74-100%), the positive predictive value was 100% (95% CI: 59-100%), and the negative predictive value was 100% (95% CI: 74-100%). The attenuation values of the missed stone were manually plotted into the analysis sheet values and correctly classiﬁed as containing UA. Conclusion: The ﬁrst ex-vivo experience indicates that the differentiation between UA-containing and non-UA-containing stones can be accurately performed using dual-source dual-energy CT.

B-485

11:06

Purpose: To prospectively determine whether blood oxygenation level-dependent (BOLD)-MRI is able to assess changes in intrarenal oxygenation before and after release of ureteral obstruction. Methods and Materials: Nine male patients with a median age of 44 years (range 20-73 years) with unilateral acute ﬂank pain due to a distal ureteral calculus diagnosed by unenhanced CT underwent BOLD-MRI of the kidneys on a 1.5 Tesla MR unit on admission to the hospital and at a median time of 3 months (range: 1-6 months) after successful stone treatment. Total examination time including three morphological sequences was 15 minutes. Circular regions of interest (ROIs) were delineated on the R2* maps for cortex and medulla. R2* is related to the tissue content of deoxyhemoglobin which in turn is negatively related to the pO2 of blood. Thus, R2* is an indirect estimation of pO2. Results: During acute obstruction, R2* was signiﬁcantly lower in the medulla (p=0.004) and in the cortex (p=0.004) of the affected kidney compared with the contralateral unobstructed kidney suggesting increased oxygenation levels during obstruction. After release of ureteral obstruction by successful stone treatment, R2* increased signiﬁcantly in all patients from 11.0p2.0 to 16.1p0.9 in medulla (p=0.004) and from 10.4p0.7 to 12.2p0.8 in cortex (p=0.004), values similar to those found in the unobstructed kidneys. Thus, following release of ureteral obstruction, the oxygenation of the obstructed kidney turned normal. Conclusion: BOLD-MRI has the potential to noninvasively assess changes in intrarenal oxygenation during acute obstruction and to assure its normalization after successful treatment.

B-486

11:15

Intraindividual comparison of image quality in MR-urography at 1.5 and 3 Tesla using an animal model M. Regier, C. Nolte-Ernsting, G. Adam, J. Kemper; Hamburg/DE ([email protected]) Purpose: To compare image quality and anatomical depiction of the upper urinary tract using MR-urography at 1.5 and 3 Tesla in a porcine model. Methods and Materials: Using a T1w 3D-GRE sequence (TR: 5.1 ms; TE: 2.3 ms; FOV: 34.9x29.2 cm; matrix: 256x179) and a high-resolution (HR) T1w 3D-GRE sequence (TR: 5.4 ms; TE: 2.4 ms; FOV: 32.9x29.8 cm; matrix: 464x354), four healthy domestic pigs were examined at 1.5 and 3 Tesla. The MR-urographic scans were

B-487

11:24

Excretory MR-urography at 1.5 and 3 Tesla: Comparison with MDCT-urography M. Regier, C. Nolte-Ernsting, G. Adam, J. Kemper; Hamburg/DE ([email protected]) Purpose: To intraindividually compare the morphologic accuracies of contrast-enhanced MR-urography (MRU) and MDCT-urography (MDCTU) in a porcine model. Methods and Materials: After intravenous injection of low-dose furosemide, urographic imaging was performed in four pigs using a 1.5 and 3 T MR-unit and a 16-row MDCT-scanner. A T1w 3D-GRE (TR/TE: 5.1 ms/2.3 ms; matrix: 256x179) and a highresolution (HR) T1w 3D-GRE (TR/TE: 5.4 ms/2.4 ms; matrix: 464x354) sequence were used at both ﬁeld-strengths. SENSE-imaging was additionally utilized at 3 T. For MDCTU (collimation: 16x1.5 mm; tube-voltage: 120 kV), current-time product was set at 4 gradiations (200, 125, 70 and 30 eff.mAs). A 4-point-grading-scale was applied for comparing both modalities regarding the depiction of the urinary tract. Statistics included kappa-analysis and Wilcoxon`s test. Results: Anatomical depiction of the upper urinary tract was superior at 1.5 and 3 T MRU compared to MDCTU at 30 eff.mAs. At 70 eff.mAs, MDCTU was superior to MRU at 1.5 T using a T1w 3D-GRE sequence (mean 2.56; p=0.012), but slightly inferior to HR T1w 3D-GRE MRU at 3 T (mean 3.17; p 0.05). With additional use of SENSE at 3 T, the HR T1w 3D-GRE sequence was rated signiﬁcantly higher than MDCTU at 70 eff.mAs (mean 3.44; p=0.014). With higher current-time products of 125 and 200 eff.mAs, MDCTU showed better delineation of pelvicalices and ureters, irrespective of sequence parameters or ﬁeld-strength. Conclusion: MRU with parallel image acquisition at 3 T enables superior delineation of the upper urinary tract compared to low-dose MDCTU. Only at higher tube current-time products, MDCTU allows for a detailed depiction of the urinary tract and is superior to MRU at 1.5 and 3 T.

B-488

11:33

Age dependence of T1 perfusion MRI based hemodynamic parameters in normal human kidneys M.I. Dujardin1, R. Luypaert1, P. Van der Niepen1, S. Sourbron2, D. Verbeelen1, T. Stadnik1, J. de Mey1; 1Brussels/BE, 2Munich/DE ([email protected]) Purpose: The aim of this study was to determine the association between cortical perfusion parameters from T1 DCE-MRI and age in healthy human kidney. Methods and Materials: 36 patients (19-74 yr) were imaged using inversion recovery prepared FLASH during the injection of 10 ml of Gadolinium-DTPA. Tissue concentration-time courses were deconvolved with an inﬂow-corrected arterial input function (AIF); renal blood ﬂow (RBF), volume of distribution (RVD) and mean transit time (MTT) were derived from the resulting impulse response function. Results: Mean RBF, RVD and MTT were 130 ml/min/100 ml (SD 82 ml/min/100 ml), 40 ml/100 ml (SD 23 ml/100 ml) and 22 s (SD 9 s). RBF values were lower than those found in the literature, probably due to residual AIF inﬂow effects. A signiﬁcant negative correlation was found between RBF and patient age (R=-0.486, p=0.003) and RVD and patient age (R=-0.390, p=0.019). MTT and age did not show a signiﬁcant correlation (R=0.049, p=0.775). Repeating this analysis for each gender separately, we found comparable mean values but different aging patterns between both. A strong age dependence of RBF (R=-0.646 with p=0.003) and RVD (R=-0.479 with p=0.038) was found in the male, but not in the female group. Conclusion: Quantitative perfusion MRI was sufﬁciently sensitive to demonstrate a signiﬁcant negative correlation of RBF and RVD with patient age. This was due to a strong age dependence of these quantities in males that seems to be absent in females.

A

B

C

D

E

F

G

251

Sunday

Monitoring response to treatment after acute ureteral obstruction by functional magnetic resonance imaging H.C. Thoeny, T.M. Kessler, S.C. Simon-Zoula, T. Binser, U.E. Studer, P. Vermathen; Berne/CH ([email protected])

performed after intravenous injection of gadolinium-DTPA (0.1 mmol/kg) and lowdose furosemide (0.1 mg/kg). Image evaluation was performed by two independent radiologists blinded to applied sequence parameters and ﬁeld-strength. Image analysis included grading of image quality of the segmented collecting system based on a 5-point-grading scale regarding the anatomical depiction and degree of artefacts observed. The signal-to-noise ratio (SNR) was determined. Statistical evaluation included kappa-analysis and Wilcoxon`s test. Results: The mean scores for MR-urographies at 1.5 Tesla were 2.83 for the T1w 3D-GRE and 3.48 for the HR T1w 3D-GRE sequence. Signiﬁcantly higher values were determined using the corresponding sequences at 3 Tesla, averaging 3.19 for the T1w 3D-GRE (p=0.047) and 3.92 for the HR T1w 3D-GRE (p=0.023) sequence. Delineation of the pelvicaliceal system was rated signiﬁcantly higher at 3 Tesla compared to 1.5 Tesla (T1w 3D-GRE: p=0.015; HR T1w 3D-GRE: p=0.006). At 3 Tesla the mean SNR was 45.6% (p=0.039) higher. A Kappa of 0.67 indicated a good correlation among the two observers. Conclusion: In an experimental setup, MR-urography at 3 Tesla allowed for a signiﬁcantly higher image quality and SNR compared to 1.5 Tesla.

Scientific Sessions B-489

10:30 - 12:00

11:42

Feasibility of intraluminal optical coherence tomography of the human upper urinary tract in vivo: Preliminary experience U.L. Mueller-Lisse, M. Bader, Y. Hocaoglu, M. Pühls, C.G. Stief, M.F. Reiser, U.G. Mueller-Lisse; Munich/DE ([email protected]) Purpose: Catheter-guided optical coherence tomography (OCT) is a new means of intraluminal microstructural imaging, with a spatial resolution of 10-20 μm that has previously been applied to human coronary arteries in vivo, but not in the human upper urinary tract. We compared delineation of tissue layers of human ureters in vivo in OCT images. Methods and Materials: The study protocol was approved by the local hospital ethics committee. OCT was performed during endoscopy in eight patients with upper urinary tract lesions (urolithiasis, urothelial cancer, urinary obstruction). The OCT probe (diameter, 0.014 inch; LightLab Inc., Westport, MA, USA) was introduced into the human ureter either via a ureter catheter or an ureterenoscope. Single-slice cross-sectional images were obtained from within the ureter by means of OCT (M2, LightLab Inc., Westport, MA, USA). Delineation of urothelium, lamina propria, and muscle layer (inner and outer layer) was rated as possible (1) or not possible (0). Results: In all patients, OCT was performed without intraoperative complications. Cross-sectional OCT images were obtained in 92 different catheter positions (368 image quadrants). Image quality was sufﬁcient for evaluation in 305/368 quadrants (82.9%). Different wall layers were delineated in 300/305 (98.4%) quadrants, with distinction of urothelium/lamina propria in 274 (89.9%), lamina propria/muscle layer in 292 (95.7%), and inner/outer muscle layer in 160 (52.5%) quadrants, respectively. Conclusion: Intraluminal OCT of the upper urinary tract is feasible. OCT delineates different wall layers of human ureters in vivo. In principle, distinction between healthy human ureter and ureter lesions appears possible.

B-490

11:51

Can dual energy virtually unenhanced images replace non contrast scans in patients with renal masses? A. Graser, T.R. Johnson, M. Staehler, C.G. Stief, M.F. Reiser, K. Nikolaou, C.R. Becker; Munich/DE ([email protected]) Purpose: To evaluate whether dual energy MDCT virtually unenhanced images generated on a dual source CT scanner can replace non contrast acquisitions in patients with renal masses. Methods and Materials: 120 patients with renal masses (renal cell cancer, n=85; angiomyolipoma, n=5; hemorrhagic cyst, n=12; other entities, n=18) underwent preoperative dual energy CT (DECT, Somatom Deﬁnition, Siemens Medical Solutions, Forchheim, Germany). The scan protocol included unenhanced, nephrographic dual energy and delayed phases. Scan parameters for DECT were: tube potential, 80 and 140 kV; tube current, 404 and 90 mAs; collimation, 14x1.2 mm. For DECT images, image quality and level of image noise were rated on a 5-point scale. CT attenuation values for single energy pre contrast and dual energy virtually unenhanced images were compared using circular regions of interest. Measurements were taken in ﬁve anatomical regions: aorta, liver parenchyma, renal cortex, renal tumor, and psoas muscle. Differences in measurements were tested for statistical signiﬁcance using the Student’s t test. Results: Mean image quality for DECT was 1.5, mean image noise 1.3. Mean single energy non contrast and dual energy virtually unenhanced Hounsﬁeld units were as follows: renal parenchyma, 31.4p4.2 and 31.7p7.4, p=0.66; aorta, 42.2p4.2 and 43.9p8.0, p=0.16; liver, 55.8p8.7 and 57.9p9.6, p=0.08; psoas, 47.2p6.5 and 48.2p9.3 HU, p=0.40. Differences were not statistically signiﬁcant. Conclusion: In patients with renal masses, DECT scanning allows for exact assessment of attenuation in organ parenchyma and tumor tissue. Virtually unenhanced images can be used for baseline density measurements and replace a non contrast scan.

252

A

B

C

D

E

F

G

Room F2

Breast

SS 1002 MR: New clinical ﬁndings Moderators: R.A. Jong; Toronto, ON/CA J. Veltman; Nijmegen/NL

B-491

10:30

Analysis of ﬂuid in MR-mammography: What kind of additional information can the T2-weighted sequences provide? D.M. Renz, P.A.T. Baltzer, M. Herold, A.B. Herzog, M. Dietzel, M. Gajda, O. Camara, W.A. Kaiser; Jena/DE ([email protected]) Purpose: MR-mammography has the highest sensitivity in breast cancer diagnosis combined with varying observer-dependent speciﬁcity levels. In order to optimise accuracy, several dynamic and morphological parameters have been analysed. However, information presented by T2-weighted sequences has not been evaluated extensively. This study was performed to assess the potential of T2-images towards further improvement of speciﬁcity, with a special focus on oedema and intraductal ﬂuid. Methods and Materials: MR examinations (1.5 T; 0.1 mmol Gd-DTPA/kg bw) of 126 patients with 133 histologically veriﬁed lesions were evaluated. None of the subjects had breast surgery or radiation therapy before. Histopathology revealed: 81 invasive carcinomas (55 poorly differentiated, 26 moderately differentiated) and 52 benign lesions (18 papillomas, 13 ﬁbroadenomas, 11 mastopathic proliferation, ﬁve mastitis, three atypical ductal hyperplasia, one hemangioma, one hamartoma). Results: Evaluating turbo-spin-echo T2-sequences, 70.4% of all malignomas presented unilateral oedema in the following localisations: 69.1% perifocal, 25.9% diffuse, 6.2% cutaneous/subcutaneous, 12.3% perimamillar, 38.3% prepectoral, 7.4% intramuscular pectoral. Of 52 benign lesions, oedema occurred only in 15.4% (four mastitis, one ﬁbroadenoma, one papilloma, two ADH; p 0.05), mainly in a diffuse localisation (15.4% diffuse, 7.7% perifocal, 3.8% prepectoral). In 83.3% of all papillomas, ﬂuid was observed within the ducts whereas only 16.0% of all carcinomas presented intraductal ﬂuid (p 0.05). Conclusion: According to our ﬁndings, breast oedema is strongly associated with malignomas, especially occurring in a perifocal localisation. Except for mastitis, benign lesions rarely present oedema. However, in most papillomas, intraductal ﬂuid can be detected. T2-weighted images therefore provide useful additional information for clinical routine.

B-492

10:39

Role of magnetic resonance mammography to detect DCIS-associated invasive component underestimated at percutaneous biopsy N. Martini, V. Londero, A. Dal Col, M. Bazzocchi, C. Zuiani; Udine/IT ([email protected]) Purpose: To evaluate usefulness of Magnetic Resonance Mammography (MRM) in pre-surgical assessment of Ductal Carcinoma In Situ (DCIS) of the breast, particularly to reveal associated invasive component not evident at biopsy. Methods and Materials: We enrolled 48 women with 44 pure DCIS and four with suspected invasion at vacuum-assisted biopsy (11G). All these lesions presented contrast uptake at MRM performed with 1.5 T scanner. In 31 of these, histological proof on surgical specimens conﬁrmed the diagnosis; in 17 cases it revealed associated invasive component (microinvasive in ﬁve patients). Results: All 31 cases of pure DCIS presented non-masslike enhancement at MRM, eight with focal/regional distribution, 13 linear-branching, 10 segmental. In 11 out of 17 DCIS with invasive component, MRM documented mass-like enhancement, with features suggesting malignancy (ill-deﬁned/spiculated margins (all masses), “rim” enhancement (one case), washout (nine out of 11 lesions), enhancement rate 100% (nine out of 11 DCIS)); in ﬁve out of these 11 cases a non-masslike enhancement component also coexisted. The remaining six of the 17 DCIS with invasive component presented a pure non-masslike enhancement; four of these lesions were DCIS with microinvasion only. The MRM positive predictive value (PPV) for DCIS-associated invasive component was 100%, the negative predictive value (NPV) was 84%. Conclusion: Mass-like pattern at MRM exhibits high PPV for DCIS-associated invasive component. Non-masslike enhancement does not exclude invasive component, but more often this is microinvasive.

Scientific Sessions B-493

10:48

The adjacent-vessel-sign in MR-mammography: New data from a study of 366 cases M. Dietzel, P.A.T. Baltzer, D.M. Renz, A.B. Herzog, M. Gajda, O. Camara, W.A. Kaiser; Jena/DE Purpose: The adjacent-vessel-sign (AVS) has recently been introduced as a new feature in MR-mammography (MRM). This study was conducted to (a) analyze AVS in a larger collective, (b) evaluate the performance of AVS in the most common breast malignancies, and (c) assess the correlation between AVS and tumour size. Methods and Materials: Over 26 months all histologically veriﬁed lesions imaged by MRM at our institution were prospectively evaluated by experienced radiologists ( 500 MRM) according to standard protocols and study design (2D-FLASH sequences, 0.1 mmol/kgBWGd-DTPA, T2-TSE). There were 183 malignant (26 CIS; 95 invasive ductal carcinomas) and 183 benign lesions. It was investigated whether an adjacent vessel leading to an enhancing lesion was seen in subtraction images. The occurrence of AVS was assessed and statistical indices were calculated for 3 categories: (a) overall prevalence (b) histological subgroups (invasive vs in situ carcinomas; invasive ductal carcinomas), and (c) different size categories. Results: (a) Overall prevalence: sensitivity (41.5%), speciﬁcity (86.9%) and PPV (76.0%). (b) Invasive cancers: sensitivity (43.3%), PPV (73.9%) - In situ: Acc. (79.9%), PPV (25.0%) - Invasive ductal cancers: sensitivity (47.4%), PPV (65.2%) and Acc. (73.4%). (c) PPV and speciﬁcity of AVS was best in small carcinomas ( 10 mm: speciﬁcity: 96.5%, PPV: 76.9%; Acc: 79.5%), whereas sensitivity was the highest in advanced tumors ( 10 mm: 32.3%; 30-50 mm: 56.5%). Conclusion: This study showed the potential of adjacent-vessel-sign (AVS) to differentiate malignant from benign lesions by MR-mammography. Yet, the performance of AVS highly depends on several factors, such as the size of a lesion and histology of the malignant tumor.

B-494

10:57

Purpose: An asymmetric increase of whole breast vascularity (WBV) ipsilateral to a unilateral cancer has been previously reported on CE-MRI with a frequency from 68 to 88% in series of mainly invasive cancers (DCIS 0-20% of cancers). We evaluated the WBV in patients with unilateral DCIS. Methods and Materials: Eight centers reviewed 75 CE-MRI examinations (1 or 1.5 T, 3D gradient-echo sequence, 0.1 mmol/kg of a Gd-chelate) in 75 patients (age 55.4p10.4 years) with pathologically demonstrated unilateral DCIS (diameter 22.4p18.2, median 15 mm). Vessels reaching r2 mm in thickness and r3 cm in length were counted on axial/coronal MIPs for both breasts and a difference r2 in these vessel was considered as WBV asymmetric increase. The frequencies of WBV symmetry, increased WBV ipsilateral to the DCIS (ipsilateral asymmetry), and increased WBV contralateral to the DCIS (contralateral asymmetry) were obtained; 95% CIs were calculated from the exact binomial distribution. MannWhitney U test was used. Results: WBV symmetry was found in 56/75 (74%; 95% CI, 63-84%), ipsilateral asymmetry in 18/75 (24%; 95% CI, 15-35%), and contralateral asymmetry in 1/75 (1%; 95% CI, 0-7%). No signiﬁcant difference was found between DCIS diameter of the cases with WBV symmetry (median 16 mm) and that of the cases with ipsilateral asymmetry (median 14 mm); the same was for pathological grading (2.4p0.7 versus 2.5p0.6, respectively). Conclusion: Unlike invasive cancers, DCIS resulted in frequent (74%) association with WBV symmetry. As a consequence, the asymmetric WBV increase can be considered as a predictor of invasiveness of an ipsilateral and unilateral tumor.

11:06

Magnetic resonance in the evaluation of pure ductal carcinoma in situ of the breast M. Forment Navarro, J. Camps-Herrero, V. Ricart Selma, J. Cano Gimeno, A. Caño Gómez, A. Escobar Hoyos, C. Martínez Rubio, A. Torregrosa Andres; Alzira/ES ([email protected]) Purpose: To evaluate breast MRI sensitivity for pure ductal in situ carcinoma and to quantify the usefulness of parametric images in the estimation of the tumor’s maximum diameter. Methods and Materials: We retrospectively reviewed 44 histologically diagnosed pure ductal in situ carcinoma patients. These patients had the diagnosis of pure DCIS with a large core needle biopsy (14 gauge) and a presurgical MRI was performed for accurate cancer staging. We used a dynamic 3D study after endovenous contrast administration. Images were analyzed morphologically and semiquantitatively following Fischer’s Scale. We obtained parametric maximum enhancement and maximum enhancement speed images and measured the largest tumour diameter. Largest diameter obtained with parametric images, and the one obtained with histologic study were correlated using Pearson’s coefﬁcient. The correlation was also obtained for the different histologic grades. Results: Breast MRI sensitivity for pure DCIS was 86.4%. Correlation coefﬁcients were 0.764 and 0.851 for the correlation between histologic diameter, and maximum enhancement and maximum enhancement speed images, respectively. DIN1c, DIN2 and DIN3 tumours showed correlation coefﬁcients of 0.120, 0.399 and 0.924 for maximum enhancement speed parametric images, and 0.053, 0.332 and 0.889 for maximum enhancement parametric images, respectively. Conclusion: Breast MRI is a sensitive technique for DCIS detection, and it showed a good correlation in the largest tumour diameter estimation. This correlation was only signiﬁcant in undifferentiated DCIS (DIN3 tumors) and was better in maximum enhancement speed images.

B-496

11:15

Inﬂammatory breast carcinoma in MR-mammography: Characteristic ﬁndings for dynamic and morphological parameters in comparison with locally advanced breast cancer D.M. Renz, P.A.T. Baltzer, J. Böttcher, S.O.R. Pﬂeiderer, M. Gajda, O. Camara, W.A. Kaiser; Jena/DE ([email protected]) Purpose: In comparison with other forms of invasive breast carcinoma, the present knowledge about the typical appearance of inﬂammatory cancer in MRmammography is fragmentary. The aim of this study was to identify typical MRI features of this highly malignant tumour in comparison with noninﬂammatory locally advanced cancer. Methods and Materials: MR examinations (1.5 T; 0.1 mmol Gd-DPTA/kg bw) of 48 patients with inﬂammatory carcinoma were compared with 52 subjects with noninﬂammatory locally advanced cancer. Age and histopathological subtype were equivalent between the two groups. To delineate characteristic imaging features, multiple dynamic and morphological parameters were evaluated. Results: No signiﬁcant differences of prevalences were found for the following criteria (p 0.05): dynamic signal characteristics, prominent vessels, perifocal oedema, axillary node involvement, morphology of focal masses and morphological pattern of non-mass like enhancement. Otherwise, the quantity of focal masses and the spatial distribution of the tumoural inﬁltration signiﬁcantly differed between the two cancer groups. The following parameters occurred more frequently in inﬂammatory carcinomas: oedema (cutaneous/subcutaneous 81.3%, perimamillar 70.8%, diffuse 89.6%, prepectoral 72.9%, intramuscular pectoral 41.7%), thickening (75.0%) and pathological enhancement (60.4%) of Cooper's ligaments, skin thickening (83.3%), punched-out sign (initially strong, focal increase of some parts followed by slowcontinuous enhancement of the surrounding skin; 56.3%). Conclusion: Inﬂammatory breast carcinoma seems to represent a speciﬁc biological entity resulting in distinct MRI characteristics. Some of the parameters (e.g. punched-out sign, thickening and pathological enhancement of Cooper´s ligaments) are supposed to visualise the extensive lymphovascular inﬁltration and therefore may improve the diagnosis of inﬂammatory carcinoma in MRI.

B-497

11:24

The evaluation of the characteristic ﬁdings of ductal carcinoma in situ on delayed contrast-enhanced coronal images with fat-suppression of MRI S. Kobayashi, H. Sugimoto; Shimotsuke/JP ([email protected]) Purpose: To evaluate the characteristic ﬁndings of ductal carcinoma in situ (DCIS) on delayed contrast-enhanced coronal images with fat-suppression (DCECIs) of MRI. Methods and Materials: From June 2005 to January 2007, 36 consecutive pathologi-

A

B

C

D

E

F

G

253

Sunday

Ductal carcinoma in-situ (DCIS): Symmetric or asymmetric whole breast vascularity on contrast-enhanced MR imaging (CE-MRI)? A multicenter study F. Sardanelli1, G. Di Leo1, A. Fausto1, A. Esseridou1, C. Zuiani2, S. Corcione3, E. Zanon4, M. Calabrese5, L. Preda1, J. Nori6, G. Valeri7, G. Rescinito5, A. Pecchi8, L. Pescarini9, M. Bazzocchi1; 1Milan/IT, 2Udine/IT, 3Ferrara/IT, 4Turin/IT, 5 Genoa/IT, 6Firenze/IT, 7Ancona/IT, 8Modena/IT, 9Padova/IT ([email protected])

B-495

Scientific Sessions cally diagnosed DCIS with or without minimal invasion were retrospectively reviewed.The pathological diagnosis was established by core needle biopsy (n=1) or surgical specimens (n=35). MRI was performed using a 1.5-T system (Symphony, Siemens, or Achieva, Philips). Affected side was examined in a prone position using either a breast array coil in Symphony or dual ﬂex-M coils in Achieva. DCECIs were obtained about six minutes after the injection of 0.1 mmol/kg gadolinum-chelate contrast material with saline ﬂash by using an automatic injector.The parameters were as follows: TR/TE/FA, 540-650/8-11/90; FOV, 18-20 cm; matrix, 256x256; thickness, 5 mm; gap, 1-1.5 mm. Retrospective image analysis was performed on PACS (Synapse, Fujiﬁlm). The following criteria were scored based on DCECIs: (1) mass- or non-mass-like enhancement; (2) distribution modiﬁers for non-mass-like enhancement (focus, focal area, linear, ductal, segmental, regional, multiple regions, or diffuse); (3) internal enhancement (homogeneous, heterogeneous, rim, stippled, clumped, reticular, or clustered ring enhancement). Results: Two of 36 (5.6%) mass-like enhancements were observed (one heterogenous and one clustered ring enhancement). 34 (94.4%) non-mass-like enhancements were observed and all of them showed segmental distribution (28 clustered ring enhancement, four clumped, and two stippled). Conclusion: Clustered ring enhancement of segmental distribution was seen most and was thought to be a characteristic ﬁnding of DCIS on DCECIs of MRI.

B-498

11:33

Feasibility of T2-weighted sequences in the differential diagnosis of masslike enhancements at magnetic resonance imaging of the breast L. Ballesio, S. Savelli, M. Angeletti, I. D’Ambrosio, A. Francioso, C. De Felice, L. Manganaro; Rome/IT ([email protected]) Purpose: To evaluate the potential role of signal intensities calculated in T2 images as an adjunctive parameter in the analysis of mass-like enhancements classiﬁed as BI-RADS assessment categories 3, 4 or 5 with the standard T1 criteria. Methods and Materials: After a retrospective review of 212 breast Magnetic Resonance Imaging (MRI) performed for the evaluation of a suspicious lesion, we selected a group of 53 mass-like-enhancements ranging from ﬁve to 20 mm, classiﬁed as 3, 4 or 5 BI-RADS assessment categories, histologically proved. In all cases we calculated the ratio between nodule and pectoral major muscle signal intensity (LMSIR) with a multiROIs analysis on T2 images. An ROC analysis was performed to test the ability of both the diagnostic parameters (BI-RADS and LMSIR) separately and combined in a new mono-dimensional variable whose values were obtained by a computerized discriminant function. Results: Histological examination assessed 35 cases with malignant lesions (66%) and 18 cases with benign lesions (34%). The separate evaluation of ROC curves gave the following results: BI-RADS area under the curve (0.838p0.054), LMSIR (0.817p0.067), combined BI-RADS-LMSIR (0.916p0.038): an increment of the area under the ROC curve was observed even if no signiﬁcant statistical difference was assessed. Conclusion: T2-weighted SI assessment with LMSIR measurement may improve the diagnostic information content of standard breast MRI and can be considered a potential tool in the differential diagnosis of mass-like enhancements judged as borderline lesions (BI-RADS 3 and 4).

B-499

11:42

Breast lesions: Correlation of contrast medium enhancement parameters with histopathologic ﬁndings and tumor angiogenesis F.-H. Li; Shanghai/CN (proﬂ[email protected]) Purpose: To study the quantitative analysis of breast tumors with real-time grayscale contrast-enhanced ultrasound (CEUS), and the relationships between enhancement parameters and histological microvessel density (MVD). Methods and Materials: Fifty-two patients with breast masses (25 benign, 27 malignant) underwent CEUS examination with Technos DU8 contrast turned imaging (CnTI) technique and SonoVue. Intratumoral vascularity was assessed by time-intensity curve analysis software to measure the quantitative parameters. Immunohistochemical staining using anti-factor CD34 was performed to evaluate the microvessel density in surgical specimens. Results: The enhancement parameters showed that the peak intensity, enhancement amplitude, areas under begin-90 s of the time-intensity curves and begin-90 s mean intensity were signiﬁcantly higher in malignant lesions than those in other benign tumors (P 0.01). MVD also showed signiﬁcant difference between the two groups (p 0.001). Malignant lesions contained more microvessels than benign lesions. It showed signiﬁcantly positive correlations between the enhancement parameters and MVD, and areas under begin-90 s of the time-intensity curves had the best correlation with MVD (r=0.89, p 0.001). Conclusion: The enhancement parameters with real-time gray-scale contrastenhanced sonography of breast lesions are well correlated with MVD, which can be used to estimate the angiogenesis in breast neoplasms in vivo.

254

A

B

C

D

E

F

G

B-500

11:51

Does the destroyed ligament sign provide further information for differential diagnosis in MR-mammography? M. Dietzel, P.A.T. Baltzer, A.B. Herzog, M. Gajda, D.M. Renz, O. Camara, W.A. Kaiser; Jena/DE Purpose: Destroyed-ligament-sign (DLS) is a new feature in MR-Mammography (MRM). We conducted this study to evaluate the potential of DLS for differential diagnosis. Methods and Materials: Over 26 months, all histologically veriﬁed lesions imaged by MRM at our institution were prospectively evaluated by experienced radiologists ( 500 MRM-examinations) according to standard protocols and study design (2DFLASH, 0.1 mmol/kg BWGd-DTPA, T2-TSE). There were 183 malignant (26 CIS, 95 invasive-ductal-carcinomas etc.) and 183 benign lesions (two phylloid-tumors, two radial-scars, 34 ﬁbroadenomas, etc.). DLS was assessed, if the perifocal Cooper-ligaments could be visualized by MRM in T2-w. It was rated positive, if these ligaments were dissolved or destroyed by a lesion, i.e. they could not be depicted within a lesion. Prevalence of DLS was documented in a database and statistical indices were calculated. Additionally, the performance of DLS was analyzed in dependence of histological outcome and size of the lesions. Results: Prevalence was 22.4% for benign lesions (phylloid, radial-scar: 50%; ﬁbroadenoma 29.4%). It was visualized in 59.6% of the malignant lesions (CIS 30.8%; invasive-carcinomas 64.3%; invasive-ductal-carcinoma 69.5%). Overall sensitivity reached 59.6%, speciﬁcity 77.6% and PPV 72.7%. NPV was best for CIS (88.8%). Regarding tumor size, DLS reached the highest sensitivity (81.5%), speciﬁcity (81.3%) and PPV (88.0%) in tumors between 30 and 50 mm in diameter. In small cancers ( 10 mm) performance was slightly worse regarding sensitivity (32.4%) and PPV (50.0%), yet speciﬁcity was still high (86.0%). Conclusion: Destroyed-ligament-sign is a potent tool for differential diagnosis by MR-Mammography. It seems to be particularly useful in the assessment of CIS, advanced lesions and radial scars.

10:30 - 12:00

Room G/H

Computer Applications

SS 1005 Radiological reporting and teaching Moderators: G. Pärtan; Vienna/AT G. Turóczy; Budapest/HU

B-501

10:30

Evaluating the impact of voice-recognition reporting on radiology training V. Shah, J.L. Hart, D. Blunt; London/UK ([email protected]) Purpose: Whilst there is evidence to demonstrate the profound impact of voice recognition reporting (VRR) on report completion times and departmental workﬂow, there has been no adequate evaluation of the effect on training in either quantitative or qualitative terms. Methods and Materials: Individual RIS data were gathered for trainee radiologists for a six-month period pre- and post-introduction of VRR (December 2004). These were grouped and analysed according to imaging modality. Results: During the study period, the mean total number of studies per month was 17,342 of which a mean of 2352 (14%) were reported by trainees. No signiﬁcant change in the overall number or percentage of reports produced by trainees was noted following the introduction of VRR. However, the mean number of cross-sectional (CT/MR) reports increased signiﬁcantly from 32 to 59 per trainee per month (2-tailed Student’s t-test; p=0.012). Conclusion: VRR has not reduced the number of studies reported by trainees, contradicting available (anecdotal) evidence. Trainee cross-sectional (CT/MR) modality report numbers have increased signiﬁcantly. In addition, there are unmeasured qualitative beneﬁts. With VRR, studies are reported, amended and checked rapidly in real time. This facilitates “parallel reporting” which is of particular value in crosssectional imaging lists. The ability to discuss cases and review reports in real time improves trainer-trainee interaction, allowing added emphasis on report quality in addition to analysis of radiological ﬁndings. Clarity of expression and appropriate emphasis in radiology reports is an important (and possibly neglected) aspect of radiological training which can be improved with VRR.

Scientific Sessions B-502

10:39

Cross language interpretation of spine imaging based on structured reporting M. Fatehi1, S. Akhlaghpoor1, A. Arjmand Shabestari1, M. Bagheri2, A. Marashi3, A. Poli4, H. Alizadeh1; 1Tehran/IR, 2Shiraz/IR, 3Darmstadt/DE, 4Trieste/IT ([email protected]) Purpose: Structured reporting (SR) is one of the modern trends in radiology interpretation which has limitations and advantages. Potential role of SR in language-independent radiology interpretation is the major idea behind our work. We believe that if radiology images are interpreted using a multi-level SR platform using a graphical interface and point-and-click methods, the need for translation will be eliminated and the report can be generated in any predeﬁned language in the system. Methods and Materials: An SR platform was designed to cover interpretation of all types of spine imaging including conventional X-ray, CT and MRI. Graphical inteface was developed to provide a point-and-click basis for selecting the area to be described. The model includes Persian, German, English and Italian interfaces. Results: Every radiologist can interpret the images in his/her own language using this SR system. The report can be reproduced in any desired but predeﬁned language for viewing, printing and sending. In this system, SR provides a controlled terminology by which no words or phrases are stored but a tree-structure of ﬁndings that can be displayed according to a dictionary embedded in software interface. No time is needed for translation. Conclusion: We believe that structured reporting can be a part of integrated digital radiology workstation providing added value to interpretation in modern global medical imaging. There exist reasons for availability of reports in native local languages. Legal issues, social security organization rules, medical records and patients’ awareness are among these reasons.

B-503

10:48

Purpose: ePACS concept as a modern e-learning approach presented in ECR 2007 indicated that PACS has educational potential for radiology departments. Addition of structured reporting methods to management of report content will help join daily routine practice with e-learning and so the PACS archive will become a teaching ﬁle. Methods and Materials: A structured reporting platform was designed for interpretation of HRCT of lung. The system is capable of storing the interpretation of attending radiologists (or those approved by him/her) as the “reference report” for educational purposes. Any resident of radiology can access these reports along with PACS archive. The resident will be asked to interpret the study using the same SR system. After description of the study the system will compare the report of the resident with reference report and give a feedback on details of description. Results: Our model for SR of HRCT of lung can be used to train residents of radiology or participants of post-graduate courses on CT. They can use structured reporting platform of the system to compare their reports with approved ones simulating a supervised interpretation in real department environment. Conclusion: Combination of PACS and SR in a university hospital environment will enhance training potential in a way that residents can use PACS/RIS archive to improve their interpretation and description skills without disturbing daily routine, for repeated times and also from distant areas. The training will be tailored to speciﬁc needs of each resident according to his/her report content.

B-504

10:57

Virtual anatomy enabled by high resolution computed tomography and magnetic resonance imaging and post processing tools: A new way of teaching and learning? F.L. Giesel1, S. Doll1, J. Mews2, A. Mehndiratta1, H. von Tengg-Kobligk1, M. Owsijewitsch1, J. Kirsch1, H.-U. Kauczor1; 1Heidelberg/DE, 2Neuss/DE ([email protected]) Purpose: Learning and teaching in a medical school, students are affected by visual memorisation of anatomical structures dissected on embalmed cadavers in the gross anatomy lab. This study investigates a new approach of anatomical learning, combining the traditional lab dissection work and side-by-side new imaging and post processing techniques from radiology. Methods and Materials: 29 students who were participating in the gross anatomy course were also enrolled in a ‘Virtual Anatomy Seminar’ in order to

B-505

11:06

The multimedia, virtual MR imaging department P.B. Marthinsen1, J. Hald1, R. Bergstrøm2, J. Jakobsen1, D. Christensen1, H. Roterud1, F. Lærum1; 1Oslo/NO, 2Trondheim/NO ([email protected]) Purpose: Evaluate the use of integrated online PACS/RIS and multi-channel HDTV videoconferencing for running multiple remote MR-laboratories. Methods and Materials: Lack of peak competent MR-radiologists may be a temporary or permanent problem in some hospitals. Special diagnostic problems or critical time windows for onset of therapy may require special competence and direct communication between physicians, including simultaneous interpretation of MR images. Instead of using conventional teleradiology, we established a solution with online multi-vendor PACS/RIS and simultaneously open videoconference to several MR laboratories located in remote hospitals. A studio has been established inside our department, with a four-channel HDTV videoconference-system to allow direct communication between the MR-radiographer and the radiologist. The remaining three channels were used for a document camera, a patient camera and a camera for communication between the remote clinician and the central radiologist. Results: Several technical problems had to be solved during the setup process, to get the system feasible. Once running, the bi-directional multichannel-HDTV videoconference and the integrated PACS/RIS gave a recognizable environmental setting for communication, almost as if the central radiologist was present at the remote department. The images to be read were of high quality, and one could get access to and switch between different remote MR-laboratories. Conclusion: The intention to ease the access to highly competent MR-expertise in centralised institutions may be achieved in our virtual MR imaging department. Utility and efﬁciency will be discussed.

B-506

11:15

Podcasts for radiology teaching: Initial experience E.E.J.G. Coche, P. Meurrens; Brussels/BE ([email protected]) Purpose: To explain the practical realization of podcasts for radiological teaching delivery. Methods and Materials: Teaching ﬁles are presented orally every week by residents in the library of the department of Medical Imaging. Oral presentations are performed with powerpoint ﬁles using a personal computer connected to the network. Audio is captured using high frequency Clip-on microphone (Omni, Sennheiser, Wedemark, Germany) and a bodypack transmitter and receiver (Sennheiser, Wedemark, Germany). The teaching ﬁle in powerpoint format is transferred to a portable computer (Macintosh Powerbook 2GhZ Intel Core Duo, Apple Computer, Cupertino, CA, USA) equipped with Mac Os X 10.4.2 and Wirecast 3.0 (Varasoftware London, England) for broadcast. Sharing ﬁles between the two computers is performed with Desktop Presenter (Varasoftware London, England) and connection to the network. This software streams the computer’s desktop to Wirecast. Wirecast incorporates that desktop into the ﬁnal broadcast. The podcast is loaded on an external server using internet. An internet domain with the following address is activated: http://radiologypodcasteur.dk. Results: Quality of podcasts is excellent. Powerpoint ﬁles are presented with audio synchronisation. A podcast of 10 minutes duration represents approximately 12 megabytes. Podcasts are available via the itunes platform and downloadable by medical students on PC, Mac or ipodvideo from the web server. Conclusion: Podcasts appear as a promising technique for disseminating knowledge and teaching radiology via the web.

A

B

C

D

E

F

G

255

Sunday

Added value of structured reporting for ePACS concept: A model for lung HRCT interpretation M. Fatehi1, A. Arjmand Shabestari1, S. Akhlaghpoor1, M. Bagheri2, K. Yazdanifard1; 1Tehran/IR, 2Shiraz/IR ([email protected])

enhance their knowledge with the help of high-resolution computed tomography (CT) and magnetic resonance imaging (MRI) modalities. The pictures were presented subsequently on a post processing workstation (MIPAV, NIH, Bethesda and Vitrea®, Vital®, Minnesota, USA). After the course students were asked to answer 15 questions regarding the pros and cons of this new approach. Results: 27 students scored the new approach as an ideal combination. Only six preferred simply computer-based learning without any practical dissection. Half of the group indicated another added value: they appreciated the possibility to utilize their anatomical knowledge from dissection work to correlate with medical imaging modalities simultaneously. Conclusion: Imaging and post processing technologies in radiology are improving and are opening new horizons for other medical disciplines. 3D-teaching courses facilitate the spatial understanding of structures, it helps to transformation written knowledge onto a virtual plane, and hence on another level of understanding. This is as well eminently important for medical students as the technical progress of examination methods proceeds and they therefore have to learn how to rapidly interpret data such as CT and MRI pictures.

Scientific Sessions B-507

11:24

Seamless integration of the radiological teaching database MIRC with PACS clients according to IHE TCE C.M. Ahlers, K. Wunder, J. Schneider, C. Düber, P. Mildenberger; Mainz/DE ([email protected]) Purpose: Purpose of the software development project was the vendor independent integration of the teaching database MIRC with PACS systems. Methods and Materials: A software tool according to the IHE TCE integration proﬁle was developed as interface between PACS clients and the radiological teaching database MIRC. Linked to any DICOM send capable PACS client, the tool allows semiautomated creation of radiological teaching ﬁles within clinical workﬂow. The tool provides a graphical user interface for categorization and image transfer, entering of keywords and clinical information. Implementation of the RadLex categorization system allows use of standardized terminology and therefore permits efﬁcient keyword based retrieval of MIRC documents. Hereby a signiﬁcant repository for teaching and education can be built up at minimal effort using MIRC or other teaching ﬁle systems. Selected cases will be duplicated to a virtualized MIRC instance to serve as public image repository for university wide e-learning projects. Results: The tool was developed and intensively tested at our institution for about six months. Up to 50 signiﬁcant teaching cases per week could be created within this period of time. Evaluation revealed high user acceptance and comfort. The tool will be made available to the public as freeware for multiple platforms before the end of 2007. Virtualization techniques are used for multiplatform MIRC implementation, case publication and integration with an e-learning platform. Conclusion: Our implementation of IHE TCE links all DICOM send capable PACS clients to the MIRC system for semiautomated comfortable creation of signiﬁcant radiological teaching ﬁles.

B-508

11:33

Tele-medicine and tele-collaboration: How MDCT is changing the rules of the workﬂow game H.E. Gregersen; Aalborg/DK ([email protected]) Purpose: As MDCT scanners become mainstream modalities, the requirements for tele-medicine and tele-collaboration systems are changing. In this study, several systems/methods are examined regarding the support provided for the workﬂow of typical orthopedic radiologists, speciﬁcally the clinical need to remotely access patient data and the need for consultation with other clinicians, in the context of trauma cases. Methods and Materials: Three systems/methods are evaluated for providing access to MDCT data, both to the radiologist and to consulting clinicians: (1) traditional distribution/access techniques; (2) standard web-based PACS; and (3) client-server-based distribution using server-side computing and thin-clients. The effectiveness of each system/method is analyzed regarding: (1) overall time to reach a diagnostic conclusion; and (2) subjective difﬁculty/effectiveness of the overall workﬂow. Twelve trauma cases are used - six occurring during work-hours and six on-call cases - all involving whole-body MDCT scans. Results: For “routine” injuries such as simple fractures, for which sub-samples of full MDCT data are sufﬁcient, all three systems/methods are effective, although timeto-diagnosis varies. For complex traumas, which often require access to full MDCT data, the system/method utilizing server-side computing and thin-clients performs signiﬁcantly better, both for time-to-diagnosis and effectiveness of workﬂow. Conclusion: As use of MDCT increases, especially for trauma cases, so do the needs of radiologists to access full-ﬁdelity MDCT data in a timely and facile manner, especially for tele-medicine and tele-consultation purposes. Traditional methods of access are not sufﬁcient, and new thin-client techniques appear promising in resolving challenges posed by the sheer size of MDCT data.

B-509

11:42

Singapore-India teleradiology link: Our experience in the primary health care setting A. Kalyanpur1, N. Sreenivasan1, J. Singh1, L. Tan2, T. Goh2; 1Bangalore/IN, 2 Singapore/SG ([email protected]) Purpose: Prior to teleradiology implementation, radiographs performed at primary health care facilities managed by the National Healthcare Group, Singapore had a mean report turnaround time of three days, using a workﬂow that involved physical transportation of ﬁlms for reporting. Following the implementation of teleradiology, this workﬂow was modiﬁed to transmit images to a reading facility in Bangalore, India for interpretation. The purpose of this study was to evaluate the beneﬁts of this teleradiology model.

256

A

B

C

D

E

F

G

Methods and Materials: Over the period January-December 2006, outpatient radiographs performed at four primary health care facilities administered by the National Healthcare Group, Singapore, were transmitted via teleradiology to a reading facility in Bangalore, India, for interpretation. We evaluated report turnaround times, quality assurance statistics and physician satisfaction with the new reporting workﬂow. Results: 35,531 radiographic examinations were reviewed via teleradiology over the study period. Report turnaround time was under one hour in 93% of studies and under 30 minutes in 69% studies. There was a 0.02% error rate for radiologic misses, indicating a report accuracy of 99.98%. Increased physician satisfaction, indicated by increased referral volumes (140%) and cost savings were also noted. Patients were given their radiologic reports at the initial visit and thereby saved a repeat visit to the polyclinic, resulting in improved manpower productivity. Conclusion: Utilization of international teleradiology by primary health care facilities in Singapore has resulted in signiﬁcantly improved report turnaround time and enhanced physician satisfaction, as also improved manpower productivity in the population being served.

B-510

11:51

Are mobile home-radiology workplaces as good as ﬁxed radiology workplaces? Analysis after one year of experience P. Soegner, R. Frank, A. Klauser, H. Zunterer, C. Cihak, F. Koppelstaetter, D. zur Nedden; Innsbruck/AT ([email protected]) Purpose: Do mobile home-radiology workstations offer more time ﬂexibility to radiologists with special demands like single parents or radiologist with long travel times to work at home? We evaluated the home-radiology scenery about their concordance with the Austrian medicolegal requests and the routine demands of radiologists. Methods and Materials: Five radiologists used a MacBookPro (2.6-MHz Intel CoreProcessor) in combination with a 30-inch Apple Monitor. Virtual Private Network (VPN) to the MUI (Medical University of Innsbruck) was realized with different local network providers (bandwidth from 2 to 6 Mbit). The radiology-information-system (RIS) used was an online client of Magic-SAS (Siemens-Medical, Austria, AG) and the medicolegal certiﬁed version of OSIRIX (PACS) (Aycan-GmbH, Germany) with a real-time connection to AIM (icoserve-GmbH) and MAGIC-STORE (Siemens Medical Austria-AG). Results: 398 studies (CT, Chest-X-rays, Ultrasound and MRI) were transmitted over VPN, 574 studies were copied during daily routine over local area network (LAN). The home workstations showed a sufﬁcient diagnostic accuracy. The mean time for training was 68 minutes; the mean reading times vary from two (Chest-Xrays) to 17 minutes (large CT studies with over 2500 images). The diagnostic tools show no signiﬁcant difference between the workstation at the MUI or at home. The online connection over the VPN networks were stable and very sufﬁcient with low error rates. Total investment for one workplace was 750,000 Euro, average costs for VPN 6900 Euro/month. The release time of ﬁnal reports was reduced into hours by ﬁnalizing in the evening from home. Conclusion: The home-radiology workstations offered radiologists new opportunities for higher ﬂexibility employment contracts, e.g. single parents won more ﬂexibility to be online at their departments from home.

10:30 - 12:00

Room I

Physics in Radiology

SS 1013 Dose, risk assessment and optimisation Moderators: F.R. Verdun; Lausanne/CH U. Zdesar; Ljubljana/SI

B-511

10:30

A comprehensive risk assessment method in diagnostic X-ray imaging by means of multiple-decrement life tables J. Geleijns1, M. Dewey2, J. Kievit1, J. Sont1, Y.L. O1; 1Leiden/NL, 2Berlin/DE ([email protected]) Purpose: Assessment of radiation-related health risks in patients undergoing diagnostic X-ray imaging routinely focuses on single risk factors and is therefore inaccurate. A comprehensive method, based on multiple-decrement life tables, integrates all relevant characteristics and risk factors of speciﬁc patient populations. Methods and Materials: Risk assessment was applied to diagnosis of coronary

Scientific Sessions artery disease and follow-up of endovascular abdominal aortic aneurysm (AAA) repair based on life tables that represent age- and gender-related functions pertaining to mortality. Mortality from radiation exposure and excess mortality from the disease and complications were included. The BEIR VII risk model was used to estimate long-term mortality related to radiation exposure. All risks were expressed as reduction of life expectancy. Results: Under clinical conditions, the acute risk of mortality from heart catheterization in CAG, although small (0.1%), substantially exceeds radiation risks from either CAG (5 mSv) or coronary CT angiography (15 mSv). The high disease-related mortality for patients suffering AAA (6%/y) not only by far exceeds the radiation risks from CT follow-up (18 mSv/y) but also reduces radiation risk estimates substantially when compared to current methodologies for radiation risk assessment. Conclusion: Comprehensive risk assessment can be performed by application of multiple-decrement life tables that allow for integration of short and long term excess mortality related to the disease and treatment. Current approaches for radiation risk assessment (late effects) that do not take short term complications into account (including mortality) for the disease and treatment underestimate health risks in general and overestimate radiation risk.

B-512

10:39

Dedicated breast CT at the same exposure level as two-view mammography: Inﬂuence of detector pixel size and focal spot size on image quality S.V. Vollmar, W.A. Kalender; Erlangen/DE ([email protected])

B-513

10:48

Usefulness of computed tomography dose index (CTDI) and practical dosimetry for ﬂat-detector CT Y. Kyriakou, P. Deak, O. Langner, W.A. Kalender; Erlangen/DE ([email protected]) Purpose: We investigated the usefulness of the CTDI and practical dosimetry approaches for ﬂat-detector CT (FD-CT). Methods and Materials: Dose assessment according to the current deﬁnition of the CTDIL, where L is the integration length, would demand larger ionization chambers and phantoms. A Monte Carlo (MC) simulation tool (ImpactMC, VAMP GmbH, Erlangen, Germany) was used to assess dose characteristics. Measurements were performed on a C-arm system (Axiom Artis, Siemens Medical Solutions, Forchheim, Germany) using a ﬂat-detector of 40x30 cm². Dose was assessed for 70 and 125 kV in cylindrical PMMA phantoms of 160 and 320 mm diameters with a varying z-extent from 150 to 900 mm. Simulated values were compared to the values obtained with ionization chambers of 100 and 250 mm lengths and to TLD dose proﬁles. The MC simulation was used to calculate respective conversion factors of CTDIL to the demanded CTDIinf. Results: The MC results were in good agreement with the measurements to within 6%. Standard phantoms which have a z-extent of 150 mm underestimate the dose at the center by up to 55%, whereas a z-extent of 400 mm appears to be sufﬁcient for FD-CT. An increased integration length of 400 mm is needed to approximate CTDIinf to within 2%. Employing MC generated conversion factors allows for an exact estimate (to within 5%) of the CTDIinf. Conclusion: MC simulations can be used to assess conversion factors to provide estimates of the CTDIinf without the need for extra-long phantoms and chambers.

10:57

Real-time capability of an optimal reconstruction phase determination during cardiac CT scanning D. Ertel1, T. Pﬂederer1, S. Ertel2, S. Achenbach1, W.A. Kalender1; 1Erlangen/DE, 2 Bayreuth/DE ([email protected]) Purpose: To validate the capability of an online determination of the optimal reconstruction phase during cardiac CT scanning in order to optimize tube-current-modulation (TCM) with the potential for dose reduction. Methods and Materials: A rawdata-based center-of-mass tracking [Med. Phys., 29 (7)] provides a motion function of the heart, the so-called kymogram. The optimal reconstruction phase represented by the least heart motion is taken from the least variation of the center-of-mass. Real-time capability was achieved by a combination of an update computation procedure and an individually generated template function based on sequential-quadratic-programming (SQP). For evaluation, 100 patient scans with a standard cardiac protocol were used: 2.32*0.6 mm collimation, 0.33 s rotation time (Sensation 64, Siemens Medical Solutions, Forchheim, Germany). For each patient the optimal reconstruction phase was identiﬁed by visual assessment as a reference standard. For validation we used the mean difference between the online computed kymogram-based and the visually identiﬁed optimal reconstruction phase. Results: The reliability of the online determination increased with the scan time with a mean phase difference of 20% directly at scan start which decreased to 10% after a third of the complete scan. In 10% of all cases our method yielded degraded results with a phase difference of more than 20% at scan end. Online capability can be ensured by a computation time far below the cardiac cycle time. Conclusion: A real-time kymogram-based determination of the optimal reconstruction phase was proven. The method has the potential to optimize TCM and therefore reduce patient dose in cardiac CT.

B-515

11:06

Measured and calculated tissue air ratios in CT W. Huda1, K.M. Ogden2, S. Edyvean3, D.G. Frey1; 1Charleston, SC/US, 2 Syracuse, NY/US, 3London/UK ([email protected]) Purpose: To compare measured with calculated Tissue Air Ratios (TAR) for selected organs of adult patients undergoing head and body CT examinations. Methods and Materials: The TAR is the mean organ dose divided by the corresponding isocenter (free-in-air) dose obtained under the same scan conditions. In this study, TAR values were measured for the male Rando phantom undergoing a whole body scan on a Siemens Sensation 16 scanner operated at 120 kV. Eight radiosensitive organs were investigated (i.e., colon, lung, stomach, bladder, esophagus, liver, thyroid, and brain). TAR measurements were obtained using Lithium Fluoride dosimeters, which were compared with the ratios of mean organ dose to the ImPACT mathematical anthropomorphic phantom divided by the corresponding isocenter (free-in-air) absorbed dose to soft tissue. Results: Measured TAR values ranged from 0.59 (bladder) to 0.93 (thyroid). The average measured TAR of all eight organs was 0.67p0.11. The ratio of measured to calculated TAR ranged from 1.19 (colon) to 0.97 (bladder and esophagus). The average ratio of measured to calculated TAR was 1.05p0.07, demonstrating excellent agreement between measured and calculated TAR values. Conclusion: On average, measured TAR values in CT (0.67) are 5% higher than calculated values. Combining TAR values with isocenter (free-in-air) doses (mGy/mAs) permits users to obtain reliable organ doses to patients undergoing CT examinations.

B-516

11:15

Measuring blood ﬂow speed in CT raw data S. Prevrhal, J. Rinkel, B. Yeh, J. Leach, R.S. Gould; San Francisco, CA/US ([email protected]) Purpose: Currently, Computed Tomography (CT) Angiography is used to generate static images of blood vessels, notably for differential diagnosis of cardiovascular disease. Some recent efforts describe dynamic CT image-based studies of contrast agent ﬂow. We hypothesized that the ﬂow speed of contrast agent is also quantiﬁable using the raw projectional data acquired by the CT scanner. Because a modern CT scanner typically collects about 1000 such raw projections per reconstructed image, a temporal resolution gain in ﬂow speed measurement in the same order is hypothetically possible. Methods and Materials: Water was circulated through a simple vascular phantom with an unobstructed vessel lumen of d=6.3 mm including a small section simulating a mild 20% stenosis and was optically determined to ﬂow at 190 mm/s in the unobstructed and computed to ﬂow at 290 mm/s in the stenosed sections, respec-

A

B

C

D

E

F

G

257

Sunday

Purpose: To evaluate the inﬂuence of detector pixel size and focal spot size on image quality, resolution and noise, in pendant-geometry cone-beam breast CT aiming at a resolution of at least 50 lp/cm. Methods and Materials: Simulations of dose and image quality were performed for dedicated breast CT. We used cylindrical phantoms with 6-14 cm diameter with softtissue tumour-equivalent inserts, micro-calciﬁcation clusters and wires for modulation transfer function (MFT) determination. Simulations were made for monochromatic X-rays at optimized photon energy at a constant exposure level with the software tools ImapctMC and ImpactSim (VAMP GmbH, Erlangen, Germany). We measured spatial resolution, contrast and image noise to assess the performance. Results: Detector pixel size and focal spot size have a major impact on noise and spatial resolution of dedicated breast CT. Noise increases by a factor of 5-6 while decreasing detector pixel size from 200 to 80 μm at the same exposure level. Spatial resolution at the centre of rotation (10% of MTF) increases from 35 to 60 lp/cm while decreasing the focal spot size from 800 to 200 μm for a detector pixel size of 80 μm. Conclusion: For desired image quality in dedicated breast CT at the same exposure level as in two view mammography, technical parameters such as focal spot in the order of 200 μm or less and detector pixel size in the order of 100 μm or less have to be made available.

B-514

Scientific Sessions tively. After an axial CT scan covering the stenosis was triggered, a small dose of iodinated contrast was auto-injected. Raw projectional data were processed ofﬂine to determine the angular distance of projections with minimum square difference of relative contrast enhancement between detector rows. Results: Flow speed measurements averaged 185-290 mm/s in the unobstructed and 290 mm/s in the stenosed sections. Accuracy errors were found to be dependent on the spatial distance between detector rows used for the computation. Conclusion: Raw data-based ﬂow speed measurement was remarkably accurate and after further characterization, is expected to serve as powerful adjuvant measurement to static anatomy-targeted CT Angiography.

B-517

11:24

Reduction of dose to the female breast as a side-effect of spectral optimization for thoracic CT S.V. Vollmar, W.A. Kalender; Erlangen/DE ([email protected]) Purpose: Various approaches to reduce dose to the female breast in thoracic CT have been investigated. We here evaluate the potential for reduction of dose to the breast by optimal choice of the X-ray spectra. Methods and Materials: We examined the effect of energy variation on dose to the female breast in thoracic CT by simulations and measurements of contrast, noise and dose. We used a standard thorax phantom with various extension rings (QRM GmbH, Möhrendorf, Germany) with breasts added and the following contrast inserts: iodine, calcium hydroxyapatite and pure soft tissue density difference. 3D dose distributions were determined by a validated Monte Carlo tool (ImpactMC, VAMP GmbH, Erlangen, Germany). Contrast-to-noise ratio per unit dose (CNRD) was determined for 60-140 kV on a clinical CT scanner (SOMATOM Deﬁnition, Siemens, Forchheim, Germany). Results: CNRD curves did not show signiﬁcant variations for soft tissue density contrast, but considerable optimization potential for iodine and skeletal imaging at reduced energies. Exact values depend on the patient cross-section and X-ray ﬁltration. For example, by reducing tube voltage from 120 kV to 80 kV on the given scanner, reduction of dose to the female breast by typically 60-70% was achieved without a loss in image quality. This step exceeds the dose reduction potential of other measures. Conclusion: Tube voltages in thoracic CT can be lowered for contrast medium and skeletal imaging in the thorax without a loss in image quality but with a signiﬁcant decrease of dose to the female breast.

B-518

11:33

Clinical applications of dual energy CT T.R. Johnson, A. Graser, D. Morhard, C. Fink, H. Michaely, S. Weckbach, M. Sedlmair, M.F. Reiser, C.R. Becker; Munich/DE ([email protected]) Purpose: The aim of this study was to assess the feasibility of a spectral analysis using dual energy CT to differentiate iodine from other materials, different body tissues or kidney stones. Methods and Materials: Fifty patients were scanned on a Dual Source CT (DSCT) system in dual energy mode at tube voltages of 140 and 80 kVp and a ratio of 1:4.5 between tube currents. Scans included angiographies of head and neck, pulmonary arteries or runoff vessels, abdominal scans in venous phase and unenhanced scans of peripheral joints. The dose equaled that of a routine examination. For spectral analysis by three material decomposition, different algorithms were applied to depict iodine content, extract vascular structures, identify collagen or differentiate kidney stones. Results: The results showed that a differentiation and depiction of contrast material distribution is possible in the lung, the liver and the kidneys with or without the underlying tissue of the organ. Thus, it is possible to visualize lung perfusion and to characterize focal liver or kidney lesions. In angiographies, bone structures can be removed from the dataset to ease the evaluation of the vessels. The differentiation of collagen makes it possible to depict tendons and ligaments. Kidney stones can be differentiated into uric acid and other types of calculi. Conclusion: Dual energy CT can be regarded as a promising technique with a lot of clinical potential. The diagnostic value of the speciﬁc applications requires further investigation.

258

A

B

C

D

E

F

G

B-519

11:42

Quality control and dose optimisations for computed radiography: A ﬁve year experience in Luxembourg A. Schreiner, A. Jahnen, M. Grelot, C. Bokou, O. Kaphammel, C. Back; Luxembourg/LU ([email protected]) Purpose: The passage from conventional radiography to computed radiography in all Luxembourg hospitals has created the necessity to update the current legislation as far as acceptance testing and constancy testing of computed radiography systems are concerned and to establish new diagnostic reference levels. Methods and Materials: A study was performed to establish acceptance testing protocols for computed radiography systems. These protocols were carried out on every new computed radiography system that was installed. An automated quality control software for constancy testing based on DIN 6868-13 and the commercially available corresponding phantoms was developed. Installations of the software have been performed. Several patient dose measurement campaigns were carried out for three most frequently asked X-ray examinations in all the hospitals. Results: The KCARE protocols were found to be very efﬁcient for the acceptance testing of computed radiography systems. The automated quality control software proved to be easy to use, gave good results and saved time for the radiographers who performed the testing. The patient dose measurements showed that for systems where no proper acceptance testing had been done due to lack of protocols, the doses had gone up when compared with patient doses from conventional systems. Patient dose measurements carried out on the same systems after proper optimisation showed a net reduction in the dose. Conclusion: This ﬁve year experience has shown that the proper setting up of computed radiography systems and the acceptance testing of these systems is extremely important.

B-520

11:51

Kerma-area products (KAP) in head CT W. Huda1, K.M. Ogden2, Z. Rumboldt1, R. Coursey1; 1Charleston, SC/US, 2 Syracuse, NY/US ([email protected]) Purpose: To obtain the Kerma-Area Product (KAP) for adult patients undergoing head CT examinations on 16-slice (GE LightSpeed 16) scanner, and compare these with diagnostic and therapeutic head examinations. Methods and Materials: Adult head examinations were performed using techniques of 120 kV/150 mA/2 s in axial mode, a beam width of 20 mm, and a total scan length of 150 mm. The average incident air kerma on the head of an adult patient was multiplied by the corresponding cross-sectional area that intercepts the patient to obtain the projection KAP. All projections were summed to generate the CT examination KAP. Measurements of CTDIair, together with the beam proﬁle in the fan beam plane, were used to obtain values of the average incident air kerma. The corresponding cross-sectional area was determined by combining the focus to isocenter distance (54.1 cm) with the head dimensions modeled as an ellipse (radii 9.47 and 7.38 cm). Results: The CT scanner CTDIair was 96.6 mGy/300 mAs, and the KAP for one rotation of the X-ray tube (20 mm beam width) was 2.44 Gy-cm2. For the complete head CT, the patient KAP was thus 18.3 Gy-cm2. This value may be compared with UNSCEAR 2000 values of KAP for skull radiographs (~1 Gy-cm2), diagnostic cerebral angiography (~50 Gy-cm2), and interventional cerebral embolization procedures (~100 Gy-cm2). Conclusion: The patient KAP in CT examinations may be determined from air kerma and patient irradiation geometry. Use of KAP in CT could unify dosimetry for this imaging modality with radiography and ﬂuoroscopy.

Scientific Sessions 10:30 - 12:00

Room K

SS 1012 Chest and abdomen Moderators: A. Paterson; Belfast/UK B. Silberman; Paris/FR

B-521

10:30

Outcome after mechanical aortic valve replacement in children and adolescents: Long term results evaluated by echocardiography and magnetic resonance tomography J. Ley-Zaporozhan1, R. Arnold2, S. Ley1, T. Loukanov1, C. Sebening1, J.-B. Kleber1, M. Gorenﬂo1, H.-U. Kauczor1; 1Heidelberg/DE, 2Freiburg/DE ([email protected])

10:48

10:39

Role of contrast enhanced MR angiography of pulmonary artery in patients with pulmonary atresia A.A.A. Abdel Razek, G. Gaballa, H. Almarsafawy; Mansoura/EG ([email protected]) Purpose: To evaluate the role of contrast enhanced MR angiography (CE-MRA) of pulmonary artery in patients with pulmonary atresia. Methods and Materials: Thirty three patients (17 M, 16 F aged 1 m-13 years; mean=2 ys) with pulmonary atresia examined on 1.5 Tesla MR unit (Symphony; Siemens). Eleven of these patients were after surgery. Bolus infusion of Gadolinium-DTPA was administrated in a dose of 0.4 ml mol/kg BW after determining scanning delayed time with test bolus technique. Parameters used were: TR=4.5 ms, TE=1.5 ms, slice thickness=0.8 mm, NEX=1 & FA=30 degrees. Correlation of CEMRA ﬁndings with angiocardiography & surgery was done. Results: CE-MRA is more accurate than angiocardiography in identiﬁcation of hypoplastic, stenotic and/or non-conﬂuent pulmonary arteries. Overall test parameters for CE-MRA and angiocardiography were: sensitivity 94 vs 90%, speciﬁcity 99 vs 95%, accuracy 97 vs 94%. There was correlation between the diameters of pulmonary arteries measured with CE-MRA and angiocardiography (r=0.93). CE-MRA was better for visualization of the origin, size, number, course & insertion of collaterals as well as multifocal pulmonary blood supply. There is complete agreement between the two methods in detecting patent ductus arteriosus. Patency of the vascular shunts was better evaluated with CE-MRA than angiocardiography. Combined analysis of source images & three dimensional CE-MRA images increased diagnostic performance of CE-MRA. Conclusion: CE-MRA angiography is a fast, accurate non-invasive technique for pre-operative mapping of the pulmonary artery and delineation of collaterals and pulmonary blood supply in patients with pulmonary atresia. It gives anatomical information essential for planning of treatment.

Purpose: To explore the clinical value of ECG-gated DSCT angiography in the diagnosis of congenital heart disease (CHD). Methods and Materials: DSCT angiography was performed on 17 patients diagnosed with CHD. Two-dimensional and three-dimensional reconstructions were performed in all cases. In the meantime all patients underwent CDFI examination. 12 patients were given surgery treatment. Results: 17 patients with CHD underwent DSCT angiography. The pulmonary arteries, coronary arteries, cardiac structure, and aorta of the patients were detected within single breath-hold. Nine patients were diagnosed with tetralogy of fallot, one with double outlet ventricle with ventricular septal defect, two and four cases with atrial septal defect and ventricular septal defect, respectively. One case was diagnosed with single ventricle, six cases with coronary artery heteroplasia, seven cases with pulmonary stenosis. Four cases with pulmonary severity stenosis showed pulmonary side-branches originating from aorta. Six cases were detected with dilatated bronchus artery. 77 malformations were detected by DSCT angiography. 67 malformations were detected by CDFI. 12 operated cases showed 57 malformations. The malformations were correctly identiﬁed with DSCT and CDFI in 54 and 50 malformations, respectively. This proves a diagnostic accuracy of 94.74% for DSCT and 87.72% for CDFI. Conclusion: With current techniques, DSCT angiography of the entire chest has very good image quality. It not only shows heart structure, but also shows coronary artery, aorta and pulmonary artery in lower dose. It is extremely valuable for planning the operative procedure on the patients with complex CHD.

B-524

10:57

Chest tuberculosis in children: CT ﬁndings in active and post-treatment cases R. Gupta, A.K. Gupta, A. Seith, S. Kabra, M. Gupta; Delhi/IN ([email protected]) Purpose: To determine the role of CECT in diagnosis of disease activity in chest tuberculosis in children. Methods and Materials: Forty-ﬁve patients (age 17 yrs) with clinically diagnosed tuberculosis or positive chest radiographs were evaluated. Chest radiograph, followed by a CECT chest (low dose) was done within 30 days of anti tubercular treatment (ATT). Follow-up CECT was done in all patients after six months of ATT. Spectrum of ﬁndings of active and inactive tuberculosis was documented. Statistical analysis was done using McNemar test and Wilcoxon signed ranks test. Results: In pretreatment cases, lymphadenopathy was the most common abnormality (96.7%), followed by parenchymal lesions (70.3%), bronchial compression (35.2%), and pleural involvement (19.8%). Nodes were conglomerate in 56.8% and discrete in 43.2%. The enhancement was homogenous in 34.1%, ghost-like in 29.5% and peripheral in 13.6%, and mixed in the rest. The perinodal fat was obscured in 84.1%. In post-treatment CT, all visible nodes were discrete, homogenous and 1 cm in size. The differences in the incidence of peripheral, inhomogenous, and mixed patterns of enhancement, conglomerate nodes and obscuration of perinodal fat were statistically signiﬁcant, and were thus taken as indicators of disease activity. Other indicators of activity included consolidation, nodules, tree in bud lesions, miliary disease, bronchial narrowing, and empyema. Conclusion: CECT is an essential tool both for diagnosis and monitoring the response to ATT. Our observation of homogenous nodes, conglomerate nodes and obliteration of perinodal fat as important indicators of disease activity has not been described in any pediatric study.

B-525

11:06

Usefulness of 64-slice MDCT for follow-up of young children with coronary artery aneurysm due to Kawasaki disease: Initial experience Y. Peng, J. Zeng; Beijing/CN ([email protected]) Purpose: To evaluate the initial application value of 64-slice MDCT in the follow-up of young children with coronary artery aneurysm due to Kawasaki disease. Methods and Materials: Twelve boys (mean age 5.1 years, range 1.8-7.8 years) for follow-up (time range from 1.1 to 5.1 years) of known coronary artery aneurysm underwent 64-slice MDCT ECG-gated coronary angiography without breath holding. Two radiologists independently assessed image quality and measured the diameters of all coronary segments. Each coronary artery aneurysm was observed and compared with those of ECHO performed previously. Results: A total of 114/156 segments permitted visualization with diagnostic acceptability; the measurements showed good interobserver and intraobserver reli-

A

B

C

D

E

F

G

259

Sunday

Purpose: Aortic valve replacement (AVR) using a mechanical prosthesis represents one of the options currently available for children presenting with a severe dysplastic aortic valve not suitable for valve sparing surgical procedures. Methods and Materials: In this monocentric descriptive study we performed a prospective clinical follow-up examination in 30 patients who underwent surgery for AVR including conventional and tissue Doppler echocardiography and magnetic resonance imaging (MRI). Results: AVR was performed at 14 [8-24] years (median [range]) using mechanical prosthesis (St. Jude Medical, diameter: 23 [17-27] mm). Indications were severe aortic stenosis in 18/30, aortic regurgitation in 5/30 or a combination of aortic stenosis and regurgitation (7/30). 20/30 patients showed a dilation of the left ventricle (LV). Follow-up was 6 [1-15] years. 29/30 patients were in functional class NYHA I. Re-replacement of mechanical prosthesis was necessary in one patient due to valve thrombosis while on low-molecular-weight heparin. None of the patients suffered from major bleeding or cerebrovascular accidents during follow-up. Echocardiography showed an LV-ejection-fraction of 55 [24-67]%. 4/30 showed a dilated LV. MRI could be obtained in 25 patients. LV-ejection-fraction was 53 [33-75]% and LV-mass was 56 [34-75] g/m2. 21/25 showed normal LV diameter and 23/25 had normal myocardial mass. There was good correlation between echocardiography and MRI for LV-mass (r=0.6) and LV-ejection-fraction (r=0.7). Conclusion: AVR with mechanical prostheses offers a good option for children and adolescents presenting with severe dysplasia of the aortic valve leading to normalization of LV size and function in most patients.

B-522

B-523

The importance of DSCT in the diagnosis of congenital heart disease X.-M. Wang, L.-B. Wu, C. Sun, C. Liu; Jinan/CN ([email protected])

Pediatric

Scientific Sessions ability, coefﬁcients were 0.93 and 0.88, respectively. A total of 30 coronary artery aneurysms were identiﬁed with measured mean of 7.5p3.8 mm diameter, and 12.4p9.1 mm longitudinal length. 10 tumors were small, 8 tumors were medium and 12 tumors were giant aneurysms. The affected segments included LM 7, 9 of LAD1, 4 of LAD2, 2 of LCX1; 6 of RCA1, 9 of RCA2 and 4 of RCA3. Calciﬁcations were found in 5 aneurysms and 3/5 with thrombosis; six stenotic segments were found. ECHO failed to detect 8 tumors with 2 in LAD, 1 in LCX and 5 in RCA, and those included 4 small aneurysms. Conclusion: The use of 64-slice MDCT angiography proved valuable for monitoring young children with coronary artery aneurysm due to Kawasaki disease. However, further study is necessary to specify the sensitivity and speciﬁcity of MDCT in the follow-up.

B-526

11:15

Tumour volume doubling times of paediatric metastatic pulmonary nodules E.J. Helm, L. Vidarsson, P.S. Babyn; Toronto, ON/CA ([email protected]) Purpose: To determine the range of tumour volume doubling times (TVDT) of metastatic paediatric pulmonary nodules. To determine the optimum interval for repeat imaging to detect nodule growth. Methods and Materials: A retrospective study was performed of patients with known primary malignancies and pulmonary nodules who underwent at least two CT scans at least seven days apart. Nodule volume was calculated using either CAD software or an area-based elliptical method. The TVDT of each nodule was calculated, and used to determine the minimum scan interval necessary to detect a 19% increase in volume (previously shown to be the minimum detectable increase). Results: A total of 10 patients underwent betweentwo and ﬁve serial CT examinations at an interval of 9-187 (mean 38.5) days. 73 doubling times were calculated for 47 nodules. The mean TVDT was 40 days (SD 35, range 2.8-148). The shortest calculated interval for detection of nodule growth was one day, and the longest interval was 36 days. The average interval for detection of growth was 10 days. Repeat scanning at 27 days (approximately four weeks) would be sufﬁcient to detect growth in 95% of nodules. Conclusion: Paediatric metastatic pulmonary nodules have rapid growth rates compared with adult primary lung nodules. Repeat scanning after four weeks should detect growth in 95% of malignant lung nodules. Repeat scanning after a short interval may be a practical alternative to biopsy in paediatric patients with known primary malignancy and indeterminate lung nodules.

B-527

11:24

Contrast enhanced ultrasound in children with blunt abdominal trauma: Diagnostic performance M. Valentino, L. Barozzi, P. Pavlica, M. Lima, S. Baroncini; Bologna/IT ([email protected]) Purpose: The purpose of our study was to prospectively compare the sensitivity and speciﬁcity of US and CE-US for depiction of solid organ injuries in children with blunt abdominal trauma, using CE-CT as the reference standard. Methods and Materials: The study protocol was approved by the ethics board of our institution and written informed consent was obtained from parents. US, CE-US and CE-CT were carried out in 27 consecutive children (19 males, 8 females; mean age 8.9p2.8 years) with blunt abdominal trauma to assess the presence of solid abdominal organ injuries. Sensitivity, speciﬁcity, positive and negative predictive values, number of lesions correctly identiﬁed, agreement and accuracy of US and CE-US ﬁndings versus CE-CT were calculated. Results: Fifteen patients were CE-CT negative. Twelve patients had fourteen solid organ injuries depicted on CE-CT: seven spleen, four liver, one right kidney, one right adrenal gland and one pancreas lesion. CE-US showed 13 out of 14 lesions in the 12 patients with CE-CT conﬁrmed lesions and no lesions in the CE-CT negative patients, while US depicted free ﬂuid in two of the 15 CE-CT negative patients and free ﬂuid and/or parenchymal lesions in eight of 12 CE-CT positive patients. Overall, CE-US showed a higher diagnostic performance than US: sensitivity, speciﬁcity, positive and negative predictive values were 92.2, 100, 100 and 93.8%, respectively. Conclusion: Our study demonstrated that CE-US was almost as accurate as CT in depicting solid organ injuries in children.

260

A

B

C

D

E

F

G

B-528

11:33

Role of diffusion-weighted MR imaging in quantiﬁcation of hepatic ﬁbrosis in pediatric patients with chronic hepatitis A.A.A. Abdel Razek, A. Abdalla, E. Omran, K. Zalata; Mansoura/EG ([email protected]) Purpose: To evaluate the role of diffusion-weighted MR imaging in quantiﬁcation of liver ﬁbrosis in pediatric patients with chronic hepatitis. Methods and Materials: Forty ﬁve consecutive patients (32M, 13 F aged 3-14ys; mean 8 ys) with chronic hepatitis considered for antiviral therapy underwent diffusion-weighted MR imaging of the liver. Diffusion-weighted MR imaging was done on 1.5 Tesla unit (Symphony, Siemens) using a single shot spin echo type of echo planar imaging. The parameters used were TR/TE: 4200/38 msec, B value = 0, 250, and 500 sec/mm2. Percutaneous liver biopsy was obtained in all patients with a mean delay of 21 days. METAVIR ﬁbrosis score was calculated and graded as mild (F1 & F2) or advanced (F3 & F4) scores. The ADC value of the liver was calculated and correlated with METAVIR ﬁbrosis score. A total of 15 aged matched volunteers underwent the same MR study. Results: There was statistically signiﬁcant difference in the mean ADC value between volunteers and patients with chronic hepatitis (p 0.01), patients with mild or advanced METAVIR ﬁbrosis scores (p 0.002). There is signiﬁcant positive correlation between the mean ADC value and METAVIR ﬁbrosis score (r=0.78, p 0.01). Selection of 1.08X10-3 mm2/sec for ADC value as a threshold value to predict ﬁbrosis revealed 89% accuracy, 93% sensitivity and 91% speciﬁcity. Conclusion: We concluded that apparent diffusion coefﬁcient value is a new promising quantitative parameter used for quantiﬁcation of liver ﬁbrosis in pediatric patients with chronic hepatitis.

B-529

11:42

Color Doppler imaging of the liver in the neonates with biliary atresia M. Lee, M.-J. Kim, C. Yoon, S. Han, J. Oh, Y. Park; Seoul/KR ([email protected]) Purpose: To describe color Doppler ﬂow imaging ﬁndings of the liver in the neonates with biliary atresia and compare them with the ﬁndings of non-biliary atresia. Methods and Materials: From March 2000 to July 2007, among patients of pathologically conﬁrmed biliary atresia, we selected 32 patients (age range, 3-268 days: mean ages, 64 days) for preoperatively evaluated color Doppler ultrasonography. We also performed color Doppler ultrasonography in 34 patients (age range, 1-150 days: mean age, 43 days) with non-biliary atresia. Their clinical diagnoses were as follows: neonatal hepatitis (n=8), TPN induced cholestasis (n=5), Alagille syndrome (n=2), non-syndromic paucity of interlobular bile duct (n=5), portal vein thrombosis (n=1), idiopathic hyperbilirubinemia (n=13). Results: All patients with biliary atresia demonstrated hepatic arterial ﬂow extending to the hepatic surface on color Doppler imaging. 29 patients with non-biliary atresia did not demonstrate hepatic subcapsular ﬂow. Five patients with non-biliary atresia had hepatic subcapsular aterial ﬂow, four patients underwent total parenteral nutrition more than one month, and one patient was CMV IgM positive. Conclusion: The arterial ﬂow extending to the hepatic surface on color Doppler imaging could be a characteristic ﬁnding of biliary atresia. The presence of hepatic arterial subcapsular ﬂow is useful to differentiate biliary atresia from other causes of neonatal jaundice when Grey scale ultrasonographic ﬁndings are ambiguous.

B-530

11:51

The use of dynamic MR renography for the evaluation of congenital urinary tract dilatation A.Y. Kandeel1, M. Bazeed1, W. Abd Elghaffar1, A. Mohamram2, S. El Halwagy1; 1 Mansoura/EG, 2Cairo/EG ([email protected]) Purpose: To evaluate the diagnostic value of dynamic MR renography as an alternative to diuretic renal scintigraphy (DRS), for the differentiation of obstructed from non-obstructed kidneys and the evaluation of split renal function in infants and children with congenital urinary tract (UT) dilatation. Methods and Materials: 20 patients with congenital UT dilatation were evaluated using coronal 3D spoiled gradient echo T1-weighted sequence covering the kidneys, ureters and bladder in six seconds. Images were acquired before the injection of Gd-DTPA (0.02 mmol/kg) and then it was repeated continuously for ﬁve minutes and then once every 30 seconds for another 25 minutes after contrast injection. Acquired images were used to generate two types of time-intensity curves; one from the renal parenchyma and the other from the whole kidney. Split renal function was calculated from the parenchymal curve while urinary excretion was assessed on the whole kidney curve. Results of MR renography were compared with those of DRS in all cases.

Scientific Sessions Results: Parenchymal and whole kidney MR renographic curves of diagnostic quality were obtained in all cases. Values of split renal function calculated from MR renography showed good correlation (r=0.923, p 0.001) with values calculated from DRS. MR renography correctly evaluated the pattern of urinary excretion in 38 (95%) of the 40 kidney units evaluated. In the remaining two kidneys MR renography overestimated the disturbance of urinary excretion as obstruction. Conclusion: Dynamic MR renography is a promising alternative to radionuclide renography for the evaluation of infants and children with congenital UT dilatation.

10:30 - 12:00

Room L/M

Radiographers

SS 1014 Development of the professional role Moderators: L. Dalla Palma; Trieste/IT I.-H. Tassiopoulos; Athens/GR

B-531

10:30

Routine T2 relaxometry measurement by MR radiographers G. Darwent, D.A. Capener, R.A. Grunewald, L.I. Wallis, N. Hoggard, P.D. Grifﬁths, I.D. Wilkinson; Shefﬁeld/UK ([email protected])

B-532

10:39

Radiographers’ role in the development of MR-guided transperineal brachytherapy protocol C. Vandulek, F. Lakosi, G. Antal, O. Petnehazy, P. Bogner, J. Hadijiev, I. Repa; Kaposvar/HU ([email protected]) Purpose: MRI provides excellent visualization of the prostate and surrounding tissues, making it the choice for guiding and monitoring brachytherapy. The purpose of this pilot study was to develop a feasible protocol for transperineal MR-guided prostate brachytherapy. Methods and Materials: The procedures were performed in a low ﬁeld, open, 0.35T MRI scanner (GE Healthcare). For interventions, an MR-compatible custom-made device was designed. The study volunteers were scanned in different positions (supine, prone, feet ﬁrst, right lateral decubitus position). Tested coils were the body, abdomen and GP coil. The tested pulse sequences incorporated standard and modiﬁed sequences. MRI-compatible coaxial needles and catheters were used. A dedicated software was used for template reconstruction, trajectory planning, target and organ at risk delineation. Results: The radiographers successfully tested and modiﬁed the imaging protocol allowing clear deﬁnition of the prostate, periprostatic tissues as also the coaxial needles and catheters. The GP coil with the patient in the right lateral decubitus position provided good image quality and access to transperineum. The MRcompatible device was modiﬁed to improve stability and needle accuracy. Needle position veriﬁcation study of 37 implaced needles conﬁrmed 97% were under 4 mm compared to target position.

B-533

10:48

Commenting on trauma radiographs: Developing and formalising the radiographer’s role M. Hardy1, B. Snaith2; 1Bradford/UK, 2Wakeﬁeld/UK ([email protected]) Purpose: Radiographer commenting schemes were ﬁrst introduced in the mid 1990s and are promoted as an extension to the traditional ‘red dot’ practice by radiographers working in the trauma environment in the UK. In order to standardise practice and to inform national guidance, a cross-sectional survey was undertaken to establish current radiographer contribution to trauma decision making. Methods and Materials: A postal questionnaire was distributed to every hospital that offered both radiography and trauma services in the UK, Channel Islands and the Isle of Man. A total of 456 sites were identiﬁed from the British Association of Emergency Medicine online directory. The questionnaire was designed to elicit information relating to the scope, education, audit and mandatory/voluntary nature of radiographer commenting practice. Results: A total of 306 (n=306/456; 74%) responses were received. Although the majority of respondents (n=284/306; 92.8%) indicated that a radiographic abnormality detection scheme (RADS) was in operation, only 61 (n=61/306; 19.9%) sites indicated that a radiographer commenting scheme was in place. Importantly, no consistent method of indicating a radiographer recognised abnormality existed. 20 methods of highlighting abnormal images and 8 separate commenting approaches were identiﬁed. Additionally, although participation in RADS was often included in radiographer job descriptions, mandatory commenting supported by education was only identiﬁed at 13 sites (n=13/61; 21.3%). Conclusion: This study conﬁrms the continued widespread support of radiographer contribution to trauma diagnosis. The informal nature of the systems, inconsistent approaches to audit and education and variations in the methods of radiographer communication employed are issues that require professional guidance.

B-534

10:57

Cranial computed tomography reporting by radiographers: The impact on service delivery and professional development L.M. Clarkson; Bradford/UK ([email protected]) Purpose: Radiographer reporting of CT head scans has been a necessary evolution within radiology. It has been a response to an acknowledged national shortage of radiologists and to the increase in the number of cranial CT scans being requested due to the introduction of new guidance. This study looks at the impact on local service delivery following the qualiﬁcation of postgraduate radiographers from the University of Bradford. It also examines the radiographer’s personal professional development. Methods and Materials: Four cohorts of students, (n=55), who undertook a Postgraduate certiﬁcate/diploma in Cranial CT reporting from the University, were followed up post qualiﬁcation, to determine if they were reporting and if this had impacted on service delivery/waiting times. They were also questioned on the effect on their personal development. A questionnaire was initially posted to the radiographers and followed up by a telephone interview. Results: 66% were reporting independently in different areas of the UK. Improvements were seen in service delivery: 36% reduction in out-patients waiting lists, stroke and T.I.A. clinic provision increased, 32% immediate reports returned with in-patients and a more efﬁcient use of radiologist’s time in 48% of hospitals. Personal development: 50% had been given a reporting allowance, 46% had been upgraded. Conclusion: Radiographer cranial CT reporting by certiﬁed practitioners has resulted in improvements in service delivery and personal development. Such initiatives are important ‘skill mix’ developments in response to new clinical guidelines on head injury and stroke at a time of national radiologist shortages.

A

B

C

D

E

F

G

261

Sunday

Purpose: T2 relaxometry has been shown to be a reliable measurement sensitive to hippocampal disease and should therefore be included in a standard MR imaging protocol for patients referred with possible hippocampal sclerosis. To ensure reproducibility of measurement of T2 relaxation times, it should be performed by one or two trained operators. In Shefﬁeld, this has historically been performed by a senior neurologist. Increase in patients referred for epilepsy imaging and demands on neurologists’ time have made this unworkable. Consequently, two experienced senior MR radiographers were trained to perform the measurement during the full epilepsy imaging protocol. Methods and Materials: 50 patients were analysed retrospectively by the senior neurologist and two senior MR radiographers. All were blinded to the results. All scans were performed on a Philips Medical Systems Intera 3 T MR scanner. The T2 sequence is a multi-echo, single slice sequence (CPMG) placed through the head of both hippocampi, perpendicular to the body. Slice position is chosen by the MR radiographer after review of a T2HIRES coronal sequence in the same angulation. The ROI centred on the head of the hippocampi avoiding any CSF spaces. The mean value was calculated using Philips proprietary software. Results: There was good intraclass correlation between the three operators. Further analysis of results will be presented including outcomes of patients with increased T2 values. Conclusion: This was a workable and accurate method of performing T2 relaxometry as part of a full epilepsy imaging protocol. 265 patients have now been analysed by the MR radiographers.

Conclusion: A system for transperineal MR-guided prostate brachytherapy has been developed and tested successfully. The radiographers’ experience was an important factor contributing to the success of the pilot study. Our results facilitate us to introduce MR-guided brachytherapy into the clinical practice in the near future. The radiographers contributed to the success of the study.

Scientific Sessions B-535

11:06

Radiographer reporting in the UK: Education to support changes in radiographer roles and responsibilites M. Hardy1, B. Snaith2; 1Bradford/UK, 2 Wakeﬁeld/UK ([email protected]) Purpose: The reporting of radiographic images by radiographers is well established in the UK. It is the expectation of the professional, regulatory and academic bodies that graduate radiographers will possess image interpretation skills. Additionally, the UK College of Radiographers has asserted that its aspiration is for all radiographers to be able to provide an immediate written interpretation on skeletal trauma radiographs by 2010. This paper explores the readiness of radiography education programmes to deliver this expectation. Methods and Materials: A questionnaire was distributed to all Higher Education Institutions in the UK (identiﬁed from the UK Society & College of Radiographers register) that provide pre-registration radiography education. Information was sought relating to the type of image interpretation education delivered; the anatomical range of image interpretation education; and education delivery styles. Results: A total of 19 responses (n=19/25; 76.0%) were received. Image interpretation education was included as part of all pre-registration programmes and offered at post-registration level at 12 academic centres (n=12/19; 63.2%). The anatomical areas and educational delivery methods varied across institutions. However, at both pre- and post registration levels, the focus was predominantly musculoskeletal image interpretation. Conclusion: Radiography education providers have embraced the need for image interpretation education within both pre- and post-registration radiography programmes. As a result, UK education programmes do support the 2010 College of Radiographers aspiration. However, the context of image interpretation education and its application within the clinical environment may need to be re-visited to ensure education is effectively supporting the changing roles and responsibilities of radiographers.

B-536

11:15

PGMI classiﬁcation of full-ﬁeld digital mammograms in the Norwegian population-based breast cancer screening program according to the radiographers’ experience R. Gullien, A.E. Haakull, J.G. Andersen, P. Skaane, E. Rostad; Oslo/NO ([email protected]) Purpose: To retrospectively evaluate the image quality of full-ﬁeld digital mammography in population-based screening according to PGMI classiﬁcation. Methods and Materials: PGMI model (P=perfect, G=good, M=moderate, I=inadequate) is used in screening for monitoring quality of mammograms: 75% should be P+G, 22% M, 3% I. Two experienced radiographers reviewed 4800 mammograms from 1200 women examined in 2005-2007. Four standard images (CC and MLO of each breast) from each woman were evaluated. The 16 radiographers were divided according to experience into group A ( 2 years) and B ( 2 years). Examinations including more than four standard images, carried out by students, and implants were excluded from analysis. Results: PGMI of all mammograms classiﬁed 84% as P+G, 11% as M, and 4% as I. Of MLOs, 75% were P+G, 22% M, and 3% I. Corresponding values for CCs were 94% P+G, 1% M, and 5% I. There was practically no difference in PGMI between the groups A and B for MLOs (P+G 75%, M 20%, I 3% versus 74, 22 and 3%, respectively). The corresponding values for CCs were 93% P+G, 1% M, 6% I for group A versus 96% P+G, 0% M, 4% I for group B. Classiﬁcation of the MLO varied from 47 to 93% and CCs from 53 to 100% among radiographers. Five radiographers had no M or I groups. Conclusion: PGMI image quality was found to be in accordance with the guidelines. There were considerable differences among radiographers. Years of experience was of little importance as compared to individual competence.

B-537

11:24

Findings from a community based study on radiological procedures for people with cancer S.A. Mathers, R.A. Chesson, G.A. McKenzie, E.M. Robertson; Aberdeen/UK ([email protected]) Purpose: The main aim of the study was to explore the experience of patients with cancer undergoing imaging procedures. In this presentation we will present general ﬁndings prior to focusing on patients with prostate cancer. Methods and Materials: The majority of participants were recruited from six voluntary sector Cancer Support Groups, based in urban and rural areas of North-east Scotland. Forty-one participants were interviewed in their own homes. Interviews were audio recorded. Full text transcriptions were produced and analysed following the recommendations of Miles and Hubermann (1994).

262

A

B

C

D

E

F

G

Results: Fifty interviews were completed (23 men; 27 women); the mean age of participants was 59.7 years, with approximately half living in rural areas. Nine men had prostate cancer, with the diagnosis made between two and 25 years. The mean number of imaging procedures was 2.5, with ultrasound the most commonly performed, followed by MRI. Three main themes related to imaging emerged from the data: i) lack of knowledge of procedure and processes e.g. type of scan, and implications for care; ii) response to the imaging procedure itself e.g. anxieties, and discomfort, and iii) communication e.g. information-giving regarding reason for having the scan, and giving of results. Imaging procedures were rarely described in isolation, but were seen as part of the patient journey. Conclusion: For health professionals to engage with patients in decision making, information giving and methods of communication need to be enhanced.

B-538

11:33

Retrospective analysis of radiographer QA in digital chest radiography W.J. Sehnert1, D.H. Foos1, B. Reiner2, E. Siegel2; 1Rochester, NY/US, 2 Baltimore, MD/US ([email protected]) Purpose: A study was implemented to compare radiographers’ QA in digital chest radiography against a retrospective QA analysis performed by a panel of their peers. Methods and Materials: Computed radiographs of the chest, 292 upright posteroanterior (PA) and 456 portable antero-posterior (AP), were collected at a large community hospital for use in a QA study. Both QA-accepted and QA-rejected images were included. Observers, six radiographers, each evaluated a constrainedrandom subset of images. During softcopy review, an image was displayed, with clinical indication, and the observer was asked to decide if they would accept it for interpretation. Results were combined to deduce panel consensus for classiﬁcation: an image was labeled “panel-reject” if a majority of the decisions were to reject it; otherwise it was labeled “panel-accept.” Results: McNemar’s test was used to compare differences in accept/reject classiﬁcation between the original QA decisions versus the panel’s retrospective assessments. Where there was disagreement, a signiﬁcant bias in the ratio of images that were panel-rejected but QA-accepted was found for both classes of chest images. For PA chests, this ratio was 39/55 (p 10-3), whereas for portable chests, the ratio was 145/149 (p 10-10). These ratios were also found to be signiﬁcantly different (p 10-8), indicating a stronger rejection rate by the panel for QA-accepted portable images. Conclusion: Current QA practice in digital chest radiography passes a signiﬁcant number of low-quality images on for diagnostic interpretation. The fraction of low-quality images passing through QA is signiﬁcantly higher for portable chest radiography than for traditional, upright chest radiography.

B-539

11:42

Emergency department workﬂow: Portable vs. stationary digital radiography A. Scherrer, E. Mathieu; Suresnes/FR ([email protected]) Purpose: To quantify the impact of portable digital radiography (DR) on patient workﬂow in an emergency department (ED). Methods and Materials: Time elapsed between patient entering the ED, and between request of radiographic examination and image availability on PACS was recorded with a portable DR unit (41 cm² CsI-aSi ﬂat panel detector) for 96 patients and compared with equivalent historical data from 141 patients imaged in a stationary DR unit. Portable DR examinations were rated in terms of anatomic positioning, brightness perception and overall image quality using a 3-step scale (3=highest score). Results: The average time spent by a patient in ER is 142 min. Using a portable DR system, the average time interval between patient admission and image availability was 88 min (sd:44 min).The average time intervals between the request of examination and image availability were 34 (sd:33 min) and 78 min, respectively, for portable and stationary systems. Average examination quality ratings with portable DR were 2.38 for positioning, 2.88 for brightness perception and 2.67 overall. Conclusion: Portable digital radiography should signiﬁcantly accelerate patient workﬂow in the emergency department while maintaining high examination quality standards.

Scientific Sessions B-540

B-542

11:51

10:39

Technical feasibility of 18 F-FDG PET/CT with integrated CT-perfusion T. Berthold, S. Knöfel, M. de Bloeme, M. Leuze, P. Veit-Haibach, V. Treyer; Zurich/CH

Brain spatial characteristics of human numerical inductive reasoning process: An fMRI study Y. Yang, K. Li, P. Liang, N. Zhong, S. Lu; Beijing/CN ([email protected])

Purpose: FDG-PET/CT is widely accepted in therapy assessment in oncology. However, for chemotherapy evaluation, integrated perfusion measurements are desirable. We assessed the technical feasibility of a CT-Perfusion protocol integrated into a clinical contrast-enhanced PET/CT-protocol (cePET/CT). Methods and Materials: 15 patients with different liver metastases underwent the integrated perfusion protocol on a Discovery VCT (GE Medical Systems). The patients had to fast for 4 h, a positive oral contrast media was applied during the 60 minutes uptake time. An attenuation CT was performed and emission data acquired consecutively. The position of the liver lesion was then deﬁned on the non-cePET/CT. For CT-perfusion (1 sec rotation time, 5 mm slice thickness), 40 ml of contrast media (Ultravist 370) was applied with a ﬂow of 4 ml/s. After 35 seconds, another bolus of contrast media was applied (50 ml, 3 ml/s) to assure full diagnostic ceCT-data. A routine abdominal ceCT was performed after another 70 sec. Results: CT-Perfusion acquisition of the liver metastases was feasible based on the PET/CT-guided localisation. Furthermore, the integrated perfusion protocol provided full diagnostic cePET/CT data. Thus, perfusion parameters could be evaluated in addition to the routine anatomical evaluation as well as glucose metabolism measurements. Conclusion: PET/CT with integrated CT-perfusion for liver metastases is technically feasible and easy to perform within 25 minutes. No additional contrast media compared to a routine cePET/CT is required. The integrated cePET/CT perfusion protocol might serve as a valuable tool to assess tumor perfusion and the corresponding glucose metabolism for therapy follow-up studies.

Purpose: Inductive reasoning is one of the most important higher level cognitive functions of the human brain, and it is still in its infancy for us to understand the neural mechanism so far. In this study, event-related fMRI was used to explore the spatial characteristics of inductive reasoning process. Methods and Materials: 20 undergraduate or graduate male students participated in the experiment. The subjects ranged in age from 20 to 26 years. All subjects are right-handed and none of them reported any history of neurological or psychiatric diseases. A typical task of inductive reasoning, function-ﬁnding, was adopted. The basal element of the task is a reverse triangle, in which three numbers located in three different positions may constitute a calculation rule, i.e., an equation like Z=X+Y. Induction tasks and simple calculation tasks (baseline) were performed in the experiments. Scanning was performed on a 3 T Siemens MR system. Functional data were acquired using a gradient echo EPI pulse sequence. Data were analyzed using Statistical Parametric Mapping (SPM2). Results: fMRI results showed many activations, including bilateral prefrontal cortex (BA 6), left precuneus (BA 7), inferior parietal lobule (BA 40), right superior and inferior parietal lobules (BA 7, 40), and bilateral occipital lobes (BA 18). The left dorsolateral prefrontal cortex (DLPFC) plays a critical role, and may relate to access of corresponding background knowledge from the long-term memory. Conclusion: This study preliminarily depicts the spatial characteristics of human inductive reasoning process, which will be helpful to characterize the information processing course of inductive reasoning process.

10:30 - 12:00

Room N/O

Neuro

Functional MRI Moderators: B. Góraj; Nijmegen/NL S. Sunaert; Leuven/BE

B-541

10:30

Detailed fMRI investigation of multiple cognitive domains in patients with amnesic MCI D. Lenzi1, L. Serra1, P. Pantano1, E. Paulesu2, F. Giubilei1, R. Perri1, G. Lenzi1, C. Caltagirone1, E. Macaluso1, M. Bozzali1; 1Rome/IT, 2Milan/IT ([email protected]) Purpose: Patients with amnesic mild cognitive impairment (a-MCI) (a frequent prodromal state of Alzheimer’s disease [AD]) typically show a relative preservation in most of the cognitive domains. Nevertheless, a relevant percentage of them is likely to develop a multi-domain cognitive impairment in a short time. Aim of this study was to investigate, using fMRI, the patterns of activation in a-MCI patients when performing tasks that selectively engage speciﬁc cognitive domains. Methods and Materials: Eight patients with MCI and a group of healthy controls were enrolled, and studied on a 3 T-MR system by echo planar T2*-weighted sequence using BOLD contrast. The fMRI tasks were aimed at investigating semantic abilities (encoding), memory functions (sentence recognition), spatial attention (spatial cues symmetry identiﬁcation), and empathic ability (observation of emotional faces). Results: FMRI analysis (SPM5) showed similar patterns of activation in both groups. We highlighted three separate networks: i) the lateral temporal lobes (leftright), the right inferior frontal gyrus and the occipital lobes for encoding and recognition; ii) a dorsal fronto-parietal network, plus insula for spatial attention; iii) fronto-parietal motor and mirror areas for the empathy task. The direct comparison between the two groups indicated greater activation of these networks in patients than in controls. The most robust effects were observed in regions of the temporal cortex during encoding and recognition. Conclusion: These ﬁndings are consistent with memory deﬁcits in a-MCI and demonstrate that fMRI can be used to investigate the patho-physiological substrate of AD evolution and to monitor clinical trials.

10:48

Neurocognitive processing of socially relevant negative emotions: An fMRI study S. Modi, S. Khushu, M. Bhattacharya, S.B. Raj, M.K. Mandal; Delhi/IN ([email protected]) Purpose: The ‘primitive emotional contagion’ is how one gets affected by the emotions of others. Calamity of any sort invokes negative emotions like sadness, fear, anger or disgust. Besides those who are directly affected, others in contact with them without being personally affected are also susceptible to emotional trauma. With an idea to test ‘primitive emotional contagion’, neurocognitive mapping for the processing of socially relevant negative emotions has been carried out in healthy Indian women, by fMRI. Methods and Materials: fMRI was carried out on six healthy Indian women. Block paradigm was chosen in which the active phase consisted of actual photographs taken after any calamity or disaster. The baseline consisted of mosaic photographs. Subjects were asked to rate the photographs in a ﬁve-point rating scale for the intensity of negative emotions invoked in them. Results: Sadness was primarily invoked in the subjects as reported by them in the interview following the fMRI session. Group analysis showed signiﬁcant activations in the fusiform gyrus, middle and superior temporal gyrus and occipital lobe. The fusiform gyrus is particularly implicated in face recognition and is a part of the network involved in the processing of socially salient stimuli. Temporal lobe cortex is involved in the explicit processing of social information from facial expressions while middle temporal gyrus is involved in empathic judgements. Conclusion: Study shows the role of fusiform gyrus and occipitotemporal circuitry in the processing of negative emotions by an individual when he gets affected by the emotions of others.

B-544

10:57

Functional magnetic resonance imaging study investigating mechanisms of memory suppression in stress-related psychiatric disorders: Preliminary results N. Agarwal, G. Como, P. Brambilla, M. Baiano, M. Maieron, M. Balestrieri, L. Perini, R. Zuliani, R. Ciano, M. Bazzocchi; Udine/IT ([email protected]) Purpose: Functional Magnetic Resonance Imaging (fMRI) technique called BOLD (Blood-Oxygen-Level-Dependent) detects signal intensity changes caused by increased paramagnetic deoxygenated haemoglobin in activated brain structures. Patients with psychiatric disorders such as Major Depression (MD) and Borderline Personality Disorder (BPD) who might have suffered recurrent early sexual/physical abuse or dysfunctional relationship are unable to actively suppress memory of undesired events. Alterations in the prefrontal cortex (PFC) and the hippocampus have been found previously. We used fMRI-BOLD technique to study dysfunctions in the PFC-hippocampus circuitry.

A

B

C

D

E

F

G

263

Sunday

SS 1011

B-543

Scientific Sessions Methods and Materials: 10 patients (ﬁve with MD and ﬁve with BPD; mean age: 40.9p11.5 years; 3M/7 F; all right-handed; all Caucasians) and 11 healthy controls (HC) (mean age: 34.2p13.9 years; 3M/8 F; 10 right-handed/one left-handed; 10 Caucasians/one Indian) underwent a speciﬁc go/no-go test (suppression and retrieval phase; Anderson test) using a 1.5 T MR scanner (Avanto, Siemens). A qualitative cluster analysis, based on Talairach parameters, was performed by visually detecting the activation of PFC and hippocampus. Results: PFC was activated during the suppression phase (SP) and the retrieval phase (RP) in both HC and patients. However, PFC was activated more in HC than in patients during both SP and RP. The hippocampus of the patients shows a trend towards greater activation during the SP while HC shows a greater activation during the RP. Conclusion: These preliminary data suggest the presence of a dysfunctional PFC-hippocampus circuitry during memory suppression in these patients. The high temporal and spatial resolution of the fMRI-BOLD technique permits investigation of transient cognitive functions in small brain areas.

B-545

11:06

The study of functional MRI in ﬁrst episode depressed patients with reversible affective defect H. Liu, X. Feng, J. Sun; Shanghai/CN ([email protected]) Purpose: The purpose of this study was to use functional magnetic resonance imaging (fMRI) to probe the neural circuitry associated with reactivity to negative and positive affective stimuli in patients with major depressive disorder before and after treatment with venlafaxine. Methods and Materials: Patients with ﬁrst episode major depressive disorder (N=10) and healthy comparison subjects (N=10) were scanned on three occasions, during which trials of alternating blocks of affective and neutral pictorial visual stimuli were presented. Symptoms were evaluated at each testing occasion, and both groups completed self-report measures of mood. Statistical parametric mapping was used to examine the fMRI data with a focus on the group-by-time interactions. Results: Patients showed a signiﬁcant reduction in depressive symptoms with treatment. Group-by-time interactions in response to the negative versus neutral stimuli were found in the left insular cortex and the left anterior cingulate. At baseline, both groups showed bilateral activation in the visual cortices, lateral prefrontal cortex, and amygdala in response to the negative versus neutral stimuli, with patients showing greater activation in the visual cortex and less activation in the left lateral prefrontal cortex. Patients with greater relative anterior cingulate activation at baseline in response to the negative versus neutral stimuli showed the most robust treatment response. Conclusion: The components of the neural circuitry activated by negative affect in depression can be changed with antidepressant medication. fMRI is a promising method to diagnose depression and valuate therapeutic effect.

B-546

11:15

Assessment of visuospatial symmetry judgment deﬁcit in alcohol dependants through fMRI and voxel based morphometry R.P. Tripathi, M. Bhattacharya, S. Modi, M. Devi, S.N. Deshpande, S. Khushu; Delhi/IN ([email protected]) Purpose: Alcoholism induces changes in morphology, metabolic proﬁle and results in deﬁcit neurocognitive functioning particularly in tasks requiring spatial information processing. In the present study, we addressed a possible deﬁcit in spatial symmetry judgment in alcoholics due to altered visual assimilation competence using fMRI and associated morphological changes using Voxel Based Morphometry (VBM) techniques. Methods and Materials: fMRI and VBM studies were carried out on eight controls and 11 alcohol dependents (diagnosed on DSM IV criterion) using 1.5 Tesla wholebody MRI system. For fMRI, the spatial symmetry judgment task was presented using a Block Paradigm. EPI sequence with an optimized inter-scan interval of 7.2 seconds was used. For VBM, T1 weighted 3D-MPRAGE sequence with 140 thin slices was performed in the sagittal plane. Pre- and post-processing of the data were performed using SPM2 software for group analysis. Results: fMRI showed signiﬁcantly reduced activations in alcoholics in the frontal-parietal-occipital regions involved in spatial decision making as compared to controls. Analysis of the VBM data showed decrease in gray and white matter volume primarily in the frontal, parts of the parietal lobule and temporo-limbic regions in alcoholics. The two studies in conjunction clearly showed that spatial memory encoding and decision making are affected in alcohol dependents thus revealing the inability of alcohol dependents to perform a visual assimilation task involving symmetry judgment.

264

A

B

C

D

E

F

G

Conclusion: Our ﬁndings thus suggest a possibly weakened visuoperceptual and memory encoding ability in alcoholics resulting in diminished symmetry decision making skills due to an affected parieto-limbic and frontal network.

B-547

11:24

Functional neuroimaging in treatment evaluation of “malignant” anxietyobsessive disorders (AOD) A.A. Stanzhevsky, A.V. Korzenev, L. Tyutin, A. Pozdnyakov, V.K. Shamrey, R.A. Shalek; St. Petersburg/RU ([email protected]) Purpose: To evaluate possibilities of 18FDG PET and +HMRS in treatment evaluation of resistant forms of obsessive compulsive disorders (OCD) and Tourette syndrome (TS). Methods and Materials: 18FDG PET carried out in 24 patients with resistant OCD and Tourette syndrome before and after stereotactic surgical effect. Three examinations were performed on each patient before surgical treatment (during conventional therapy, therapy cancellation, and ampliﬁed load of anxiolytics and SSRI). Besides, +HMRS was carried out in 16 patients before and after surgical treatment. Then PET and +HMRS results were compared with the clinical state severity (Y-BOCS and Spilberger scale (SS). Results: 18FDG PET data showed metabolism increase in prefrontal and orbitofrontal cortex, anterior cingulate and metabolism decrease in nuclei caudate before surgical treatment. In patients with TS metabolism decrease in thalami was observed. +HMRS revealed N-acetylaspartate decrease, increase of choline and creatine peaks in nuclei caudate in 14 patients before treatment. Catamnestic monitoring (from 2 to 6 years) after stereotactic surgery (including radiosurgical effect by a narrow photon beam with cumulative dose up to 120 Gy) showed a reduction of obsessive symptoms. Metabolic and clinical data signiﬁcantly intercorrelated (r Spearman=0.59 for PET and +HMRS, p 0.05, r=0.48 and 0.41, p 0.05, for PET, + HMRS and clinical state, respectively). Conclusion: This results revealed signiﬁcant correlation between imaging and clinical data. Neuroimaging ﬁndings offer the possibility to prove a choice of stereotactic structures-targets in patients with resistant forms of AOD. Our results prove participation of limbicostriatal system in pathogenesis of AOD.

B-548

11:33

A direct demonstration of both structure and function of motor and sensor after hemispherectomy combining diffusion tensor imaging with functional magnetic resonance imaging J. Qu, S. Li, J. Dai, X. Zhang; Beijing/CN ([email protected]) Purpose: To display the brain’s plasticity after hemispherectomy modiﬁed for infantile hemiplegic epilepsy combining diffusion tensor imaging with functional magnetic resonance imaging (fMRI). Methods and Materials: In this fMRI experiment, six patients with hemispherectomy modiﬁed for refractory epilepsy performed a series of movement with their damaged hand, and their damaged hand was brushed, then 3D tractorgraphy was fused with activation sites with a high ﬁeld MR system (Siemens TrioTim 3 T, Germany). Results: The new areas of the cortex responsible for these new functions in the remaining hemisphere are associative, and include the typical primary motor and somatosensory regions and other new activation sites, and their white matter tracts were demonstrated directly. Conclusion: Combining diffusion tensor imaging with functional magnetic resonance imaging is an effective tool in the study of regionalized sensorimotor plasticity after hemispherectomy.

B-549

11:42

The pre- and post-operative assessment of SM1 area in patients with brain tumor using DTI and FMRI on a 3.0 T MRI scanner E.J. Lin, Q.F. Meng, L.J. Mao, B.N. Luo; GuangZhou/CN ([email protected]) Purpose: An evaluation of the assessment of SM1 area brain tumor using DTI and FMRI on a 3.0 T MRI scanner, and its clinical application in pre- and postoperative investigation. Methods and Materials: Routine MRI, DTI and FMRI were performed in 16 patients with brain tumor a week before the surgical procedure and 40 days after the operation. The scan was done on a Siemens Trio 3.0 T MRI scanner and data was post-proceeded on a Siemens Leonardo workstation. The SM1 area was activated by the patients’ six phasic alternate movements of their hands. A sequence of 12 directional EPI-diffusion was used to perform the DTI. The images of the bilateral corticospinal tract were reconstructed, and FA value measured, providing a preoperative guidance for surgeons. Same imaging was performed post-operatively, assessing the effects of the operation, when the images were compared with those

Scientific Sessions obtained pre-operatively to evaluate the imaging assessment itself. Results: Images helped surgeons maximally preserve the functional areas and projective ﬁbers when excising the tumor. Only two patients had worse KPS scale post-operatively; the other 14 patients improved their KPS scale after tumors were removed, in which eight patients were restored to 100 on the KPS scale. Conclusion: Assessment of the SM1 area with DTI and FMRI using a 3.0 T MRI scanner is a feasible imaging modality, and might be a helpful guidance for surgical planning and post-operative evaluation in patients with brain tumors.

B-550

11:51

Motor related brain activity in the restless legs syndrome L.G. Astrakas, P.N. Margariti, S. Konitsiotis, S. Tsouli, L.C. Tzarouhi, M.I. Argyropoulou; Ioannina/GR ([email protected])

10:30 - 12:00

Room R2

Cardiac

SS 1003 Myocardial and pericardial diseases Moderators: Z.M. Metafratzi; Ioannina/GR D. Piotrowska-Kownacka; Warsaw/PL

B-551

10:30

Variables associated with gadolinium-enhanced cardiovascular magnetic resonance in hypertrophic cardiomyopathy E. Feliu1, F. Marín2, J. González-Carrillo2, E. Payá1, V. Climent1, V. Roldán2, I. Al-Assir1; 1Alicante/ES, 2Murcia/ES ([email protected]) Purpose: Hypertrophic cardiomyopathy (HCM) shows increased myocardial collagen and disarray. Delayed enhancement (DE) in cardiovascular magnetic resonance (CMR) is observed in regions of increased myocardial collagen. The extent of hyperenhancement has been associated with increased risk of sudden death. We describe the relation between DE, clinical characteristics and presence of risk factors for sudden death in a series of patients from two referral centres for HCM. Methods and Materials: 120 patients (47p16 years; 84 males) were included. All patients fulﬁlled conventional criteria for HCM. Risk factors for sudden death were evaluated. CMR was performed obtaining DE images in the left ventricular (LV) short-axis orientation. NT-proBNP and C reactive protein were determined in serum samples. Results: 83 patients (69%) showed DE areas. Patients with myocardial enhancement had higher maximal LV wall thickness (22p5 vs 17p3 mm, p 0.001), more frequent obstruction (42 vs 16%, p=0.006), non-sustained ventricular tachycardia (38 vs 8%, p=0.001), worse exercise capacity (8p4 vs 10p4 METs, p=0.003) and increased NT-proBNP levels (656 [300-1948] vs 290 [122-948] pg/mL, p=0.020). On multivariate analysis, maximal LV wall thickness (p 0.001) and non-sustained ventricular tachycardia (p=0.011), and the number of risk factors for sudden death

B-552

10:39

Value of cardiac magnetic resonance imaging in patients with suspected acute myocarditis with a follow-up period of 1 year K.F. Kreitner, K. Koch, F. Steyns, N. Abegunewardene, M. Todt, G. Horstick, C. Dueber; Mainz/DE ([email protected]) Purpose: To assess the value of CMRI in patients with clinical suspicion of acute myocarditis. Methods and Materials: 49 patients with suspected acute myocarditis and 11 control patients could be included. Cardiac catheterization was available in 21/49 patients. The study protocol included fat-suppressed STIR imaging for assessment of myocardial edema, cine imaging for assessment of global and regional myocardial function, and delayed contrast-enhanced MRI for assessment of myocardial tissue damage. Follow-up studies were done at 6 and 12 months in 20 and 11 patients, respectively. Results: In the patient group, myocardial edema was present in all patients, in 47 cases it involved the left ventricular (LV) myocardium, in 28 patients the right ventricular (RV) myocardium. LV ejection fraction was normal in 22, slightly reduced in 16, moderate reduced in six and severely reduced in ﬁve patients. Regional wall motion abnormalities of LV segments were seen in 30, and of RV segments in nine patients. Areas of late enhancement were detected in 30 patients in the left ventricle, and in seven patients in the right ventricle. Follow-up studies documented an improvement of LV function and regional wall motion abormalities, and myocardial edema. Areas of late enhancement showed a tendency to decrease. The control group did not show any of the ﬁndings of the patient group. Conclusion: Myocardial edema, regional wall motion abnormalities and areas of subepicardial late enhancement are characteristic ﬁndings in acute myocarditis and can be detected up to one year after the onset of symptoms.

B-553

10:48

Systemic sclerosis: Detection of myocardial ﬁbrosis by contrast-enhanced magnetic resonance imaging F. Breuckmann1, K. Nassenstein1, A. Kreuter2, P. Altmeyer2, R. Erbel1, J. Barkhausen1; 1Essen/DE, 2Bochum/DE ([email protected]) Purpose: To assess the prevalence of abnormal cardiac MRI ﬁndings in patients suffering from systemic sclerosis (SSc). Methods and Materials: Cardiac MRI was successfully performed in 34 patients (30 female, 4 male, mean age of 54p14 years) with proven SSc and an age, sex and cardiovascular risk factors matched control group. All examinations were performed on a 1.5 T MR system (Avanto, Siemens, Germany). The scan protocol included cine steady state free precession sequences in long and short axes views to assess myocardial function, T2-weighted fat-suppressed images in identical slice positions to visualize myocardial edema and inversion recovery fast low angle shot sequences 10 to 15 minutes after injection of 0.2 mmol Gd-DTPA per kg body weight to detect myocardial late enhancement. Results: Magnetic resonance imaging detected cardiac abnormalities in 50% of our patients and in 24% of the control group. Late gadolinium enhancement (LGE) as in-vivo correlate of myocardial ﬁbrosis was observed in 15% (5/34), a pericardial effusion in 18% (6/34), a reduced ejection fraction in 21% (7/34) and valvular pathologies in 29% (10/34) of our SSc patients. In the control group, LGE was detected in only 3% (1/34), pericardial effusion in 3% (1/34), and valvular pathologies in 12% (5/34). Conclusion: Our results demonstrate a high frequency of abnormal cardiac MRI ﬁndings in SSc patients and the ability of contrast-enhanced MRI to detect myocardial ﬁbrosis in vivo.

B-554

10:57

Role of coronary MR angiography as a non-invasive imaging modality for sequential follow-up of a paediatric population with Kawasaki disease I. Carbone, E. Algeri, H. Grazhdani, F. Ciolina, M. Basood, N. Galea, C. Catalano, R. Passariello; Rome/IT ([email protected]) Purpose: Kawasaki disease (KD) is an acute systemic vasculitis, often involving coronary arteries (CA) with the development of aneurysms which may evolve to rupture, thrombosis, up to complete occlusion (O). As patients with KD need to be periodically followed up from the time of diagnosis to evaluate the evolution of CA

A

B

C

D

E

F

G

265

Sunday

Purpose: Restless Legs Syndrome (RLS) is a sensorimotor disorder characterized by an urge to move the legs or other parts of the body, accompanied with uncomfortable sensations. The characteristic night-time worsening of the symptoms frequently leads to insomnia. Our purpose was to study the functional anatomy of RLS using functional magnetic resonance imaging. Methods and Materials: Eighteen patients (10 women, 8 men; mean age 66.5p8.9 yrs; range 55-82 yrs) and 12 sex- and age-matched controls were studied using a 1.5 T scanner. A T1-weighted high resolution three-dimensional spoiled gradient echo sequence (TR/TE=25/4.6) was used for structural imaging and a single shot multislice gradient echo planar sequence was used for functional imaging (TR/TE=4000/50, 40° ﬂip angle, 64X64 matrix, 5 mm slice thickness, 0 mm gap). The motor paradigm consisted of alternating epochs of rest and movement of both feet, each 40 sec long. SPM5 was used for statistical analysis. Results: Within-group analysis showed that both patients and controls activated mostly the middle cerebellar peduncle, the culmen, the paracentral lobule, the precentral gyrus and the superior temporal gyrus. In addition, patients activated the thalamus, putamen, middle frontal gyrus, middle temporal gyrus and cingulate gyrus. Between-group analysis showed signiﬁcant differences mainly in frontal lobe areas in agreement with previous neuropsychological tests. Large variability of patients’ activation prevented additional signiﬁcant differences with controls. Conclusion: RLS patients show enhanced activation of frontal lobe areas during a simple motor task probably explained by their inability to focus due to chronic sleep deprivation.

were associated with DE (OR 2.18, 95% CI: 1.45-3.20, p 0.001). Conclusion: CMR-DE is a common ﬁnding in HCM patients. Increased maximal LV wall thickness and non-sustained ventricular tachycardia are associated with this ﬁnding. Furthermore, a clear association with the number of risk factors for sudden death is observed.

Scientific Sessions disease, a non invasive method, without any ionizing radiation employment, is highly desirable. The aim of our study was to test the ability of Coronary-MR-Angiography to identify and follow-up CA lesions in patients with KD. Methods and Materials: Eleven patients with a previous diagnosis of KD underwent a Coronary-MRA. First a whole heart acquisition was performed, then, after the intravenous injection of 8 mL of a blood-pool contrast agent (Vasovist 0.25 mMol/ mL, Schering), a further acquisition targeted on the vessel of interest was done, as also a “steady state” whole heart acquisition. Images were reconstructed on a dedicated software, using MPR, curved MPR and MIP. Results were compared with ﬁndings at previously performed selective coronary angiography (SCA) and/ or multi detector computed tomography (MDCT). Results: Coronary-MRA produced diagnostic images in 9/11 patients. Eight giant aneurysms (GA) were identiﬁed by MRA; seven non GA, one stenosis and three O were found. Information provided by Coronary-MRA in the nine diagnostic exams correlated well with SCA and MDCT ﬁndings. Conclusion: Coronary-MRA is a very useful diagnostic tool for the follow-up of patients with KD. In order to reduce radiation dose the decision to perform SCA can be based on Coronary-MRA ﬁndings.

B-555

11:06

Isolated left ventricular noncompaction: MRI ﬁndings P. Klimeczek, R. Banys, M. Pasowicz; Krakow/PL ([email protected]) Purpose: Isolated non-compaction of the left ventricle (LVNC) is associated with altered myocardial wall morphology showing prominent trabeculae and deep intertrabecular recesses. Methods and Materials: We examined six patients (four M, age 13-57 yrs) with clinical suspicion of LVNC, made by echocardiography: multiple trabeculations (six cases), multiple deep trabecular recesses (two cases), two-layered structure of myocardium (two cases), and increased compacted-to-non compacted ratio (four cases). CMR examinations were performed using 1.5 T MR scanner (Siemens, Germany). Results: In 5/6 patients both LV and RV functions were impaired LVEF below 35 and 40%, respectively, in all six patients we found dilatation of LV with contractile abnormalities in non-compaction area. Other ﬁndings corresponded with echocardiography multiple trabeculations in LV (six cases), multiple deep trabecular recesses (one case); two-layered structure of myocardium was not visualised. In non-compacted area the decreased diastolic wall thickness, impaired thickening and wall motion abnormalities (hipokinesia to dyskinesia) were visualised (all six cases). In four patients RV- involving was detected - the multiple trabeculations in apical area of RV were visualised. Delayed enhancement (DE) was detected in four among six examined patients; localisation of changes was diffused with focal linear pattern. The DE areas were located both in normal myocardium and non-compacted area. Two patients with DE were qualiﬁed for implantation of cardioverter-deﬁbrilator device due to complex arrythmia (NSVT). Conclusion: CMR is a useful tool for detection of LVNC. DE ﬁndings may be linked with more frequent occurrence of complex arrythmia.

B-556

11:15

Dual-energy CT of the heart for coronary CTA and myocardial blood pool analysis: Initial experience B. Ruzsics1, P. Suranyi1, H. Lee1, S. Nguyen1, P. Zwerner1, T.G. Flohr2, P. Costello1, U.J. Schoepf1; 1Charleston, SC/US, 2Tübingen/DE ([email protected]) Purpose: To comprehensively assess coronary anatomy and the myocardial bloodpool using dual-energy CT (DECT). Methods and Materials: Ten patients underwent contrast-enhanced, retrospectively ECG-gated DECT coronary angiography using dual-source CT with 330 msec rotation, pitch 0.2, and 64x0.6 mm collimation. One tube of the dual-source system was operated with 150 mAs/rot at 140 kV, the second tube with 165 mAs/rot at 100 or 80 kV, depending on body habitus. DECT image analysis was performed by mapping the myocardial iodine distribution from low kV spectra on virtual non-contrast enhanced multiplanar reformats of the myocardium in long- and short-axis views. Two independent observers analyzed all DECT studies for myocardial blood-pool deﬁcits using the AHA 17 segmental model. Segmental comparison was performed between coronary CTA and DECT myocardial segments. Results were also compared with ﬁndings at SPECT and/or invasive catheter angiography. Data were analyzed using the Bayesian model. Results: 170 myocardial segments were evaluated. 21 segments in three patients showed blood-pool deﬁcits on DECT. In all of these cases, CTA showed signiﬁcant stenosis in the corresponding coronary arteries. In all territories where DECT showed blood-pool deﬁcits, SPECT or catheter angiography conﬁrmed ischemia

266

A

B

C

D

E

F

G

or critical stenosis. Overall, DECT had 85% sensitivity and 98% speciﬁcity with 96.5% accuracy for detecting ischemic myocardial segments. In patients without signiﬁcant stenosis and negative SPECT (n=7), DECT showed no sign of bloodpool deﬁcits. Conclusion: Initial experience suggests that DECT may be promising for the comprehensive analysis of coronary artery morphology and the myocardial blood pool and is in good agreement with myocardial perfusion imaging and catheter angiography.

B-557

11:24

Three dimensional cine MRI in infants and children with congenital heart disease M. Fenchel, A. Seeger, G.F. Greil, P. Martirosian, J. Doering, L. Sieverding, U. Kramer, B. Klumpp, C.D. Claussen, S. Miller; Tübingen/DE ([email protected]) Purpose: Three-dimensional (3D) steady-state-free-precession (SSFP) techniques have been successfully used for detailed depiction of cardiovascular anatomy in small infants. However, it would be beneﬁcial to acquire morphologic and functional data with high spatial and temporal resolution with reduced planning effort. First results using 3D-SSFP cine sequences in examinations of small children with congenital-heart-disease (CHD) are presented. Methods and Materials: Twelve patients (ﬁve male; age 0.04-3.9 y) were examined at 1.5 T. The clinically required MR examination comprised the acquisition of standard sequences (dark-blood-SE, 2D-GRE-Cine) and acquisition of 3D-SSFP cine sequences (voxel-size 0.85-3.3 mm3; temp-res 25 ms). Image quality was rated on a 4-point-scale. Vascular structures were measured in 3D-SSFP cine and SE images. Additionally, left ventricular ejection-fraction (EF) was determined from reformatted 3D-SSFP cine and 2D-cine images. The observers recorded any functional and morphologic pathology that was encountered, having reference data from patient history. Results: Image quality was rated as good or excellent in 10/12 datasets. SNR in 3D-SSFP cine of left ventricular blood and myocardium was 27.1p9.1 and 10.8p3.7, respectively. Mean difference for measurements of vascular structures was less than 7%, whereas mean difference between EF was 9%. In all patients with diagnostic image quality the primary diagnosis of the patients could be correctly made by assessment of the 3D-SSFP cine datasets. Conclusion: 3D-SSFP cine imaging yields comprehensive anatomic and functional information in infants and children with CHD. The almost operator independent image acquisition and the post processing capabilities of the 4D data allow the detailed assessment of cardiovascular pathologies.

B-558

11:33

Cardiac magnetic resonance imaging of marathon runners in the elderly: Prevalence of myocardial late enhancement F. Breuckmann, K. Nassenstein, B. Sievers, T. Schlosser, D. Kiefer, R. Erbel, S. Möhlenkamp, J. Barkhausen; Essen/DE ([email protected]) Purpose: Cardiac magnetic resonance imaging (cMRI) is an established noninvasive imaging modality for the assessment of myocardial disease, including ischemia, myocardial ﬁbrosis and scarring. Our study aimed to investigate the prevalence of myocardial late gadolinium enhancement (LGE) in male marathon runners aged 50 years and a control group. Methods and Materials: A total of 108 asymptomatic healthy male subjects aged 50-72 years with a minimum of ﬁve completed marathon races during the past three years and an age- and risk factors-matched control group were enrolled. Participants with known cardiovascular disease or diabetes mellitus were excluded. CINE MRI sequences and inversion recovery fast low angle shot sequences after gadolinium contrast application were collected. Individuals with LGE underwent additional adenosine stress perfusion imaging. Results: The cMRI exam could be completed in 102 individuals. Of those, 12 participants (12%) showed LGE. Five participants (5%) had subendocardial LGE suggestive of ischemic myocardial scarring and adenosine stress perfusion showed stress induced defects in four of those subjects. Seven individuals (7%) exhibited midmyocardial LGE suggestive of non-ischemic origin. During 21.3p2.8 months of follow-up, event-free-survival was worse in those runners with myocardial LGE as compared to those without LGE (log-rank, p 0.0001). A higher prevalence in the marathon cohort than in age-matched controls did not reach statistical signiﬁcance (12 versus 4%, p=0.08). Conclusion: Presumably healthy marathon runners have an unexpectedly high rate of myocardial LGE. Its presence is suggested to have a prognostic impact. LGE assessment may contribute to identify marathon runners at risk of cardiac events.

Scientific Sessions B-559

11:42

Differences in perfusion of malignant and benign cardiac tumors K.U. Bauner, S. Sourbron, M. Schmoeckel, M.F. Reiser, A.M. Huber; Munich/DE ([email protected]) Purpose: To determine whether dynamic contrast material-enhanced magnetic resonance (MR) imaging with use of kinetic parameters reveals statistically signiﬁcant differences between benign and malignant cardiac tumors. Methods and Materials: 23 patients with cardiac tumors were examined (Myxoma (n=7), Lipomatous hypertrophy of the interatrial septum (n=2), Fibroma (n=1), Rhabdomyoma (n=2), Angiomyolipoma (n=1), Angiosarcoma (n=4), Rhabdomyosarcoma (n=1), Hemeangiomyosarcoma (n=1), Myoliposarcoma (n=2), Lymphoma (n=1), Metastasis (n=1). For measurement of cardiac and tumor perfusion, a T1 weighted turboFLASH sequence was acquired. Postprocessing was performed by using software written in-house in IDL 5.4. Contrast enhancement ratios, maximum slope of contrast enhancement ratio curve (%/sec) and the area under the curve (%*sec) were calculated. Results: The contrast enhancement ratio (CER, %) was signiﬁcantly higher in malignant cardiac tumors (153.9p82.5, p=0.0035) compared to benign lesions (57.8p52.5) as was the maximum slope of CER (%/sec) (p=0.01; 10.5p5.9 vs 4.7p3.9). The values for the area under the curve (%*sec) were signiﬁcantly higher in malignant (p=0.0011, 4204.5p2122.3) in comparison to benign masses (1253.2p1135.2). Conclusion: Dynamic MR imaging delineates signiﬁcant differences in perfusion between malignant and benign cardiac tumors and can contibute to discriminate them by determination of contrast enhancement ratios, calculation of the maximum slope of the CER curve and the area under the curve.

B-560

11:51

Purpose: In the past ﬁve years, numerous studies have evaluated late gadolinium enhancement sequences with cardiovascular magnetic resonance (CMR) imaging in the assessment of patients with hypertrophic cardiomyopathy. These sequences allow us to in vivo identify myocardial ﬁbrosis and study its implications in the patient’s prognosis. Our objective is to study the association between presence of ﬁbrosis in magnetic resonance and exercise capacity in these patients. Methods and Materials: We performed CMR, echocardiography, exercise testing and Holter, and the usual clinical assessment, in 98 patients from two referral centres for hypertrophic cardiomyopathy (age 46.3p15.4 years, 71.4% male). CMR studies included ﬁbrosis quantiﬁcation (percentage of enhanced myocardium after contrast) 10 minutes after infusion of gadolinium. Symptom-limited exercise testing allowed us to determinate exercise capacity (metabolic equivalents, MET units). In 71 patients, basal NT-proBNP levels were also available. Results: 67 (68.4%) patients presented delayed gadolinium enhancement. Patients with gadolinium enhancement showed less exercise capacity (8.04 vs 10.41 METS, p=0.003). Presence of enhanced myocardium was inversely correlated to METS achieved (r=-0.21, p=0.044). Basal NTproBNP levels were the best predictors for exercise capacity (r=-0.51, p 0.0001). Multivariate analysis conﬁrmed age, previous history of atrial ﬁbrillation, NT-proBNP levels and presence of gadolinium enhancement on CMR as independent predictors of exercise capacity (model's r2=0.460). Conclusion: Presence of myocardial late gadolinium enhancement areas in CMR correlates with impairment of exercise capacity in patients with hypertrophic cardiomyopathy.

Room R3

Interventional Radiology

SS 1009 Abdomen and pelvis Moderators: S. Fabiano; Rome/IT H.-M. Hoogewoud; Fribourg/CH

B-561

10:30

ePTFE/FEP-covered metallic stents versus uncovered mesh stents for palliation of malignant jaundice due to cholangiocarcinoma. Clinical results in 50 patients M. Krokidis1, F. Fanelli2, A. Hatzidakis1, G. Orgera2, M. Bezzi2, N. Gourtsoyiannis1, R. Pasariello2; 1Iraklion/GR, 2Rome/IT ([email protected]) Purpose: To compare clinical effectiveness of ePTFE/FEP-covered-metallic stents with uncovered mesh-stents for palliative treatment of malignant jaundice due to cholangiocarcinoma. Methods and Materials: Fifty patients with inoperable cholangiocarcinoma Bismuth types I-II and no hepatic metastasis were admitted for palliation. They were 28 men and 22 women (age 39-86 years, mean 62.3). Stricture was located in upper CHD (without inﬁltrating the duct conﬂuence) in four patients, lower CHD in 13, upper CBD in 22 and in lower CBD (without inﬁltrating the papilla) in 11. There was no inﬁltration of the cystic duct. We used 36 Wallstents (Boston Scientiﬁc) 6-9 cm long and 10 mm wide in 30 patients (Group A) and 21 Viabil (W. L. Gore) covered stents with and without side holes in 20 patients (Group B). All patients were followed-up until death. Results: Technical success was 100% in both groups. Early stent occlusion was noted in one Group B patient, due to sludge formation, and was treated by dilatation. Complication rate was 13.3% in Group A and 10% in Group B. Mean survival was 155.8 days for Group A and 170.3 days for Group B. One year primary patency was 54% for Group A and 77.2% for Group B. Conclusion: Viabil stents offer a signiﬁcantly better 12-month primary patency than mesh stents in the palliation of malignant strictures due to cholangiocarcinoma without hepatic metastasis.

B-562

10:39

Whole liver drainage versus partial liver drainage: A survival analysis of patients with malignant biliary strictures treated by percutaneous metallic stenting N. Ptohis, E. Brountzos, I. Panagiotou, K. Malagari, E. Alexopoulou, A. Kelekis, N. Kelekis, A. Gouliamos; Athens/GR ([email protected]) Purpose: Percutaneous metal stenting is an accepted palliative treatment for malignant biliary obstruction. Nevertheless, factors predicting survival are not known. Methods and Materials: Ninety-ﬁve patients with inoperable malignant biliary obstruction were treated with percutaneous placement of metallic stents. Twenty ﬁve patients had non-hilar lesions. Seventy patients had hilar lesions classiﬁed as Bismuth type I (n=18 patients), type II (n=31), type III (n=15), or type IV (n=6 patients). Clinical success rates, patency, and survival rates were compared in patients treated with complete (n=52) versus partial (n=43) liver parenchyma drainage. Results: Stenting was technically successful in all patients with an overall signiﬁcant reduction of the post-intervention bilirubin levels (p 0.001), resulting in a clinical success rate of 97.9%. Clinical success rates were similar in patients treated with whole-liver drainage versus partial liver drainage. Survival was similar in the complete and partial drainage groups. The post-intervention serum bilirubin level was an independent predictor of survival (p 0.001), which dichotomized patients with good versus poor prognosis. Conclusion: Partial liver drainage achieved results as good as those after complete liver drainage. A serum bilirubin level of less than 4 mg/dl after stenting is the most important independent predictor of survival, while increasing age and Bismuth IV lesions represent dismal prognostic factors.

A

B

C

D

E

F

G

267

Sunday

Myocardial late gadolinium enhancement cardiovascular magnetic resonance in the assessment of exercise capacity in hypertrophic cardiomyopathy E. Feliu1, F. Marin2, J. Gonzalez-Carrillo2, A. García-Honrubia1, M. García-Carbonell1, S. López-Celada1; 1Alicante/ES, 2Murcia/ES ([email protected])

10:30 - 12:00

Scientific Sessions B-563

10:48

Malignant biliary strictures palliation: Clinical results in 155 patients using e-PTFE covered metallic stents G. Orgera1, F. Fanelli1, A. Hatzidakis2, M. Krokidis2, M. Bezzi1, R. Passariello1, N. Gourtsoyiannis2; 1Rome/IT, 2Iraklion/GR ([email protected]) Purpose: To evaluate the efﬁcacy of e-PTFE covered stent (VIABIL-W.L.-Gore) in the treatment of malignant biliary strictures. Methods and Materials: We perform a retrospective analysis on 155 patients with malignant common bile duct (CBD) strictures treated at University hospital of Rome (n=102) and of Iraklion (n=53), (mean age 69.6 y) with Viabil. Viabil is a self-expandable stent-graft with an inner e-PTFE/FEP lining and an outer supporting structure of nitinol wire. Biliary obstructions were secondary to pancreatic carcinoma (n=94), cholangiocarcinoma (n=18), gallbladder cancer (n=11) and metastatic lymphadenopathy (n=32). Stricture was located in the upper third of the CBD in 24 patients, in the middle portion in 67 and in the lower third in 64. Stent-grafts implanted were 159. Evaluation of serum bilirubin and liver enzyme levels were done at 1, 3, 6, 9 and 12 months. Results: Technical success was 99.4%. The 30-day mortality rate was 16%. Survival rates were 38 and 19.8% at 6 and 12 months, respectively. The stent-graft patency was 91, 79 and 78% at 3, 6 and 12 months, respectively. Acute cholecystitis occurred in 11 cases (7%). Complications occurred in 14 patients (9%). Conclusion: e-PTFE stent-graft seems to be an effective tool in malignant strictures of the CBD.

B-564

10:57

Iatrogenic bile duct transection: Long-term follow-up of percutaneous transhepatic management G. Orgera, F. Salvatori, F. Fanelli, G. Nardis, E. Boatta, P. Rabufﬁ, R. Passariello; Rome/IT ([email protected]) Purpose: To describe the technique and evaluate the results of combined radiologicendoscopic intervention in the re-establishment of interrupted biliary ducts. Methods and Materials: A total of 28 patients with complete transection of the biliary tract were treated with a combined radiological-endoscopic technique. Traumatic interruption of the biliary tree manifests either with acute symptoms related to bile leak or with progressive jaundice due to bile duct obstruction. Biliary damage was secondary to trauma in two cases and to laparoscopic/laparotomic surgical complications in 26 cases. In all cases, a Rendez-Vous technique was used to retrieve a guidewire crossing the interrupted tract, and dilatation and long-period stenting were performed. Results: Recovery of the biliary tree was achieved in all patients. Biliary drainage was successfully carried out, allowing the cessation of acute symptoms and bile leak. Three patients underwent subsequent surgical repair. In the other 25 patients, the biliary catheters were replaced by plastic endoprostheses. Complete healing of the biliary damage after endoprosthesis removal was seen in 16 patients and they remain asymptomatic at 9-36 months. Nine patients have not still completed the treatment and prostheses are still in situ (asymptomatic at 4-18 months). Conclusion: The radiological-endoscopic approach is often the only therapeutic option in those patients with complete interruption of the biliary tract. It could potentially avoid or delay surgery for acute symptoms.

B-565

11:06

Percutaneous transluminal angioplasty and stent placement in patients with hepatic artery stenosis or thrombosis after liver transplantation F. Boyvat1, A. Harman1, C. Aytekin1, B. Gümüs2, M. Haberal1; 1Ankara/TR, 2 Istanbul/TR ([email protected]) Purpose: To assess feasibility of percutaneous transluminal angioplasty (PTA) and stent placement in the treatment of hepatic artery stenosis (HAS) or hepatic artery thrombosis (HAT) after liver transplantation. Methods and Materials: Between March 2000 and February 2007, 26 patients (12 children, 14 adults, mean age 21 years) with transplant hepatic artery problems were treated. Endovascular treatment indications for these patients were HAT (12), HAS (12), active bleeding from hepatic artery (2). PTA was effective in four patients but in 22 patients, stents were also used. Stents were placed postoperatively 0-7 days in 11 patients, 8-30 days in ﬁve patients and more than 30 days in six patients. Stents were bare (n=13) and graft covered (n=9). Main indication to use graft covered stents were rupture of the hepatic artery during endovascular intervention and post operative active bleeding. Results: A technical success was achieved in 24 of the 26 endovascular procedures. Six ruptures were seen during PTA of the hepatic artery but only one of them was reoperated. Other patients were managed with graft covered stents.

268

A

B

C

D

E

F

G

The follow-up ranged from one month to 4.5 years. Nine patients expired during this period unrelated to hepatic artery intervention with patent hepatic arteries. In two patients, stents were occluded at three and ﬁve months later without any clinical symptoms. Conclusion: Early and late postoperative PTA or stent placement to the graft hepatic artery is technically feasible. Especially, graft covered stents can be performed safely and effectively in the early postoperative period.

B-566

11:15

Hepatic encephalopathy in patients treated with e-PTFE covered stentgrafts: Initial results of a randomized trial of 8 mm versus 10 mm VIATORR stent-graft G. Orgera, F. Fanelli, G. Luppi, P. Nardis, M. Corona, M. Allegritti, R. Passariello; Rome/IT ([email protected]) Purpose: To report initial results of a randomized trial comparing 8 vs 10 mm Viatorr stent-graft (WL-Gore) to evaluate clinical efﬁcacy and post-procedural encephalopathy (HE) rate. Methods and Materials: Twenty-one consecutive cirrhotic patients candidate to TIPS for variceal bleeding (uncontrollable with medical treatment) or refractory ascites were enrolled and randomized. All patients were followed-up according to a prospective protocol for a correct diagnostic work-up. Results: The two groups were comparable for age, sex, etiology and severity of cirrhosis, pre-TIPS HE stage. After TIPS the porto-systemic pressure gradient (PSG) signiﬁcantly decreased in boths groups (K2-test p=0.89); however, PSG value was higher in the 8 mm group (8.3p2.1 vs 5.5p1.9 mmHg). The incidence of HE was similar between the two groups. The recurrence of persistent ascites was signiﬁcantly higher in patients with the small stent (4/9 cases). In these patients, PSG after TIPS was higher (9.7p2.5 vs 6.3p2 mmHg). In the 8 mm-Viatorr-group, two patients had persistence of large varices not observed in the 10 mm-group. Shunt dysfunction and rebleeding episodes did not occur in both groups. Conclusion: Preliminary results suggest that TIPS created with 8 mm e-PTFE stent-graft seems to have minor clinical efﬁcacy, probably related to an inadequate PSG decrease without any advantage in reduction of post-TIPS HE.

B-567

11:24

TIPS with e-PTFE-stentgrafts in patients with refractory ascites or recurrent variceal bleeding C. Schneider, S.C.A. Herber, G. Otto, J. Schattenberg, M. Schuchmann, K. Oberholzer, C. Düber, M.B. Pitton; Mainz/DE Purpose: To evaluate the technical and clinical outcome of e-PTFE-covered stentgrafts in refractory ascites or acute variceal bleeding. Methods and Materials: From July 2002 to July 2006, 58 patients (31 male, 27 female, age 54.7p12.4) were treated with e-PTFE stentgrafts for refractory ascites (n=41) and recurrent variceal bleeding (n=17). Underlying disease included alcoholism (n=33), hepatitis C (n=8), Budd-Chiari-Syndrome (n=5), hepatitis B (n=4), and cryptogenic/others (n=8). 11/58 (18.9%) were treated on an emergency basis. Follow-up included clinical visits, laboratory ﬁndings, Duplex sonography, and CT or MRI. Results: Portosytemic pressure gradient (PSG) was signiﬁcantly reduced from 21.2p4.2 mmHg (range 12-32) to 7.2p3.3 mmHg (range 0-15), mean PSG-reduction was 14.2p4.7 mmHg (range 4-25). 30-day mortality was 10.3% (6/58). Six patients were subsequently admitted for liver transplantation. Follow-up is 13.1p15.8 months (range 0-61). 22 TIPS revisions were performed in 18 patients (31.1%), including TIPS reduction in four patients (two encephalopathy and two deteriorated liver function). Restoration of sufﬁcient TIPS diameters were necessary in 14 patients: 18 procedures (11 ballon dilatations, seven additional stent placements) because of persisiting ascites (n=8), TIPS-stenoses (n=6), elevated PSG (n=3), and one persisting bleeding. 25 patients died during follow-up. Mortality was signiﬁcantly lower in cases with refractory ascites (16/41 (40%), follow-up 14.6p16.3 months) compared to variceal bleeding (9/17 (52.9%) follow-up 8.0p13.4 months). Emergency TIPS cases demonstrated higher mortality than elective cases. Conclusion: TIPS procedure using e-PTFE stentgrafts were performed with acceptable 30-day mortality. However, regular follow-up is mandatory in order to maintain adequate shunt function over time and to perform secondary interventions betimes.

Scientific Sessions B-568

11:33

Endovascular treatment of vascular complications of acute pancreatitis: 6 year experience of a tertiary referral centre S. Athreya, N. Mathias, G.H. Roditi; Glasgow/UK ([email protected]) Purpose: Vascular complications of pancreatitis are a major cause of morbidity and mortality. Arterial complications include haemorrhage from direct arterial erosion or pseudo aneurysm formation, and visceral ischaemia. The aim of the study is to review our experience of endovascular treatment of visceral pseudo aneurysms and haemorrhage following acute pancreatitis over a 6-year period in our tertiary referral centre. Methods and Materials: Details of patients who underwent endovascular treatment for pseudo aneurysms and haemorrhage from January 2001 until December 2006 were obtained from the radiology information system. The indications for endovascular treatment, the site of bleeding/pseudo aneurysm, cross sectional imaging and any immediate complications were noted from radiology reports. Any late complications and ﬁnal outcome were obtained from patient case records. Results: A total of 43 visceral angiograms were performed on 38 patients. The ages ranged from 28 to 75 years. All the patients had pre-procedure CT scans. Twenty-nine patients had pseudo aneurysms of the visceral artery. Twenty-ﬁve patients were treated successfully by coil embolisation, while the procedure failed in four patients. There was no cause of bleeding in the remaining nine patients on diagnostic visceral angiography. One patient with early recurrence of pseudo aneurysm needed percutaneous thrombin injection. There were no late complications. Two patients died within 30 days. Conclusion: Endovascular treatment of visceral pseudo aneurysms following acute pancreatitis is a safe and efﬁcient technique. Follow-up by CT is needed to conﬁrm resolution and also to monitor patients without any source of bleeding on angiogram.

B-569

B-570

11:51

Uterine ﬁbroid embolization: First results from the Embozene™ microsphere study J. Kettenbach1, J. Moss2, L. Lampmann3, J. Falco-Fages4, R.W. Günther5, A. Rampoldi6, E.-P. Strecker7, F. Orsi6, T. Helmberger8, D. Vorwerk9, G.M. Richter10, J. Moyses1, J. Lammer1; 1Vienna/AT, 2Glasgow/UK, 3Tilburg/NL, 4Sabadell/ES, 5 Aachen/DE, 6Milan/IT, 7Karlsruhe/DE, 8Munich/DE, 9Ingolstadt/DE, 10Heidelberg/DE ([email protected]) Purpose: To present preliminary data of a multicenter study, following uterine-ﬁbroid embolization (UFE) using novel designed microspheres. Methods and Materials: In 11 European centers, 63 patients (mean age 43p20 years) were treated using Polyzene®-F coated microspheres with a hydrogel core (EmbozeneTM Microspheres, CeloNova BioSciences, Newnan, GA, USA). The mean diameter of dominant ﬁbroids was 7.3p3.5 cm (submucosal 19%, intramural 52%, and subserous 13%). One ﬁbroid was present in 21% of patients, up to three ﬁbroids in 28% and more than three ﬁbroids in 51%. Primary endpoints were menstrual bleeding score (MBS), pain-related symptoms (PRS), and quality-of-life (QoL) evaluated before and after embolization. Results: A total volume of 6.7 ml EmbozeneTM microspheres (500-, 700-, or 900-μm in diameter) was used per patient. Mean procedure time was 74p31 minutes, mean ﬂuoroscopy time was 19p10 minutes. Primary technical success was achieved in 97%. No signiﬁcant adverse events were observed. Three- and six-month follow-up were completed in 34 of 63 (54%) and in 12 of 63 (19%) patients, respectively. Compared to pre-UFE, at 3-6 months after UFE, median MBS declined signiﬁcantly (p=0.001) from 432 to 80 and 81, respectively; PRS reduced from 3.5p1.5 to 2.1p1 and 2.0p1.0, respectively; QoL improved from 70 to 80. Conclusion: Using EmbozeneTM microspheres with Polyzene®-F coating, a signiﬁcant improvement of primary endpoints was observed.

11:42

Sunday

Imaging of complications after percutaneous intrahepatic islet transplantation: Long-term monitoring after 112 transplants C. Losio, M. Venturini, A. Zanasi, P. Mafﬁ, F. De Taddeo, R. Nano, A. Secchi, A. Del Maschio; Milan/IT ([email protected]) Purpose: Percutaneous intrahepatic islet transplantation (PIT) is a feasible alternative to surgical pancreas for the care of type I diabetes, also in previously kidney-transplanted patients (PIT-K). Our aim was to retrospectively assess the rate, severity and diagnosis of complications in the islet-transplanted population at our center, after a long-term follow-up. Methods and Materials: From 1989 to 2007, 112 US-ﬂuoroscopy-guided intraportal islet transplantations were performed in 30 non-uremic (PIT) and 37 kidney-transplanted (PIT-K) diabetics (mean 1.7 procedures/patient), comparable for pre-transplant characteristics, except for steroids and aspirin in PIT-K. Follow-up was based on clinical parameters and CD-Ultrasound, with CT and MRI in selected cases. Results: Mean follow-up was 6.2 years. All procedures were well tolerated. Main procedure-related complications were: 14 intraperitoneal bleeding, two hemothorax, one arterioportal ﬁstula, two segmental portal thrombosis, one bilioma, all recognized and monitored with CDU, with conservative treatment. Major immunosuppression-related complications were: two viral myocarditis and one sirolimus-induced arthritis, diagnosed with MRI; one sirolimus-induced pneumonitis detected with CT; two irreversible renal impairment (PIT) and two kidney-graft rejections simultaneous with islet transplant (PIT-K), conﬁrmed with CDU and biopsy. No deaths related to islet transplantation or GVHD occurred. In kidney-transplanted patients we excluded complications due to pre-existing immunosuppression. No signiﬁcant differences in complications between PIT and PIT-K were observed. Conclusion: Percutaneous intrahepatic islet transplantation using US-ﬂuoroscopic guidance is a safe procedure, with low short- and long-term complications, also in kidney-transplanted patients. Extensive use of CDU, and in some cases MRI, associated with clinical markers, allows early detection and conservative management of major complications.

A

B

C

D

E

F

G

269

Scientific Sessions

270

A

B

C

D

E

F

G

Scientific Sessions

A

B

C

D

E

F

G

271

Monday

Monday, March 10

Scientific Sessions room A 2nd level

room B 2nd level

room C 2nd level

room E1 entr. level

room E2 entr. level

room F1 entr. level

room F2 entr. level

room G/H lower level

room I lower level

07:00

07:00

07:30

07:30

08:00

08:00

08:30

08:30 CC 1317 Multidetector CT Made Easy CT of the brain, head and neck: Is there anything new? (p. 86)

CC 1316 MRI Abdomen: Is MRI the ultimate tool? (p. 86)

09:00

09:30

SF 13 Special Focus Session The radiologist in image-guided surgery: Just a tour guide? (p. 87)

RC 1310 Musculoskeletal Shoulder (p. 88)

E3 1320 Interactive Teaching Session The imaging management of lung cancer (p. 89)

RC 1308 Head and Neck Imaging of the thyroid and parathyroid (p. 89)

RC 1302 Breast Changing approaches to breast diseases? (p. 90)

RC 1311 Neuro Pediatric neuroradiology (p. 90)

RC 1313 Physics in Radiology Quality control and dose reduction in digital radiology (p. 91)

10:30

10:30 SS 1401a Abdominal Viscera Checking out the pancreato-biliary system (p. 276)

SS 1410 Musculoskeletal Cartilage (p. 274)

11:00

11:30

SS 1401b GI Tract Esophageal and gastric cancer: New approaches (p. 278)

ESR General Assembly

E3 1420 Interactive Teaching Session The imaging management of stroke (p. 96)

SS 1407 Genitourinary Imaging of renal lesions: Perfusion diffusion and more (p. 280)

SS 1402 Breast MR: Know-how for new approaches (p. 282)

SS 1406b Molecular Imaging Clinical applications (p. 285)

SS 1413 Physics in Radiology New advances in imaging techniques and applications (p. 287)

12:00 Honorary Lecture 3 (p. 96)

13:00

13:30

13:30

14:00

14:00 CC 1517 Multidetector CT Made Easy MDCT angiography: Do we need anything else? (p. 98)

EM 4 ESR meets India Tuberculosis (p. 97)

14:30

15:00

SA 15 State-of-the-Art Symposium Small bowel: Old problems, new tricks (p. 98)

SF 15a Special Focus Session Frontiers in liver imaging (p. 100)

E3 1520 Interactive Teaching Session The imaging management of back pain (p. 101)

RC 1507 Genitourinary Percutaneous stone extraction: Success by team work (p. 101)

RC 1502 Breast Ultrasound and breast disease (p. 102)

RC 1511 Neuro Brain tumors (p. 102)

RC 1513 Physics in Radiology Managing patient dose (p. 103)

15:30

15:00

16:00

16:30

EPOS™- scientiﬁc exhibition

technical exhibition

registration

PC 16 Professional Challenges Session Women in radiology: How to maximise their input (p. 109)

CC 1616 MRI Cardiac MRI: An affair of the heart (p. 108)

SF 16a Special Focus Session Emergency imaging: Avoiding serious mistakes! (p. 110)

RC 1610 Musculoskeletal New trends in musculoskeletal radiology (p. 111)

E3 1620 Interactive Teaching Session Traps in imaging of cirrhosis and portal hypertension (p. 112)

SF 16b Special Focus Session Molecular imaging (p. 112)

RC 1602 Breast Diagnosis and staging of breast cancer (p. 113)

RC 1611 Neuro Grey and white matter disease (p. 114)

RC 1613 Physics in Radiology Monitor reporting: Quality and diagnostic accuracy (p. 115)

16:30

17:00

17:30

18:00

18:30

19:00

272

14:30

15:30

16:00

18:30

11:30

12:30 SY 19 Toshiba Satellite Symposium (p. 545)

13:00

18:00

11:00

12:00

12:30

17:30

09:30

10:00

10:00

17:00

09:00

19:00

A

B

C

D

E

F

G

Scientific Sessions room K lower level

room L/M 1st level

room N/O 1st level

room R1 1st level

room R2 1st level

room R3 1st level

room U 2nd level

room X 2nd level

room XI, Z & Y 2nd level

room Z 2nd level

07:00

07:00

07:30

07:30

08:00

08:00

08:30

09:00

08:30

RC 1312 Pediatric Chest imaging (p. 92)

RC 1314 Radiographers Trauma (p. 93)

09:30

PC 13 Professional Challenges Session Capturing the market: How quality helps (p. 93)

RC 1303 Cardiac Imaging of myocardial infarction (p. 94)

RC 1301 Abdominal and Gastrointestinal CT colonography: How to use it (p. 95)

RC 1309 Interventional Radiology Interventions in non-malignant liver diseases (p. 95)

09:00

09:30

10:00

10:00

10:30

11:00

11:30

10:30 SS 1415 Vascular MR: Angiography, microcirculation and lymphography (p. 289)

SS 1414 Radiographers Dose reduction strategies (p. 291)

SS 1411 Neuro MRI at 7 T (p. 293)

SS 1404 Chest Advances in CT (p. 295)

SS 1403 Cardiac MR/CT in treatment evaluation (p. 297)

SS 1409 Interventional Radiology Tumor ablation/ Liver regeneration (p. 299)

WS 1421 Image Guided Breast Biopsy: How to do it

WS 1424 Test-Driving your Workstation

11:30

12:00

12:00

12:30

12:30 SY 20 Guerbet Satellite Symposium (p. 545)

13:00

13:00

13:30

13:30

14:00

14:30

14:00 RC 1514 Radiographers Computed tomography: CT and contrast media (p. 105)

SF 15b Special Focus Session Advances in tumor ablation (p. 105)

RC 1504 Chest Diagnostic mimics in chest diseases (p. 106)

CC 1518 Breast: From Basics to Advanced Imaging Interventional procedures (p. 107)

RC 1509 Interventional Radiology Treatment of acute hemorrhage (p. 107)

WS 24D1 Test-Driving your Workstation

WS 1521 Image Guided Breast Biopsy: How to do it

15:00 WS 24D2 Test-Driving your Workstation

15:30

16:00

17:00

RC 1612 Pediatric Safety considerations in imaging (p. 115)

RC 1614 Radiographers Non medical applications (p. 116)

RC 1606 Molecular Imaging and Contrast Media US contrast agents (p. 116)

RC 1604 Chest Computer-aided diagnosis in the chest (p. 117)

CC 1618 Breast: From Basics to Advanced Imaging Advanced imaging: Fullﬁeld digital mammography (FFDM)/MRI (p. 118)

14:30

15:30

16:00 RC 1609 Interventional Radiology Carotid interventions (p. 118)

WS 1621 Image Guided Breast Biopsy: How to do it

16:30

17:00

17:30

17:30

18:00

18:00

18:30

18:30

19:00

19:00

A

B

C

D

E

F

G

273

Monday

RC 1512 Pediatric Neuro imaging (p. 104)

15:00

16:30

11:00

Scientific Sessions 10:30 - 12:00

Room A

SS 1410 Cartilage Moderators: M.A. Cova; Trieste/IT V. Jevtic; Ljubljana/SI

10:30

Evaluation of biomechanical properties of articular cartilage with quantitative magnetic resonance imaging at two different ﬁeld strengths X. Li, J. Qi, L. Xia, D. Hu; Wuhan/CN ([email protected]) Purpose: Quantitative MRI provides a non-invasive means for biochemical assays of articular cartilage. Biomechanical characteristics of cartilage were determined by the biochemical state. The purpose of this study is to demonstrate that cartilage mechanical properties can be assessed by quantitative MRI at two different ﬁeld strengths. Methods and Materials: The adjacent cartilage disk pairs (n=40) were prepared from bovine proximal humerus and patellofemoral surfaces. For one sample, the cartilage tissue’s Young’s modulus, aggregate modulus, dynamic modulus and Poisson’s ratio were evaluated in unconﬁned compression. The adjacent cartilage disk was scanned at 4.7T and 1.5 T to measure the tissue T (2) relaxation time, sensitive to the integrity of the collagen network, and dGEMRIC index (T1Gd relaxation time), reﬂecting cartilage proteoglycan content. Results: Statistically signiﬁcant linear correlations were observed between MRI and mechanical parameters at both ﬁeld strengths, especially between T2 and Young’s modulus. T2 and T1Gd relaxation times at 4.7T and 1.5 T were correlated with the compressive moduli (R=0.914 and 0.871, respectively). Topographical variation of T2 and T1Gd values at both ﬁeld strengths was similar to that of Young’s moduli. No signiﬁcant difference was found between the T2 measurements at different ﬁeld strengths in predicting mechanical properties of the tissue. Conclusion: The present results demonstrate the feasibility of quantitative MRI to reﬂect the mechanical properties of cartilage at both ﬁeld strengths. This research is supported by NSFC (30670605).

B-572

10:39

Regional measurements of T1R and T2 relaxation for detecting onset degeneration of articular cartilage Z. Zhou1, X. Zou1, S. Ringgaard1, H. Li1, L. Zou1, M. Bendtsen1, H. Stødkilde-Jørgensen1, H. Shan2, C. Bünger1; 1Aarhus/DK, 2Guangzhou/CN ([email protected]) Purpose: The purpose of the present study was to assess changes in spin-lattice relaxation in the rotating frame (T1R) and transverse relaxation (T2) for detecting on-set degeneration of articular cartilages induced by IL-1B in a porcine model. Methods and Materials: Eight 3-month-old female Danish landrace pigs were employed in the study. Each knee cavity of hind limbs was given IL-1B or saline by means of local injection. After 6 hours, the pigs were killed and the patellar samples were collected. T1R- and T2-images were collected with a fast spin-echo sequence pre-encoded with a spin-lock pulse cluster and a spin-echo sequence with multiple echoes on a 7T scanner. Using a home-built analysis program, T1R- and T2-maps were obtained and the cartilage from each sample was manually segmented by drawing regions-of-interest. This segmentation separated the patellar cartilage into superﬁcial, medium, deep, calciﬁed zones followed by total cartilage layer. Results: T1R- and T2-weighted imagings demonstrate that the patellar cartilage could be identiﬁed into superﬁcial, medium, deep, calciﬁed zones with a laminar structure. After 6 hours of local injection of IL-1B, increase in T1R relaxation was signiﬁcantly on the superﬁcial and calciﬁed zones of the patellar cartilage compared to normal controls (p=0.034); however, no difference in T2 relaxation was found on all zones of patellar cartilage between the IL-1B-treated patellae and normal controls. Conclusion: This study demonstrates that regional measurements of T1R relaxation are more sensitive to detect on-set degeneration of articular cartilage compared to measurement of T2 relaxation.

274

A

B

10:48

Accuracy and precision of cartilage T2-relaxation time quantiﬁcation: Impact of the ﬁtting algorithm J.G. Raya, O. Dietrich, A. Horng, J. Dinges, M.F. Reiser, C. Glaser; Munich/DE ([email protected])

Musculoskeletal

B-571

B-573

C

D

E

F

G

Purpose: Short T2-times and high resolution lead to low SNR data in cartilage imaging. Noise makes the measured signal intensity not decaying exponentially with the echo time (TE). Our purpose is to introduce a new noise-corrected ﬁt (NC-ﬁt) for T2 calculation, to compare results to traditional T2 ﬁtting and to assess its impact on in vivo T2 reproducibility. Methods and Materials: Synthetic and phantom image data (T2: 10-110 ms, SNR: 15-120 at TE=0 s) were used to compare the new NC-ﬁt to standard exponential and linear regression ﬁts. Accuracy (mean) and precision (standard deviation, SD) were assessed. The impact of the different ﬁts on in vivo voxelwise reproducibility of patellar cartilage T2 was assessed in six healthy volunteers (seven consecutive measurements, SNR~25). Results: In simulations and phantom measurements best accuracy of T2 was achieved by the NC-ﬁt for all SNRs, as mean T2 coincides with the reference T2 (P 0.001). For T2=10 ms and SNR=15 average deviation from reference T2 was 3.8% (NC-ﬁt, P=0.03), 420% (exponential, P 0.001) and 790% (linear, P 0.001). The NC-ﬁt showed highest precision for all SNRs (P 0.001, e.g. for T2=21 ms and SNR=25, precision was 3.5, 4.5 and 7.2 ms with NC-ﬁt, exponential and linear ﬁt). In vivo voxel-based reproducibility was better (P 0.001) with the NC-ﬁt (mean SD of 3.8 ms) than the exponential (4.8 ms) and linear (6.9 ms) ﬁts. Conclusion: Noise correction signiﬁcantly improves accuracy and in vivo precision of patellar cartilage T2 calculation. This technique may contribute to improve data scattering in studies on OA based on cartilage T2.

B-574

10:57

4.7 T MR microimaging of single chondrocyte labeled with SPIO X. Li, H. Li, J. Guo, J. Qi; Wuhan/CN ([email protected]) Purpose: To assess the detectability of single chondrocyte with transferrin-receptor overexpressed and labeled SPIO in vitro on 4.7T MR microimaging. Methods and Materials: The adenoassociated virus expression vector was constructed and transfected into rabbit primary cultured chondrocytes in vitro to overexpress transferrin receptor. The chondrocytes with overexpress transferrin receptor were magnetically labeled with superparamagnetic iron oxide particles (SPIO) by incubating with iron (560 μg Fe/ml) for 48 h. The labeling efﬁciency was evaluated by Prussian blue staining and electron microscopy. MRI microimaging with isotropic spatial resolution of 12 μm was conducted on a single labeled chondrocyte in a culture ﬂask. All scanning sequences of SE T2WI, FSE T2WI, and GRE T2 * with a 35-mm Bruker birdcage coil were used for image acquisition. The comparison of MRI and corresponding histological ﬁndings was made. Results: Prussian blue staining showed numerous blue-stained particles in the cytoplasm of the labeled cells. The labeling efﬁciency of chondrocytes with SPIO was 97%. The comparison of MRI and corresponding histological ﬁndings veriﬁed that single chondrocyte labeled with SPIO on MR images showed as remarkably low signal dot (black dot) with diameter of 25-35 μm. MR images with all sequences can show labeled chondrocyte, especially in the condition of scanning sequence of GRET2. Conclusion: MicroMRI is feasible for detecting single chondrocyte labeled with SPIO in vitro. (This research is supported by NSFC (30670605)).

B-575

11:06

Evaluation of repair cartilage after micro fracturing procedure using dGEMRIC and T2 mapping A. Watanabe1, C. Boesch2, P.M. Varlet2, S.E. Anderson3; 1Ichihara/JP, 2Berne/CH, 3 Melbourne/AU ([email protected]) Purpose: To investigate the ability of delayed gadolinium-enhanced magnetic resonance imaging of cartilage (dGEMRIC) and T2 mapping to evaluate the quality of repair cartilage after micro fracturing procedure. Methods and Materials: Six knees from six goats were studied. An osteochondral defect (6 mm in diameter) with micro fracturing procedure was created at both medial and lateral femoral condyles of the left hind leg. Six months after the surgery, MR imaging was performed using a 3 Tesla magnet. dGEMRIC was performed to evaluate glycosaminoglycan (GAG) distribution, and T2 mapping was performed to evaluate collagen arrangement and hydroxyproline (HP) concentration in cartilage. Immediately after MR imaging, the goats were euthanized and the repair cartilage and adjacent native cartilage were extracted. The cartilage samples were assessed for general histology using modiﬁed O’Driscol score including assessment of col-

Scientific Sessions lagen orientation by means of polarized light microscopy (PLM). The concentrations of GAG and HP were also studied. Results: Repair cartilage was classiﬁed as mixed cartilage or ﬁbrocartilage. Concentration of GAG and HP in repair cartilage was signiﬁcantly lower than that in native cartilage. A signiﬁcant correlation was observed between GAG concentration and T1 value as also HP concentration and T2 value in repair cartilage (p 0.05). However, no signiﬁcant correlation was observed between histological grading and T1 value as also T2 value in repair cartilage. There was no signiﬁcant correlation between T2 and collagen orientation assessed by PLM. Conclusion: dGEMRIC and T2 mapping might correlate closely with macromolecular concentration, but not with comprehensive histological grading.

B-576

11:15

Three-dimensional delayed gadolinium enhanced MRI of cartilage (dGEMRIC) at 3 Tesla for in vivo differentiation of normal hyaline cartilage and reparative tissue in patients after different cartilage repair procedures: Preliminary results S. Trattnig1, T.C. Mamisch2, K. Pinker1, P. Szomolanyi1, S. Marlovits1, S. Kutscha-Lissberg1, G.H. Welsch1; 1Vienna/AT, 2Berne/CH ([email protected]) Purpose: To use a newly developed short 3D-GRE sequence with two ﬂip angle excitation pulses for dGEMRIC to evaluate relative glycosaminoglycan (GAG) content of repair tissue in patients after microfracturing (MFX) and matrix-associated autologous chondrocyte transplantation (MACT) of the knee joint. Methods and Materials: In a phantom study, T1-mapping based on a 3D-GRE sequence with two ﬂip angle excitation pulses was compared to standard inversion recovery (IR) sequence at 3.0 T. Twenty patients treated with MFX or MACT were enrolled. For comparability, patients of each group were matched by age (MFX: 37.1p15.4 years; MACT: 37.7p8.9 years) and post-operative interval (MFX: 33.0p5.2 months; MACT: 32.0p13.1 months). The delta relaxation rate for repair tissue and normal hyaline cartilage and relative delta relaxation rate were calculated and mean values were compared between both groups using analysis of variance. Results: The phantom study demonstrated a good correlation between dual ﬂip angle excitation pulse 3D GRE and IR. Mean delta relaxation rate for MFX was 1.07p0.34 s-1 versus 0.32p0.20 s-1 at the intact control site and for MACT 1.90p0.49 s-1 compared to 0.87p0.44 s-1 resulting in relative delta relaxation rate of 3.39 for MFX and 2.18 for MACT. The difference between the cartilage repair groups was statistically signiﬁcant. Conclusion: The new 3D dual ﬂip angle excitation pulse GRE based dGEMRIC technique is comparable to standard T1 IR technique for T1 mapping but reduces scan time to four minutes. Furthermore, the preliminary in vivo study demonstrates the feasibility of the technique in the evaluation of relative GAG content in patients after different cartilage repair surgeries.

B-577

11:24

Purpose: To evaluate the zonal distribution of glycosaminoglycans in normal weightbearing cartilage and repair tissue by dGEMRIC and to monitor the development of GAG content in MACT over one year. Methods and Materials: Fifteen patients were examined after MACT in the knee joint using an eight-channel knee-coil at 3 T. The dGEMRIC technique for T1-mapping was performed using a 3D-GRE-sequence with a 35/10° ﬂip angle excitation pulse combination. The D relaxation rate (DR1) was calculated for repair tissue and normal cartilage, which was divided into two zones (deep and superﬁcial). Zonal mean values as well as total mean values were compared using analysis of variance. Results: The mean DR1 for the repair-cartilage decreased signiﬁcantly from 2.2 to 1.71 in a one-year-follow-up period, with a DR1 of 1.34-1.07 for reference cartilage. For the zones there was an increase of values (decrease of GAG) from the deep to the superﬁcial layers in reference cartilage (DR1 1.14-1.60) and the transplant cartilage ($R1 2.10-2.38). In the follow-up DR1 changed to 1.80 for the deep and 1.66 for the superﬁcial zones for cartilage transplant. The decrease of DR1 for the superﬁcial zone of the cartilage transplant with a drop from 2.38 to 1.66 was statistically signiﬁcant. Conclusion: dGEMRIC is feasible for the zonal T1-mapping of normal hyaline cartilage and repair tissue. Global GAG as well as zonal GAG content demon-

B-578

11:33

Collagenous ﬁbre architecture of articular cartilage: Comparison between high ﬁeld DTI and scanning electron microscopy L. Filidoro1, D.L. Weber2, A.P. Arnoldi1, J. Weber1, J.G. Raya1, P.M. Jakob2, M.F. Reiser1, R. Putz1, E. Muetzel1, C. Glaser1; 1Munich/DE, 2Würzburg/DE (lucianna.ﬁ[email protected]) Purpose: Disruption of the collagenous ﬁbre architecture (CFA) is considered the hallmark of irreversible damage to cartilage with consecutive progression to open osteoarthritis (OA). Our purpose was to evaluate whether DTI is able to depict the CFA of cartilage and whether DTI-based zonal heights correlate to those measured by scanning electron microscopy (SEM). Methods and Materials: DTI data (six directions) were acquired from 20 human patellar cartilage on bone samples at 750 MHz (1500×62.5×250 μm³, b: 0, 500 s/ mm²). ADC, fractional anisotropy (FA) and ﬁrst eigenvector (EV) were plotted against cartilage depth from surface. SEM was performed on a Jeol35CF (critical point drying (CO2), freeze fracturing (N2), 400 A gold sputtering). Heights of tangential (TGZ), transitional (TRZ) and radial (RZ) zones were compared between SEM and EV. Results: All DTI parameters varied with increasing depth from surface consistent with different cartilage zones. FA was 0.1 for TGZ, 0.09 for TRZ and 0.18 for RZ. Heights of TGZ, TRZ and RZ measured with DTI and SEM showed a Pearson correlation coefﬁcient of 0.11, 0.89 and 0.87, respectively. TRZ (SEM) exceeded TRZ (DTI) by 20% whereas the opposite was found for RZ. Conclusion: Zonal variation and zonal height of EV correlated well with SEM of the CFA. Shrinkage due to preparation for SEM and partial volume effects in DTI may explain the observed differences in zonal height. DTI parameters such as ADC, FA and EV are able to reﬂect the zonal CFA of cartilage suggesting high potential of DTI for the workup of cartilage in OA.

B-579

11:42

Diffusion-weighted imaging of knee cartilage at 3.0 T: Follow-up of patients after matrix-associated autologous chondrocyte transplantation K.M. Friedrich1, C.M. Plank1, G.H. Welsch1, G. Langs1, E. Salomonowitz2, S. Trattnig1, T.C. Mamisch3; 1Vienna/AT, 2St. Poelten/AT, 3Berne/CH Purpose: To evaluate the use of Diffusion-Weighted-Imaging (DWI) for the assessment of cartilage maturation in patients after Matrix-Associated Autologous Chondrocyte Transplantation (MACT). Methods and Materials: Fifteen patients after MACT were examined by 3-T Magnetic-Resonance-Tomography; the examination was up to 13 months after surgery in group 1, and later than 13 months after surgery in group 2. Both groups had a follow-up one year later. DWI was acquired using a steady-state gradientecho sequence. Mean values of the diffusion-quotients of regions-of-interest within cartilage-repair-tissue and of reference-regions were assessed. Each region-ofinterest was subdivided into a deep, middle, and superﬁcial area. Results: Mean diffusion-quotients of cartilage-repair-tissues were 1.44 (baseline) and 1.44 (follow-up). Mean diffusion-quotients of reference-tissues were 1.29 (baseline) and 1.28 (follow-up). At the follow-up diffusion-quotients of cartilagerepair-tissue were signiﬁcantly higher than those of reference-cartilage. In group 1 the diffusion-quotients were signiﬁcantly lower at the follow-up (1.45-1.65); in group 2 no statistically signiﬁcant differences between follow-up (1.39) and baseline (1.41) were found. Reference-cartilages and cartilage-repair-tissues of group 2 showed a decrease of diffusion-quotients from the deep to the superﬁcial area being stable at the follow-up. In group 1 initially a signiﬁcant increase (1.49-1.78) of the diffusion-quotients from deep to superﬁcial area of the cartilage-repair-tissue was found which then decreased (1.65-1.52) at the follow-up. Conclusion: DWI is able to detect changes of diffusion within cartilage-repair-tissue that may reﬂect cartilage maturation. Zonal variations within cartilage can be measured. Changes in diffusity seem to occur only within two years after surgery.

B-580

11:51

T2-mapping versus diffusion-weighted imaging for the assessment of cartilage repair tissue after matrix-associated autologous chondrocyte transplantation at 3.0 T C.M. Plank1, K.M. Friedrich1, G. Langs1, G.H. Welsch1, T.C. Mamisch2, S. Trattnig1; 1 Vienna/AT, 2Berne/CH ([email protected]) Purpose: To compare the results of T2-mapping and Diffusion-Weighted Imaging (DWI) in monitoring cartilage repair tissue at different time points after MatrixAssociated Autologous Chondrocyte Transplantation (MACT).

A

B

C

D

E

F

G

275

Monday

3D GRE sequence with two different ﬂip angle excitation pulses for zonal T1- mapping of articular cartilage at 3 T: One-year follow-up in patients after matrix-associated autologous chondrocyte transplantation (MACT) of the knee joint K. Pinker1, P. Szomolanyi2, G.H. Welsch1, T.C. Mamisch3, S. Marlovits1, S. Trattnig1; 1Vienna/AT, 2Bratislava/SK, 3Berne/CH ([email protected])

strate an increase over a follow-up period of one year in patients after MACT of the knee joint.

Scientific Sessions Methods and Materials: In total, 28 condyles after MACT were examined on a 3.0-T MR unit. Based on postoperative interval patients were divided into two groups: group1 up to 13 months (n=12), group2 more than 19 months (n=16). For T2-mapping a 2D SE-MC sequence with six echoes was used, for DWI a 3D balanced steady-state gradient-echo pulse sequence (3D-DW PSIF) with a maximum b-value of 75 mT ms m-1. Mean signal-intensities of regions of interest within cartilage-repair-tissue and within a reference-region of healthy cartilage were assessed. Additionally, those regions of interest were subdivided into a deep area and a superﬁcial area. Finally, values of T2-mapping and DWI were compared. Results: Both techniques showed signiﬁcant differences for cartilage-repair-tissue and reference cartilage in both groups. T2-mapping: T2-values showed a signiﬁcant decrease from group1 to group2 (65-56 ms); zonal variation (deep-superﬁcial area) was signiﬁcantly different compared to healthy cartilage in group1, and more alike in group2. DWI: the diffusivity decreased from group1 to group2 (1.48-1.41); the zonal variation changed - there was an increase from the deep to the superﬁcial area in group1, and a decrease in group 2. Conclusion: Both T2-mapping and DWI are able to monitor changes in cartilage maturation after MACT. Based on zonal variation different results were observed, which could reﬂect different abilities of the techniques in imaging of macromolecular components of cartilage.

10:30 - 12:00

Room B

Abdominal Viscera (Solid Organs)

SS 1401a Checking out the pancreato-biliary system Moderators: P.E. Huppert; Darmstadt/DE F. Takis; Keratea/GR

B-581

B-582

Purpose: To determine the diagnostic value of the loss of low signal intensity on T2-weighted magnetic resonance images (MRI) in differentiating rT1b lesions from bT1a lesions in gallbladder (GB) cancers. Methods and Materials: Between 2003 and 2007, 35 patients with surgically resected GB carcinomas which were conﬁrmed as less than pT2 lesion (bpT1a (n=11), rpT1b (n=24)) were enrolled in this study. All of these patients were examined preoperatively, T2-weighted HASTE MR images with 1.5 T units, phased-array coil. The MR images were independently evaluated by two observers for the "loss of low signal intensity line of muscle layer" which represents tumor inﬁltration at muscle layer, suggests rT1b GB cancer. The reviewers graded their diagnostic conﬁdence for loss of low SI on T2WI using a 5-point scale, and receiver operating characteristic curve analysis was performed. Interobserver agreement was evaluated with kappa statistics. With consensus reading, the sensitivity, speciﬁcity, and positive and negative predictive values of the "loss of low SI on T2WI" for diagnosing rT1b lesions were evaluated. Results: Az values of the two observers for making the correct diagnosis of rT1b GB cancer were 0.886 and 0.856. The interobserver agreement was good (weighted kappa value 0.68). The sensitivity, speciﬁcity and positive and negative predictive values of the sign were 91.7, 72.7, 88.0, and 80.0%, respectively. Conclusion: The loss of low signal intensity on T2-weighted HASTE MRI can be a useful sign for differentiating rT1b lesions from bT1a lesions in early GB cancers.

B-583

10:30

Evaluation of potential living liver donor’s biliary anatomy: Comparison of optimized high-resolution magnetic resonance cholangiography and contrast-enhanced multislice-CT cholangiography S. Kinner, J. Barkhausen, A. Radtke, S. Lunkeit, C. Valentìn-Gamazo, T. Schröder; Essen/DE ([email protected]) Purpose: One focus of preharvest imaging of potential living liver donors is to determine biliary variations or abnormalities. Aim of this study was to compare the diagnostic performance of optimized non-enhanced high-resolution MR cholangiography and contrast-enhanced multi-slice computer tomographic cholangiography (ce-MSCT-CA). Methods and Materials: 20 consecutive potential living donors underwent both preoperative CT and MR imaging. MRI was performed on a 1.5 T scanner (Magnetom Avanto®, Siemens) involving specially adapted thin slice T2-HASTE/T2-TSE sequences, high-strength gradients and optimized temporal/spatial resolution. Thinslice CE-ms-CT-CA was performed on a 16-row multislice scanner (Sensation16®, Siemens) and involved i.v. administration of a biliary contrast agent (Biliscopin®, Schering). Image analysis was performed by two experienced radiologists and focused on the quantiﬁcation of the displayed biliary branches (1=visualization limited to bigger branches/upper biliary conﬂuent, up to 4=visualization of small, r3rd order intrahepatic branches) and the qualiﬁcation of the achieved image quality. Statistical comparison of CT and MRI employed the Wilcoxon test. Results: ce-MSCT-CA showed signiﬁcantly better delineation of the biliary tree than MRC (p=0.003). While CTC frequently displayed the intrahepatic bile ducts up to the 3rd branch (mean 2.8), the display by MRC was mostly limited by the 2nd branch (mean 2.1). As a variant, three donors presented a trifurcation at the upper conﬂuent, which could, however, be detected in both examinations. Conclusion: For preoperative evaluation of potential living liver donors, CE-msCT-CA is still superior to optimized MRC. However, the differences are decreasing and new MRC techniques now are regarded sufﬁcient for the display of operationrelevant anatomy.

10:39

MR differentiation between rT1b and bT1a lesions in early gallbladder carcinomas: Signiﬁcance of "loss of low signal intensity" on T2 weighted HASTE images S. Kim, J. Lee, J. Lee, S. Kim, J. Han, B. Choi; Seoul/KR ([email protected])

10:48

MR imaging for diagnosing cystobiliary ﬁstulas in cystic echinococcosis T. Heye1, M. Stojkovic1, T. Jänisch1, J. Werner1, H. Friess2, G.W. Kauffmann1, T. Junghanss1, W. Hosch1; 1Heidelberg/DE, 2Munich/DE ([email protected]) Purpose: To evaluate the sensitivity and speciﬁcity of magnetic resonance imaging (MR imaging) including MR cholangiography for the identiﬁcation of cystobiliary ﬁstulas in patients with hepatic hydatid disease. Methods and Materials: Retrospective analysis of three groups of patients (20 patients with 51 echinococcal cysts) in a cohort of 103 patients with cystic echinococcosis with different pretest probabilities for cysto-biliary ﬁstulas. Patients who had MR imaging/MR cholangiography with symptoms and signs of biliary obstruction (ﬁve patients with 16 cysts, group I), before surgery for reasons other than biliary obstruction (nine patients with 14 cysts, group II) and for cyst staging (six patients with 21 cysts, group III). All MR images were evaluated before surgery for the presence of cyst wall defects and hydatid debris in bile ducts. In groups I and II, MR results were compared with surgical, parasitological, and biochemical ﬁndings of each individual cyst. Results: Based on direct (i.e. defects in the cyst wall and continuity of dilated biliary ducts into adjacent cysts) and indirect MR imaging/MR cholangiography imaging signs (i.e. intraluminal debris) our best estimate of sensitivity and speciﬁcity (stage-speciﬁc, WHO types CE3 and CE4) for cystobiliary ﬁstulas were 75 and 95%, respectively. Conclusion: MR imaging with MR cholangiography is a valuable non-invasive imaging technique to assess the risk of cystobiliary ﬁstula-related complications and for planning of surgery.

B-584

10:57

Quantitative assessment of pancreatic exocrine reserve, in patients with chronic pancreatitis, by means of MRCP during secretin stimulation, compared to fecal elastase-1 R. Manfredi, S. Perandini, M. Motton, S. Mehrabi, W. Mantovani, R. Pozzi Mucelli; Verona/IT ([email protected]) Purpose: To perform a quantitative assessment of the pancreatic exocrine reserve by means of dynamic MR cholangiopancreatography (MRCP) during secretin stimulation compared to fecal elastase-1 (FE-1). Methods and Materials: Thirty-ﬁve patients with chronic pancreatitis/recurrent episodes of acute pancreatitis were enrolled in this IRB approved, retrospective study. All patients underwent clinical determination of the pancreatic exocrine reserve by means of FE-1 and it was classiﬁed as severe 100 μg/g; moderate 100-200 μg/g; normal 200 μg/g. All underwent MRCP during secretin stimulation

276

A

B

C

D

E

F

G

Scientific Sessions with mean interval of 23.6 days. Both thick-slab MRCP images and Half-Fourier thin slice T2-WI were acquired. MRCP image analysis included: quantiﬁcation of duodenal ﬁlling both on thick slab MRCP images 10 min after administration of secretin, and on thin slice coronal T2-WI Half-Fourier images. The duodenal ﬂuid present before administration of secretin was calculated and subtracted. Correlation between duodenal ﬁlling, measured both on thick-slab MRCP and thin slice coronal T2-WI, and FE-1 were calculated. Results: According to FE-1, 6/35 (17%) patients showed severe and 3/35 (9%) moderate pancreatic insufﬁciency; 26/35 (74%) showed normal pancreatic reserve. Median FE-1 was 331 μg/g (range 0.60-587.6). Median duodenal ﬁlling, calculated on thick-slab MRCP images was 168.06 cc (range 73.1-431.57 cc). Median duodenal ﬁlling calculated on thin-slice Half-Fourier T2-WI was 331.0 cc (range 0.10-587.60 cc). Thick-slab MRCP images showed a 0.44 correlation with FE-1, whereas thin slice T2-WI showed a correlation of 0.57. Good correlation was observed for FE-1 values 500 μg/g whereas for values 500 μg/g there was low correlation. Conclusion: Dynamic MRCP, during secretin stimulation, is able to noninvasively quantify pancreatic exocrine reserve, especially for severe-moderate insufﬁciency.

B-585

11:06

Asymptomatic serum hyperamylasemia and hyperlipasemia: Spectrum of ﬁndings on secretin-stimulated MRCP F. Donati, P. Boraschi, R. Gigoni, S. Salemi, M. Cossu, F. Falaschi, C. Bartolozzi; Pisa/IT ([email protected])

B-586

11:15

CT patterns of autoimmune pancreatitis, and their changes after steroid treatment R. Manfredi, R. Graziani, C. Cicero, M. Motton, S. Mehrabi, R. Pozzi Mucelli; Verona/IT ([email protected]) Purpose: To describe CT patterns of autoimmune pancreatitis (AIP) and their changes after steroid therapy. Methods and Materials: From February 1997 to January 2006, twenty-one patients, with histology-proven AIP observed in a single institution, that underwent contrast-enhanced Computed Tomography (CT), both at diagnosis and after steroid treatment, were included in this IRB approved retrospective study. Image analysis included: presence/absence of pancreatic parenchyma enlargement and type (focal/diffuse); contrast-enhancement of pancreatic parenchyma during the dynamic study; size of main pancreatic duct (MPD) within the lesion and upstream; pancreatic parenchyma thickness (head, body, tail). Same criteria for image analysis were applied on follow-up CT examinations and data were compared to pre-treatment ones. Results: Pancreatic parenchyma showed focal enlargement in 14/21 (67%) and diffuse enlargement in 7/21 (33%) patients. Pancreatic parenchyma, affected by

B-587

11:24

Evaluation of intra-papillary mucinous neoplasms by movies of CT pancreatogram images obtained around the duct as center axis: Can it help in characterization of IPMNs? A.K. Singh, W. Cai, D.V. Sahani, O. Catalano, G. Harris, H. Yoshida; Boston, MA/US ([email protected]) Purpose: To assess the diagnostic performance of CT pancreatogram (CTp) images obtained around pancreatic duct axis for characterizing IPMNs. Methods and Materials: CT data sets of 20 patients (1.25 mm axial sections for pancreatic phase) operated for IPMNs (side branch duct (SBD)=10, main pancreatic duct (MPD)=5, combined=5) were used. A centerline was created by plotting points along the pancreatic duct on axial sections and CT images every 5-degree around it recording the time. Two radiologists (R1, R2) reviewed CTp movies following which other CT dataset review was done to assess morphologic details of IPMN. Sensitivity and accuracy for predic-ting malignancy, kappa values and diagnostic conﬁdence (DC) (5-point scale) was calculated with surgical pathology being reference standard. Results: Sensitivity, speciﬁcity, and accuracy for detection of malignancy by radiologists on reviewing SACTp movies were 87, 83 and 82%, respectively. Cyst communication with MPD was detected in 9/10 and 7/10 cases of side-branch IPMNs by R1 and R2, respectively, and tumor nodules in 5/7 malignant IPMNs by both (inter-observer agreement [0.7-1.00]). Axial and 3D reformats did not improve detection except orientation. Mean DC was 4.3 and time taken to generate SACTp movies was 6.4 minutes. Conclusion: CT pancreatogram image obtained around pancreatic-duct center axis is a good 3D tool for outlining IPMN tumor-duct communications and mural nodules and is useful conjunct to axial CT dataset.

B-588

11:33

Magnetic resonance imaging (MRI) and MR cholangiopancreatography (MRCP) of multifocal intraductal papillary mucinous tumors (IPMT) of the side branches: MR pattern and its evolution R. Manfredi, S. Mehrabi, M. Motton, R. Graziani, S. Perandini, R. Pozzi Mucelli; Verona/IT ([email protected]) Purpose: To describe the MRI/MRCP pattern of IPMT and assess their evolution during follow-up. Methods and Materials: 26 patients with multifocal IPMT of side branches (SB) were included in this IRB approved, retrospective study. Inclusion criteria: r2 ectasic pancreatic SB, presence of communication duct with main pancreatic duct (MPD), r2 MR/MRCP examination. Exclusion criteria: IPMT involving both MPD and SB, previous surgery, absence of follow-up MRI examination. Median follow-up 27 months (range 6-59). Qualitative MRI analysis: number of cystic dilations of SB; communication morphology between cystic dilations of SB and MPD; presence of endoluminal ﬁlling defects within cystic dilations of SB; presence of parietal nodules, enhancement of wall of cystic dilations of SB. Quantitative MRI analysis: mean maximal diameter of cystic dilations of SB; mean communication neck length. Results: Diagnosis. Mean number of cystic dilations of SB: 7.5; communication neck was detected in 16/26 (60%) patients; endoluminal ﬁlling defects within SB in 6/26 (23%); parietal nodules in 1/26 (4%). Mean diameter of the cystic dilations of SB=18.8 mm. Main communication neck length=6.9 mm. Follow-up. Mean number of the cystic dilations of SB=8.4. Communication neck was detected in 20/26 (77%) patients. Endoluminal ﬁlling defects within SB were detected in 7/26 (27%); parietal nodules in 2/26 (8%). Enhancement of SB wall was detected in 2/26 (8%). Mean diameter of cystic dilations of SD increased to 22.3 mm (p 0.05), main communication neck length was 8.6 mm. Conclusion: Multifocal IPMT of SB shows a progression in time, but very slow. In our series only 2/26 patients showed parietal nodules.

A

B

C

D

E

F

G

277

Monday

Purpose: To assess the Secretin-stimulated MRCP ﬁndings in asymptomatic patients with slight/mild elevations of serum amylase and lipase levels and to determine whether there might be a pathoanatomic cause for these laboratory abnormalities. Methods and Materials: Eighty asymptomatic patients with serum hyperamylasemia and hyperlipasemia underwent MRI at 1.5 T-device (Signa EXCITE, GE Healthcare). After the acquisition of axial T1w/T2w sequences, MR pancreatography was performed using a single-slice coronal breath-hold, thick-slab (40-60 mm), SSFSE T2w sequence, repeated every 30 seconds up to 15 minutes following intravenous injection of Secretin (Secrelux®, Sanochemia). The presence of parenchymal and ductal abnormalities of pancreas, the changes of pancreatic ductal system and the ﬁlling of duodenum after Secretin-stimulation were evaluated by two observers in conference. Results: MRI showed pancreatic abnormalities in 55 (68.7%) patients, including pancreas divisum (n=22) (27.5%), morphologic changes compatible with mild chronic pancreatitis (n=12), juxtapapillary duodenal diverticulum (n=1), papillary sclerosis (n=3). Small cystic lesions ( 1 cm) within the pancreas were also seen in 22 (27.5%) patients and were associated with pancreas divisum in ﬁve of these cases. No malignancy was identiﬁed. After Secretin injection, a normal duodenal ﬁlling was observed in 75 patients, whereas pancreatic ﬂuid outﬂow appeared reduced in three patients with papillary sclerosis (2/3 also showed ductal "acinar ﬁlling" effect) and in two with mild chronic pancreatitis. Conclusion: Investigation of asymptomatic patients with nonspeciﬁc hyperamylasemia and hyperlipasemia by means of Secretin-stimulated MRCP yielded pancreatic ﬁndings in 68.7% of our series. Pancreas divisum was found more often than expected in the general population.

AIP, appeared hypoattenuating in 19/21 (90%), and isoattenuating in 2/21 (10%). During portal venous phase, pancreatic parenchyma showed contrast retention in 18/21 (86%), and wash out in 3/21 patients (14%). The main pancreatic duct was never visible within the lesion. After treatment, 15/21 (71%) patients showed normal enhancement pattern of pancreatic parenchyma, whereas 6/21 (29%) remained hypovascular. The MPD returned to its normal size within the lesion in all patients on follow-up CT. In focal forms of AIP, upstream MPD remained dilated in 1/8 patients. Conclusion: AIP appears as pancreatic parenchyma enlargement, with MPD stenosis within the lesion and upstream dilation, in focal forms. The normalization of these ﬁndings after steroid treatment is a helpful clue for the differential diagnosis with other pancreatic diseases.

Scientific Sessions B-589

10:30 - 12:00

11:42

Patients with suspected pancreatic malignancy: 64-row MDCT vs. gadobenate-enhanced 3.0 T MRI C. Kölblinger, A. Ba-Ssalamah, C. Kulinna-Cosentini, G. Schüller, P. Götzinger, W. Schima; Vienna/AT ([email protected]) Purpose: To prospectively compare the diagnostic performance of 64-row MDCT and gadobenate-enhanced 3.0 T MR imaging for detection and characterization of pancreatic masses. Methods and Materials: 39 patients (19 m, 20 f; mean age 65 y) were examined within one week with 64-row MDCT (Sensation 64, Siemens) and gadobenateenhanced 3.0 T-MRI (Trio, Siemens) in this prospective study. The CT protocol consisted of unenhanced, pancreas-parenchymal and portal-venous phase images, acquired after injection of 150 ml Iomeprol® (300 mg iodine/ml; ﬂow-rate 5 ml/s). MRI included T1w-GRE, T2w-TSE with and without fatsat, MRCP, dynamic and 60 min-delayed gadobenate (MultiHance®, at 0.1 mmol/kg)-enhanced T1w-3D-VIBE sequences. The standard-of-reference (surgery, n=17; biopsy, n=16; follow-up, n=6;) revealed 24 adenocarcinomas, seven metastases in two patients, one cystadenoma, two IPMT, one CCC, one neuroendocrine tumor, two inﬂammatory masses, one focal steatosis and normal ﬁndings in ﬁve patients. Images were evaluated by one blinded reader. Results: Both CT and MRI detected 38/39 tumors (97%). All 24 adenocarcinomas were visible with both modalities. Compared to CT, MRI showed one additional inﬂammatory mass, which was classiﬁed as probably malignant; another mass was incorrectly identiﬁed by MRI (false-positive). One gastric cancer metastasis was detected only with CT. One benign focal steatosis was classiﬁed incorrectly as malignant by CT. Accuracy for lesion classiﬁcation was 94% for CT and 92% for MRI. Accuracy for neoplastic vessel inﬁltration was 94% for CT and 97% for MRI (n.s). Conclusion: There is no signiﬁcant difference between 64-row MDCT and gadobenate-enhanced 3 T-MRI regarding detection and classiﬁcation of pancreatic tumors.

B-590

11:51

Diagnostic impact of F-18 FDG-PET-CT evaluating solid pancreatic lesions versus endosonography, ERCP with intraductal ultrasonography, and abdominal ultrasound K.U. Juergens, V. Schick, C. Franzius, O. Schober, T. Pohle, W.L. Heindel; Münster/DE ([email protected]) Purpose: This prospective single-center phase-II-trial evaluated the diagnostic impact of PET-CT assessing solid pancreatic lesions of unknown dignity (Ør1 cm) compared with clinically established modalities endosonography (EUS), ERCP with intraductal ultrasonography (IDUS), and ultrasonography (US). Methods and Materials: 46 patients (32m/14f, Ø=61.7 yrs). with clinically suspected pancreatic malignoma underwent F-18-FDG-PET-CT including contrastenhanced biphasic multidetector-spiral CT of the abdomen followed by a diagnostic work-up with EUS, ERCP with IDUS and US within 3 weeks’ time. PET-CT data sets were analyzed by two expert readers in a consensus reading. Histology from surgery (N=25), biopsy/FNA (N=18), and/or clinical follow-up r12 months (N=3) served as standard of reference. Results: In all patients, FDG-PET-CT was technically feasible: 36/46 pancreatic masses (78%) were detected in the pancreatic head, 7/46 and 3/46 in the body and the pancreas tail; 27 malignant tumors were diagnosed. The sensitivity and speciﬁcity of PET-CT were 89 and 74%, PPV/NPV were 83 and 82%. Sensitivity (81-89%), speciﬁcity (74-88%), PPV (83-90%), and NPV (77-82%) achieved by EUS, ERCP, and US were not signiﬁcantly different (all p 0.05). In 16 patients the result of at least one modality was false-positive or false-negative. In one patient, PET-CT detected a second lesion in the colon that appeared to be a 5 cm tubular adenoma with intermediate dysplasia on polypectomy. Conclusion: F-18 FDG-PET-CT has full technical feasibility and achieves comparable high sensitivity, speciﬁcity, PPV, and NPV evaluating small solid pancreatic lesions compared with clinically established imaging modalities. Additional clinical diagnoses are derived from concurrent imaging of the whole body.

278

A

B

C

D

E

F

G

Room C

GI Tract

SS 1401b Esophageal and gastric cancer: New approaches Moderators: M.A. Bali; Brussels/BE N. Papanicolaou; Philadelphia, PA/US

B-591

10:30

Histological evaluation of N-staging of esophageal cancer by diffusionweighted magnetic resonance imaging in comparison with PET K. Shuto, T. Aoyagi, G. Ohira, K. Narushima, H. Saito, A. Sato, T. Ota, T. Kazama, S. Okazumi; Chiba/JP Purpose: To evaluate the accuracy of diffusion-weighted magnetic resonance imaging (DWI) and PET for detecting lymph node metastasis of esophageal cancer with histological investigation. Methods and Materials: Thirty-seven consecutive patients with esophageal squamous cell cancer who underwent radical esophagectomy with cervicothoraco-abdominal lymphadenectomy were enrolled. Prior to surgery, both MRI and FDG-PET were performed. MR images were acquired using a 1.5 T scanner with following parameters: T2-weighted, fast spin-echo sequence; DWI, single shot spin-echo type of echo planar sequence. PET imaging was performed after injection of 370 MBq of F18-DG tracer. Deﬁnition as metastasis was made visually by using T2-WI/DWI fusion image on MRI and was made as SUV 3 on PET images. We compared both imagings with histological results and assessed the percent area of the cancer nest in histologically positive nodes. Results: 499 lymph node groups were examined. 51 (10%) were proven histologically positive. The sensitivity, speciﬁcity and accuracy rate of DWI were 73, 96 and 94%, respectively. Those of PET were 51, 98 and 93%. There was a signiﬁcant difference in sensitivity (p=0.030). In the entire histologically positive nodes, the percent area of the cancer nest of true-positive nodes on DWI was 52% and falsenegative was 17% (p=0.0218). However, no signiﬁcant difference of percent area was found between true-positive and false-negative on PET. Conclusion: Noninvasive DWI, which may depend on the density of cancer nest, will yield a better predictive power than PET and may be recommended in initial evaluation of suspected nodal metastasis for esophageal cancer patients.

B-592

10:39

MR-ﬂuoroscopy as follow-up examination in patients with narrow gastric tube reconstruction after total oesophagectomy V. Panebianco, M. Osimani, M. Di Martino, D. Lisi, S. Bernardo, R. Passariello; Rome/IT ([email protected]) Purpose: To ﬁnd new evaluation elements about functionality and morphology of neo-oesophagus in subjects who underwent narrow gastric tube (NGT) reconstruction, using MR-ﬂuoroscopy with Dynamic Turbo-FLASH sequences acquired during positive oral contrast agent administration. Methods and Materials: Ten healthy volunteers without history of oesophageal motility disorders and thirty patients, who underwent NGT reconstruction after total oesophagectomy between 2004 and 2007, were studied using a 1.5 T magnet, equipped with surface phased-array coil. Dynamic imaging protocol included Turbo-FLASH sequences (TR=600 ms; TE=1.3 ms; ﬂip angle=8°; thickness=20 mm; FoV=350; matrix=90x126; N. acquisition=45; TA=25 s; trigger off) acquired on sagittal, coronal and axial planes during oral administration of positive contrast agent boluses (yoghurt+Gd-DTPA 0.5 M, 1:100). Normality patterns of contrast agent transit time, oesophageal clearance and oesophageal lumen calibre, obtained from volunteers, were subsequently used as a reference in treated patients. Results: Good quality images were obtained in all patients, with adequate lumen contrast. Twenty patients had completely re-established motility of NGT (tube calibre below 50 mm, re-established motility with transit time below 9 seconds, no reﬂux, and no stasis); eight patients had mild alterations such as altered calibre ( 25 mm), raised transit time ( 20 sec), reﬂux and contrast agent stasis; two patients had severe alterations with grossly dilated NGT ( 45 mm), severe reﬂux and stasis with high transit time ( 20 sec). Conclusion: MR-ﬂuoroscopy approach represents a promising radiation-free modality in the follow-up of functionality and morphology of NGT.

Scientific Sessions B-593

10:48

Staging of gastric cancer: A comparison between 64-MDCT and MRI M. Anzidei, A. Napoli, B. Cavallo Marincola, C. Zini, P.L. Di Paolo, K. Lanciotti, C. Catalano, R. Passariello; Rome/IT ([email protected]) Purpose: To compare the usefulness of MRI with T2-weighted turbo spin-echo and fast T1-weighted 3D gradient-echo sequences with 64-DCT in the local staging of gastric carcinoma. Methods and Materials: Thirty patients with endoscopic diagnosis of gastric carcinoma underwent preoperative MRI and MDCT. MR (1.5 T) and 64-MDCT imaging (200 mA, 120 kv, thickness 1.25 mm, rec. int. 0.6 mm) were performed after i.m. injection of scopolamine and water distension of the stomach. In the MRI protocol True-FISP T2-weighted sequences in all three planes, turbo spin-echo T2-weighted axial sequences (TR 4000 ms, TE 100 ms, thickness 3 mm, matrix 384x384, TA 30) and gadolinium-enhanced fat-suppressed 3D gradient-echo axial sequences (TR 5.8 ms, TE 2.75 ms, thickness 3 mm, matrix 384x384, TA 18) were included. 64-MDCT was performed 60 sec after contrast agent injection (3.5-4 ml/sec). Two groups each of two radiologists independently analyzed MRI and 64-MDCT images; results were compared with pathologic ﬁndings. Results: MR imaging accuracy was higher than that of 64-MDCT (85 and 79%, respectively) in the differentiation of T1-T2 stages; the accuracy of MR and CT were not signiﬁcantly different in the differentiation of T2-T3 and T3-T4 stages (p 0.05). Overstaging was noted in 6% of cases with MRI and 10% with 64-MDCT. Understaging was noted in 15% of cases with MR imaging and 18% with 64-MDCT. Conclusion: MRI and 64-MDCT accuracies were not different in advanced stages of disease, while MRI was superior in the differentiation of early stages of gastric cancer.

B-594

10:57

Preoperative local staging of gastric cancer: Comparison of 3D multidetector row CT with air versus water distension H. Park, J. Lee, S. Kim, J. Lee, J. Han, B. Choi; Seoul/KR ([email protected])

B-595

11:06

Lymph nodes metastasis of gastric cancer: A comparison study of CT with pathological N staging L. Tang, X.P. Zhang, Y.S. Sun, J. Li, N. Wang, K. Cao; Beijing/CN ([email protected]) Purpose: To explore the correlation between CT ﬁndings of lymph nodes and pathological N (pN) staging in patients with gastric cancer. Methods and Materials: Eighty-nine patients with gastric cancer underwent CT examinations prior to radical resection. Number and the sum of size of lymph nodes detected on CT in each patient were recorded and compared with pN staging. The pN staging (pN0-3) was evaluated according to UICC/AJCC TNM staging system. The differences of number and the sum of size between metastasis and non-metastasis cases were analyzed. ROC curves were employed to choose the cutoff point and calculate diagnostic speciﬁcity and sensitivity.

B-596

11:15

Diffusion-weighted MR imaging of gastric cancer: Preliminary studies X.P. Zhang, L. Tang, Y.S. Sun, K. Cao, J. Li, N. Wang; Beijing/CN ([email protected]) Purpose: To evaluate the application of diffusion-weighted MR imaging (DWI) in displaying gastric cancer, and explore the usage of ADC values. Methods and Materials: MRI was performed on 101 patients with gastric cancer on a 1.5 T scanner. EPI-DWI was performed combined with sensitivity encoding technique. The parameters were: TR 3000 ms, TE 65 ms, matrix 128×128, slice thickness 5 mm, NEX=4, segmented breath-holds. The b values were 0 and 1000 s/mm2. FSPGR T1WI and SSFSE T2WI were also performed. Two radiologists measured the ADC of cancers, and calculated the CNR. Results: Ninety gastric cancers demonstrated discernable higher signal to nearby normal gastric wall on DWI. Forty-four gastric cancers displayed uniformly high signal; and 46 gastric cancers displayed nonuniform signal, mainly as delamination (27 cases) and mixing signals (19 cases). Ten Borrmann4 gastric cancers displayed “sandwich” sign on DWI, which appeared high on mucosa and serosa sides, with a low signal band immediately, and the display rate was higher than that of SSFSE T2WI (10/16 vs 3/16, p 0.05). The mean ADC of gastric cancers was 1.20p0.31×10-3 mm2/s. The ADCs of mucinous cancers were higher than that of non-mucinous ones (1.47p0.46×10-3 vs 1.18p0.30×10-3 mm2/s, p 0.05), and ADCs of diffuse-type gastric cancers were lower than that of conﬁning-type (1.45p0.31×10-3 vs 1.12p0.27×10-3 mm2/s, p 0.05). The CNR of DWI was higher than that of FSPGR and SSFSE with statistical differences (p 0.05). Conclusion: DWI has better contrast than conventional MRI, and it can provide useful information for the diagnosis and typing of gastric cancer.

B-597

11:24

Is there an incremental value for CT direct multiplanar reconstruction in T staging of gastric cancer? A comparison with conventional axial images M. Anzidei, A. Napoli, B. Cavallo Marincola, C. Zini, P.L. Di Paolo, K. Lanciotti, C. Catalano, R. Passariello; Rome/IT ([email protected]) Purpose: To compare the diagnostic accuracy of MDCT axial images and multiplanar reconstruction (MPR) obtained from raw data in the T staging of gastric cancer. Methods and Materials: Twenty-ﬁve patients with endoscopic diagnosis of gastric carcinoma underwent preoperative staging with MDCT. All examinations were performed on a 64-MDCT after i.m. injection of scopolamine and oral administration of 800 ml of tap water. Imaging protocol was performed pre and 60 s after i.v. contrast agent injection (ﬂow rate: 3.5-4 ml/sec, concentration 400 mg I/mL). MPRs were obtained directly from raw data for improved image quality. Two radiologists independently performed a blind analysis ﬁrst reading the axial images and then the multiplanar reconstructions; the results were compared with the pathologic ﬁndings. Results: Overall sensitivity of axial images alone was 82% in the ﬁrst reader and 85% in the second reader with 87% and 89% accuracy. Overall MPR sensitivity was 91% in the ﬁrst reader and 93% in the second reader with 90% and 94% accuracy. Interobserver agreement was high (weighted k-value 0.85) and there was signiﬁcant difference between the two modalities (p 0.01). Understaging on axial images was noted in 10% of cases for the ﬁrst reader and in 7% for the second; understaging on MPR images occurred, respectively, in 7% and 4% of cases. Conclusion: MPR images directly obtained from raw data are superior to axial images and are of pivotal importance for the staging of T parameter in patients affected from gastric cancer.

A

B

C

D

E

F

G

279

Monday

Purpose: To compare the accuracy of multi-detector row CT (MDCT) images with air versus water distension for preoperative T staging of gastric cancer based on surgical and histopathologic ﬁndings. Methods and Materials: During a four-month period, 112 consecutive patients with surgically conﬁrmed gastric cancer (70 men, 42 women, age range, 23-85 years; mean age, 58 years) who underwent preoperative MDCT were included. Fifty-ﬁve patients received effervescent granules orally before CT scanning for gastric distension, and CT included acquisition of CT gastrographic images using surface-shaded technique, as also transverse and multiplanar reformation (MPR) images. Fifty-seven patients drank 1 L of tap water before CT scanning and transverse and MPR images were obtained. Images were reviewed by two reviewers and differences in accuracy of CT images with air versus water distension for T staging were assessed with Fisher’s exact test. Results: For T1 cancer (n=59), detection rates of tumors with air-distended and water-distended images were 93.1% (27/29) and 66.6% (20/30) for reviewer 1, 75.9% (22/29) and 43.3% (13/30) for reviewer 2. Difference between two techniques was signiﬁcant in both reviewers (P=0.021 and 0.017). In assessment of T staging in T1 cancer, CT images with air versus water distension was 58.6% (17/29) versus 73.3% (22/30) and 55.2% (16/29) versus 73.3% (22/30) for each reviewer, and this difference was not signiﬁcant (P 0.05). Detection rates of primary tumor and T staging of advanced gastric cancer (T 2, n=54) showed similar trend but the differences were not signiﬁcant (P 0.05). Conclusion: MDCT with air-distension technique is comparable to that with waterdistension in preoperative local staging of gastric cancer.

Results: The mean numbers of lymph nodes detected by CT were 4.16p3.78, 10.90p3.21, 12.79p4.15, 19.20p3.99, and the sum of size (cm) were 2.68p1.88, 8.08p2.96, 11.0p4.97, 18.05p10.79 for pN0, pN1, pN2 and pN3, respectively. Both number and the sum of size showed statistical differences between non-metastasis and metastasis groups (p 0.01). Taking n=8 as cutoff point for number to judge metastasis on ROC, the Az value was 0.932, and diagnostic sensitivity and speciﬁcity were 81.4% and 89.5%. Taking 4.25 cm as cutoff point for the sum of size, the Az value was 0.983, and diagnostic speciﬁcity and sensitivity were 95.7% and 89.5%, respectively. Statistical differences were found between pN1 and pN3 (p 0.01), pN2 and pN3 (p 0.01) in number; and pN1 and pN2 (p 0.01) in the sum of size. Conclusion: Number and the sum of size of lymph nodes detected on CT can provide useful information for N staging of gastric cancer.

Scientific Sessions B-598

11:33

Efﬁcacy of HMDCT in differentiation between normal and pathologic wall thickening of the gastric antrum W.K. Matzek, N. Fakhrai, N. Bastati, J. Zacherl, C.J. Herold, A. Ba-Ssalamah; Vienna/AT ([email protected]) Purpose: The purpose of this study was to determine the normal and pathologic thickness of the antral wall using hydro multidetector computed tomography (HMDCT) in patients with and without underlying gastric disease using endoscopy as gold standard. Methods and Materials: We included 40 patients who underwent endoscopy and biopsy due to suspected gastric or esophageal disease. Contrast-enhanced HMDCT scans were performed with gastric and esophageal distension using water and effervescent granulate. The images were reviewed by two radiologists in consensus to determine antral wall thickness, symmetry of the antral wall, presence of stenosis or mass and assessment of morphology of the outer border. The thickness and length of the antrum were measured and the thickness of gastric body was used as a reference. Endscopical and histological results were used as gold standard. Results: In 8/40 patients with gastric cancer the tumor was located at the antrum. In these patients, the mean thickness of the antral wall was 18.5 mm (SD 7.65, min 11, max 24), compared to a mean of 5.9 mm (SD 2.48, min 2, max 12) in patients without an antral tumor. This difference was highly signiﬁcant (p 0.01). Seven tumors consisted of a circumferential wall thickening whereas only one tumor showed an asymmetric wall thickening. Conclusion: HMDCT measurements of the normal antral wall thickness range from 2 to 12 mm whereas wall thickening of more than 11 mm signiﬁcantly indicates an antral tumor. HMDCT of the stomach is a highly sensitive and speciﬁc tool for evaluation of antral tumors.

B-599

11:42

Comparison of tumor blood ﬂow with histological structure of gastric cancer by perfusion CT: An initial study A. Satoh, K. Shuto, G. Ohira, T. Aoyagi, K. Narushima, H. Saito, T. Ohta, Y. Nabeya, N. Yanagawa, S. Okazumi; Chiba/JP ([email protected])

value of blood volume to discriminate the microcirculation of the tumor from that of normal stomach. Results: Blood volume was signiﬁcantly increased in the test group compared with the control group (20.69p14.59 ml/100 g versus 9.08p7.60 ml/100 g in total stomach, p 0.0001; 16.13p8.04 versus 7.24p7.05 ml/100 g in the cardia, p=0.001; 25.51p16.76 versus 11.48p7.29 in the body, p=0.009; and 20.74p16.00 versus 8.38p8.06 in the antrum and pylorus, p=0.002). The threshold of 8.6 ml/100 g of blood volume corresponds to a sensitivity of 88.7% and a speciﬁcity of 62.9% to discriminate the microcirculation of gastric adenocarcinoma from that of normal stomach. There were no signiﬁcant differences in other comparisons. Conclusion: The ﬁrst-pass perfusion CT with whole tumor acquisition technique is a feasible and repeatable technique for quantifying tumor vascularity and angiogenesis in gastric adenocarcinoma.

10:30 - 12:00

Room F1

Genitourinary

SS 1407 Imaging of renal lesions: Perfusion, diffusion and more Moderators: M. Claudon; Vandoeuvre-les-Nancy/FR T. Meindl; Munich/DE

B-601

10:30

MR diffusion tensor imaging of the kidneys: Clinical feasibility study M. Notohamiprodjo1, U.I. Attenberger1, J.G. Raya1, C. Glaser1, M.F. Reiser1, S.O. Schoenberg2, H.J. Michaely2; 1Munich/DE, 2Mannheim/DE ([email protected])

Purpose: The purpose of this study is to compare tumor blood ﬂow with histopathological characteristics of gastric cancer and to assess tumor stromal structure including microvessel density and stromal density using perfusion CT (p-CT). Methods and Materials: The study included 48 patients with gastric cancer who underwent surgery. P-CT was performed on a single level of the tumor where largest in size. Tumor blood ﬂow (BF, ml/min/100 g tissue) value was measured by placing ROI within the tumor. We compared BF with histopathological characteristics of gastric cancer and evaluated its correlation with microvessel density and tumor stromal density by using an analyzing software to calculate the ratio of vessels and stromal tissue in a 100x magniﬁcation ﬁeld. Results: No signiﬁcant correlation was found between BF and tumor size. There was a signiﬁcant decrease in BF in advanced tumor depth (pT2 vs pT3/4=64.1 vs 41.0, p=0.0062) and in distant metastasis (M0 vs M1=53.3 vs 36.7, p=0.0499). A signiﬁcant difference was found between differentiated and undifferentiated subtypes (65.4 vs 34.5, p=0.0007). As for the stromal structure, despite the lack of signiﬁcant correlation with microvessel density, BF decreased with increased stromal density (R=0.861, p 0.0001). Conclusion: Decreased BF value acquired from p-CT may reﬂect a progressive state of gastric cancer. Change in stromal structure with increased stromal tissue was found, indicating an increased level of malignancy as supported in other biomolecular reports. P-CT can be a better strategy to estimate malignancy levels of cancer and a predictive indicator for therapeutic outcome.

Purpose: To evaluate the feasibility of MR Diffusion Tensor Imaging (DTI) of the kidney in volunteers and patients with various renal diseases. Methods and Materials: Ten volunteers and 25 patients (mean age 56p14.3) with various renal pathologies (renal masses, renal artery stenosis) underwent breathhold coronal fat-saturated echo-planar MR-DTI (six directions, b=0 and 300 sec/mm2, TR 730 msec/TE 72 msec/GRAPPA factor 2; ﬁve slices; slice thickness six mm; spatial resolution 2.1x2.1 mm; acquisition time 26 sec) of the kidneys at 1.5 T (Siemens Avanto). Using the Syngo DTI-task-card regions of interests were placed in the cortex, medulla and in renal masses if present. Fractional anisotropy (FA) and apparent diffusion coefﬁcients (ADC) were determined and tractography was used to visualize the renal diffusion anisotropy. Results: In all volunteers FA was signiﬁcantly (p 0.001) higher in the medulla (0.36p0.03) than in the cortex (0.21p0.02), whereas the ADC was signiﬁcantly higher in the cortex (2.43p0.19) than in the medulla (2.16p0.22). Tractography typically revealed a radial preferred direction of diffusion. In one patient with unilateral renal artery stenosis the medullar FA of the affected kidney was lower than on the contralateral side (0.30/0.37). FA/ADC of simple renal cysts (n=5) was 0.17p0.15/2.85p0.18 while renal cell carcinomas (RCCs) (n=13) showed a wide FA-range from 0.15-0.49. Using tractography displacement of renal parenchyma by RCCs could be detected. Conclusion: Using parallel imaging DTI measurements of the kidneys are feasible within a single breathhold with good discrimination between cortex and medulla. The results suggest that DTI may be a useful tool to study and monitor renal ultrastructure.

B-600

B-602

11:51

Features of tumor perfusion of gastric adenocarcinoma with 64-MDCT: Pilot study J. Yao, Z.-G. Yang, T.-W. Chen; Chengdu/CN ([email protected]) Purpose: The purpose of this study was to use ﬁrst-pass perfusion CT with volume-based technique to prospectively investigate perfusion features of gastric adenocarcinoma. Methods and Materials: Perfusion CT of gastric adenocarcinoma was performed with 64-MDCT in 53 patients, who were subdivided into three subgroups according to the location of the tumor. 24 patients with normal stomach served as controls. Four perfusion parameters, including perfusion, peak enhancement, time to peak, and blood volume, were obtained and compared between the test and control groups, among the three test subgroups, and between subgroups with and without lymph-node metastases. ROC analysis was performed to determine the cutoff

280

A

B

C

D

E

F

G

10:39

DCE-CT can predict and rapidly conﬁrm the response of metastatic renal carcinoma to anti-angiogenic drugs L.S. Fournier, R. Thiam, J. Medioni, L. Trinquart, D. Balvay, G. Frija, S. Oudard, C.A. Cuenod; Paris/FR ([email protected]) Purpose: To predict and evaluate early response of metastatic renal carcinoma to anti-angiogenic drugs by DCE-CT. Methods and Materials: 51 patients were treated for metastatic renal carcinoma by anti-angiogenic drugs (sorafenib N=9, sunitinib N=17), by placebo (N=13) or by interferon (N=5), as part of two phase III clinical trials. Dynamic enhancement of a metastatic target was followed for 90 s after injection of 80 ml of iodinated contrast, with a temporal resolution of 1 image/s. Microvascular parameters were measured before and after treatment by CTPerf software (GEMS): tumor blood ﬂow (TBF), tumor blood volume (TBV), vascular permeability (PS), mean transit time (MTT), and correlated to the best overall response according to RECIST.

Scientific Sessions Results: In patients under anti-angiogenic drugs, baseline values were higher in patients who would respond to treatment (N=10, TBF = 245.3 ml/min/100 g and TBV=15.5%), than in non-responders (N=20, 119.5 ml/min/100 g and 8.2%), showing that the initially highly vascular tumours were the most sensitive to these drugs. After one cycle of treatment (6 weeks), there was a signiﬁcant decrease in TBF (162.5 vs 76.7 ml/min/100 g), TBV (9.1 vs 3.9%) and PS (9.1 vs 4.1 ml/ min/100 g) for patients under anti-angiogenic treatment, whereas there was no change for those under placebo or interferon, showing that the anti-angiogenic effect detected by DCE-CT was speciﬁc to these drugs. Conclusion: Microvascular parameters derived from DCE-CT predict response to anti-angiogenic drugs and allow early detection of this response after a single cycle of treatment.

B-603

10:48

CT perfusion in follow-up of antiangiogenic drugs in patients with renal cell carcinoma: Preliminary experience A. Tognolini, O. Pianykh, R. Bhatt, S.N. Goldberg, M.B. Atkins, V. Raptopoulos; Boston, MA/US ([email protected]) Purpose: To determine optimal use of CT perfusion for monitoring antiangiogenic treatment in Renal Cell Carcinoma (RCC). Methods and Materials: From May 2006 to May 2007, 13 patients with RCC treated with Sorafenib or Sunitinib were enrolled in this IRB approved HIPAA compliant study. CT perfusion was performed with oncology protocol prior to treatment of all patients. For six patients, 1-3 follow-ups have been obtained so far. CT perfusion was performed on a 64-MDCT scanner (40 ml IV-contrast, 120 kVp, 50 mA, 4 cm region) using 0.5 sec rotation and delay interval of 5 sec - ﬁrst 75 sec, 10 sec for additional 130 sec. Blood ﬂow (BF), blood volume (BV), mean transit time and permeability surface were measured and correlated to diameter of target lesions. Results: Effective exposure time for dynamic scan was 7.5 sec (DLP: 124 mGy/ cm). On six patients with follow-up, BF at 2-4 months decreased from 343.8p205 ml/ min/100 g to 188.78p203 (34% decrease) and BV from 21.3p14 ml/100 g to 9.3p7 (37% decrease). Tumor growth was variable: three tumors increased (67% increase), two decreased (58% decrease), one unchanged. Follow-up at 4-10 months showed progressive decreases in BF to 81.8p33 (54% decrease) and BV to 4.65p3 (45% decrease) with either stability in size (two) or growth (one). Mean transit time and permeability surface were without any signiﬁcant trend. Conclusion: BF and BV progressively drop over time. These changes may not necessarily be predictive of tumor growth patterns. Further research is necessary to determine importance of physiologic and anatomic data for guiding therapeutic decisions and for predicting outcome.

B-604

10:57

Segmental inversion of enhancement on biphasic MDCT: A characteristic ﬁnding of small renal oncocytoma J. Kim1, J. Cho1, H. Lee2, S. Kim1; 1Seoul/KR, 2Seong Nam/KR ([email protected])

11:06

Evaluation of renal focal lesions by contrast-enhanced ultrasonography (CEUS): Correlation with computed tomography and magnetic resonance imaging M. Diezhandino Gallo-Ruiz, T. Fontanilla Echeveste, R. Pérez Aranguena, C. González Hernándo, C. Cortés León, M. Pastrana Ledesma; Madrid/ES ([email protected]) Purpose: To evaluate the ability of Pulse Inversion Harmonic Imaging (PIHI) with Sonovue® (Bracco) in depicting renal parenchymal changes on focal renal lesions. Methods and Materials: We prospectively evaluated 28 patients, nine females and 19 males (15-82 years; mean age 53.2 years), with 30 focal renal lesions including solid and cystic neoplasms and acute focal pyelonephritis. All patients were studied with gray-scale, color Doppler and PIHI at baseline and following contrast injection. Results were compared with reference methods such as multiphasic helical CT and MRI. Histological results were obtained in 12 cases. Results: Detection and conspicuity of renal focal lesions on CEUS were signiﬁcantly better than those on conventional US with sensitivity values of 87.5%. The delineation of lesions and internal necrosis was remarkedly improved except when the contrast uptake of the renal lesion matched with the signals from the normal surrounding cortex (three cases). Global concordance with reference methods was 75%. Interestingly, CEUS changed diagnosis and further management in two cases of cystic hypernefroma diagnosed as Bosniak category II by CT and speciﬁed diagnosis of two cases of papillary hypernefromas whose enhancement was difﬁcult to visualize by CT. It also visualized a hypernefroma in an inﬂamed kidney, not seen on CT. Conclusion: CEUS is a sensitive technique for depicting and characterizing renal parenchymal lesions. It has important advantages when compared with other contrast-enhanced methods and can make a big impact on the management of these patients. Overall, its lack of radiation emission makes it ideal for lesions that require follow-up.

B-606

11:15

3 T MR correlation between ADC values of diffusion weighted imaging and cellularity in malignant renal tumors E. Squillaci, G. Manenti, S. Mancino, R. Mastrangeli, C. Cicciò, G. Simonetti; Rome/IT ([email protected]) Purpose: To correlate ADC values of diffusion-weighted MR imaging (DW) at 3 Tesla (3 T) with tumour cellularity in malignant renal tumours. Methods and Materials: Ten normal volunteers and 28 patients with histologically proven malignant renal lesion underwent MRI of kidney by using a 3 T superconductive magnet. Axial T1 and T2W images, with and without fat-suppression, and coronal T2W images were acquired for each patient. DW images were obtained in the axial plane during breath-hold (19 seconds) with a SE-EPI single shot sequence with max b factor of 500 s/mm2. All lesions were surgically resected and the mean tumor cellularity was calculated as the number of tumour cell nuclei in 10 high-magniﬁcation ﬁelds in area without regressive phenomena. Comparison between tumour cellularity and the mean ADC value was performed using simple linear regression analysis. Results: Mean ADC value in normal renal parenchyma was 2.35p0.31x102 mm3/ sec, while mean ADC value in malignant renal tumours was 1.69p0.24x102 mm3/ sec. In our series there wasn’t a statistically signiﬁcant difference between ADC values of different histotypes (1.73p0.8 clear cell Ca, 1.84p0.4 papillary variant Ca, 1.67p0.9 granular cell Ca, 1.71p0.3 sarcomatoid variant Ca, 1.46p0.6 transitional cell Ca). The analysis of mean ADC values showed an inverse linear correlation with cellularity in malignant renal tumours with a Pearson's correlation coefﬁcient r=-0.71 (p 0.01). Conclusion: DW MR imaging with a 3 T MR system seems to be a feasible and reliable technique to evaluate renal parenchyma and to differentiate malignant renal tumours on the basis of tumour cellularity.

B-607

11:24

Value of FLASH-STAR MRI in the diagnostic evaluation of renal masses M.M.H. Abd Ellah, L. Pallwein, C. Kremser, G. Göbel, G. Bartsch, F. Frauscher; Innsbruck/AT ([email protected]) Purpose: To assess the value of FLASH-STAR sequence in magnetic resonance imaging (MRI) in the diagnostic evaluation of renal masses. Methods and Materials: MRI examination was performed for 10 patients on a 1.5 T whole body scanner. A pulsed arterial spin labelling (ASL) technique based on STAR (signal targeting with alternating radiofrequency) was used. Results were

A

B

C

D

E

F

G

281

Monday

Purpose: To assess the usefulness of segmental inversion of enhancement between corticomedullary (CMP) and early excretory phase (EEP) in differentiating small renal oncocytoma from renal cell carcinoma (RCC) on biphasic MDCT. Methods and Materials: Ninety-eight patients with pathologically conﬁrmed renal masses smaller than four cm were enrolled in this study. We deﬁned segmental inversion of enhancement: relatively highly enhancing and less enhancing segments in the mass on CMP images were converted into less enhancing and highly enhancing segments, respectively, on EEP. Two experienced radiologists retrospectively assessed the existence of segmental inversion and measured attenuation values of each segment on CMP and EEP images. The Fisher’s exact test was used for statistical analysis. Results: Ten oncocytomas and 88 RCCs were included. A total of eight of 10 oncocytomas showed segmental inversion of enhancement between CMP and EEP; only one of 88 RCCs showed. There was statistically signiﬁcant difference (p 0.0001, odd ratios, 39.6). In differentiating oncocytoma from RCC, the segmental inversion of enhancement showed a sensitivity of 80%, a speciﬁcity of 98.9%, positive likelihood ratios of 72.7, and negative likelihood ratios of 0.20. The mean values of attenuation difference between two segments on CMP and that of same location on EEP were 60.38p40.20 and -36.88p20.02 HU, respectively, corresponding to enhancement inversion of each segment. Conclusion: The segmental inversion of enhancement between CMP and EEP is a characteristic enhancement pattern of small renal oncocytoma on biphasic MDCT, and it is a reliable predictor for differentiating small oncocytoma from RCC.

B-605

Scientific Sessions obtained by subtraction of two data sets acquired alternately with and without ﬂow inversion, which were then analysed using the free software ImageJ. The image ﬁndings were compared with pathohistology. Results: Of the 10 patients six showed RCC and four TCC on pathohistology. The mean of the FLASH-STAR measurement values for RCC was 61.180p6.475, and for TCC 82.952p1.542. The mean for the cortex of all 10 patients was 37.689p4.735 and for the medulla 65.646p6.845. We found signiﬁcant differences between TCC and RCC using FLASH-STAR measurement values (p 0.001). Conclusion: ASL MRI using FLASH-STAR sequence showed signiﬁcant difference between RCC and TCC in this small series. Therefore this new technique maybe useful in the differentiation of renal tumors.

B-608

11:33

Correlation between RECIST, WHO and volume measurement on CT in patients with renal cell carcinoma W.H. Gayed, A. Salah, A. Sahdev, J. Shamesh, T. Powles, R.H. Reznek; London/UK ([email protected]) Purpose: To compare WHO and RECIST (Response Evaluation Criteria in Solid Tumours) criteria with direct tumour volume measurements in assessing tumour response in patients with renal cancer. We have previously shown that volume measurement MDCT correlates extremely well with true tumour volume. Methods and Materials: Two radiologists independently measured 184 lesions on 44 CT scans in 18 patients using all three methods to assess tumour response in patients treated with tyrosine kinase inhibitor. Accepted standards for response categories were used for all three methods of measurement. Bland-Altman analysis (http://www-users.york.ac.uk/~mb55/meas/ba.htm) was used to statistically test the data. Results: Bland-Altman method of agreement showed good agreement between RECIST and volume (Bias=-17.5); no signiﬁcant agreement between RECIST and WHO (Bias=-1028.67); no signiﬁcant agreement between volume and WHO (Bias=1011.13). Using volume measurements, RECIST and WHO, partial response was recorded in one, two, and four patients; progressive disease in four, ﬁve, and six patients; and stable disease in 13, 11, and eight patients, respectively. Conclusion: Our results show that RECIST criteria correlate well with direct tumour volume measurements. WHO criteria correlate poorly. When volume measurement was compared with RECIST, four out of 18 (22%) patients were designated to a different response category and 12 out of 18 (67%) when compared with WHO.

B-609

Purpose: To assess the kidney medulla and cortex microstructural behavior using the diffusion parameters and to evaluate the possibilities of diffusion tensor imaging (DTI) in the assessment of the transplanted kidney allograft (XTK). Methods and Materials: 19 patients (13 males, six females, mean age 42.2 y) underwent DTI of XTK 12 hours before the biopsy after written consent; study was approved by the local ethics committee. Imaging protocol contained half-Fourier single shot T2 weighted sequence (HASTE) to plan 12-directional diffusion weighted sequence with b=700 s/mm and b=0. DTI software package (General Hospital Corporation, Boston, MA, USA) was used for calculation of apparent diffusion coefﬁcient (ADC) and fraction anisotropy (FA) separately in medulla and cortex. ADC and FA of the medulla and cortex within the same kidney were compared using paired t-test. Two groups, normal (n=9) and pathologic biopsy ﬁnding (n=10), were tested using Wilcoxon-White test to ﬁnd differences in ADC, FA, ADC medulla/ cortex ratio and FA medulla/cortex ratio. Results: A signiﬁcant difference (p=0.001) was found in FA between medulla and cortex, but no difference between ADC values. Signiﬁcant differences of FA medulla/ cortex ratio of (p=0.001) and FA of the medulla (p=0.05) were found in patients with normal kidneys. Other differences were statistically insigniﬁcant. Conclusion: The medulla differs from cortex signiﬁcantly in FA values. FA medulla/ cortex ratio is promising in the assessment of the normal or pathological condition of the kidney allograft. Supported by the research project MSM-0021620819 of the Czech government.

A

B

C

D

E

F

G

11:51

Non-invasive assessment and quantitation of renal parenchyma perfusion by using contrast-enhanced ultrasound after sulfur hexaﬂuoride-ﬁlled microbubble injection: A proposal for a new mathematical model E. Quaia, L. Torelli, A. Nocentini, M.A. Cova; Trieste/IT ([email protected]) Purpose: To assess the capability of a new mathematical model for the calculation of numerical parameters related to renal perfusion at contrast-enhanced US. Methods and Materials: Twelve healthy volunteers (7 male, 5 female; mean age 45 years) without any undergoing pharmacologic treatment underwent renal US scan after sulfur hexaﬂuoride-ﬁlled microbubbles i.v. injection (2.4 ml in 60 seconds). Volunteers were divided into two groups according to age: 27-48 years (n=8; group 1) and 61-80 years (n=4; group 2). In both kidneys, the progressive reﬁll of renal parenchyma after microbubble destruction was monitored by low transmit power (mechanical index, 0.12). The reﬁll kinetics of renal cortex was approximated by the standard negative exponential function and by piecewise linear function previously developed from a phantom model (dialysis tube with 8600 capillaries presenting an internal diameter of 200 μm). Results: Piecewise linear function provided a curve composed of 1-2 (volunteer group 1) or 3-6 lineary tracts (volunteer group 2). The piecewise linear vs negative exponential function provided a better data approximation both in the ﬁrst ascending tract- to blood velocity-related (Mean Square Error MSE: 250 vs 450 arbitrary units; p 0.05) and plateau phase of the curve- to renal fractional blood volume-related (MSE: 300 vs 750 arbitrary units; p 0.05, Wilcoxon test). Conclusion: The more complex reﬁll kinetics consisting of 3-6 lineary tracts suggests the presence of nephroangiosclerosis. The piecewise linear function improves data approximation and, consequently, the calculation of numerical parameters related to renal perfusion in comparison to standard negative exponential function.

10:30 - 12:00

Room F2

Breast

SS 1402 MR: Know-how for new approaches Moderators: M. Müller-Schimpfle; Frankfurt a. Main/DE K. Vassiou; Larissa/GR

B-611

11:42

Diffusion tensor imaging of the transplanted kidney, ﬁrst experience J. Ferda, T. Reischig, O. Hes, H. Mirka, M. Matejovic, B. Kreuzberg; Plzen/CZ ([email protected])

282

B-610

10:30

Single USPIO-enhanced MR imaging in the assessment of axillary lymph node metastases in patients with breast carcinomas M. Memarsadeghi1, C.C. Riedl1, A. Kaneider2, A. Galid1, M. Rudas1, M. Weber1, T.H. Helbich3; 1Vienna/AT, 2Horn/AT, 3Toronto, ON/CA ([email protected]) Purpose: To retrospectively assess the value of single ultrasmall superparamagnetic iron oxide (USPIO)-enhanced MR imaging for the depiction of axillary lymph node metastases in patients with breast carcinomas. Methods and Materials: Single USPIO-enhanced images of axillary lymph nodes were analyzed in 22 female patients (mean age, 60 years; age range, 40-79 years) with breast carcinomas by two radiologists. Single USPIO-enhanced (2.6 mg Fe/kg intravenously) MR images were evaluated on axial T1-weighted, axial T2-weighted, and axial T2*-weighted sequences. Visual analysis of MR images was performed on a node-by-node and patient-by-patient basis using histology as reference standard. Sensitivity, speciﬁcity, and accuracy values and the corresponding 95% conﬁdence intervals (CI) were calculated. Results: On USPIO-enhanced MR imaging, 13/133 lymph nodes were rated as true positive, 120/133 as true negative. No false negative or false positive lymph node was diagnosed. USPIO-enhanced MR imaging revealed a node-by-node sensitivity, speciﬁcity, and accuracy of 100% each. USPIO-enhanced MR imaging did not miss any metastatic lymph nodes on a patient-by-patient correlation. Conclusion: Single USPIO-enhanced MR imaging appears to be highly accurate for the assessment of axillary lymph node metastases in patients with breast carcinomas. Therefore, non-enhanced MR imaging prior to USPIO-enhanced MR imaging can be avoided.

Scientific Sessions B-612

10:39

B-614

10:57

A novel class of macromolecular MRI contrast media: Evaluation of their potential to differentiate breast cancer from benign soft tissues C.C. Cyran1, Y. Fu1, B. Chaopathomkul1, H.J. Raatschen2, B. Sennino1, V.S. Rogut1, D.M. Shames1, M.F. Wendland1, R.C. Brasch1; 1San Francisco, CA/US, 2Berlin/DE ([email protected])

Diffusion weighted imaging (DWI) in MR-mammography (MRM): Comparison of echo planar imaging (EPI) and half-Fourier single-shot turbo spin echo (HASTE) diffusion techniques P.A.T. Baltzer, D.M. Renz, K.-H. Herrmann, J. Böttcher, J.R. Reichenbach, W.A. Kaiser; Jena/DE ([email protected])

Purpose: To evaluate three newly developed macromolecular polyethylene glycol (PEG)-core dendrimeric gadolinium (Gd)-based MRI contrast agents for their performance in quantitative assays of endothelial leakiness and tissue vascular density for the differentiation of cancer from normal soft tissues. Methods and Materials: A total of 32 athymic rats with human breast cancer xenografts (MDA-MB-435) were imaged by dynamic MRI at 2.0T using a T1-weighted SPGR sequence following enhancement with one of three new (Gd-DOTA)-conjugated PEG-core dendrimer contrast agents (effective molecular weights 161-323 kDa). Results were compared to albumin-Gd-(DTPA), a prototype macromolecular contrast agent. Estimates of tumor microvascular permeability (KPS; μl/min/100 cm3) and tumor fractional plasma volume (fPV; %) were calculated for tumor xenografts and skeletal muscle based on a two-compartment kinetic model. Results: The largest PEG-core contrast agent, PEG20,000-Gen4-(Gd-DOTA), leaked in breast tumors (KPStumor=50p23) while exhibiting no measurable transendothelial leak (KPSmuscle=0) from normal soft tissue microvessels allowing successful differentiation of cancers from normal muscle (p 0.05). The two smaller agents, PEG12,000-Gen4-(Gd-DOTA) and PEG12,000-Gen3-(Gd-DOTA), leaked from tumor and normal muscle microvessels but only PEG12,000-Gen4-(Gd-DOTA) allowed successful differentiation of cancer from normal muscle (K PStumor=51p26, KPS=21p18, p 0.05). muscle Conclusion: MRI assays of vascular endothelial leakiness enhanced with new PEG-core, (Gd-DOTA)-conjugated macromolecular contrast agents, which are intended for clinical development, proved successful for the differentiation of human breast cancer xenografts from normal soft tissue. The apparent threshold in effective molecular weight for a clear differentiation of cancer from normal muscle with no measurable leak in the muscle is between 194 and 323 kDa. Supported by NIH/NCI grant R01 CA082923.

Purpose: To compare the diagnostic use of additional EPI- and HASTE-DWI in routine MRM. Methods and Materials: Following routine MRM (1.5 T Siemens Sonata, 0.1 mmol/ bw Gd-DTPA), EPI- and HASTE-DWI (TR 3500/3000, TE 800/1000, averages 6/1, b: 0, 750, 1000/0, 800 s/mm2, acquisition time 2:48/2:04 min). with matrix 192 and FOV 350 was applied. ADC-maps were calculated. Lesion demarcation on DWI images was compared to subtracted T1w images using four ordinal categories (0=not visible, 1=just visible in knowledge of contrast enhanced images, 2=limited visibility, 3=same visibility as on T1w-images). Greatest diameter and mean ADC-values were compared using Mann-Whitney-U, Wilcoxon and ROC statistics. Results: DWI was performed in 65 patients, 61 lesions (35 malignant, size 23.6p18.2 mm and 26 benign, size 16.2p10.8 mm) were found. ADC values (10-3 mm2/2) in malignant lesions were calculated with 1.01p0.09 for EPI- and 1.06p0.17 for HASTEDWI. In benign lesions ADC values of 1.65p0.41 (EPI) and 1.74p0.37 (HASTE) were measured. Diagnostic accuracy was found to be 91.2% for EPI- and 90.7% for HASTE_DWI. Lesion demarcation was signiﬁcantly (p 0.001) lower on HASTE- (mean 2) compared to EPI-DWI (mean 2.5). No signiﬁcant differences between T1w-images, EPI- and HASTE-DWI were found for size measurements (p 0.05). Conclusion: EPI-and HASTE-DWI were found to have an equally high ( 90%) diagnostic accuracy in breast lesions, underlining the possibility of non-contrast imaging of the breast in a short acquisition time. Tumor size measurements showed no signiﬁcant differences compared to T1w-images. However, visual lesion demarcation in HASTE-DWI was lower compared to EPI-DWI.

B-613

10:48

Contrast-enhanced magnetic resonance mammography (MRM): Improvement in breast lesion detection and characterization with gadobenate dimeglumine (Gd-BOPTA) compared to gadopentate dimeglumine (Gd-DTPA) F. Pediconi, C. Catalano, S. Padula, A. Roselli, V. Dominelli, S. Cagioli, R. Passariello; Rome/IT ([email protected])

11:06

Role of diffusion weighted imaging (DWI) in monitoring patients with locally advanced breast cancer (LABC) undergoing primary chemotherapy (PCT) L. Martincich, I. Bertotto, F. Montemurro, C. Bussolino Peretto, L. Cellini, D. Regge; Candiolo/IT ([email protected]) Purpose: DWI provides a quantitative functional parameter (Apparent Diffusion Coefﬁcient, ADC) related to cellular density. The aim of the study was to evaluate if variations of ADC value occur in patients with LABC treated by PCT. Methods and Materials: 14 patients with unifocal LABC (stage II with T 3 cm or IIIA/B) underwent DWI before, after two cycles and at the end of PCT. Images were acquired by 1.5 T equipment and 8-channel coil, using EPI sequence (b-value 900 s/mm2-thickness 4 mm-acquisition time 80 sec). ADC was obtained using Functool software (GEHC) by drawing a region of interest within the lesion. Pathological response was classiﬁed according to Smith (JCO 2002); patients achieving grade 5 (pathological complete response) or 4 (small clustered cancer cells) were considered as Responders while those who scored 1-3 were considered as NonResponders. At each examination variation of ADC was evaluated for Responders versus Non-Responders. Results: At baseline no signiﬁcant difference was found in ADC mean value between the three Responders and the 11 Non-Responders (1.06p0.14 mm2/s vs 0.96p0.17 mm2/s). Mean ADC value during PCT was 1.49p0.08 mm2/s for Responders and 1.07p0.21 mm2/s for Non-Responders (p=0.01); after PCT it was 1.6p0.17 mm2/s in Responders and 1.15p0.23 mm2/s in Non-Responders (p=0.02), showing a signiﬁcant difference between the two groups during and after PCT. Conclusion: Different changes of ADC values were observed in the Responders compared to Non-Responders during and after PCT, suggesting DWI as imaging technique able to identify medical treatment effects. DWI shows advantages for clinical application with short acquisition time and execution without contrast media administration.

B-616

11:15

Feasibility and utility of magnetization transfer ratio in breast lesions V.F. Muglia, R.H.M. Bonini, D.M. Zeotti, J. Elias, Jr, C.S. Trad, J.M. Sarmento Filho, H.H.A. Carrara, J.M. Andrade, A.C. Santos; Ribeirao Preto/BR ([email protected]) Purpose: To evaluate the feasibility and utility of Magnetization Transfer Ratio (MTR) for characterization of breast lesions. Methods and Materials: Forty-nine patients with 52 breast lesions, classiﬁed as Bi-RADS 4 and 5, were prospectively evaluated using magnetization transfer sequences. MTR maps were generated using minctools software and the MTR

A

B

C

D

E

F

G

283

Monday

Purpose: To prospectively and intra-individually compare equivalent 0.1 mmol/kg doses of gadobenate dimeglumine (Gd-BOPTA) and gadopentate dimeglumine (Gd-DTPA) for accuracy of detection and characterization of breast lesions on contrast-enhanced MR mammography (MRM). Methods and Materials: Fifty-six consecutive women with one or more suspected breast tumors on conventional X-ray mammography and sonography underwent two identical CE-MRM examinations at 1.5 T separated by 48-72 h. A T1w 3D gradientecho sequence was acquired before and at 2, 4, 6, 8 and 10 minutes after injection of either Gd-DTPA or Gd-BOPTA at an identical ﬂow rate of 2 ml/s. Separate and combined assessments of non-enhanced, enhanced and subtracted images were performed blindly by two readers in consensus. The accuracy for lesion detection was determined against ﬁnal diagnosis, based on ﬁndings from conventional mammography, sonography and surgery. The sensitivity, speciﬁcity, positive and negative predictive values, and accuracy for malignant lesion identiﬁcation were determined against histology. Results: Gd-BOPTA-enhanced MRM detected signiﬁcantly (p=0.003) more lesions (66/76) than Gd-DTPA-enhanced MRM (56/66) and detected lesions were more conspicuous with Gd-BOPTA. Conﬁdence for lesion characterization was (p=0.031) greater with Gd-BOPTA. The sensitivity, speciﬁcity, PPV, NPV and overall accuracy for malignant lesion identiﬁcation were superior for Gd-BOPTA-enhanced MRM compared with Gd-DTPA-enhanced MRM. Evaluation of signal intensity-time curves revealed (p 0.0001) greater quantitative lesion enhancement with Gd-BOPTA. Conclusion: Gd-BOPTA is better than Gd-DTPA for the detection of breast lesions and for the identiﬁcation of malignant breast lesions. Gd-BOPTA could be used in breast cancer screening procedures.

B-615

Scientific Sessions values were calculated for each lesion using Dousset technique. MTR values from pectoral major muscle, adipose and ﬁbroglandular tissues were also calculated. These results were compared to histopathological diagnosis. Patients were divided into two groups: benign and malignant lesions. Results: The mean age was 50.5 years, ranging from 27 to 73. Thirty-six lesions (69.3%) were benign and 16 were (30.7%) malignant. The mean value for MTR for benign group was 14.77p2.19 and for malignant 19.02p3.40. Using the student's-t test, a highly signiﬁcant p value (p 0.001) was found. No signiﬁcant differences were found for MTR values from muscle, adipose and ﬁbroglandular tissues between the two groups. A Receiver-Operator Characteristic (ROC) curve showed an area under the curve of 0.865. Conclusion: MTR values for benign and malignant breast lesions were signiﬁcantly different. In our series MTR could be used to discriminate breast lesions.

B-617

11:24

Analysis of apparent diffusion coefﬁcient (ADC) values of malignant and benign pathology of the breast L. Ballesio, L. Manganaro, S. Savelli, M. Angeletti, I. D’Ambrosio, N. Ravazzolo, A. Francioso; Rome/IT ([email protected]) Purpose: To evaluate the potential role of apparent diffusion coefﬁcient (ADC) values calculated on ADC maps in the differential diagnosis of malignant and benign breast lesions. Methods and Materials: We prospectively evaluated 42 selected patients among all women who referred to our department for breast MRI examination for characterization of suspicious breast lesions detected at standard methodics and histologically proved. We performed routine MRI including DWI scans acquired with b factors of 0, 500 and 1,000 s/mm2. ADC maps were reconstructed and ADC values of each lesion calculated. MR ﬁndings were compared with histologic ﬁndings, and sensitivity, speciﬁcity, PPV, NPV and accuracy were obtained. Results: MRI detected 57 lesions: 19 cystic, 14 benign solid and 24 malignant solid lesions. The mean ADC value was signiﬁcantly lower for breast cancer (1.09p0.20x10-³ mm2/s) than for cystic lesions (2.10p0.21x10-³ mm2/s) and benign solid lesions (1.56p0.29x10-³ mm2/s). Despite its high contrast resolution due to a suboptimal spatial resolution, DWI showed some limits as ghost, magnetic susceptibility artefacts and difﬁculty in detecting small lesions. When an ADC threshold of 1.23x10-3 mm2/s was used for differentiating malignant from benign lesions, sensitivity was 83.3%, speciﬁcity 78.8%, PPV 74.1%, NPV 86.7% and accuracy 80.7%. Conclusion: ADC value assessment can be considered a second-level tool to be added to routine MRI evaluation as it improves the diagnostic information content of standard BIRADS parameters in distinguishing between malignant and benign breast lesions.

B-618

11:33

Quantiﬁcation of total choline compound in in-vivo breast cancer with external standard with 1H magnetic resonance spectrum L. Jiang, C. Zhou, X. Du; Beijing/CN ([email protected]) Purpose: To quantify the total choline compound (tCho) in in-vivo breast cancer with external standard with 1H MRS. Methods and Materials: Five bottles of choline and CuSO4 solution with different choline concentrations were made as test models and one bottle of choline solution as calibration model. Under the same conditions, MRS acquisition was performed on the same position on different test models and linear correlation between areas under choline peaks and choline concentrations was analyzed. MRS acquisition with different sizes of voxels on the same position on the same model was performed and linear correlation between voxel sizes and areas under choline peaks was analyzed. The uniformity of signal intensity over the volume enclosed within the coil was tested. T1 and T2 values of choline in test model and calibration model, respectively, were measured. Then the calibration strategy of quantiﬁcation was tested. At last, we performed MRS on seven breast cancers and calculated their choline concentrations using calibration equation taking calibration phantom as the external standard. Results: Linear correlation existed between subpeak areas of choline and choline concentration as well as between voxel sizes and subpeak areas. T1 and T2 values of test and calibration models were 1902.5/1032.8, and 2426.3/989.4, respectively. With the calibration strategy, the absolute concentration of choline in the calibration phantom was 2.4p0.6 mM, which had no signiﬁcant difference with the actual concentration. The average concentration of seven breast cancers was 3.64p2.52 mM. Conclusion: Quantiﬁcation of in-vivo choline with 1H magnetic resonance spectrum with external standard is a valuable and applicable method.

284

A

B

C

D

E

F

G

B-619

11:42

Do we really need CAD? A comparison between visual, region-of-interest (ROI) and computer-assisted-diagnosis (CAD) measurement methods for time curve data analysis in MR-mammography P.A.T. Baltzer, D.M. Renz, C. Freiberg, S. Beger, J. Böttcher, M. Gajda, O. Camara, W.A. Kaiser; Jena/DE ([email protected]) Purpose: Available CAD systems for MR-Mammography (MRM) offer the possibility of semiautomatic dynamic analysis. This investigation was performed to compare visual, ROI and CAD methods for dynamic data measurement in MRM. Methods and Materials: 329 patients undergoing surgery of breast lesions after MRM (1.5 T, 0.1 mmol/kg bw Gd-DTPA) were analyzed prospectively using a CAD-system (Dynacad, Invivo, USA). Dynacad classiﬁes dynamic data depending on initial signal increase (50-100%; 100%) and time curve type (continuous increase; plateau; washout). The system automatically detects the most suspect curve and visualises the percental curve type distribution in a volume of interest (VOI). Not-threshold-reaching lesions ( 50%) were counted negative. The most suspect curve was assessed before CAD measurement by two blinded observers using visual and ROI methods. Kappa statistics and diagnostic parameters were calculated for visual, ROI and CAD results. Results: Pathology revealed 469 lesions (279 malignant, 190 benign). 240 malignant (86.0%) and 88 (46.3%) benign lesions were colour-coded by CAD. Curve type measurements varied signiﬁcantly (p 0.01) between visual, ROI and CAD with kappa of 0.81 (visual/ROI), 0.78 (visual/CAD) and 0.81 (ROI/CAD). Sensitivity was found to be highest in CAD (78.8%) compared to visual (72.4%) and ROI (75.3%) measurements whereas speciﬁcity amounted to 77.4% (visual), 76.3% (ROI) and 73.2% (CAD). Conclusion: Time curve measurements by visual, manual or computer-assisted methods differ signiﬁcantly. Highest sensitivity was found using CAD whereas speciﬁcity was highest in visual curve type assessment. A second reader opinion using CAD might be achieved by integrating other criteria besides dynamic data.

B-620

11:51

Correlation of apparent diffusion coefﬁcient value of cancer breast with prognostic factors A.A.A. Abdel Razek, G. Gaballa, A. Denewer, N. Nada; Mansoura/EG ([email protected]) Purpose: To correlate the apparent diffusion coefﬁcient value of breast cancer with prognostic factors. Methods and Materials: Prospective study was conducted on 77 consecutive untreated female patients (aged 25-64 ys; mean 45 ys) with breast lesions. All patients were examined at 1.5 Tesla using dedicated bilateral breast coil. They underwent dynamic contrast enhanced MR imaging and diffusion weighted MR imaging of the breast using a single shot echo planar imaging with a b-factor of 0, 200 & 400 sec/mm2. The scanning parameters were: TR=10000 ms, TE=108 ms, NEX=2, bandwidth=125 kHz, FOV=2-25 cm, slice thickness=4 mm. Apparent diffusion coefﬁcient (ADC) maps were reconstructed. The ADC value of the breast lesion was calculated. The ADC value correlated with histologic types, classic pathologic prognostic factors (tumor size, histologic type, grade and lymph nodes) and immuno-histochemistry detected biomarkers (ER, PR and Ki-67). Pearsons correlation test was used to determine the strength of relationship between the ADC value and prognostic factors. Results: The mean ADC value of breast cancer was 1.02p0.13X10-3 mm2/sec. Lower ADC value was associated with established predictor of poor prognosis. It was associated with larger tumor size, higher histologic grade, negative ER, negative PR, lower Ki-67 and presence of lymph nodes. The ADC value correlated with tumor size (p 0.03), histologic grade (p 0.003), ER (p 0.001), PR (p 0.002), Ki-67 (p 0.04) and lymph nodes (p 0.03). Conclusion: We concluded that apparent diffusion coefﬁcient value is a new imaging parameter that may identify highly aggressive breast cancer.

Scientific Sessions 10:30 - 12:00

Room G/H

Molecular Imaging

SS 1406b Clinical applications Moderators: C.A. Cuénod; Paris/FR C.C. Riedl; New York, NY/US

B-621

10:30

Comparison of whole-body FDG-PET/CT and whole body MRI at 3 T for tumor staging in oncology E. Squillaci, O. Schillaci, S. Mancino, F. Cicciò, G. Manenti, G. Simonetti; Rome/IT ([email protected]) Purpose: To compare the accuracy of PET/CT with whole-body MRI (WBMR) in staging different malignant diseases. Methods and Materials: 20 consecutive patients with known primary tumors (colorectal, breast, lung cancer, lymphomas, HCC) underwent [18F]-2-ﬂuoro-2deoxy-D-glucose (FDG) PET/CT for tumor staging (GE Discovery ST 16). Imaging evaluation was performed for CT alone, PET+CT viewed side by side, and fused PET/CT data. WBMR was performed at 3 T (Philips Achieva, Best, The Netherlands) with a body coil and unlimited ﬁeld of view. A coronal STIR-sequence was used for imaging of the different body regions before and after administration of 20 ml of GD-DTPA at 2 ml/sec. Histopathology or clinical follow-up of 6 (p2) months served as standard of reference. Results: Fused PET/CT proved signiﬁcantly more accurate in assessing the overall TNM stage compared to CT alone (p 0.05), side-by-side CT+PET (p 0.05) and WBMR. Out of 20 patients 16 (80%) were correctly staged with PET/CT, 14 (70%) with side-by-side CT+PET, 13 (65%) with CT alone and 12 (60%) with WBMR. No statistically signiﬁcant difference could be detected between PET/CT and CT+PET in assessing M-staging. Combined PET/CT had an impact on the treatment plan in three patients compared to other PET modalities and with two patients compared to WBMR. Conclusion: FDG-PET/CT is signiﬁcantly more accurate than CT alone and sideby-side CT+PET in tumor staging. WBMR is an effective and fast method for staging cancer patients but in our experience cannot reach accuracy of FDG-PET/CT.

B-622

10:39

Detecting bone metastases in melanoma and lung cancer: Whole-body MRI or whole-body FDG-PET/CT? T. Heusner, P. Gölitz, F. Vogt, J. Barkhausen, A. Bockisch, G. Antoch; Essen/DE ([email protected])

10:48

Whole body diffusion weighted MR imaging and positron emission tomography for tumor staging in oncology S. Li, H.D. Xue, F. Sun, Z.Y. Jin; Beijing/CN ([email protected]) Purpose: To compare the accuracy in staging of various malignant tumors with whole-body diffusion-weighted MR imaging (WB-DWI) and positron emission tomography (PET). Methods and Materials: Sixteen patients (nine males, seven females, age 35-82 years) with various oncological diseases underwent tumor staging by WB-DWI and PET within three weeks. WB-DWI was performed by using short TI inversion recovery diffusion-weighted EPI sequence with free breathing and build-in body coil. Correct classiﬁcation of the primary tumor, regional lymph nodes and distant metastasis (overall TNM stage) were assessed for both modalities. The sensitivity, speciﬁcity, positive predictive value, negative predictive value, accuracy and the accordance of both modalities were evaluated with histological results and other radiological follow-up within six months as the standard of reference. Results: WB-DWI identiﬁed 80 pathological lesions in 11 cases, while PET demonstrated 77 lesions in 13 cases. Concordance between WB-DWI and PET occurred in 12 of 16 patients (75%). There was no statistically signiﬁcant difference between the two modalities (p=0.437). The sensitivity, speciﬁcity, positive predictive value, negative predictive value, accuracy of PET were 91.67 (11/12), 50 (2/4), 84.6 (11/13), 66.7 (2/3), 81.3% (13/16), those of WB-DWI were 91.67 (11/12), 100 (4/4), 100 (11/11), 80 (4/5) and 93.75% (15/16). One prostate cancer was missed by PET, while WB-DWI missed one residual lesion in one NHL case after chemotherapy. Conclusion: WB-DWI and PET are reliable imaging modalities for tumor staging. Superior performance of overall TNM staging suggests WB-DWI is a valuable technique for whole-body tumor screening.

B-624

10:57

Comparison of DCE-MRI, [18F]galacto-RGD- and [18F]FDG-PET in metastatic cancer patients S. Metz, C. Ganter, S. Lorenzen, G. Brix, M. Schwaiger, E.J. Rummeny, A.J. Beer; Munich/DE ([email protected]) Purpose: Dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) (microcirculation characteristics) and positron emission tomography (PET) using [18F]Galacto-RGD (AvB3-expression) or [18F]FDG (tumor glucose metabolism) provide functional and molecular information in addition to tumor morphology, which might be useful to monitor the response to anti-angiogenic therapies. We therefore evaluated the relationship of these three imaging modalities. Methods and Material: Twelve patients with primary or metastasized cancer (NSCLC n=9; renal cell carcinoma n=1; rectal cancer n=1; bronchus carcinoid n=1) were examined with DCE-MRI using Gd-DTPA and also with PET using [18F] Galacto-RGD and [18F]FDG. From the DCE-MRI data, the IAUGC, the relative blood volume (rBV) and the relative blood ﬂow (rBF) were estimated using an open one-compartment model and compared with standardized-uptake-values (SUVs) from PET, taken from the same tumor regions. To this end, a linear regression analysis was performed. Results: Comparing the functional MRI parameters with the molecular PET-data, a signiﬁcant (p 0.05) correlation between the IAUGC values and the SUVs for [18F] Galacto-RGD could be demonstrated. All other linear regression analyses revealed no signiﬁcant correlation. However, a positive trend could be observed for [18F]FDG versus rBF and for [18F]Galacto-RGD versus rBV. Note, there was no correlation for [18F]FDG compared to [18F]Galacto-RGD. Conclusion: A signiﬁcant correlation between IAUGC as a functional parameter from DCE-MRI and [18F]Galacto-RGD-PET could be demonstrated. These preliminary data suggest that both DCE-MRI and [18F]Galacto-RGD might be useful for evaluation of anti-angiogenic therapies, whereas [18F]Galacto-RGD-PET and [18F] FDG provide complementary information.

B-625

11:06

The clinical value of 18 F-FDG-PET/CT and mediastinoscopy in initial staging in non-small cell lung cancer D. Putzer, D. Kendler, C. Uprimny, R. Madleitner, R. Bale, P. Kovacs, T. Schmid, I.J. Virgolini; Innsbruck/AT ([email protected]) Purpose: Initial staging in patients with non-small cell lung cancer (NSCLC) is decisive for determining the most effective therapeutic management. The aim of this study was to assess the value of 18F-FDG-PET/CT in comparison to mediastinoscopy in lymph node staging in patients with NSCLC. Methods and Materials: Twenty-eight patients with NSCLC (62p10 years; 17

A

B

C

D

E

F

G

285

Monday

Purpose: To assess the accuracy of whole-body FDG-PET/CT and whole-body MRI for the detection of bone metastases at initial diagnosis in patients with malignant melanoma (MM) and non-small cell lung cancer (NSCLC). Methods and Materials: 109 patients with malignant melanoma (n=55) and nonsmall cell lung cancer (n=54) underwent whole-body FDG-PET/CT and whole-body MRI (non-enhanced T1 and T2 as well as contrast-enhanced T1). Both imaging procedures were performed within a mean time interval of 0.6 days (0-14 days). Histopathology and a mean clinical follow-up of 434 days (42-1292 days) served as the standards of reference. Differences between the imaging procedures were tested for signiﬁcance by McNemar’s test (p 0.05). Results: Based on the standard of reference three patients with malignant melanoma and eight patients with NSCLC had bone metastases. Whole-body PET/CT correctly detected bone metastases in ﬁve patients (MM/NSCLC: 0/5), whole-body MRI in seven patients (MM/NSCLC: 2/5). MRI was false-positive in six patients, FDG-PET/CT was false-positive in one patient. Sensitivities, speciﬁcities, positive and negative predictive values, and accuracies for detection of bone metastases were 0.45/0.99/0.83/0.94/0.94 with PET/CT and 0.64/0.94/0.54/0.96/0.91 with MRI. The differences between the two imaging modalities were not statistically signiﬁcant (p 0.05). Conclusion: Whole-body FDG-PET/CT and whole-body MRI are similarly accurate when assessing patients with initial diagnosis of malignant melanoma and non-small cell lung cancer for bone metastases. Higher sensitivities of MRI go along with an increased rate of false-positive ﬁndings compared to FDG-PET/CT.

B-623

Scientific Sessions male, 11 female) underwent PET/CT followed by mediastinoscopy for initial staging. 18 F-FDG PET scans were acquired 60 minutes after i.v.-application of 370 MBq of 18F-FDG. Image fusion with CT was performed in all patients, using a vacuumbased ﬁxation for optimal image alignment. Mediastinoscopy was done within ﬁve days after imaging procedures. Results: Twelve patients (43%) showed lymph node metastases on PET/CT which was proven by histopathology. One patient had a secondary cancer diagnosed by PET/CT, and in six patients image fusion led to the diagnosis of previously unknown distant metastases. Eleven patients (39%) had enhanced uptake suspicious for lymph node metastases on PET/CT, while mediastinoscopy revealed inﬂammatory disease. In one patient, distant metastasis (suprarenal gland) was diagnosed by PET/CT. Five patients (18%) had negative ﬁndings on PET/CT and histopathology. Conclusion: PET/CT is effective for accurate staging of lymph node involvement and early diagnosis of distant metastases and therefore reduces the number of non-curative thoracotomy in patients with NSCLC. If image fusion ﬁndings are negative, diagnostic mediastinoscopy could be omitted and surgery could be performed immediately. In case of non concordant PET/CT ﬁndings, veriﬁcation by mediastinoscopy leads to clear diagnostic strategy.

B-626

11:15

11C-metomidate-PET, CT and MRI for characterisation of adrenal incidentalomas A.E. Sundin, J. Hennings, H. Ahlström, P. Hellman; Uppsala/SE ([email protected]) Purpose: To evaluate the clinical use and value of 11C-metomidate-PET (MTOPET) compared to CT and MRI in the imaging work-up and characterisation of adrenal incidentalomas. Methods and Materials: First, a retrospective evaluation was performed of 20 adrenal incidentalomas in patients who had undergone CT, MRI and MTO-PET and from whom we had either an histopathological diagnosis or clinical follow-up data. Secondly, a prospective study was conducted in which 24 adrenal incidentalomas were imaged by CT, MRI and MTO-PET and the results were correlated to those from histopathology (n=8) and clinical diagnosis after follow-up (n=16). Results: In the retrospective analysis, MRI, and especially MTO-PET, correlated well with histopathology and clinical diagnosis after follow-up, whereas the speciﬁcity with CT was low. This was most likely explained by the presence of ﬁve haematomas with ﬁbrosis which, because of their low attenuation and contrast-enhancement, were misdiagnosed. In the prospective cohort, sensitivity and speciﬁcity with CT were 0.71 and 1.0, respectively. Further characterisation by MRI increased these values to 0.86 and 1.0 and maximum sensitivity and speciﬁcity (1.0 and 1.0) were reached when MTO-PET was added. Conclusion: The diagnosis of an adrenocortical adenoma may be established by CT in most patients and by MRI in an additional number of subjects. For the few remaining patients needing further characterisation, MTO-PET is advantageous as an additional imaging modality. The CT appearance of adrenal haematomas with ﬁbrosis is an interesting ﬁnding in this study.

B-627

11:24

FDG-PET/CT for detection and therapy monitoring of granulocytic sarcoma (chloroma) P. Aschoff, M. Häntschel, M.P. Lichy, M.Ö. Öksüz, C.D. Claussen, W. Vogel, C. Pfannenberg; Tübingen/DE ([email protected]) Purpose: The rare extramedullary manifestations of myeloid leukemia are called granulocytic sarcoma or chloroma. Diagnosis and therapy monitoring with morphology-based imaging methods like CT is challenging. We present ﬁrst results of combined functional and morphological imaging with FDG-PET/CT for diagnosis and therapy monitoring of chloromas. Methods and Materials: Retrospective analysis of 11 FDG-PET/CT examinations of ﬁve patients with chloromas (two patients with AML, three patients with CML). Correlation with histology, further imaging and clinical follow-up. Results: Before treatment all 40 known chloromas were FDG-avid (SUV avg. 0.8-6.4, mean 3.3, SD 1.3) and visually detectable in FDG-PET. In four of ﬁve patients additional manifestations were found. These lesions had no morphologic correlation in CT, were previously interpreted as benign (lymph nodes, sinusitis) or were not included in the ﬁeld of view of prior examinations. In one patient the detected additional manifestations changed the therapy from local irradiation to systemic chemotherapy. FDG-PET/CT after stem cell transplantation showed a metabolic response in 2/3 patients and a stable disease in one patient consistent with the clinical course and a survival of 25 to 34 weeks up to now. In the follow-up FDG-PET/CT detected a relapse in three patients proven by histology or the further clinical course of the disease.

286

A

B

C

D

E

F

G

Conclusion: According to our ﬁrst results FDG-PET/CT is a valuable diagnostic tool for staging, therapy monitoring and follow-up of patients with chloroma and has an inﬂuence on patient management.

B-628

11:33

Detection of brain metastasis by routine contrast-enhanced PET-CT of the head J.B. McCain, G.D. Marano, J.S. Carpenter; Morgantown, WV/US ([email protected]) Purpose: The purpose of this study was to evaluate the clinical usefulness of inclusion of a dedicated brain study as part of a routine protocol in PET-CT in the oncology patient. Our initial experience in the detection of brain malignancy in patients referred for oncology workup is presented. Methods and Materials: At our institution a dedicated acquisition of the brain is part of the routine protocol for all patients undergoing contrast-enhanced PET-CT. A review of patient records yielded 206 patients that carried the diagnosis of brain malignancy out of 2985 that had undergone contrasted PET-CT exams during a two-year period. Recorded parameters were the presence or absence of ﬁndings consistent with brain malignancy on PET-CT, CT and PET alone, and gadoliniumenhanced MRI. Sensitivity, speciﬁcity, and negative and positive predictive values were calculated for each modality. Results: PET scan alone had a low sensitivity for detecting brain tumors (83%), compared to those patients who had contrasted MRI (94%). However, the addition of fused contrast-enhanced CT scans to PET images increases the sensitivity of tumor detection by identifying abnormal enhancement in isometabolic or hypometabolic masses. In our series, PET-CT had a sensitivity of 94%, comparable to those patients who also underwent MRI. Conclusion: The addition of a dedicated contrast-enhanced PET-CT of the brain as part of the routine protocol for oncology patients appears useful in detecting brain malignancy due to its high sensitivity, short scan time, low additional radiation dose, and the potential for early discovery and treatment of a brain metastasis.

B-629

11:42

Residual lymphoma masses in PET/CT in patients with complete metabolic remission: An early prognostic factor for relapse? P. Veit-Haibach, K. Strobel, J. Soyka, N. Schäfer, M. Pérez Lago, K. Mende, T.F. Hany; Zurich/CH ([email protected]) Purpose: To assess the prognostic value of residual morphologic masses (RMM) after treatment of lymphoma on PET/CT in patients with complete metabolic remission. Methods and Materials: 128 patients (mean age: 40 years) with lymphoma (48 Hodgkin’s lymphoma (HD), 68 NHL, eight Burkitt lymphoma (BL)) were monitored with PET/CT for therapy response. All patients had at least two PET/CT scans. PET/ CTs after 2-4 cycles of chemotherapy (interimPET/CT) were evaluated concerning morphological and metabolic response based on newly developed response criteria for PET/CT (International Harmonization Project in Lymphoma). Patients with RMM but complete or partial metabolic response were evaluated for early relapse of lymphoma during clinical course. The mean follow-up time was 21p10 months. Results: 24/128 patients (19%) showed RMM on CT on the interimPET/CT (10 patients with HD, 14 with NHL). 21 patients had a complete metabolic response, three patients had partial metabolic remission. 4/24 patients died during the clinical follow-up. One patient died of a secondary carcinoma, two patients with partial metabolic remission on interimPET/CT died of lymphoma recurrence. Only one with RMM and complete metabolic response at the time point of interimPET/CT showed lymphoma recurrence. Conclusion: RMM on interimPET/CT and complete metabolic response are not a reliable prognostic factor for early recurrence of lymphoma. Thus, complete metabolic response, even with RMM on interimPET/CT, may be evaluated as an overall complete response. Longer follow-up is needed to deﬁne the predictive and prognostic value of those residual masses concerning the overall rate of recurrence in lymphoma patients.

B-630

11:51

Detection of macrophage activity in infected areas by USPIO-enhanced MRI in an experimental vertebral osteomyelitis model G. Bierry, F. Jehl, N. Boehm, G. Prévost, J.-L. Dietemann, S. Kremer; Strasbourg/FR ([email protected]) Purpose: To evaluate the ability of macrophage USPIO MR imaging to detect infected areas in an experimental vertebral osteomyelitis model.

Scientific Sessions Methods and Materials: Lumbar vertebral osteomyelitis was induced in 10 rabbits by intradiscal injection of bacteria at test level opposed to saline serum injection at control levels in another disc. After a mean interval of 12 days, rabbits were imaged prior to and 24 hours after administration of USPIO (Sinerem, Guerbet, France) at a dose of 45 μmol/kg. Imaging protocol consisted of sagittal T1-w SE, T2-w SE and T2*-w GRE sequences. Images were correlated with histopathologic ﬁndings (HE staining, macrophage immunostaining and Perls blue staining). Results: Twenty-four hours after USPIO administration, T1-w images of infected test levels showed a signiﬁcant increase of signal (p=0.005) in the vertebrae, whereas T2-w and T2*-w images showed no signiﬁcant changes (p=0.14 and p=0.87, respectively). Histopathologic examination of infected areas demonstrated replacement of haematopoietic bone marrow by inﬁltration of macrophages. Perls blue staining showed that some macrophages were iron-loaded. At control level, T1 (p=0.02), T2 (p=0.04) and T2* weighted sequences (p=0.04) of the vertebrae showed a signiﬁcant decrease of signal. Histopathologic examination conﬁrmed the persistence of normal bone marrow, reﬂected by a more intense Perls blue staining compared to the infected areas. Conclusion: MR imaging can identify USPIO loaded macrophage inﬁltration present in infected areas in an experimental vertebral osteomyelitis model. Infected areas present signiﬁcant T1 signal increase not present in non-infected areas correlated to the presence of iron-loaded macrophages.

10:30 - 12:00

Room I

Physics in Radiology

SS 1413 New advances in imaging techniques and applications Moderators: T.G. Maris; Iraklion/GR W.J.M. van der Putten; Galway/IE

B-631

10:30

Phase-contrast radiography in a clinical setting, using high-resolution digital imaging system with a small-focus X-ray tube instead of synchrotron radiation: I. A prerequisite study for its clinical application J. Tanaka1, H. Oda1, T. Mimura1, H. Ohara2, C. Honda2, Y. Wada1, H. Kawasaki1; 1 Iruma-gun/JP, 2Hachioji/JP ([email protected])

10:39

Image- and rawdata-based metal artifact reduction methods for ﬂatdetector CT D. Prell, Y. Kyriakou, M. Knaup, W. Kalender; Erlangen/DE ([email protected]) Purpose: To provide and compare metal artifact reduction algorithms for ﬂatdetector CT (FD-CT). Methods and Materials: Metal implants produce non-linear artifacts which often disturb image quality. We present two methods for metal artifact reduction in FD-CT for cone-beam geometry. (1) An image-based method independent of the system geometry performs a segmentation of the metal in the reconstructed images. Reprojection in parallel beam geometry provides artiﬁcial sinogram data where the metal is replaced by linear interpolation. Filtered backprojection reconstructs the corrected data. (2) A rawdata-based method which uses 3D segmentation of the metal in the reconstructed images. The sub-volumes are backprojected onto the detector identifying the 2D metal and replacing the metal by 2D linear interpolation. Subsequently standard Feldkamp reconstruction is performed. Performance was evaluated by simulations and measurements. Measurements were performed on a C-arm FD-CT system (Axiom Artis, Siemens Medical, Forchheim, Germany). Data was acquired using tissue-equivalent phantoms (QRM GmbH, Möhrendorf, Germany) with metallic inserts. The accuracy of correction was investigated in subtraction images. Results: Both methods efﬁciently reduced metal artifacts in the reconstructed images. The image-based method is scanner independent but slower (approximately twice the reconstruction time). The rawdata-based method is faster (5% of total reconstruction time) and allowed a more ﬂexible selection of correction methods. Generally better image quality was achieved using the rawdata-based method. Conclusion: Both methods improve image quality of FD-CT reconstructions containing metal artifacts at acceptable computational costs.

B-633

10:48

High resolution phase contrast microtomography of human cartilage P. Coan1, A. Bravin1, E. Mützel2, D. Habs3, M. Reiser2, C. Glaser2; 1Grenoble/FR, 2 Munich/DE, 3Garching/DE ([email protected]) Purpose: In view of osteoarthritis (OA) comprehensive assessment of both bone and cartilage with a single imaging technique would be highly desirable. The depiction of ﬁne structural detail and damage within the cartilage matrix considered to be an early precursor of OA implies high resolution cross sectional analysis. Phase contrast (PhC) imaging techniques, in particular analyzer-based (ABI) and propagation-based (PBI) imaging, have been shown to provide dramatic gains in contrast over conventional radiography. Therefore our goal was to assess whether high resolution phase contrast imaging CT is feasible in cartilage and whether the technique is able to depict structural properties of the intact cartilage matrix. Methods and Materials: Eight healthy cartilage on bone cylinders from three human patellae have been imaged by computed tomographic ABI and PBI mode at high resolution (35 keV, digital Frelon camera, voxel size: 8³ micron³ and 46³ micron³) and clinically compatible doses ( 10 mGy). Results: Phase contrast CT showed excellent depiction of the 3D architecture of subchondral bone in both ABI and PBI modes. In PBI a zonal pattern within the cartilage matrix could be visualized. In the reconstructed data PBI appeared to show architectural properties of the cartilage matrix better than ABI. Conclusion: High resolution PhC microtomography as one single imaging approach is able to reveal structural detail in both subchondral bone and cartilage suggesting a high potential of the technique for future diagnostic workup of OA.

B-634

10:57

Development of a software package for radiation dose calculation outside therapeutic photon beams M. Mazonakis, F. Zacharopoulou, J. Damilakis, N. Gourtsoyiannis; Iraklion/GR ([email protected]) Purpose: Peripheral dose (PD) received by radiosensitive structures located outside the boundaries of the primary radiation ﬁeld is of clinical importance. The aim of this study was to develop a software program for PD calculation during therapeutic external irradiation. Methods and Materials: Dose measurements were performed using an ionization chamber embedded in a water phantom for 6 and 18 MV photon beams. PD values were determined for ﬁeld sizes from 5×5 to 20×20 cm2 in increments of 1 cm at all possible distances from the ﬁeld edge and collimator orientations. Dose data were obtained to deﬁne the variation of PD with the ﬁeld shape, ﬁeld elongation, distance from X-ray source, wedge introduction, depth and off-axis displacement of

A

B

C

D

E

F

G

287

Monday

Purpose: To evaluate the image sharpness of phase-contrast radiography, which uses a high-resolution digital imaging system with a small-focus X-ray tube instead of synchrotron radiation. Methods and Materials: The system was created to achieve an edge-enhancement effect to sharpen X-ray images. The system consists of a tungsten-anode X-ray tube with a 123 μm focal spot and a computed radiography (CR) plate (sampling pitch 43.75 μm): the distances between the focal spot and object, and between the object and CR plate are 0.65 and 0.49 m, respectively, and therefore images are magniﬁed. The distance between the object and CR plate is required to capture the edge-enhancement effect of narrow width, due to a phase-shift phenomenon (refraction). Images were obtained using a bone-mineral phantom (voltage range 30-50 kVp). Conventional digital X-ray images were also obtained with a CR plate in geometrical contact with the objects. Results: In an X-ray-intensity analysis of images obtained by phase-contrast technique, black-white contrast at the object edges was observed, and corresponded to additional contrast over that expected for a normal technique. This was not observed on contact radiographs. Conclusion: Phase-contrast radiography can be conducted using small-focus Xray tube and a high-resolution CR system under appropriate conditions instead of synchrotron radiation. This system gave radiographs of signiﬁcantly high resolution for clinical purposes, such as for the early diagnosis of rheumatoid arthritis.

B-632

Scientific Sessions the measurement point. All measured dose values were compiled into a database. A PC software program running under Windows was written in Labview 7.1 and Matlab 6.5 to compute PD utilizing the out-of-beam dosimetric database. Software calculations were veriﬁed by measurements obtained using an anthropomorphic phantom and thermoluniscence dosimetry (TLD). Results: The software was designed to be simple and user-friendly. The treatment parameters, tumor dose and distance between the point of interest and ﬁeld edge are easily entered on the input screen. The PD is computed in a few seconds and displayed on the output screen with a summary of input data. The mean difference between software calculations and TLD measurements was 12.1%. Conclusion: The developed software program allows the direct and quick PD calculation from therapeutic irradiation with sufﬁcient accuracy.

B-635

11:06

Single-source dual-energy MDCT: Effect on density and contrast volume in an experimental CTA model J. Sosna1, S. Amin Spector2, D. Planer1, A. Altman2, E. Libson1; 1Jerusalem/IL, 2 Haifa/IL ([email protected]) Purpose: Dual-energy CT may be obtained with a single X-ray source with a dual detection system. Images based on low energy photons reﬂect increased attenuation of iodinated contrast material. Our purpose was to measure the degree of contrast attenuation in an experimental rabbit CTA. Methods and Materials: The study was performed on a prototype dual-energy MDCT (Philips Medical Systems). CTA studies were performed on 18 NZ rabbits using 1 mm slice thickness, 0.5 mm increment, 140 kVp, 250 mAs, and 7 cc (1 cc/ sec) of nonionic contrast (300 mgI/ml). In three other rabbits, increasing doses of contrast (3, 5, 7 cc) were injected at 10-min intervals. Three regions-of-interest were drawn on the aorta on low energy and regular CT images. Density differences (HU) and percent change for each of the measured regions were calculated. Results: Average abdominal aorta density in 18 rabbits was 498p28 HU. Density increased to 806p36 HU in low energy images, with a mean difference in attenuation of 306 HU (average density increase of 61.4%). In the other rabbits, densities with 3, 5, and 7 cc of contrast were 228, 324, 482 HU in the clinical image and 320, 471, 720 HU in the low energy images. Low energy density at 5 cc was similar to that of the regular image with 7 cc of contrast (P=NS), enabling a 28.5% reduction contrast volume. Conclusion: Signiﬁcant iodine density augmentation can be obtained using the low energy layer versus the clinical image. A reduction in contrast volume may be feasible with similar densities at CTA studies.

B-636

11:15

Musculoskeletal imaging by a prototype of a digital imaging X-ray machine based on direct capture of X-ray photons with pixel detectors coupled to photon counting readout electronics M. Gruber1, P. Homolka1, M. Tortajada2, R. Martinez2, M. Pretterklieber1, F. Kainberger1; 1Vienna/AT, 2Barcelona/ES ([email protected]) Purpose: The aim of the project is to develop a digital imaging X-ray machine based on direct capture of X-ray photons with room temperature solid-state pixel detectors coupled with photon counting readout electronics, Medipix-2 chip. The direct capture approach has a much higher SNR compared to a system where photons are captured in scintillating crystal coupled to Amorphous-Si. Methods and Materials: The Medipix-2 chip is a matrix of 256x256 pixels with a pixel pitch of 55 μm. Thus the effective area of one chip assembly (a Medipix-2 chip bump-bonded to CdTe) is 14.08x14.08 mm2. A large monolithic image (11.2x7 cm2) was obtained by consecutive displacement approach. We took images of eight human cadaver hands at four different dose levels (0.25, 0.64, 1.25, 2.5 mAs) at 44 kV tube voltage with the new detector and also with a ﬂat-panel detector system (Philips Digital Diagnost, Philips, Hamburg, Germany). Results: The new detector delivered a subjectively better image quality especially in ultra low dose levels 0.25 or 0.64 mAs compared to the ﬂat-panel detector system. However, we noticed some punctiform (from 55 μm to 5 mm) artifacts of different origin. Conclusion: The prototype of a X-ray machine based on direct capture of Xray photons with solid-state pixel detectors coupled to photon counting readout electronics has big potential to obtain excellent image quality in musculoskeletal images at very low dose levels due to a much higher SNR compared to established detector systems.

288

A

B

C

D

E

F

G

B-637

11:24

CT perfusion evaluation: Accuracy of physiological parameters in tumors determined with different pharmacologic models V. Hietschold, P. Gohl, A. Abramyuk, A. Koch, M. Laniado, N.D. Abolmaali; Dresden/DE ([email protected]) Purpose: Comparison of four two-compartment models with respect to bias and reliability by means of computer simulations. Methods and Materials: The Lee-Harvey, generalized Kety, Johnson-Wilson and Patlack approaches were applied to six pairs of contrast medium concentration curves and Arterial Input Functions after adding Gaussian noise (S=1, 3, 6 HU). Results: Optimum lengths of data acquisition times are about 40 s (Patlack), 60 s (Lee-Harvey), 140-180 s (Kety) and 600 s (Johnson-Wilson). KTrans: Most stable results - though biased by down to -50% - were obtained with the Patlack approach. Using the Kety model, the mean results tend to be more correct at the cost of large stochastic errors. The Lee-Harvey model appeared not to be suited for determination of KTrans due to stochastic errors regularly 100%. rBV: Both the Patlack and the generalized Kety models allow for good differentiation between high and low rBVs. ve: With both the Lee-Harvey and Kety models, differentiation between high and low interstitial volumes is possible only at noise levels about 1 HU. E and ve determined using the Johnson-Wilson model scatter by 50% or more even with given KTrans and rBV. Conclusion: In routine imaging, sufﬁciently stable perfusion parameters can be obtained using models with b3 parameters (plus time shift of the AIF). At acceptable levels of stochastic errors, the Kety model gives the least biased KTrans and rBV results, whereas the Patlack approach distinguishes itself by best reproducibility at signiﬁcantly biased KTrans results.

B-638

11:33

Correlation based noise reduction for CT data R. Raupach, A. Borsdorf, B. Schmidt, T.G. Flohr; Forchheim/DE ([email protected]) Purpose: To clinically exploit the advantages of routinely offered isotropic submillimeter resolution of present MSCT systems at a sensible dose level noise reduction. Methods and Materials: We present a novel edge preserving noise reduction based on the evaluation of correlations. Two spatially identical sets of images or volumes with uncorrelated noise between them served as input, denoted as A and B in the following. They were generated by reconstructions of even and odd projections, separately, or using a dual source CT system which generically provides two independent scans of the same object or patient. Both data, A and B, were transformed into wavelet domain where different levels represented different frequency bands. Due to the construction, “information” was revealed by the correlation of corresponding wavelet coefﬁcients while noise reduces correlation between A and B. The degree of correlation was used to weight the respective coefﬁcients. Final images with reduced noise were obtained by a subsequent inverse wavelet transformation. The approach was applied to simulated data, phantom and clinical CT data. Results: The presented approach allows for a noise reduction of up to 40% without affecting the delineation of small structures or edges. Therefore, sharpness to noise ratio can be improved to the same degree. For clinical data, dose could be reduced by up to 30% without loss of diagnostically relevant information. Conclusion: The developed algorithm represents an efﬁcient noise reduction system with a high potential for dose reduction in CT.

B-639

11:42

Ratio dosimetry: A practical method for the assessment of radiation exposure to medical staff during interventional procedures F. de Lange, W.P. Moerman, K.J. Renema, L.J. Oostveen, P.A.J. Jonkergouw, H. van Langen, L.J. Schultze Kool, R.L. Kamman; Nijmegen/NL ([email protected]) Purpose: To develop methodology allowing the assessment of radiologist eye and hand dose levels based on regular dosimeter badge readings and dose recordings of an interventional procedure expected to yield the highest occupational dose levels. Methods and Materials: Dose levels were recorded using thermoluminescent dosimeter (TLD) chips placed on both hands, on both sides of the radiation protection glasses, and on the dosimeter badge during 15 ethanol embolization procedures of arteriovenous malformations. All procedures were performed by the same right-handed radiologist. To obtain “worst-case” measures for the occupational dose to the eyes and hands, relative dose levels were calculated from

Scientific Sessions the ratio of the TLD dose readings at the designated locations and that obtained at the location of the dosimeter badge (ratio dosimetry). The dosimeter badge was attached externally to the lead apron at chest level. Dose-area products (DAP) were recorded simultaneously. Results: The range in recorded TLD dose levels was broad. Also, the mean dose values were threefold larger than the median dose values. In contrast, using ratio dosimetry the range was about tenfold smaller while the mean and median relative doses agreed within 25%. Median relative dose values [range] for the right and left eyes and hands were 0.65 [0.41-2.03], 1.23 [0.04-2.42], 2.06 [0.44-4.02] and 3.25 [0.31-7.66], respectively. The radiologist’s DAP-normalized dose was 6.2 [0.3-49] μSv/(Gy cm2). Conclusion: We obtained badge-to-hand and eye dose conversion factors allowing a practical and more accurate assessment of the upper value of occupational dose based on regular dosimeter badge recordings.

age quality (sharpness of vessel wall, signal homogeneity, artefacts), delineation of aortic root and supra-aortic vessels. Results: Non-CE-MRA failed in 8/58 due to irregular diaphragmatic motion while CE-MRA was successful in all volunteers. Imaging time was 14p5 min vs 22p2 s. Overall image quality was rated excellent in 58%, moderate in 34%, poor in 8% for non-CE-MRA; 26, 69, 5, respectively, for CE-MRA. The aortic root was rated excellent in (non-CE-MRA/CE-MRA) 86/15%, moderate 14/30% and poor 0/13%. Supra-aortic vessels were rated (non-CE-MRA/CE-MRA) excellent 45/47%, moderate 30/49%, poor 13/5%. 12% of supra-aortic vessels at non-CE-MRA were visualized 1 cm. Only minor artefacts occurred with both sequences. Conclusion: Diagnostic imaging of the thoracic aorta without application of contrast media is feasible. Image quality with non-CE-MRA is superior to CE-MRA for aortic root and equal for rest of thoracic aorta. Drawback is a long acquisition time and requirement of a regular breathing pattern.

B-640

B-642

11:51

Motion-compensated reconstruction for rotational X-ray coronary angiography D. Schaefer1, U. Jandt1, B. Movassaghi2, J. Garcia3, J.D. Carroll3, M. Grass1; 1 Hamburg/DE, 2Bothell, WA/US, 3Denver, CO/US ([email protected]) Purpose: To enable in-room automated coronary artery 3-D reconstructions that are made from easily acquired selective X-ray angiograms. Methods and Materials: For each cardiac phase, 3D coronary centerline models are calculated from nearest neighbour gated projections. The motion ﬁeld is deﬁned by the correspondences introduced between centerline points by correlation measures throughout the whole cardiac cycle. A ﬁltered back-projection algorithm enabling motion-compensation is used to reconstruct 4D sequences of the coronary arteries. The method is applied to a beating cardiac phantom with contrast agent ﬁlled vessels and clinical data (N=12 patients) acquired with a frame rate of 15 Hz or 30 Hz along a circular arc with 180° angular coverage in 7 seconds. Volumetric 3D and 4D data sets are reconstructed at voxel sizes of 0.35 mm3 and displayed as volume rendered images or maximum intensity projections. Results: The diameters of the phantom vessels vary between 1.5 (stenosis) and 4.4 mm, corresponding to lumen of 1.8 and 15 mm2. The relative error of the reconstructed lumen is 9%. Excellent image quality for the main vessels (LM, LAD, Cx, RCA) can be achieved in 83% (N=10). Decreased quality in two cases is mainly due to strong respiratory motion. Conclusion: Excellent image quality is obtained and signal-to-noise ratio is improved compared to standard gating techniques. These results can be achieved not only for late diastole but especially also for systolic phases of strong motion. The presented method is a major step towards quantitative analysis of the coronary artery tree and can now be validated and evaluated in the clinical arena.

10:30 - 12:00

Room K

Vascular

MR: Angiography, microcirculation and lymphography Moderators: F. De Cobelli; Milan/IT R.W. Günther; Aachen/DE

B-641

10:30

Assessment of the thoracic aorta: Intraindividual comparison between contrast-enhanced and non-contrast-enhanced ECG-triggered MR angiography H. von Tengg-Kobligk, J. Ley-Zaporozhan, V. Henninger, K. Ruf, H.-U. Kauczor, S. Ley; Heidelberg/DE ([email protected]) Purpose: To avoid contrast media application for MRA, new sequences using inherent blood contrast are available. The clinical potential of these non-contrast-enhanced MRA (non-CE-MRA) sequences are not yet evaluated for the aorta. Thus, the goal was to compare a standard CE-MRA with a non-CE ECG-triggered MRA for the thoracic aorta. Advantages and disadvantages of each method were assessed. Methods and Material: CE-MRA and non-CE-MRA were performed in the same 58 healthy volunteers (mean age: 50 years, range: 18-76 years) at a 1.5 T clinical scanner. CE-MRA: GRE Turbo-Flash 3D (1.2x1.2x1.6 mm³), 0.15 mmol Gd/kg, 3 ml/s. Non-CE-MRA: Respiratory- and cardiac-gated, T2-prepared 3D-trueFISP (1.2x1.2x1.3 mm³). Assessment included (three readers, consensus): overall im-

Purpose: To evaluate contrast-enhanced MR angiography (CE-MRA) as a screening procedure for detection of pulmonary AV-malformations (PAVM) in patients with hereditary hemorrhagic telangiectasia (HTT). Methods and Materials: 203 consecutive patients (87 men, mean age 47.4p17.6 years; 116 women, 46.1p15 years) with conﬁrmed HHT according to Curaçao criteria underwent screening pulmonary CE-MRA with gadobenate dimeglumine (0.1 mmol/kg bodyweight) for the presence of PAVMs. PAVM presence was scored as 0 (=none present), 1 (=deﬁnitely present) or 2 (=uncertain) and was evaluated by patient gender, age (0-24; 25-65; 65 years) and PAVM size (0-4; 5-10; 11-15; 16-20; 20 mm). Patients scored as 1 or 2 with at least one PAVM of r5 mm were referred for conventional pulmonary angiography (PA) for possible embolization. Results: Overall 56/203 (27.6%) patients were scored as 1 (deﬁnite) and 1/203 (0.5%) as 2 (uncertain) for PAVM presence on CE-MRA. Overall, CE-MRA detected 156 PAVMs in these 57 patients (65 in 27 men; 91 in 30 women of which 82 were in 25 women of 26-65 years). Two or three PAVMs were detected in 21 patients and four or more in 11 patients. Most PAVMs detected on CE-MRA were small ( 5 mm: n=49; 5-10 mm, n=61). 40 of 57 patients with 119 evaluable PAVMs detected on CE-MRA underwent global or selective PA. Signiﬁcantly (p 0.001) fewer pAVMs (92/119 [77.3%]) were demonstrated on PA of which 82 were embolized. Conclusion: CE-MRA is a suitable screening procedure for patients with HHT, permitting accurate detection and staging of PAVMs and appropriate differentiation of lesions requiring embolization.

B-643

10:48

Diagnostic accuracy of high-spatial resolution renal MRA at 3.0 T H.J. Michaely1, U.I. Attenberger2, H. Kramer2, K.-P. Lodemann3, M.F. Reiser2, S.O. Schoenberg1; 1Mannheim/DE, 2Munich/DE, 3Konstanz/DE ([email protected]) Purpose: To assess the diagnostic accuracy of high-spatial resolution renal MRA at 3 T. Methods and Materials: After IRB approval, 30 consecutive patients with suspected renal artery disease (suspected renal artery stenosis, renal transplant dysfunction) were examined with a high-spatial resolution MRA on a 3.0 T MR-scanner (Siemens Tim Trio) after the bolus injection of 0.1 mmol/kg BW gadobenate-dimeglumine at 2.5 ml/s. An arterial-phase fast 3D-FLASH sequence with a spatial resolution of 0.9x0.8x0.9 mm³ was acquired in 18 s using parallel imaging GRAPPA factor 3 (TR/ TE/FA - 3.11 ms/1.09 ms/23°, 96 partitions, slice thickness 0.9 mm, oblique coronal slab orientation). The MRA data were read by two radiologists in consensus using the source data and thin MIPs. The clinical performance of renal MRA was compared to the ﬁnal clinical diagnosis based on the clinical course of the patient, ultrasound and scintigraphy as well as digital subtraction angiography results and operation reports. Results: The 3.0 T renal MRA could be performed successfully in all patients. The ﬁndings included renal artery stenoses, iliac artery dissections in renal transplant recipients, renal transplant artery kinking and ﬁbromuscular dysplasia. One case of vasculitis was not diagnosed correctly. The overall sensitivity and speciﬁcity of 3.0 T renal MRA were 93 and 100%, respectively, the positive and negative predictive value were 100 and 94%, respectively. Conclusion: The high spatial resolution of 3 T renal MRA allows for exact diagnosis of a variety of renal artery lesions with a contrast dose of only 0.1 mmol/kg BW gadobenate-dimeglumine.

A

B

C

D

E

F

G

289

Monday

SS 1415

10:39

Contrast enhanced MR angiography for detection of pulmonary arterio-venous malformations in patients with hereditary hemorrhagic telangiectasia (HHT) G.K. Schneider1, A. Massmann1, P. Fries1, M.A. Kirchin2, U.W. Geisthoff1, A. Bücker1; 1Homburg a.d. Saar/DE, 2Milan/IT ([email protected])

Scientific Sessions B-644

10:57

Peripheral vascular malformations: Imaging with high resolution three dimensional contrast enhanced MR angiography A.A.A. Abdel Razek, S. Samir, H. Sharaf, A. Megahed; Mansoura/EG ([email protected]) Purpose: To demonstrate the role of high-resolution three-dimensional contrastenhanced MR angiography in classiﬁcation of peripheral vascular malformations. Methods and Materials: Three-dimensional contrast-enhanced MR angiography (CE-MRA) was performed for 37 patients (18M, 17 F; mean age 21 ys) with peripheral vascular malformations. This study was conducted on 1.5 Tesla MR unit (Symphony-Siemens). The parameters used were: TR=7 msec, TE=3 msec, Flip angle=40 deg, FOV=230 mm & matrix =256X256. Bolus of gadolinium-DTPA was injected in the forearm (0.2 mmol/kg at rate of 3 ml/sec followed by 20 ml saline ﬂush at the same rate). The ﬁrst acquisition served as a mask, which was subtracted from the other acquisitions, the contrast run. Conventional angiography was done for 14 patients within 5-7 days after CE-MRA. Results: All CE-MRA examinations were technically adequate for analysis with high signal-noise ratio. Vascular malformations may be of high or low ﬂow. Venous malformation (n=18) appeared as multiple dilated stagnant venous spaces or dilated veins. Cavernovenous malformation (n=7) appeared as abnormal capillaries in arterial phase with contrast pooling in dilated venous phase. The feeding arteries, nidus and draining veins of arterio-venous malformation (n=5) were clearly detected with CE-MRA. Low ﬂow vascular malformation of Klippel-Trenaunay syndrome (n=2) can be differentiated from high ﬂow malformation in Parks-Weber syndrome (n=1). Hemangiomas (n=4) appeared as highly vascular lesions in young patients. Conclusion: High-resolution three-dimensional contrast enhanced MR angiography is a promising diagnostic imaging modality for classiﬁcation, assessment of angioarchitecture and treatment planning of peripheral vascular malformations.

B-645

11:06

Peripheral magnetic resonance angiography (MRA) with parallel imaging and a dedicated 36 element matrix-coil at 3 Tesla H. Kramer1, H.J. Michaely2, B.J. Wintersperger1, C. Glaser1, M.F. Reiser1; 1 Munich/DE, 2Mannheim/DE ([email protected]) Purpose: Matrix-coils, parallel imaging (PI) and high ﬁeld strength facilitate high spatial resolution MRA. Our purpose was to evaluate the advantages of a dedicated peripheral angio matrix coil (PAM) in terms of SNR gain, increase in spatial resolution and decrease of acquisition time (AT) in MRA. Methods and Materials: We examined 15 healthy volunteers with a PAM on a 3 T 32-channel MR system. Three different MRA protocols with different PI acceleration factors (3 and 4) were evaluated in ﬁve volunteers each. Acquired volumes and AT were similar whereas spatial resolution was increased with the acceleration factor. To evaluate the acceleration, AT/voxel was calculated. Image quality (vessel conspicuity, venous overlay, artifacts) was rated independently by two readers. Interreader agreement was calculated by kappa-statistics. SNR and CNR were evaluated. Results: None of the exams showed venous overlay. In 70% of the exams excellent vessel conspicuity was obtained. However, artifacts were present in 50% of the exams. All of these artifacts were rated as not degrading diagnostic accuracy. Interreader agreement was good (kappa: 0.65-0.74). SNR and CNR did not show signiﬁcant differences (p=0.069) in PI 3 and 4. AT per voxel could be reduced by up to 58%. Conclusion: The implementation of dedicated coils for peripheral MRA contributes to increase image quality in terms of spatial resolution and SNR. Venous overlay can be effectively avoided despite the use of high resolution scans. High PI factors at high ﬁeld strength offer the opportunity of high spatial resolution within reasonable acquisition time at good image quality.

B-646

11:15

Time-resolved contrast enhanced MR angiography (CE-MRA) in diabetic patients: Improved accuracy and time saving in the infrapopliteal district G. Savino, A. Cina, A. Moscariello, R. Iezzi, R. Marano, L. Natale, L. Bonomo; Rome/IT ([email protected]) Purpose: The purpose of the study is to assess the improvement of time resolved acquisition (TR) employed on the calves in comparison with the standard boluschase acquisition in detecting signiﬁcant stenosis and depicting the exact degree of arterial involvement. Methods and Materials: Twenty consecutive arteriopathic diabetics patients (11 M, 9 F, mean age 78Y) underwent Gadolinium subtracted mobitrak MRA (Signa Echospeed 1.5 T, GE) implemented with TR on calves and feet (ASSET subtracted

290

A

B

C

D

E

F

G

TRICKS, GE). Selective DSA performed within 30 days during the endovascular treatment was used as reference standard. Degree of stenosis ( 50%; 50-70%; 70; 100%) and diagnostic conﬁdence (1 poor-3 good) were evaluated in the infrapopliteal district by three blind radiologists for standard examination alone (ST), TR alone and ST plus TR sequence. Acquisition and reporting time were also compared. Results: Twenty-four limbs and 168 segments were assessed. The overall mean time for the examinations was 20.1 minutes (14.1 for ST and 6.0 for the TR). Accuracy was 75% for ST; 85.6% for TR and 86.1% for ST plus TR. The mean conﬁdences were respectively 2.4 for ST; 2.8 for TR and 2.85 for ST plus TR (ST vs TR, p 0.01). In ST, 17% of segments showed inadequate conﬁdence for the treatment planning, while only 9% were inadequate in TR and 6% in ST plus TR. Conclusion: Our results in diabetic patients suggest signiﬁcant improvement in accuracy and time saving using time resolved acquisition on the calves.

B-647

11:24

MRA of the hands using a blood pool agent, in evaluation of the cold immersion test M. Olszycki, B. Wozniakowski, P. Pietrzak, P. Grzelak, M. Sapieha, L. Stefanczyk; Lodz/PL ([email protected]) Purpose: Blood pool contrast agents offer important advantages in MRA such as a higher signal and a longer half-life, allowing for evaluations at higher spatial resolutions. Thermography is an acknowledged procedure for monitoring function tests; however, its accessibility is limited. The aim of this study was to evaluate MRA of the small vessels of the hand acquired during a cold immersion test, with the use of a blood pool contrast agent, and to compare the method to thermography. Methods and Materials: Two groups of women with differing thermographic proﬁles were examined: efﬁcient reperfusion (N=10), delayed reperfusion (N=12). We performed four MRAs of the hands (T1, vibe, we, cor), before and in intervals (3, 6, 12 min) after a cold immersion test. The MRA results were compared with images of temperature distribution after the same immersion test, obtained using a thermovisory camera. Results: The MRA images clearly demonstrated the vessels at the level of the ﬁngers, enabling the observation of their reactivity in response to the test. There were differences between the groups at all stages of the immersion test (Group1 - efﬁcient reversal of temperature to initial values in six minutes, Group2 - delayed reversal of more than 12 minutes). Conclusion: Changes in MRA images of the hand and changes in thermovisory images following an immersion test demonstrate similar progression, which suggests the two methods are complementary. MRA with the use of a blood pool contrast agent may serve both as a method for visualizing blood vessels and as a method for monitoring function tests to establish predisposition to vasospasm.

B-648

11:33

MRI is a predictive biomarker for tumor angiogenesis response to AvastinTM H.-J. Raatschen1, G.H. Simon2, Y. Fu2, D.M. Shames2, K.-J. Wolf1, R.C. Brasch2; 1 Berlin/DE, 2San Francisco, CA/US ([email protected]) Purpose: To determine if the change in MRI-assayed tumor vascular leakiness after a single dose of an angiogenesis inhibitor can serve as a biomarker for tumor growth response. Methods and Materials: MDA-MB 435 cancer cells were injected subcutaneously into 15 nude rats. When tumors reached a diameter of 1 cm, albumin-(GdDTPA)30-enhanced MRI was performed. Immediately after baseline imaging, angiogenesis-inhibition was initiated with doses given every third day for a total of four doses of bevacizumab (Avastin) at three dose levels: 0.1 mg (n=3), 0.5 mg (n=5) and 1.0 mg (n=3). Control animals (n=4) received saline. The second MRI was performed 24 hours after treatment initiation. Baseline and post-treatment tumor vascular leakiness (KPS) were calculated using a pharmacokinetic model. Tumor volume was measured every other day, tumor growth rate was determined by monoexponential data ﬁtting. Results: The MRI-assayed change in tumor microvascular leakiness ($KPS) correlated positively and signiﬁcantly with tumor growth rate (r2=0.75, p 0.0001). KPS increased 23% in the control group and decreased signiﬁcantly (p 0.05) in the three different Avastin-treated groups 24 hours after a single injection of bevacizumab (-62 %, -88% and -95%, respectively). After 10 days, the 0.5 and 1.0 mg groups showed a signiﬁcantly (p 0.05) slower tumor growth rate (0.12p0.01 and 0.14p0.02, respectively) than the control group (0.20p0.03). Conclusion: The dynamic MRI-assayed tumor response to Avastin, reﬂected in the suppression of vascular leakiness 24 hours after a single dose, strongly correlated with tumor growth suppression at 10 days of treatment and thus was a successful biomarker.

Scientific Sessions B-649

11:42

Longitudinal evaluation matrix metalloproteinases (MMPs) to monitor atherosclerosis progression and regression in vivo using molecular MRI V. Amirbekian1, F. Hyaﬁl2, S. Amirbekian3, E. Vucic2, E. Lancelot4, C. Corot4, Z.A. Fayad2; 1Boston, MA/US, 2New York, NY/US, 3Atlanta, GA/US, 4Paris/FR ([email protected]) Purpose: MMPs are involved in atherosclerotic plaque growth and rupture. P947 (Guerbet) is an MRI contrast-agent formed of a peptide with a high afﬁnity for MMPs bound to a molecule of gadolinium-chelate. We hypothesized that MMP activity can be monitored in atherosclerosis of hypercholesterolemic rabbits with P947-enhanced MRI. Methods and Materials: Using atherosclerotic New-Zealand-White rabbits (n=8), MRI was done before and 90-min after IV-injection of 50 μM Gd/kg P947. The same procedure was repeated after injection of the standard MRI contrastagent Gd-DOTA. After baseline MRI-scans, four rabbits were maintained on a high-cholesterol diet (“progression-group”) and four were switched to a chow diet (“regression-group”). A second set of scans was performed four months later in both groups. Enhancement of atherosclerotic plaques was measured using [(CNR after contrast agent/CNR before)x100]-100. The MMP activity was quantiﬁed using aortic specimens and SDS-PAGE gelatin-zymography. Results: On the baseline P947-enhanced MRI, signal-intensity of atherosclerosis increased similarly in the “progression” group rabbits (36p3%) and in the "regression" group (41p5%). After four months, signal-enhancement in atherosclerotic plaques in the "progression" group was similar to the baseline MRI (38p2%), whereas a signiﬁcantly lower enhancement was found in the "regression" group (14p3%; p=0.05). No signiﬁcant difference in signal-intensity was found in atherosclerotic plaques on MRI using Gd-DOTA. The mean gelatinolytic activity of MMP2 was signiﬁcantly higher in atherosclerotic plaques from the "progression-group" as compared with the "regression-group" (42Kp19K vs 23Kp19K; p 0.05). Conclusion: P947-enhanced MRI allows for the monitoring of MMP activity in atherosclerotic plaques representing a promising non-invasive technique for the evaluation of therapies aimed at plaque stabilization.

B-650

11:51

MR lymphangiography at 3 T: Feasibility study M. Notohamiprodjo1, K.A. Herrmann1, S.O. Schoenberg2, R. Baumeister1, C. Glaser1, M.F. Reiser1, T.F. Jakobs1; 1Munich/DE, 2Mannheim/DE ([email protected])

Room L/M

Radiographers

SS 1414 Dose reduction strategies Moderators: L. Lönn; Copenhagen/DK S. Willis; Exeter/UK

B-651

10:30

The inﬂuence of exposure parameters and beam ﬁltration in ESD: A phantom study using a CR system L. Lanca, J.F. Serra, J.M. Serra, R. Pereira, T. Gomes; Lisbon/PT ([email protected]) Purpose: European guidelines on quality criteria for diagnostic radiographic images provide recommendations for good radiographic techniques by which the diagnostic requirements and the dose criteria can be achieved. This study aims to evaluate the inﬂuence of exposure parameters variation in entrance skin dose (ESD) in chest and lumbar spine radiography in a CR system. Methods and Materials: Radiographs were obtained using two anthropomorphic phantoms. A total of 25 images were taken. Tube potentials of 70 and 90 kV for the lumbar spine AP and 77 and 93 kV for the lateral projection were applied. In chest PA 90-133 kV tube potential was applied. For each exposure combination four images were obtained: no added ﬁltration, 0 mm Cu; 0.1 mm Cu; 0.2 mm Cu and 0.3 mm Cu. This resulted in different technique groups for each projection. ESD values were calculated from the DAP meter chamber and from the tube output. Results: The use of Cu ﬁlters allows an ESD reduction of 50 and 40% in lumbar spine AP projection for exposures obtained with 70 and 90 kV, respectively. A reduction of 50% is found in lumbar lateral projection for both tube potentials (77 and 93 kV). In chest PA a reduction of 52% is found at 125 kV. Conclusion: A considerable ESD reduction is achievable when using the most adequate exposure parameters. Our study conﬁrms that ESD could be signiﬁcantly reduced in a CR system when using beam ﬁltering. Findings should be further evaluated in a wider variety of clinical studies.

B-652

10:39

Balancing diagnostic image quality with entrance dose and ﬁltration in computed radiography: A European perspective T. Lehnert, H. Korkusuz, F. Khan, V. Jacobi, M.G. Mack, T.J. Vogl; Frankfurt a. Main/DE ([email protected]) Purpose: In this study, image quality was based on required clinical criteria, in order to investigate to what degree entrance dose could be lowered and what kind of added ﬁltration can be used without impinging on radiologist conﬁdence levels in diagnosing. Methods and Materials: Images were taken of extremities from a cadaver using stepwise decreasing dose levels [DL] (100%, 75%, 50%, 25%) and variation of added ﬁltration (no added ﬁltration, aluminum, aluminum/copper) under digital projection radiography (Kodak DirectView CR950). The starting point dose level for all body parts imaged was the current X-ray technique. The images were presented to four radiologists in a blinded fashion on a clinical diagnostic workstation and they rated each with an image quality score from 1 to 9, with 9-very satisﬁed and 1-very unsatisﬁed indicating loss of diagnostic value. Results: Without added ﬁltration, image quality mean score was rated as 6.3 (DL 100%), 5.5 (DL 75%), 5.1 (DL 50%) and 4.3 (DL 25%). An added aluminum ﬁltration induced an image quality mean score of 5.8 (DL 100%), 5.9 (DL 75%), 4.7 (DL 50%) and 3.4 (DL 25%). Using aluminum/copper ﬁltration, image quality mean score was rated as 5.9 (DL 100%), 5.8 (DL 75%), 4.6 (DL 50%) and 2.9 (DL 25%). Conclusion: It is possible, in the case of extremities, to lower entrance doses up to 75% of the normal value, a reduction of 25% in dose, under simultaneous use of added aluminum or aluminum/copper ﬁltration, without comprising the diagnostic value required.

A

B

C

D

E

F

G

291

Monday

Purpose: To evaluate the feasibility of contrast-enhanced MR-lymphangiography (MRL) at 3 T in patients with lymphedema of the lower extremity. Methods and Materials: Eight consecutive patients with unilateral and two patients with bilateral lymphedema of the lower extremity underwent MRL. MRL was performed after intracutaneous injection of 0.9 ml Gd-DTPA and 0.1 ml Mepivacaine hydrochloride 1% into three interdigital spaces prior to image acquisition. A high-resolution (HR) T1w-3D-GRE-(FLASH) sequence (TR 3.76 msec/TE 1.45 msec; FOV 500 mm with spectral fat saturation and an isotropic voxel size of 0.8x0.8x0.8 mm3) was acquired on a 3.0 T scanner. Three regions were examined: lower leg, upper leg and the pelvis. The patients were mobilized after 40 minutes. After 20 minutes of walking late phase images were acquired. Images were analyzed using 3Dmaximum-intensity projection techniques to depict the lymphatic vessels. Results: Depiction of contrast enhanced lymphatic vessels was successful in all patients in both affected and unaffected lower extremities. Normal lymphatic drainage was observed in unaffected extremities shortly after contrast material application reaching the pelvic level after approximately 30 min. In extremities with lymphedema, ectatic lymphatic vessels and lacunae were depicted. Lymphatic drainage was considerably delayed. In two patients a lymphocele was identiﬁed and a ﬁne reticular pattern of the lymphatic system could be visualized. Image acquisition after patient mobilization did not provide additional information. Conclusion: HR-MR lymphangiography at 3 T is technically feasible providing higher spatial resolution with isotropic voxel size. The technique is promising in increasing the detection of ﬁne reticular lymphatic vessels.

10:30 - 12:00

Scientific Sessions B-653

10:48

Exposure parameters and ESD in chest radiography: A comparative study from two DR systems L. Lanca, D. Sousa, M. Freitas, O. Felicio, M.I. Ramalho; Lisbon/PT ([email protected]) Purpose: DR systems based on solid state detectors have the potential to achieve a considerable dose reduction in chest radiography. However, wide variations in patient dose for the same radiographic examinations are described in literature. This study aims to evaluate exposure parameters and entrance skin dose (ESD) in chest PA and lateral radiography. A comparative study was conducted at two hospitals using the same detector technology (a-Si:H/TFT) from the same manufacturer. Methods and Materials: A total of 627 chest exposures were evaluated (67 in hospital A; 560 in hospital B). Exposure parameters for each radiograph were obtained from the DICOM Log ﬁle available from both ﬂat-panel systems. ESD values for PA and lateral projections were estimated from DAP and ﬁeld size relation and the backscatter factor. Results: In hospital A 43% of cases are higher than the European reference of 0.3 mGy. In lateral projection 83% are below the reference (1.5 mGy). In hospital B ESD values are signiﬁcantly below the references (p 0.05), both in PA and lateral projections. Exposure time in both hospitals is below the reference of 20 (PA) and 40 ms (lateral). Mean exposure time is higher in hospital A (8.59 ms) when compared to hospital B (3.33 ms) for PA projection. In lateral projection differences were also identiﬁed: 9.53 ms (hospital A), 14.83 ms (hospital B). Conclusion: A considerable variation in ESD and exposure time was observed. Results conﬁrm that wide variations in patient dose and exposure time can be found at the same DR detector technology. Performance evaluation programmes should be implemented.

B-654

10:57

Dose reduction strategies for CT cardiac imaging R. Raupach1, H. Bruder1, A.N. Primak2, X. Liu2, C.H. McCollough2, T.G. Flohr1; 1 Forchheim/DE, 2Rochester, MN/US (thomas.ﬂ[email protected]) Purpose: To reduce radiation dose in ECG-controlled cardiac CT to the lowest achievable level without compromising image quality. Methods and Materials: We compare ﬁve different approaches for cardiac CT examinations: (A) helical scan with constant tube current, (B) helical scan with ECGcontrolled tube current modulation (20% of the maximum power applied outside the desired cardiac phase [CP]), (C) scan as B but with reduction to 4% outside the CP of interest, (D) triggered sequential acquisition, and (E) as D but with dynamic temporal windows (triggering and scan length prospectively adjusted according to the ECG trend and variability). For each method, the CT scan was simulated using 60 ECGs recorded from non-beta blocked patient coronary CTA examinations (heart rate between 40 and 103 bpm). Image quality and dose reduction were then compared. Results: As a reference, method A allows reconstructing at arbitrary CP with maximum image quality (CP reliability), but at highest dose. Methods B, C and E show the same reliability (98% of desired CP able to be reconstructed at high tube current) while method D missed the desired CP in 18%, resulting in reduced image quality. Compared with approach A, radiation dose was reduced by 45% (B), 54% (C), 75% (D) and 68% (E). Conclusion: The sequential scan (D) yields the lowest dose, but suffers from inconsistent availability of the desired CP which will generally degrade image quality. Low tube current helical (C) and adaptively triggered sequential scan (E) provide the lowest patient doses without compromising image quality reliability.

B-655

11:06

DAP in patients submitted to interventional cardiac catheterization procedures G. Paulo, M. Costa, C. Gama, M. Rocha, P. Lavandeira, L. Marques; Coimbra/PT ([email protected]) Purpose: The object of the study is to start the implementation of diagnostic reference levels (DRLs) through the measurement of DAP (dose area product) in Portugal on cardiac catheterization procedures. Methods and Materials: The study involved the measurement of DAP, exposure time and other patient data (sex, age, weight, height, AP and lateral diameter) corresponding to each of the 91 patients that were referred for diagnostic cardiac catheterism. DAP was measured with resource to an ionization chamber (DIAMENTOR M4-KDK) that was located in the exit of collimators. Results: The medium value of DAP was 25.71 Gy cm2 for the left ventriculography and coronariography, and 38.22 Gy cm2 for left ventriculography, coronariography and aortography. The value of DAP was statistically signiﬁcant for men compared to

292

A

B

C

D

E

F

G

that obtained for women (p 0.05). The value of DRL calculated in function of the weight reference group (65-75 kg) was 31.9 Gy cm2 and in function of BMI (Body Mass Index) was 48.85 Gy cm2. Conclusion: The results are lower compared with other international studies. Taking into consideration the previous studies and considering that the sex inﬂuences the value of DAP, in agreement with the deﬁnition of “standard patient” (70p5 kg), we concluded that the weight by itself does not establish any comparison criterion between different people and that DRLs deﬁned in function of BMI will allow to compare results of different populations more easily.

B-656

11:15

Snapshot pulse: Low-dose cardiac imaging with prospective ECG-triggering P. Spagnolo, M. Giglio, M. Romano, A. Mazzarola, T. Khouri, S. Sironi; Monza/IT ([email protected]) Purpose: Over the last few years MDCT has become an important tool for noninvasive cardiac imaging. The major drawback of this technology is the high radiation dose caused by data acquisition with retrospective-ECG gating. Our purpose was to evaluate image quality and radiation dose reduction in 64-slice coronary CTA with prospective ECG-triggering. Methods and Materials: Thirty-seven consecutive patients (28 men, mean age 57.4, mean BMI 26.6p3.5) were included in this study and underwent MDCT examination with prospective ECG triggering to rule out CAD (SnapShot Pulse, GE Healthcare). Eleven patients were revascularized with bypass. In all subjects, heart rates were evaluated on ECG before CT examination. Beta-blockers were administered to patients with heart rate above 65 bpm. Patients with high heart rates, not responsive to betablockers, were excluded and underwent conventional gated helical scan retrospectively. Results: The average radiation dose was 6.8p2.7 mSv. Patients' heart rate ranged from 48 to 67 bpm (average 57p6.2 bpm). Diagnostic images of all coronary artery segments were obtained in all patients. Conclusion: This new technique, by synchronizing the acquisition to the patient’s heart rate through a prospectively triggered axial method, provides signiﬁcant X-ray exposure and dose reduction. High image quality for a wide range of patient sizes was obtained.

B-657

11:24

Patient dose survey from coronary angiography and percutaneous transluminal coronary angioplasty in a Chinese cardiac centre B. Liu; Beijing/CN ([email protected]) Purpose: To analyze dose parameters in coronary angiography (CA) and percutaneous transluminal coronary angioplasty (PTCA) and to assess the radiation exposure received by patients in Beijing Xuanwu Hospital. Methods and Materials: 100 cases of CA procedures and 51 cases of PTCA procedures undertaken on a Siemens AXIOM Artis dFC X-ray equipment were monitored in the present study by measuring peak skin dose (PSD), dose-area product (DAP) and other operational parameters. Thermoluminescent dosimetries (TLDs), matrix of dosimetry, transmission ion chambers, PCXMC (PC-based Monte Carlo program) and a humanshaped phantom are used to measure and to calculate dose related quantities. Results: The DAP values ranged from 7.6 to 60.5 Gy cm2 for CA and 16.4-162 Gy cm2 for PTCA. The effective dose to patient was determined to be in the range of 1.1-6.9 mSv for CA and 2.3-20.1 mSv for PTCA. Conversion factors between effective dose and DAP were derived for CA and PTCA, respectively. During interventional procedure, different projection angles required different exposure doses. Conclusion: PTCA had the potential to produce clinically signiﬁcant radiation doses. The results of organ dose from TLDs agreed with those from PCXMC software. Comparing with TLDs, DAP measurement based on Monte Carlo simulation is found to be more convenient and reliable for the effective dose calculation of patient. However, dose area product (DAP) and cumulative dose (CD) have a poor relationship with PSD, and cannot serve as a proper indicator to monitor patient’s deterministic radiation risk.

B-658

11:33

Impact of heart rate on radiation dose in dual source cardiac CT U. Saueressig, G. Pache, A. Frydrychowicz, A. Eberl, M. Langer, T.A. Bley; Freiburg/DE ([email protected]) Purpose: To demonstrate the effect of heart rate on the dose performance of a dual source CT scan of the coronary arteries. Methods and Materials: Sixty patients without frequency control were examined for coronary artery disease on a DSCT-scanner (Siemens Deﬁniton) with the manufacturer supplied protocol (ECG-gated tube current, variable pitch, 120 kV,

Scientific Sessions 2x350 mAs maximum tube current). Width of time of maximum tube current was automatically selected depending on heart rate. Patients with heart rates below 75 bpm were reconstructed at a narrow window at 70% of RR cycle (n=32), and with heart rates between 75 and 90 (n=28) with an interval between 35 and 70%. Dose was compared between the groups. In the second group, an analysis of best reconstruction window for coronary artery assessment was performed. Results: Dose values of patients with a heart frequency below 75 were signiﬁcantly lower than in the other group (0.86p0.18 vs 1.02p0.13 mSv/cm). The largest number of assessable segments of the coronary arteries in the group with higher heart frequency could be found in the 40% reconstruction, followed by 70%. Conclusion: Dose of cardiac CT in patients with a heart frequency below 75 was lower than for patients with high heart rates. The longer reconstruction interval cannot be omitted for patients with higher heart rates, because in some coronary segments early diastolic image reconstruction is superior. Therefore, a pharmaceutical lowering of heart rate is advisable to lower patient radiation dose exposure in DS-cardiac CT.

B-659

11:42

Dual-source coronary artery computed tomography angiography: Positioning of pulsing windows determined by heart rate dependent shift of the optimal reconstruction window within the cardiac cycle K. Anders, S. Achenbach, U. Ropers, D. Ropers, A. Kuettner, W.A. Bautz; Erlangen/DE ([email protected]) Purpose: To evaluate reconstruction window placement for coronary CT angiography in patients with low and high heart rates using dual source (DS) CT with a temporal resolution of 83 ms in order to determine heart rate dependent pulsing windows. Methods and Materials: 150 patients (heart rate 37-116 bpm) were studied by DSCT with heart rate independent temporal resolution of 83 ms for halfscan reconstruction. Datasets were evaluated for occurrence of motion artifacts and positioning of the optimal reconstruction window within the cardiac cycle in relation to the heart rate. Reliable phase prediction enables an effective use of ECG-pulsing with preferably narrow pulsing windows or of step-and-shoot techniques. Results: All scans yielded diagnostic images. In 114 patients, a single reconstruction was sufﬁcient to evaluate the entire coronary tree. At b75 bpm (91 patients), diastolic reconstruction yielded the best results: 65% R-R in 10%, 70% R-R in 57%, 75% R-R in 24%. At 75 bpm, reconstructions were spread between 35% and 75% of the cardiac cycle, in 5/8 patients 90 bpm, 45% R-R was used. In 36 patients, more than one reconstruction was required. Conclusion: A narrow diastolic pulsing window is feasible in patients with controlled heart rates. In patients with higher heart rates, a rather wide pulsing window (e.g. 35-75% R-R) seems mandatory. However, this should be reﬁned by larger studies focussing on high heart rate exams.

B-660

11:51

Rotational tomography of the normal and reconstructed middle ear J. Kromeier, C. Offergeld, M. Langer; Freiburg/DE ([email protected])

Room N/O

Neuro

SS 1411 MRI at 7 T Moderators: M. Cirillo; Naples/IT E.-M.B. Larsson; Aalborg/DK

B-661

10:30

First clinical study on high ﬁeld MRI in patients with multiple sclerosis: Comparison of 1.5 T and 7.0 T K. Kollia, S. Maderwald, N. Putzki, M. Forsting, M.E. Ladd, I. Wanke; Essen/DE ([email protected]) Purpose: The purpose of this study was to investigate the potential of high resolution imaging of multiple sclerosis lesions in vivo comparing 7 T with conventional 1.5 T. Methods and Materials: Ten patients with clinically deﬁnite MS were scanned on a 7 T whole-body scanner (Magnetom 7 T, Siemens, Erlangen, Germany) with an 8-channel phased-array transmit-receive-head-coil and on a 1.5 T Avanto (Siemens, Erlangen, Germany) with a 12-channel phased-array-head-coil. The imaging protocol consisted of high-resolution axial PD+T2-weighted turbo spin echo, T2*-gradient echo, slice thickness 2 mm, matrix 512x384 and sagittal T1-weighted 3D-MPRAGE, matrix 512x384, 0.6x0.5x0.6 mm. Results: The sequence parameters at 7 T had to be modiﬁed due to speciﬁc absorption rate restrictions but we were able to keep pulse sequence parameters equivalent to the imaging protocols at 1.5 T. White matter lesions were better detected and delineated from adjacent structures at 7 T compared to 1.5 T. The perivascular migration of MS lesions was very well visible on T2*-GRE sequences. In larger lesions (10 mm) there was a multilayer structure revealed on T2*-GRE not seen at 1.5 T. Due to the higher resolution it was possible to differentiate between juxtacortical white matter lesions and cortical lesions. Conclusion: High ﬁeld imaging of MS patients at 7 T was well tolerated and provided better visualisation of MS lesions in the grey matter and better demonstrated structural abnormalities within the MS lesion itself.

B-662

10:39

Promising insights into patients’ hippocampi at 7 Tesla J.M. Theysohn, O. Kraff, S. Maderwald, A. de Greiff, L. Schäfer, S. Lohbeck, M. Forsting, M.E. Ladd, S.C. Ladd, E.R. Gizewski; Essen/DE ([email protected]) Purpose: The human hippocampus plays a central role in various neuropsychiatric disorders. High-resolution evaluation of morphology and inner structure are of interest, as changes are often subtle. New high ﬁeld MRI systems (7 T) for human imaging provide more signal and thus higher-resolution images. The purpose of this study was to depict hippocampal features of volunteers and patients with cutting-edge resolution. Methods and Materials: Measurements were performed on a whole-body 7 T scanner (Magnetom 7 T, Siemens, Germany) using an 8-channel transmit-receive head coil (Rapid, Germany). Coronal PD-, T2-, T2*-, FLAIR-, and T1-3D-MPRAGE sequences were optimized on eight patients with respect to contrast and resolution before examining six patients (e.g. epilepsy, dementia, post-trauma). Results were compared to 1.5 T images. Results: Depiction of the hippocampal area was achieved in all 14 subjects with all sequences at 1.5 T and 7 T. Imaging time per acquisition ranged from 2:30 to 8:42 minutes, and allowed complete coverage in 48 minutes for all sequences together. Resolutions were T2*: 0.22x0.22x3.0 mm³, PD/T2: 0.25x0.25x3.0 mm³, FLAIR: 0.3x0.3x4 mm³, and MPRAGE: 0.5x0.5x0.5 mm³. All sequences visualized structural hippocampal features (cornu ammonis, dentate gyrus, subiculum, hippocampal sulcus) of the volunteers and patients in high detail not seen at 1.5 T. Conclusion: Acquisition of high-resolution images of the human hippocampus at 7 T is possible in a reasonable time. Visualization of hippocampal substructures not depictable at 1.5 T will provide additional information regarding changes to speciﬁc subregions. Thus, in the future 7 T might help in the diagnosis of subtle hippocampal changes e.g. in patients with focal epilepsy.

A

B

C

D

E

F

G

293

Monday

Purpose: Imaging is essential in reconstructive middle ear surgery, especially in pre-operative planning. The improvement of imaging quality and the development of new imaging techniques such as rotational tomography (RT) enables postoperative follow-up and evaluation of surgical results. The aim of this study was to evaluate RT as a new tool for postoperative determination of middle ear anatomy and implant position. Methods and Materials: RT was performed in 10 temporal bone specimens after insertion of different middle ear prostheses concerning material, shape and length (PORP; TORP; Stapes piston). An implantable hearing device (Symphonix Soundbridge) was also implanted and visualized. For comparison some specimens additionally underwent conventional computed tomography (CT). We used an Axiom Artis angiography unit and a Sensation 64 CT scanner (Siemens, Erlangen, Germany). Results: Characterization of anatomical structures of the temporal bone using RT was of comparable quality to conventional CT-scans in all investigated specimens while requiring approximately 40% of the CT’s irradiation exposure. RT showed almost no problems due to metallic artefacts of the implanted prostheses. Furthermore RT enabled angle determination of inserted prostheses towards the tympanic membrane and/or the malleus handle. Detailed imaging of the prostheses allowed determination of shape and localization within the reconstructed middle ear. Conclusion: RT allows precise presentation of anatomical structures and middle ear implants within the temporal bone region. Therefore this method could be suitable not only for pre- but also for post-operative evaluation of the middle ear after reconstructive surgery to decide whether revision surgery would be promising and/ or necessary.

10:30 - 12:00

Scientific Sessions B-663

10:48

Vascular dementia at 7 Tesla: New insights into microbleeds? J.M. Theysohn, O. Kraff, M. Barth, S. Maderwald, W.P. Becker, M. Forsting, M.E. Ladd, S.C. Ladd, E.R. Gizewski; Essen/DE ([email protected]) Purpose: Cerebral microbleeds as seen in vascular dementia (V) might be indicative of higher risk for future intracerebral hemorrhage. The introduction of high ﬁeld MRI systems (7 Tesla) for human imaging provides enhanced tissue signal and sensitivity to susceptibility differences. The purpose of this work is to demonstrate initial results visualizing cerebral hemosiderin deposits and surrounding microangiopathy with high-ﬁeld MRI. Methods and Materials: All measurements were performed on a whole-body 7 Tesla scanner (Magnetom 7T, Siemens, Germany). T2*, SWI, and FLAIR sequences were optimized regarding contrast and resolution in eight volunteers before examining eight patients with known microangiopathy and/or microbleeds. An eight-channel transmit-receive head coil (Rapid Biomedical, Germany) permitted parallel imaging (GRAPPA). Parameters were chosen to optimize the demarcation of white matter lesions and hemosiderin deposits. Images were compared to 1.5 Tesla. Results: In all eight patients known pathologies from 1.5 T examinations were revealed in 7T. Imaging time per acquisition ranged from 2:30 to 7:21 minutes, and allowed coverage of the whole head in 45 minutes. Compared to 1.5 T, image resolutions were higher at 7T: 0.22x0.22x3.0 mm³ in T2*, 0.44x0.44x1.0 mm3 in SWI, and 0.3x0.3x4 mm³ in FLAIR. T2* and SWI revealed additional microbleeds not obvious in 1.5 T in 6/8 patients. FLAIR sharply visualized white matter lesions similar to 1.5 T. Conclusion: Improved detection of microbleeds at 7T might in the future have signiﬁcant impact on the therapy of patients showing cerebral vascular disease (e.g. in VD) by providing additional criteria to optimize antithrombotic treatment.

B-664

10:57

New approaches for MS lesions assessment with ultrahigh ﬁeld MRI: Comparison of T2*/phase susceptibility weighted images, with T2- and inversion recovery fast spin echo sequences J. Shah, K. Rammohan, S. Sammet, E. Bourekas, D.W. Chakeres, M.V. Knopp, P. Schmalbrock; Columbus, OH/US ([email protected]) Purpose: To evaluate 7T susceptibility weighted imaging for characterization of MS lesions and to compare with lesion depiction in T2-weighted and white-matter attenuated inversion recovery sequences. Methods and Materials: Ten MS patients (33-53 y) were studied at 7T (Philips, Achieva) using 2D-gradient echoes with TR/TE/ﬂip=1600 ms/12 ms/50o, FOV 230, 512 matrix, 2.5 mm slices, 10 min scan time. High pass ﬁltered phase (SWI-phase) images were reconstructed off-line, and compared to magnitude (SWI-T2*) images. T2-GraSE images (TR/TE 4000 ms/70 ms, 6 spin echoes, 3 gradient echoes) and inversion recovery-TSE (TR/TI/TE 8000 ms/500 ms/14 ms, 10 echoes) were acquired. All images were spatially matched, compared and visually inspected for lesion features. Results: Lesion contrast was higher on white matter-suppressed IR-TSE and T2-weighted images compared to SWI-T2* magnitude images which nevertheless showed 85-100% of lesions seen with T2 and IR. Only a fraction of lesions seen on magnitude images was seen on phase images (average 55%, range 34-97%). Characterization of lesion morphology was best with SWI-phase images, where lesions were seen as dark focal regions, or as hypointense rim. A majority of lesions were associated with small veins. Conclusion: SWI-phase images add considerable information to MS lesion morphology. Lesions seen on magnitude, but not phase images are likely due to increased free water from myelin injury, inﬂammation or subtle edema. Lesions seen on SWI-phase images have a paramagnetic component due to locally increased tissue iron, or deoxyhemoglobin from increased vasculature (angiogenesis). Further studies are needed to fully understand the contrast features in SWI-phase images.

B-665

11:06

Evaluation of intracranial aneurysms with 7 T versus 1.5 T MR angiography C. Moenninghoff; Essen/DE Purpose: The approximately 4.7 times higher signal at 7T compared with that at 1.5 T may improve image quality of MR angiograms and should provide a superior visualization of intracranial aneurysms. The purpose of this study was to determine whether 7T time-of-ﬂight (TOF) MR angiography (MRA) is superior to 1.5 T TOF MRA in the depiction of intracranial aneurysms. Methods and Materials: Ten patients with intracranial aneurysms underwent 7T

294

A

B

C

D

E

F

G

TOF MRA (Magnetom 7T, Siemens, Erlangen, Germany), 1.5 T TOF MRA (Magnetom Espree or Sonata; Siemens, Erlangen, Germany) and DSA. Both scanner types were used in combination with CP or 8-channel head coils. Two blinded neuroradiologists compared the image quality of aneurysms in both TOF MRAs by using a four-point scale. Wilcoxon rank test was used to compare the mean values of image quality determined by both readers. Results: Twelve intracranial aneurysms in ten patients depicted by conventional angiography were correctly identiﬁed in 7T and 1.5 T TOF MRA by both readers. Intracranial 3D TOF MRA at 7T revealed superior image quality of the aneurysmal dome (58% of 12 aneurysms) and the aneurysmal neck (50% of 12 aneurysms). Conclusion: Our preliminary results indicate that 7T TOF MRA offers superior image quality of the aneurysmal dome and neck compared with that of a standard 1.5 T TOF MRA. Further investigations and optimized head coils are needed to convert the signal gain of ultrahigh ﬁeld MRI into a further improvement of image quality of TOF MRAs.

B-666

11:15

Assessment of brain iron in MS patients and healthy controls by ﬁeld dependent R2 increase at 7 T compared to 3 T MRI J. Shah, G. Mihai, K. Rammohan, S. Sammet, M.V. Knopp, P. Schmalbrock; Columbus, OH/US ([email protected]) Purpose: To assess T2 measurements at 3 T and 7T with different delay times between refocusing pulses for evaluation of brain iron differences in MS patients compared to healthy controls. Methods and Materials: Ten MS patients (33-53 y) and ten healthy controls (32-56 y) underwent T2 measurements at 3 T and 7T (Philips, Achieva) with dual spin echo (tau=40 ms between refocusing pulses) and with 8-echo gradient spin echo (GraSE, tau=9 ms). ROIs were manually traced in normal-appearing frontal WM and GM, thalamus, caudate, putamen, and globus pallidus. T2 were calculated for both tau and ﬁeld strengths, and used to calculate average ﬁeld-dependent R2 increase, FDRI=R2 (7T)-R2 (3 T). Results: T2 were similar for long tau dual spin echoes at 7T which are theoretically most sensitive to tissue iron. T2 were slightly longer in MS for short tau GraSE at 7T and for dual and GraSE at 3 T (though statistically signiﬁcant only for 7T GraSE, and 3 T dual SE). In healthy brain, FDRI correlated with the estimated regional iron content. In MS patients, average FDRI were 40-80% higher than in controls for all normal-appearing brain regions. Conclusion: Apparent T2 measured in our study reﬂect combined effects from water content and tissue iron. Thus the observed T2 increase in MS may be attributed to increased water content from diffuse myelin injury, inﬂammation and/or subtle edema, which seems to mask contributions from brain iron. Conversely in FDRI, water content T2 effects are removed, and thus FDRI is more sensitive for detecting iron increase in MS.

B-667

11:24

SENSE makes sense for diffusion tensor imaging and ﬁbertracking at 7 T S. Sammet, M.O. Irfanoglu, R.M. Koch, P. Schmalbrock, M.V. Knopp; Columbus, OH/US ([email protected]) Purpose: Diffusion Tensor Imaging (DTI) allows the observation of molecular diffusion in tissues in vivo and therefore assessment of the structural organization. DTI with single-excitation protocols still faces serious challenges: limited spatial resolution, susceptibility to magnetic ﬁeld inhomogeneity, and low signal-to-noise ratio (SNR). The aim of this study was to investigate the inﬂuence of SENSitivity Encoding (SENSE) on in vivo DTI and ﬁbertracking of the human brain at 7T. Methods and Materials: DTI of the human brain was performed on six volunteers on a 7T whole body MR-scanner (Achieva, Philips Medical Systems) using a prototype 8-channel SENSE head coil. DTI was performed with different SENSE factors: 1, 1.5, 2 and 3. Acquisition parameters: TR=3600 ms; TE=74 ms; FOV=230 mm2; matrix size=128x128; b=0, 250, 500, 750, 1000 s/mm2; slice thickness=3 mm and NSA=2. Results: DTI at 7T provides high quality diffusion weighted images when using SENSE. A second order shim before the data acquisition is necessary to reduce susceptibility artifacts. SENSE reduced susceptibility and distortion artifacts in 7T DTI. A SENSE factor of 3 provided images with the least distortions and allowed to create high quality Fractional Anisotropy maps and to create ﬁbertracks of the human brain in vivo at 7T. Conclusion: SENSE makes sense for ultra-high magnetic ﬁelds diffusion tensor imaging. 7T ﬁber tractography using high-quality SENSE-DTI data provides a promising method for exploring the neuronal connectivity of the brain. DTI at ultra high ﬁelds can be applied to study diseases that affect the integrity of white matter structures.

Scientific Sessions B-668

11:33

First experience with fMRI at 7 T in patients with brain tumours B. Rumberg, C. Mönninghoff, S. Maderwald, D. Stolke, M. Forsting, M. Ladd, E. Gizewski; Essen/DE ([email protected]) Purpose: The advantages of high ﬁeld fMRI derive from an increased BOLD contrast, which could lead to single event fMRI with beneﬁts in uncooperative patients, for example. However, enhanced sensitivity to susceptibility leads to increased artefacts. Our purpose was to evaluate fMRI at 7T for identifying eloquent motor areas for presurgical planning in tumour patients. Methods and Materials: Six patients with brain tumours near the central area were examined at 1.5 and 7T using ﬁnger tapping of the contralateral hand in a block design fashion. An EPI sequence with short TE was chosen for whole-brain coverage using a CP transmit/receive head coil at 7T in two patients. Four patients were examined with an 8-channel head coil at 7T. The sequence with PAT was improved for coverage of the supratentorial brain. Results: The signal change was signiﬁcantly higher at 7T than at 1.5 T (factor 2-3) in all patients. A well-ﬁtted response curve could be detected at 7T in all cerebral areas involved in ﬁnger-tapping even in single event analysis. Signiﬁcantly reduced artefacts and improved co-registration were noticed when using the 8-channel head coil versus the CP head coil at 7T while retaining the beneﬁt of increased signal. Conclusion: Our results indicate that fMRI beneﬁts from high ﬁeld MRI in tumour patients, but is accompanied by more severe susceptibility artefacts. Using an 8-channel head coil leads to a reduction of artefacts. This technique may be a useful approach to detect vulnerable areas preoperatively even with a single event.

B-669

11:42

Multi-parametric MRI in the investigation of animal models of brain gliomas: An in vivo study at 7 T S. Capuani, P. Porcari, M. Lecce, E. D’Amore, F.S. Pastore, B. Maraviglia; Rome/IT ([email protected])

B-670

11:51

Evaluation of brain tumors with 7 T versus 1.5 T MRI C. Moenninghoff; Essen/DE Purpose: The objective of our study was to characterize primary and secondary brain tumors at 7T MRI in comparison to 1.5 T MRI. By taking advantage of the increased SNR, spatial resolution, and susceptibility of a 7T whole-body MR system a superior visualization of intracranial tumors compared with that of 1.5 T systems should be achieved. Methods and Materials: Thirteen patients with untreated and histologically conﬁrmed astrocytic brain tumors or brain metastases detected at 1.5 T MRI (Magnetom Espree or Sonata; Siemens, Erlangen, Germany) underwent 7T MRI (Magnetom 7T, Siemens, Erlangen, Germany). 7T images were performed using either a CP head coil or an 8-channel head coil. 2D and 3D spin echo and gradient sequences were performed. A clinical standard protocol at 1.5 T was compared to a scientiﬁc, high-resolution protocol at 7T. Results: All gliomas (n=10, WHO grade II-IV) and all brain metastases (n=5) that had been documented by 1.5 T were also visualized by 7T MRI. Besides small

10:30 - 12:00

Room R1

Chest

SS 1404 Advances in CT Moderators: P. Goodman; Durham, NC/US R. McDermott; Dublin/IE

B-671

10:30

Pulmonary alveolar proteinosis versus exogenous lipoid pneumonia: CT ﬁndings H. Choi, C. Park, J. Goo, H. Lee, C. Lee, J.-G. Im; Seoul/KR ([email protected]) Purpose: To compare CT ﬁndings of pulmonary alveolar proteinosis (PAP) and exogenous lipoid pneumonia (ELP). Methods and Materials: This study consisted of 16 patients with pathologically proven PAP (n=8, four men and four women; mean age, 53 years) and ELP (n=8, ﬁve men and three women; mean age, 53 years). Two radiologists reviewed all chest CT images in terms of disease extent, distribution of disease (focal, geographic, diffuse) and zonal distribution of disease (upper, middle, lower, random). We also semi-quantitatively graded the extent and severity of centrilobular nodules, consolidations, ground-glass opacity (GGO), reticulations, and areas showing crazy-paving appearance. Results: PAP involved more lung segments than ELP (14.9p4.2 versus 10.1p4.6, p=0.015). PAP showed geographic and random distribution whereas ELP showed mid-lower lung predominance. ELP showed consolidation and centrilobular nodules more frequently than PAP (p=0.0011). Centrilobular nodules were present in 83% of patients with ELP, 14% of patients with PAP (p 0.05). There were no statistical differences between the two diseases in terms of areas showing GGO, reticulation, crazy-paving appearance. Conclusion: The extent of disease, distribution of lung abnormality and presence of centrilobular nodules could be useful in differentiating between PAP and ELP.

B-672

10:39

HRCT correlation with galactomannan levels in neutropenic immunocompromised patients T.D. Suaris, S.R. Doffman, S.J. Vinnicombe, S.G. Agrawal, S. Ellis; London/UK ([email protected]) Purpose: To assess correlation between Galactomannan (GM) detection in bronchoalveolar lavage (BAL) ﬂuid and early HRCT changes in invasive pulmonary aspergillosis (IPA) in immunocompromised neutropenic patients. Methods and Materials: Neutropenic febrile patients at high risk for invasive fungal disease were recruited and had HRCTs carried out according to study protocol (febrile neutropenia for 72 hours, unresponsive to broad spectrum antibiotics). HRCTs were assessed for abnormal ﬁndings, including well-described characteristic features of invasive fungal infection. CT-directed BAL was performed, where possible within 24 hours of CT. GM, a component of the cell wall of Aspergillus spp., was measured in the BAL ﬂuid. Analyses of the HRCT ﬁndings were correlated with the GM levels. Results: Twenty patients underwent HRCT and image-guided BAL. GM levels were categorised into 0.5, 0.5, 0.7 and 1.0. At present, there are no current agreed-upon cut-offs for positivity for GM in BAL ﬂuid. Ten patients had GM levels 0.5. The commonly identiﬁed HRCT abnormality in this group was poorly deﬁned nodules. One patient had GM level 0.5 demonstrating atelectasis only. Three patients had GM level 0.7. The most prevalent HRCT abnormality was poorly deﬁned nodules. Six patients had GM scores of 1.0. This group demonstrated a range of well-described features of fungal infection, such as halo sign and cavitation within consolidation. Poorly deﬁned or spiculated nodules were also seen. Conclusion: A BAL GM level of 1.0 demonstrated typical features of fungal infection on HRCT. Levels 1.0 demonstrated a number of non speciﬁc ﬁndings.

A

B

C

D

E

F

G

295

Monday

Purpose: Animal models of brain tumours may provide useful information to better understand human pathology. Aim of the present MRI study is to investigate, using a rat model, the development of brain gliomas in different anatomical locations. A multi-parametric MRI approach was used to characterize C6-glioma cells implanted in the caudate-putamen (CP) and in the frontal lobe. Methods and Materials: Eight male Wistar rats (300-350 g average weight) were anesthetized and a C6 cell suspension (106 cells in 10 μl) was stereotatically implanted in the right CP (four rats) or in the right frontal lobe (four rats). The rats were investigated at 7T using the following acquisition protocol: T1- and T2-weighted scans, Diffusion Weighted Imaging (DWI) using a Pulse Gradient Stimulated Echo (PGSTE) sequence with six “b values” (179-2594 s/mm2), and Diffusion Tensor Imaging (DTI) using PGSTE with gradients applied along seven non-collinear directions. T1- and T2, slice-direction Apparent Diffusion Coefﬁcient (ADC), mean diffusivity (MD), and fractional anisotropy (FA) values were measured from each tumor. Results: T2 and ADC values were signiﬁcantly lower in CP than in frontal lobe tumors, thus reﬂecting a higher vascularization in the former cases as conﬁrmed by histopathological examination. Signiﬁcant differences of FA and MD were also observed, and interpreted on the basis of histopathological assessment as indirect measures of preponderant edema or necrosis in the two different tumor locations. Conclusion: This preliminary study shows the ability of a multiparametric MRI approach to assess speciﬁc pathological features of brain gliomas, which are consistent with the anatomical location.

transmedullar vessels, 7T could superiorly or even solely detect irregular areas of signal intensity loss in the tumoral bed, which might represent areas of increased microvascularity. Especially, 7T T2* weighted images were superior in depicting the assumed disorganized vessels in high-grade gliomas and microhemorrhages in brain metastases. Conclusion: Our preliminary results indicate that 7T MRI is useful for the visualization of primary and secondary brain tumors. Ultrahigh ﬁeld MRI at 7T may ease the characterization of different tumor constituents and the depiction of microvascularity.

Scientific Sessions B-673

10:48

Thin-section CT ﬁndings in the ageing lung: A comparative study between 75- and 55-year-old individual controls S. Copley, K. Hawtin, D. Gibson, J. Hodson, A. Wenaden, A.U. Wells, D.M. Hansell; London/UK ([email protected]) Purpose: To document the thin-section CT ﬁndings in the lung parenchyma of older individuals and compare the prevalence of these observations with a younger age group. Methods and Materials: The CT requests of ambulatory outpatients referred for a clinically indicated CT of the abdomen or head were reviewed. Individuals over 75 years and controls under the age of 55 years were identiﬁed. Exclusion criteria included current smokers, known pulmonary disease, asbestos exposure, neoplastic disease and previous chemotherapy or radiotherapy. Fifty-three consecutive asymptomatic individuals (38 elderly and 15 controls) were prospectively identiﬁed. Prone thin-section CT of the thorax was performed. Two blinded observers independently scored images for the presence and extent of a subpleural reticular pattern, ground glass opacity, cysts, bronchial dilatation and bronchial wall thickening. Results: The observer agreement was good for subpleural reticular pattern (Kw 0.63), acceptable for ground glass opacity, cysts and bronchial dilatation (Kw 0.43-0.51) but poor for bronchial wall thickness (Kw 0.17). A subpleural reticular pattern was strikingly more prevalent in the elderly (80% observations) than in the younger group (10% observations) (p 0.0005). Cysts were seen in 29% of the older group compared with none in the younger group (p 0.0001). Ground glass opacity was seen in 38% of the older group compared with 7% of younger controls (p 0.001). Bronchial dilatation and wall thickening was also signiﬁcantly more frequent in the older group (p 0.0005, p 0.015, respectively). Conclusion: Several parenchymal ﬁndings on thin-section CT are signiﬁcantly more frequent in asymptomatic elderly compared with younger individuals.

B-674

10:57

HRCT imaging of lung parenchyma in chronic heart failure (CHF) patients: Correlation with pulmonary function tests E. Tavernaraki, S. Drakos, I. Nikas, L. Moulopulou, Z. Margari- Siafaca, C. Kostopoulos; Athens/GR ([email protected]) Purpose: Lung parenchyma functional and structural abnormalities have been reported in CHF due to multiple cardiac-lung interactions. Nevertheless, limited data exist concerning imaging of those abnormalities, while most relative studies deal with acute heart failure. We studied lung parenchyma HRCT ﬁndings in CHF patients. Methods and Materials: We performed lung HRCT in 49 patients (37M/12 F, mean age 57.3 yrs), with stable CHF, with no history of respiratory disease. HRCT score was calculated for each patient by semiquantitative evaluation of ﬁndings concerning extent and severity of ground glass, reticular pattern, honeycombing, subpleural plaques and bullae. We also performed pulmonary function tests (PFTs) namely spirometry, lung volume and diffusion capacity measurements, cardiac ultrasound and right sided cardiac catheterization. Results: Abnormal HRCT ﬁndings were demonstrated in 33 (67%) patients, while abnormal PFTs were found in 36 (73%) patients (p=NS). HRCT score was signiﬁcantly higher in patients with abnormal compared to normal PCWP (p 0.05). There was signiﬁcant negative correlation between HRCT score and FVC (r=-307, p 0.032), TLCO (r=-386, p 0.006) and positive correlation between HRCT score and PAP (r=516, p 0.014) and PCWP (r=418, p 0.05). Conclusion: HRCT lung parenchyma abnormalities commonly exist in CHF patients. Our results suggest that HRCT ﬁndings associate with severity of pulmonary and cardiac dysfunction.

B-675

11:06

Pulmonary ﬁbrosis in heavy cigarette smokers: Spectrum of HRCT ﬁndings A. Oikonomou, A.G. Spanoudaki, A. Pataka, S. Anevlavis, D. Bouros, P.K. Prassopoulos; Alexandroupolis/GR ([email protected]) Purpose: To describe the spectrum of HRCT ﬁndings in heavy cigarette smokers with evidence of pulmonary ﬁbrosis on HRCT. Methods and Materials: Eighty-one consecutive heavy cigarette smokers ( 20 packs/yr, mean packs/yr: 60) with dyspnea and no clinical evidence of infection, collagen vascular disease, drug toxicity or environmental exposure, underwent inspiratory/expiratory HRCT of the chest and pulmonary function tests. The presence and distribution of HRCT ﬁndings were analyzed. Results: Twenty-one male patients (mean packs/yr: 77) had evidence of pulmonary ﬁbrosis on HRCT characterized by a lower-zone predominant subpleural reticular pattern with traction bronchiectasis and bronchiolectasis. Emphysema was present

296

A

B

C

D

E

F

G

in 17 and ground glass in 18 patients. Honeycombing was present in six HRCT scans, being only minimal in four. In six cases there were hazy centrilobular micronodules and in only one case there was bronchial wall thickening associated with air-trapping on expiration. Annual follow-up HRCT examination showed mild deterioration in two patients. Patients with ﬁbrosis on HRCT showed a more restrictive functional proﬁle and lower carbon monoxide transfer coefﬁcient (p 0.05). Conclusion: Dyspneic heavy smokers may demonstrate ﬁbrosis on HRCT - characterized by subpleural reticular pattern - in about one fourth of the cases. Associated ﬁndings may include ground glass, hazy micronodules and minimal or absence of honeycombing. Emphysema coexists in the majority of the cases. HRCT ﬁndings in smokers’ related pulmonary ﬁbrosis resembles more an NSIP rather than a UIP pattern.

B-676

11:15

To compare ultra low dose and low dose CT in detecting the presence and assessing the extent of lung abnormalities in patients with interstitial pneumonias C. Beigelman-Aubry, R. Khayat, M.-J. Dallaire, P. Coulon, S. Trad, P. Grenier; Paris/FR Purpose: To compare ultra low dose and low dose CT in assessing the extent of lung abnormalities in patients with interstitial pneumonias. Methods and Materials: The study included 20 patients with interstitial pneumonias (8 collagen disorders, 2 drug induced, 1 asbestosis, 9 idiopathic). CT equipment was a Phillips Brilliance 64. CT scans were performed with thin collimation, and reconstructed with a lung kernel. The dose parameters were performed on supine position at 120 kV with a dose modulation system (mean DLP: 348 mGyxcm (257-476 mGyxcm), or mAs at the value of the weight of the patient (mean: 179 mGyxcm (105-263 mGyxcm). A second acquisition was performed in prone position, at 120 kV and between 11-30 mAs (mean DLP: 51 mGyxcm (22-84 mGyxcm) and reconstructed with a soft kernel. Five most representative slices of the disease were selected, in the right lung, on the low dose exam. Five slices at the same anatomic level were selected in ultra low dose CT. Two blinded radiologists independently scored the pulmonary ﬁbrosis overall extent, ground glass opacity, reticulation and honeycombing. Results: The interclass correlation coefﬁcient values for pulmonary ﬁbrosis overall extent, GGO %, reticulation % and honeycombing % were respectively 0.82, 0.79, 0.54 and 0.75 for radiologist 1, and 0.84, 0.85, 0.77 and 0.75 for radiologist 2.The interobserver interclass correlation coefﬁcient for pulmonary ﬁbrosis extent, GGO %, reticulation % and honeycombing % were, respectively, 0.74, 0.61, 0.63 and 0.72 for low dose CT and 0.91, 0.84, 0.62 and 0.85 for ultralow dose CT. No statistical difference was observed between overall extent of disease, GGO, reticulation and honeycombing % in low dose versus ultra low dose for each observer. No interobserver statistical difference was observed in ultra low dose. Conclusion: No statistical difference between low dose and ultra low dose was observed in the assesment of pulmonary ﬁbrosis.

B-677

11:24

Incidence of excessive tracheal collapsibility in moderate to severe COPD patients: Automatic quantiﬁcation of tracheal area in cine-CT M. Owsijewitsch, S. Ley, A. Mehndiratta, J. Ley-Zaporozhan, D. Simon, F.L. Giesel, C.P. Heussel, H.-U. Kauczor; Heidelberg/DE ([email protected]) Purpose: COPD is characterized by irreversible airﬂow limitation in the peripheral airways. The potential contribution of central airway collapsibility is unclear. The purpose of our study was to determine the incidence of excessive tracheal collapsibility during forced expiration in patients with COPD by means of cine-CT. Methods and Materials: 19 patients with moderate to severe COPD (GOLD Stages III-IV) and without previously diagnosed tracheomalacia underwent cine-CT of the trachea (4 cm above the carina). The maximal change of the tracheal cross sectional area (CSA) during a forced expiration maneuver was measured. An in house developed tool for automatic tracheal area analysis on cine-CT sequences based on a MeVisLab development environment was used. We considered a decrease in CSA of 50% or more as indicative of excessive tracheal collapsibility. The individual CSA decrease was compared to FEV1. Results: The mean decrease in CSA was 60% (p12). 14 of 19 (74%) patients showed a decrease of 50%. A trend towards higher tracheal collapsibility was observed in GOLD IV patients (p=0.4). There was only moderate correlation of the decrease of CSA with the FEV1 (r=0.33). Conclusion: A considerable fraction of 74% of patients with moderate to severe COPD showed relevant tracheal collapsibility of more than 50% during forced expiration. However, this increased collapsibility was not a major factor of airﬂow obstruction as determined by FEV1.

Scientific Sessions B-678

11:33

Respiratory gated chest CT in pulmonary compromised patients: Correlation with lung function tests P. Rogalla, N. Rogalla, B. Schmidt, B. Hamm, A. Lembcke, P. Hein; Berlin/DE ([email protected]) Purpose: To evaluate respiratory gated CT of the lung in comparison to lung function tests in patients who were potential candidates for volume reduction surgery. Methods and Materials: Twenty-three pulmonary with clinically known emphysematous disease underwent a 64-multislice-CT (Aquilion 64, Toshiba) with a scanning pitch of 5.4-9.8, depending on the respiratory rate. The respiratory cycle was simultaneously recorded by means of air-ﬂow detection (Breas Medical). Retrospective image reconstruction was performed similar to cardiac CT at every 10% of the respiratory loop, resulting in 10 complete volumetric datasets at 10 equidistant time points. All images were transferred onto a PC for calculation of the area-under-the-curve (AUC, [ms*HU]), density difference (delta D, [HU]), time difference (delta T, [ms]), delta D over delta T [HU/ms], and dynamic images of the moving lung. Static color-coded maps were calculated as also the tidal volume and residual capacity, and compared to lung function tests (LFT). Results: In 21 patients, complete datasets could be successfully reconstructed. Two patients were not able to comply with the breathing commands with subsequent missing axial slices at several table positions. Average tidal volume in CT and LFT ranged from 0.7 to 1.8 and 0.9 to 2.3 l, respectively (r=0.72), residual capacity from 2.2 to 4.4 and 2.1 to 4.8, respectively (r=0.82). Interdisciplinary interpretation of the color-coded maps with dynamic movies served as a basis for selection of surgical candidates. Conclusion: Respiratory gated chest CT allows for combined morphologic and functional image analysis with good correlation to lung function tests.

B-679

and xenon image were obtained. Time-density curves of four images were plotted during wash-in and wash-out. The root mean square error (RMSE) of four images was obtained in order to compare the ﬁtting error among them. Results: Xenon image was successfully obtained in all subjects. Using dynamic CT data, detailed mapping of the pulmonary ventilation was successfully achieved in all subjects. The average of RMSE of xenon images was 4.93 (p2.44), which was signiﬁcantly lower than those of the 80 kV, 140 kV, and weighted average images (8.37p8.09, 7.55p6.96, and 7.96p7.56, respectively) (p=0.01 in all comparisons). Conclusion: Using a dual-energy technique, xenon component could be extracted without any inﬂuence by lung volume. Dynamic and static regional ventilation function can be displayed with an exact match to the high resolution CT image.

10:30 - 12:00

Room R2

Cardiac

SS 1403 MR/CT in treatment evaluation Moderators: A. Jacquier; Marseille/FR S.D. Qanadli; Lausanne/CH

B-681

10:30

Feasibility of 3D cardiac tissue engineering for embryonic stem cells therapy after myocardial infarction: Preliminary results M. Lepetit-Coiffe, M. Hauwel, J.-L. Daire, J.-N. Hyacinthe, M. Jaconi, J.-P. Vallee; Geneva/CH ([email protected])

B-680

B-682

11:51

Xenon ventilation CT using a dual-energy technique of dual-source CT: Initial experience E. Chae1, J. Seo1, H. Goo1, N. Kim1, K.-S. Song1, S. Lee1, S.-J. Hong1, B. Krauss2; 1 Seoul/KR, 2Forchheim/DE ([email protected]) Purpose: To assess the feasibility of xenon-ventilation CT using dual-energy technique as a diagnostic tool for the evaluation of regional ventilation in the lung. Methods and Materials: Eight healthy volunteers and four patients having respiratory diseases were enrolled. The subjects inhaled 30% stable xenon for 1.5 min. A single level near carina was chosen for a baseline non-enhanced and dynamic enhanced scanning during wash-in (every 18 s) and wash-out (every 30 s) period. CT was performed using a dual-energy technique of dual-source CT, Somatom Deﬁnition (Siemens, Forchheim, Germany), with 40/170 eff. mAs at 140/80 kV. For each dynamic scanning, four images of 80 kV, 140 kV, weighted average,

10:39

Cardiac MR (CMR) detection of intramyocardial haemorrhage (IMH) following primary percutaneous transluminal coronary angioplasty (PTCA) and rescue PTCA (R-PTCA) M. Francone, F. Calabrese, M. Mangia, P. Lucchesi, L. Agati, F. Fedele, C. Catalano, R. Passariello; Rome/IT ([email protected]) Purpose: R-PTCA is recommended in patients (pts) with acute myocardial infarction (AMI) after unsuccessful thrombolysis. Procedure is associated with high risk of IMH. CMR represents method of choice to detect its presence in vivo showing paramagnetic susceptibility effect of deoxyhemoglobin on T2w images. Present study was designed to evaluate incidence, CMR features and functional behaviour of IMH comparing two groups of pts treated with primary PTCA vs R-PTCA. Methods and Materials: Sixty pts with AMI underwent coronary reperfusion with primary PTCA (n=42) or R-PTCA (n=18). In all cases CMR protocol including TSE T2w black-blood, 1st-pass and delayed enhancement sequences was performed

A

B

C

D

E

F

G

297

Monday

Purpose: The aim of this study was to evaluate the modiﬁcations of the lung in patients treated with endobronchial valve for severe emphysema. Methods and Materials: Fifteen patients were studied before and 7, 30 days, 3 and 6 months after the endobronchial valve placement by a low-dose chest MDCT. The volume of the treated lobe and of both lungs was calculated by using a dedicated platform to establish the modiﬁcations of the parenchyma before and after treatment. Volume calculations were also compared with the results of lung function tests by using linear correlations. Results: The bronchus of the most affected lobe was chosen to place the valves. Volume measurement before treatment varied between 2500 and 4200 cc for each lung. Corresponding measurements (Total Pulmonary Capacity) were assessed by pletismography and compared by CT volume with a perfect agreement (r: 0.97). The treated lobe showed a volume reduction that varied between 10 and 37%. The other lobe(s) in the treated lung showed a re-expansion between 7 and 28%. Correlation with lung function tests demonstrated reduction of FEV1 and an increase of vital capacity (VC) after treatment. Linear correlation was found between lung volume reduction and modiﬁcations of functional tests in all patients. Conclusion: Endobronchial treatment appears a new promising option in the treatment of emphysema. The air vesiculation from the most affected lobe to other functional zones of the lung can improve the pulmonary elastic recoil and increase pulmonary function. MDCT conﬁrms the importance of this therapeutic option.

Purpose: To study, using MRI, the feasibility of 3D-cardiac tissue engineering for the cell therapy of myocardial infarction (MI) in rats. Methods and Materials: 3D-cardiac tissue consisted of ﬁbrin matrix incorporating 300,000 iron-oxide-particles-magnetofected cardiac-committed mouse embryonic stem cells (FeESC). Treated and untreated rats (N=5/group) were investigated under 1.5 Tesla MR magnet six hours after induction of MI by left coronary occlusion, 2 days (D2), 7 days (D7)and 45 days (D45) later. T2w-FFE images (TE/TR/ FA=7 ms/350 ms/50) were performed for iron oxide particles repairing. Global and regional cardiac functions were evaluated based on a FFE-cine sequence (acquired pixel size=0.4x0.4x2 mm3) and C-SPAMM-TAG preparation segmented-cine FFE sequence (interTAG-spacing=1.25 mm). Delayed enhancement-IR T1w-FFE sequence (TE/TR/TI/FA=7.6 ms/12 ms/300 ms/45) permitted infarct zone evaluation. Histology was processed 50 days later. Measures of wall thickening (WT) and circumferential strains were performed in six sectors and statistically compared in MI sectors for Untreated and Treated groups. Results: 3D-cardiac tissue was visible as hypointense signal due to iron-oxideparticles: for Treated group, the signal decreased in size but was still visible at 45 days. Global cardiac parameters presented same time evolution but at D45, end-systolic (ESV) and end-diastolic (EDV) volumes were signiﬁcantly different between Untreated and Treated groups. At D45, signiﬁcant conservation of the cardiac function was observed in the Treated versus Untreated group (p=0.02). The migration out of the 3D-ﬁbrin matrix of mouse ESC was conﬁrmed by histology as the Prussian-blue-stained-MEF2-positive-cardiomyocytes were observed in the MI. The MI size was identical between Untreated and Treated groups and remained stable over time. Conclusion: Under clinical 1.5 T MR scanner, improvement of cardiac function using 3D cardiac-tissue-engineering for the ES-cells therapy was demonstrated in treated small animals.

11:42

MDCT assessment of lung volume reduction in patients undergoing bronchial stenting for treatment of pulmonary emphysema: Correlation with respiratory tests F. Fraioli, L. Bertoletti, M. Mennini, S. Vetere, C. Catalano, R. Passariello; Rome/IT ([email protected])

Scientific Sessions within one week after acute event and after three months. IMH was identiﬁed according to T2w signal and differentiated from myocardial edema (ME). Left ventricular ejection fraction (LVEF) and myocardial wall thickness (MWT) were also quantiﬁed. Results: Fourteen pts (23%) showed IMH as a central hypointense core with peripheral hyperintense rim on T2w CMR imaging and 44 (74%) pts showed homogeneous hyperintense signal consistent with ME; IMH was observed in 8/18 (44%) R-PTCA pts and 7/42 (17%) primary PTCA. Among R-PTCA group, infarct size and MWT were larger in pts with IMH (33p11 vs 19p15%; p=0.097 and 9.1p3.5 vs 6.9p1.9 mm; p 0.05, respectively). Perfusion defects were observed in 14/14 (100%) cases with IMH and in 11/44 (25%) with ME. At follow-up, LVEF showed less recovery in pts with IMH (43p11-46p10%; p=0.47 vs 46p3-50p10%, p 0.05). Conclusion: IMH occurs more frequently in R-PTCA group as a result of previous thrombolysis. Although further studies with long term follow-up will be necessary to understand its real clinical signiﬁcance, it may represent a negative predictor for LV functional recovery.

B-683

10:48

Absolute quantiﬁcation of myocardial blood ﬂow from rest perfusion MRI can provide improved sensitivity for predicting myocardial viability in patients early after myocardial infarction M. Nagata, H. Sakuma, T. Kurita, M. Ishida, N. Ishida, Y. Mikami, H. Nakajima, M. Ito, K. Takeda; Tsu/JP ([email protected]) Purpose: A previous study demonstrated that thickness of late gadolinium enhancement (LGE) in acute myocardial infarction (MI) is substantially larger than that in chronic state, causing potential underestimation of myocardial viability in acute state. The purpose of this study was to evaluate if quantitative assessment of regional myocardial blood ﬂow (MBF) can provide more accurate prediction of functional recovery in patients with acute MI. Method and Materials: Thirty patients underwent rest myocardial perfusion MRI and LGE MRI within one week after onset of MI. After correcting blood signal saturation with dual bolus method, blood input and myocardial output curves were analyzed with a Patlak plot method to determine MBF in 16 LV myocardial segments. All patients underwent balanced TFE cine MRI 6 months after onset of MI. Results: LGE MRI revealed MI in 207 of 480 segments in acute state. The averaged MBF in acute MI was 79.9p41.7 ml/min/100 g in segments with preserved systolic wall thickening (SWT) in chronic state, and was 43.3p23.0 ml/min/100 g in those with impaired SWT (p 0.01). The sensitivity and speciﬁcity of LGE MRI in predicting functional recovery were 85 and 77%. When segments showing MBF of 67 ml/min/100 g were considered as viable regardless of transmural extent of LGE, the sensitivity of CMR in predicting functional recovery substantially improved from 85 to 94%, with nominal reduction of the speciﬁcity to 72%. Conclusion: Quantitative measurement of regional MBF is an objective method that has high sensitivity in detecting segments with functional recovery in patients with acute MI.

B-684

10:57

Update on multidetector coronary CT angiography of coronary stents: Invitro-evaluation of 27 new stent-types D.C. Maintz, M. Burg, H. Seifarth, A. Bunck, M. Özgün, K.U. Juergens, R. Fischbach, W.L. Heindel; Münster/DE Purpose: To test a large sample of the latest coronary artery stent models using four image reconstruction approaches with respect to lumen visualization, lumen attenuation, and image noise in dual-source multidetector row CT (DSCT) in vitro. Methods and Materials: Twenty-seven different coronary artery stents (17 steel, six cobalt-chromium, two tantalum, one magnesium) were examined in a coronary artery phantom (vessel diameter 3 mm, intravascular attenuation 250 HU, extravascular density -70 HU). Stents were imaged in axial orientation with standard parameters: 2x32x0.6 collimation, 680 mAs, 120 kV, rotation time 0.33 s. Image reconstructions were obtained with four different convolution kernels (soft, medium-soft, standard high-resolution, stent-dedicated). To evaluate visualization characteristics of the stent, the lumen diameter, intraluminal density and noise were measured. Results: The stent-dedicated kernel offered best average lumen visualization (54p8.6%) and most realistic lumen attenuation (219p44 HU) at the expense of increased noise (23.9p2 HU) compared to standard kernels (p 0.001 for all). The magnesium stent showed the least artifacts with a lumen visibility of 90%. The majority of stents (74%) exhibited a lumen visibility of 50-59%. In only six stents (22%) less than half of the lumen was visible. Conclusion: Stent lumen visibility largely varies depending on the stent type. Magnesium seems the most favourable stent material with regard to CT imaging

298

A

B

C

D

E

F

G

when compared to the more common materials steel, cobalt-chromium or tantalum. The majority of these stents exhibit a lumen visibility of 50-59%, a degree that still may be problematic for routine clinical stent assessment.

B-686

11:06

Are there differences between the four 64-multidetector computed tomography vendors for the evaluation of coronary stents? An in-vitro study F. Wolf1, G. Feuchtner2, P. Homolka1, H. Langenberger1, A. Stadler1, T.R. Bader1, A. Dirisamer1, M. Weber1, J. Lammer1, C. Loewe1; 1Vienna/AT, 2Innsbruck/AT (ﬂ[email protected]) Purpose: To compare the performance of the four leading 64-multidetector computed tomography (CT) vendors Siemens (Siem), Philips (Phi), GE and Toshiba (Tosh) for the evaluation of coronary stents by using a dedicated combined cardiac/ chest phantom. Methods and Materials: Twelve different coronary artery stents (diameters 2.6-5 mm) were placed into the drillings of a combined cardiac/chest phantom. The drillings were ﬁlled with contrast media diluted to a CT attenuation of 300 HU. The phantom was scanned with the 64-slice CT scanner of Siemens, Philips, Toshiba and GE by using a comparable dedicated coronary CT angiography scan protocol. Multiplanar reformations were evaluated for artiﬁcial lumen narrowing, CT-attenuation values inside compared to outside the stent and image noise inside the stent. Results: The artiﬁcial lumen narrowing was not different between the four vendors (ToshSiemPhilGE; 42 vs 41 vs 39 vs 37%, respectively). The image noise inside the stent was signiﬁcantly higher for Siem (48.5 SD) compared to the other three vendors (PhiGETosh; 25.8 vs 26.3 vs 20.9 SD, respectively, p=0.002). The difference in HU outside and inside the stent was signiﬁcantly lower for Siem (4 HU) when compared with the three other vendors (ToshGEPhi; 128.7 vs 102 vs 83.1 HU, respectively, p 0.001). Conclusion: The artiﬁcial in-stent lumen narrowing is not different between the four vendors. The image noise inside the stent was higher for Siemens when compared with the other vendors but lowest for Toshiba. Siemens provides a signiﬁcantly less artiﬁcal in-stent lumen attenuation when compared with the other vendors.

B-687

11:15

Assessment of myocardial perfusion reserve in patients with X-syndrome by dypiridamole rest-stress ﬁrst-pass magnetic resonance imaging (MRI): Preliminary results A. Esposito, F. De Cobelli, S. Ravelli, F. Arioli, G. Fragasso, M. Cera, G. Bassanelli, E. Belloni, A. Maseri, A. Del Maschio; Milan/IT ([email protected]) Purpose: Cardiac X-syndrome is characterized by effort angina with normal epicardial coronary artery. The pathogenesis of chest pain remains unclear. Two previous MRI studies (Panting JR et al. NEJM 2002. Vermeltfoort IA et al. EHJ 2007) investigated the presence of myocardial perfusion defects during adenosineinduced-stress, getting opposite results. Our aim was to evaluate the myocardial perfusion reserve of X-Syndrome patients by rest-stress MR-perfusion; unlike previous works we use dypiridamole as stressing drug, because its long half-life should assure a more stable stress during MR-perfusion study. Methods and Materials: MR-perfusion was performed in six patients with Xsyndrome at rest and during dypiridamole (0.56 mg/kg in 4 min+0.28 mg/kg in 2 min) induced stress, on a 1.5 T magnet (Philips). Perfusion images were acquired during ﬁrst-pass of 0.05 mmol/kg of Gadolinium-DTPA (Magnevist; Bayer-Schering) injected at 4 mL/s. Semi-quantitative analysis of perfusion images were based on an 18-segment model to obtain normalized-upslope of myocardial signal enhancement (upslope) and myocardial-perfusion reserve index (MPRI). Results: Chest pain occurs in all patients during dypiridamole infusion. Five of them had a clear diffuse hypoperfusion of subendocardial layer, at qualitative analysis. At semi-quantitative analysis upslope increased during stress in the subepicardium (rest: 0.23p0.08 vs stress: 0.29p0.17; P 0.001) and did not change signiﬁcantly in the subendocardium (rest: 0.27p0.08 vs stress: 0.29p0.13; P=n.s). As consequence MPRI was signiﬁcantly higher in the subepicardium when compared with subendocardium (1.30p0.68 vs 1.12p0.39; P 0.05). Conclusion: Our preliminary results conﬁrm the impairment of subendocardial perfusion reserve in X-syndrome patients suggesting microvascular ischaemia could be the cause of angina.

Scientific Sessions B-688

11:24

Autologous bone marrow stem cell mobilization induced by granulocyte colony-stimulating factor after STEMI undergoing late revascularization: Final results from the G-CSF-STEMI trial A.M. Huber, A.-E. Werle-Rüdinger, B.J. Wintersperger, K.U. Bauner, T. Sandner, M. Engelmann, W. Franz, M.F. Reiser; Munich/DE ([email protected]) Purpose: The purpose of this investigator-driven, prospective, randomized, double-blinded, placebo-controlled phase II study was to determine the effects of G-CSF on the improvement of myocardial function in patients with delayed PCI for therapy of STEMI. Methods and Materials: Forty-one patients with late revascularized subacute STEMI were treated either with G-CSF or placebo over ﬁve days after successful PCI. Primary end-points were change of global and regional myocardial function from baseline (one week after PCI) to three months after PCI assessed by magnetic resonance imaging (MRI). Secondary end-points consisted of characterization of mobilized stem cell populations, assessment of safety parameters up to 12 months including 6-month angiography, as also myocardial perfusion assessed by MRI. Results: LV-EF from baseline (one week after PCI) to three months improved in both groups and G-CSF was not superior to placebo (Delta-EF) 6.2p9.0 vs 5.3p9.8%, p=0.77). A slight but non-signiﬁcant improvement of regional function occurred in both groups. Granulocyte colony-stimulating factor resulted in mobilization of endothelial progenitor cell populations and was well tolerated with a similar rate of target lesion revascularization from in-stent restenosis. In both groups major adverse cardiovascular events occurred in a similar frequency. G-CSF resulted in signiﬁcant improvement of myocardial perfusion one week and one month after PCI. Conclusion: G-CSF treatment after PCI in subacute STEMI is feasible and safe. However, patients do not beneﬁt from G-CSF when PCI is performed late. G-CSF improves perfusion of the infarcted area, which may reﬂect enhanced neovascularization.

B-689

11:33

Evaluation of coronary stent patency and in-stent restenosis with dual source CT coronary angiography without heart rate control D. Oncel, G. Oncel, A. Tastan, B. Tamci; Izmir/TR ([email protected])

B-690

11:42

Post infarction ventricular septal defects: The novel use of computed tomography and magnetic resonance imaging to guide clinical management O.R. Tann, G. Morgan, R. Martin, M. Turner, M.C.K. Hamilton; Bristol/UK Purpose: Post-myocardial infarction ventricular septal defect (PIVSD) is an uncommon but devastating event. Surgical closure is associated with a high mortality. Early results of transcatheter closure studies compare favourably with surgery, but are associated with residual shunts. Traditional evaluation is with echocardiography alone. We hypothesised that computed tomography (CT) and magnetic resonance imaging (MRI) may better delineate the morphology of PIVSDs, and help guide appropriate treatment. Methods and Materials: Between July 2006 and September 2007, nine patients were referred to a single centre for closure of PIVSD. All patients were assessed

10:30 - 12:00

Room R3

Interventional Radiology

SS 1409 Tumor ablation/Liver regeneration Moderators: H.A. Deutschmann; Graz/AT I. Rozanes; Istanbul/TR

B-692

10:30

Experimental and clinical study on the safety of ultrasound-guided percutaneous microwave ablation for hepatic tumors adjacent to gastrointestinal tract, gallbladder and diaphragm P. Liang, Q. Shao, B. Dong, H. Liang, Y. Wang, W. Shi, Y. Gao; Beijing/CN ([email protected]) Purpose: To investigate the safety of ultrasound-guided percutaneous microwave ablation (MWA) for the treatment of hepatic tumors adjacent to gastrointestinal tract, gallbladder and diaphragm. Methods and Materials: (1) Animal study: MWAs were performed in hepatic tissue abutting the gastrointestinal tract, gallbladder and diaphragm in 23 healthy adult mongrel dogs. Temperature was monitored at the interface between the ablation zone and those structures. Thermal injuries were assessed immediately after MWA to screen out the appropriate treatment temperatures. MWAs were performed at the appropriate temperatures and the dogs were observed for 28 days. (2) Clinical Study: A total of 72 patients with hepatocellular carcinoma adjacent to the gastrointestinal tract, gallbladder and diaphragm received MWAs. Thermocouples were placed at sites abutting those structures during MWA. Results: (1) In animal study, no obvious thermal injuries were observed in those structures after 4-8 minutes’ hyperthermia at appropriate temperatures (55~65 °C for bowel, 55~60 °C for gallbladder wall and 50~60 °C for diaphragm). (2) In clinical study, the highest interface temperatures were 52~56 °C for gastrointestinal tract, 55~60 °C for gallbladder wall and 60 °C for diaphragm, which can ensure treatment adequacy while avoiding possible complications. The local recurrence rate was 9.7% (7/72) in a follow-up of 3-24 months. Conclusion: Under thermal monitoring, it is safe to perform MWA for hepatic tumors adjacent to the intestinal tract, gallbladder and diaphragm, which can obtain complete tumor necrosis and minimize potential complications.

B-693

10:39

Long term experience after laser ablation of 2862 colorectal carcinoma metastases in the liver in 974 patients M.G. Mack, K. Eichler, T. Lehnert, S. Zangos, C. Müller, T.J. Vogl; Frankfurt a. Main/DE ([email protected]) Purpose: To evaluate long term results after 14 years for MR-guided LITT for the treatment of colorectal liver metastases. Methods and Materials: MR-guided LITT was performed in 974 patients (651 male, 323 female, mean age 61.7 years) with 2862 metastases of colorectal cancer between 1993 and 2007. Survival rates (SR) were calculated using the Kaplan-Meier method. 30.7% of patients had recurrent metastases after surgery, 38.4% metastases in both liver lobes, 14.6% refused surgical resection, 3.4% had contraindications for surgery and 12.9% had metastases at difﬁcult localization for surgery. 35.4% of the metastases were smaller or equal to two cm, 31.6 were between two and three cm, 17.8% were between three and four cm and 15.2% were larger than four cm. 75.1% were treated in curative intention, 24.9% in palliative intention.

A

B

C

D

E

F

G

299

Monday

Purpose: The DSCT offers improved temporal resolution and permits better visualization of coronary vessels without heart rate control. We aim to determine the diagnostic performance of DSCT in the evaluation of coronary stent patency to search if improved temporal resolution may also help in better visualization of the stents. Methods and Materials: Institutional Review Board approval and informed consent were obtained. 35 consecutive patients (10 women, 25 men; mean age 62 years) with 48 stents were examined prospectively. Image quality and patency of the stents were evaluated by two radiologists blinded to the results of invasive coronary angiography (ICA). In-stent restenosis was deﬁned as 50% narrowing. Results: All stents were evaluated as assessable for diagnosis with DSCT and in 85% (41/48) of the stents, the image quality was good. Only two patent stents were misdiagnosed as stenosis. All other stents with stenosis and occlusion were correctly diagnosed. The sensitivity, speciﬁcity, positive and negative predictive values and accuracy for DSCT to detect in-stent restenosis and occlusion were 100, 94, 89, 100 and 96%, respectively. With McNemar test, no statistically signiﬁcant difference between DSCT and ICA was found. The kappa indices demonstrated excellent intraobserver and interobserver agreement. Conclusion: The high temporal resolution of DSCT has a positive impact on the evaluation of coronary stents without heart rate control. Further conﬁrmation of our preliminary results may broaden the clinical indications for CT angiography to be used as a screening test for the exclusion of in-stent restenosis.

with echocardiography and cross-sectional imaging. The choice of either CT or MRI was dependant on the clinical status of the patient. Percutaneous device closure was considered to be the preferred treatment choice, except in cases where the VSD was considered too large. Results: Four patients had CT alone, and four patients had MRI alone. One patient had both CT and MRI. Using these techniques, we have found that most PIVSDs are ovoid, and have complex morphology. This has aided more accurate closure device sizing and deployment. Those patients for whom device closure is inappropriate have also been recognized and referred for surgery. Six patients had successful transcatheter closure, without residual shunting. One patient had surgery, and two patients died without treatment. Conclusion: PIVSDs are often difﬁcult to delineate using echocardiography. Cross-sectional imaging can guide appropriate treatment, help size percutaneous occluders, and avoid attempted transcatheter intervention in patients with unsuitable defect morphology.

Scientific Sessions Results: Mean SR for all potentially curative treated patients was 3.6 years (95% conﬁdence interval: 3.3-3.8 years, maximum 13.1 years, 1-year survival 95%, 2-year survival 72%, 3-year survival 46%, 5-year survival 17%). The mean survival of patients who refused resection and were treated with LITT was 4.3 years, which is superior to data published for surgical resection. The mean survival of patients who had already partial liver resection before was 3.8 years; in patients with bilobar liver metastases the mean survival was 3.0 years. Patients with general contraindication for surgery had a mean survival of 2.9 years, and patients with difﬁcult localization for surgery a mean survival of 3.0 years. Conclusion: MR-guided LITT is a highly effective technique to treat patients with liver metastases of colorectal carcinoma.

B-694

10:48

Oligonodular liver metastases post laser-induced thermotherapy (LITT): Long-term results for colorectal carcinoma vs breast cancer T.J. Vogl, A. Portugall, S. Meis, K. Eichler, T. Lehnert, M. Mack; Frankfurt a. Main/DE ([email protected]) Purpose: To volumetrically analyze liver metastases of colorectal carcinoma and breast cancer after thermal ablation in a long-term evaluation using contrastenhanced MRI. Methods and Materials: 40 patients with liver metastases of colorectal cancer (n=20) (CRC) and breast cancer (n=20) (BC) were retrospectively analyzed with regard to initial tumor volume and thermal-induced necrosis post MR-guided laser-induced thermotherapy (LITT). All patients presented with oligonodular liver metastases and underwent follow-up with contrast-enhanced MRI in a 3-month interval for a period of at least 3 years. No concomitant oncological therapies were performed. Results: Volumetric MRI revealed 40 metastases with an initial tumor volume b5 ml (x=1.75), 9 metastases: 5-20 ml (x=12.35) and 8 metastases 20 ml (x=50.57). For CRC metastases (mean volume 8.2 m) the necrosis was x=499% of the initial volume, for BC (mean volume: 12.4 m) 604%. After 3 months there was a mass reduction of the necrosis (CRC) of 48.6% (6 months: 63%, 12 months: 70.2%, 36 months: 92.2%). In BC metastases the 36-month control revealed a lower volume reduction of 80.61%. Conclusion: MR-guided LITT induces a high volume of necrosis, and the highest reduction in the necrosis volume is detected in the ﬁrst year with lower values in the following period. The reduction in necrosis is greater in metastases from CRC.

B-695

10:57

The prevalence of the extrahepatic collateral artery supply for the tumor thrombi of hepatocellular carcinoma in inferior vena cava and its determined factors I. Lee, J. Chung, H.-C. Kim, H. Jae, J. Park; Seoul/KR ([email protected]) Purpose: To retrospectively evaluate the prevalence of extrahepatic collateral artery (ExCA) supply for the tumor thrombi of hepatocellular carcinoma (HCC) in the inferior vena cava (IVC) and to assess its determinant factors. Methods and Materials: From 1998 to 2007, 82 patients with IVC tumor thrombi on CT underwent angiographic evaluation of their ExCA supply and two radiologists interpreted their angiograms. As determinant factors for ExCA supply to IVC tumor thrombi, we included sex, age, Child-Pugh class, previous history of chemoembolization, tumor factors (size, number, growth pattern), the distance from primary tumor to IVC, portal vein invasion and the extent of IVC thrombi. The extent of thrombi was evaluated by using the following classiﬁcation: (a) whether the thrombi occupy more than half of IVC lumen, (b) whether the thrombi completely ﬁll and distend IVC lumen, and (c) whether the thrombi extends into right atrium. Univariate analysis and multiple logistic regression analysis were used. Results: Fifty-four (65.9%) had ExCA supply: 47 from right inferior phrenic artery, four from right adrenal artery, two from right internal mammary artery and one from right renal artery. The presence of ExCA supply for IVC tumor thrombi showed the statistically signiﬁcant relationship with the previous history of chemoembolization (p=0.001, odds=22.4) and the distension of IVC by tumor thrombi (p=0.005, odds=9.1). Occupancy of more than half of the IVC lumen by tumor thrombi had marginal statistical signiﬁcance (p=0.079, odds=4.71). Conclusion: IVC tumor thrombi of HCCs are commonly supplied by ExCAs. Its determinant factors are the previous history of chemoembolization and the extent of IVC tumor thrombi.

300

A

B

C

D

E

F

G

B-696

11:06

Radiofrequency ablation of renal tumors: Mid-term follow-up in 40 patients O. Eker, V. Perot, X. Coutouly, M. Sesay, J.M. Ferriere, P. Ballanger, N. Grenier; Bordeaux/FR ([email protected]) Purpose: To evaluate the efﬁciency of radio frequency ablation (RFA) of renal tumors, to inventory the complications and to estimate the effect on renal function. Methods and Materials: Between November 2002 and January 2007, 40 patients presenting with a renal tumor and high surgical risk or contra-indication to nephron-sparing surgery were treated by RFA with a percutaneous CT-guidance (n=38) or a laparoscopic US-guidance (n=2). Follow-up was clinical, biological and radiological (postoperatively, at two months, six months, one year, then annually). Absence of residual tumor or tumor recurrence was based on the absence of contrast enhancement. Results: Fifty lesions were treated. Their mean diameter was 26.6 mm (range 12-50 mm, median 25 mm). Forty were considered as small (b35 mm) and 10 as large ( 35 mm). Mean follow-up was 22 months. Ablation was complete in all small lesions and in 4/10 large lesions. No recurrence was noted after complete ablation. Complications were: two urinomas, one caliceal stenosis, ﬁve haematomas (one compressive), two infections, one acute pancreatitis, one local hypoesthesia, one small pneumothorax and one segmental renal infarct. Surgery was never necessary. Twenty-eight patients presented a moderate and transient increase of serum creatinine immediately after treatment. Conclusion: RFA can be considered as a simple and effective treatment for renal tumors in patients with contraindication for surgery, particularly for small lesions, without signiﬁcant effect on renal function.

B-697

11:15

Initial experimental study of cryoablation on rabbit’s VX2 brain tumor model under MRI guidance C. Li; Jinan/CN ([email protected]) Purpose: To evaluate the feasibility of cryoablation on rabbit’s VX2 brain tumor model under MRI guidance, analysis correlation among MRI, pathology, and therapeutic effect. Methods and Materials: Cryoablations were performed on 22 New Zealand white rabbits with VX2 brain tumor models. A 0.23 T open MRI device mounted with optical tracking system was used in procedural imaging and instrument guidance. Cryotherapy was performed by using an MR-compatible, argon-based cryotherapy system with 1.47 mm diameter cryoneedle. Eight rabbits were divided into A and B groups and cryotherapy was performed when diameter of brain tumor grew up to 8 mm in size. Freezing time lasted for 5 and 10 min, respectively. In each group one rabbit was executed immediately, 7 days, 14 days after procedure. The other 14 rabbits were divided into C and D groups. Rabbits of C group underwent cryotherapy with freezing time 10 min while those of group D were not treated. The survival times were recorded. If the rabbit did not die until two months after VX2 tumor tissue inoculation, survival time was taken as censored value. Results: The size and form of brain tumor is almost uniform. The correlation between MRI and pathologic appearance of tumors is good. The survival time of C group was longer than that of D group. Conclusion: The animal models of brain tumor with VX2 tumor tissue inoculation are stable. Cryoablation for brain tumor is safe and feasible.

B-698

11:24

Liver hypertrophy after portal vein embolization (PVE) before major hepatectomy: Our experience in 60 consecutive patients G.C. Anselmetti, A. Manca, G. Chiara, L. Cellini, D. Campanella, D. Regge; Candiolo/IT ([email protected]) Purpose: Evaluation of liver hypertrophy achieved by PVE to allow major hepatectomy. Methods and Materials: 60 patients (pts) (34 male), 40 with primary tumor (32 Klatskins’, six hepatocellular carcinoma) and 20 with metastases (colon cancer) non-resectable hepatobiliary malignancies, in whom preoperative spiral CT showed a remnant liver volume (RLV) 25-30%, underwent PVE under ﬂuoroscopic guidance and deep sedation. In 46 patients (Group A) PVE was performed in the right portal branches whereas in 14 (Group B) was extended to the IV segment branches. PVE was performed from 28 to 35 (mean 31.8p9.3) days before surgery injecting 1:5 mixture of cyanoacrylate (Glubram2) and iodinized oil (Lipiodol UF). RLV was measured by volumetric CT 4 weeks after PVE. Results: PVE was feasible in all patients without major complications; in two patients a small amount of embolizing material accidentally overﬂow in the left portal branch. RLV increase ranged between 2.2 and 125.5% (mean 52.7p28.4). Average

Scientific Sessions pre-procedural RLV of 305.9p115.1 cm3 signiﬁcantly increased to 455.5p159.4 cm3 (p 0.0001 paired t-test). Mean volume increase of segments II-III was similar in the two groups: group A 348.4p83.1 cm3 vs group B 391.2p78.0 cm3 (p=0.2). Major hepatectomy was feasible in 40 patients (66.7%). In 20 patients (33.3%) disease progression (17-28.3%) or insufﬁcient RLV hypertrophy (3-5%) did not allow surgery after PVE. Conclusion: PVE is effective in increasing the RLV before major hepatectomy. Embolization of segment IV portal branches could be avoided because a similar volume increase of segments II-III can be achieved by right PVE.

B-699

11:33

Role of preoperative portal vein embolization before partial or extended hepatectomy in cirrhotic liver: A compared and retrospective study L. Carpanese, G. Pizzi, G. Vallati, G.M. Ettorre, R. Santoro, M. Crecco; Rome/IT ([email protected]) Purpose: To evaluate the effectiveness and efﬁcacy of the trans hepatic preoperative portal vein embolization (PVE) for leading future liver remnant (FLR) to hypertrophy, as compared to surgical portal vein ligation before partial or extended hepatectomy and to assess the effectiveness of indications to PVE. Methods and Materials: A retrospective study of 47 pts who underwent partial and extended hepatectomy within 2004-2006 period was performed: 25 patients after PVE and 22 patients who did not previously undergo PVE. Both groups of patients were comprehensive of two homogeneous groups equally representative of patients with “healthy liver” and cirrhotic patients. The analysis evaluated the biological, morphological and volumetric changes shown after PVE and the post operative outcome and complications registered in the two groups of patients who underwent hepatectomy (with or without previous PVE). Results: PVE technique was well tolerated and improved an FLR hypertrophy of almost 30% in the subsequent 4-6 weeks. In patients with “healthy liver”, PVE adds real beneﬁt in extended hepatectomy only. In cirrhotic liver patients PVE effectively decreased the risk of extended hepatectomy (morbidity was 36 vs 83%), showed a less hospital stay in Intensive Care Unit (6p3 days compared with 15p11 days) and a whole hospital stay of 13p4 days compared with 30p15 days but did show a valid FLR hypertrophy. Conclusion: Since cirrhotic liver patients do not show a valid FLR hypertrophy after PVE, the extended hepatectomy is contraindicated; therefore PVE plays an important role in the selection of these patients.

B-700

11:42

Human growth factor (HGF) evaluation before liver intervention: Final results of a prognostic signiﬁcance for liver regeneration T. Lehnert, A. Gazis, K. Eichler, S. Zangos, M.G. Mack, T.J. Vogl; Frankfurt a. Main/DE ([email protected])

Monday

Purpose: To evaluate and compare the complexity of liver regeneration in correlation with the regeneration factor HGF, after liver resection and MR-guided laser induced thermotherapy (LITT). Methods and Materials: 100 patients were evaluated, 47 patients after surgical liver intervention and 53 patients after ablation (LITT). During a subsequent monitoring period of 18 months, the effects of therapy on the cytokine HGF were evaluated and compared to the initial, pre-interventional HGF value. In addition, the effects on liver volume after surgical and interventional radiological procedures were evaluated using unenhanced and contrast enhanced MRI and CT. Volumetric examination (CT-volumetry) of liver parenchyma pre- and post intervention was accomplished and the dynamics of liver regeneration and regional growth of liver segments were evaluated. Results: A signiﬁcant inﬂuence on the HGF process could be conﬁrmed on loss of liver parenchyma resulting form surgery or LITT induction. Pre-interventional HGF values 1500 pg/ml (n=49) led to a signiﬁcant increase of liver volume (97.13-584.07 ml). Patients with initial HGF values 2000 pg/ml (n=37) showed a decrease of liver volume (5.88-416.46 ml) after intervention. Pre-interventional HGF values between 1501 pg/ml and 1999 pg/ml (n=14) showed no deﬁnite reaction of liver parenchyma. All patients younger than 55 years showed stronger dynamics than 55+ year-old patients. Conclusion: Human growth factor (HGF) value, pre-interventional, has a signiﬁcant impact on the liver regeneration after MR-guided laser induced thermotherapy (LITT) and surgical resection. Pre-interventional low HGF values correlate with increased liver regeneration and stronger regeneration dynamics.

A

B

C

D

E

F

G

301

Scientific Sessions

302

A

B

C

D

E

F

G

Scientific Sessions

A

B

C

D

E

F

G

303

Tuesday

Tuesday, March 11

Scientific Sessions room A 2nd level

room B 2nd level

room C 2nd level

room E1 entr. level

room E2 entr. level

room F1 entr. level

room F2 entr. level

room G/H lower level

room I lower level

07:00

07:00

07:30

07:30

08:00

08:00

08:30

08:30

CC 1716 MRI The future is now! (p. 124)

09:00

09:30

SF 17a Special Focus Session CT and MR urography: Is intravenous urography dead? (p. 124)

RC 1701 Abdominal and RC 1710 Gastrointestinal Musculoskeletal Liver vasculature: Extra-spinal Radiologic interventional anatomy and procedures (p. 126) disorders (p. 125)

E3 1720 Interactive Teaching Session Detection and diagnosis of interstitial lung diseases (p. 127)

RC 1708 Head and Neck What you always wanted to hear about hearing loss (p. 127)

SF 17b Special Focus Session New radiation risk estimation: Can we relax? (p. 128)

RC 1711 Neuro Interventional neuroradiology (p. 129)

RC 1715 Vascular Imaging of critical limb ischemia (p. 129)

09:30

10:00

10:00

10:30

10:30 SS 1810 Musculoskeletal Tumors and marrow disease (p. 306)

11:00

11:30

12:30

registration

13:30

EPOS™- scientiﬁc exhibition

12:00

13:00

09:00

SS 1801a Abdominal Viscera HCC, still tantalising and bewildering (p. 308)

SS 1801b GI Tract Small bowel and appendix (p. 310)

SS 1803a Cardiac Coronary artery imaging (p. 312)

E3 1820 BI-RADS™ Olympics (p. 135)

SS 1807 Genitourinary Prostate imaging: Diagnosis, therapy and follow-up (p. 314)

SS 1802 Breast MR: Role in patient management (p. 316)

SS 1808 Head and Neck Skull base and temporal bone (p. 318)

SS 1815 Vascular Therapy, evaluation and outcome (p. 320)

11:00

11:30

12:00 JIIQ Junior Image Interpretation Quiz

12:30

13:00

Closing Ceremony

13:30

14:00

14:00

14:30

14:30

15:00

15:00

15:30

15:30

16:00

16:00

16:30

16:30

17:00

17:00

17:30

17:30

18:00

18:00

18:30

18:30

19:00

19:00

304

A

B

C

D

E

F

G

Scientific Sessions room K lower level

room L/M 1st level

room N/O 1st level

room R1 1st level

room R2 1st level

room R3 1st level

room U 2nd level

room X 2nd level

room XI, Z & Y 2nd level

room Z 2nd level

07:00

07:00

07:30

07:30

08:00

08:00

08:30

09:00

08:30

RC 1712 Pediatric Trauma (p. 130)

09:30

SF 17c Special Focus Session The upcoming “geriatric tsunami” and its consequences to imaging services (p. 131)

SF 17d Special Focus Session Undergraduate radiological education (p. 132)

RC 1703 Cardiac CT vs MRI (p. 133)

CC 1718 Breast: From Basics to Advanced Imaging Advanced imaging (p. 134)

RC 1709 Interventional Radiology Update on endovascular aortic interventions (p. 134)

09:00

09:30

10:00

10:00

10:30

11:00

11:30

10:30 SS 1812 Pediatric Advances in CT and MRI (p. 322)

SS 1806a Contrast Media From research to practice (p. 324)

SS 1811 Neuro Epilepsy and development issues (p. 327)

SS 1804 Chest Technical advances/ Pulmonary hypertension and perfusion (p. 329)

SS 1803b Cardiac Valvular and cardiac function studies (p. 331)

SS 1809 Interventional Radiology Vascular interventions (2) (p. 333)

11:00

11:30

12:00

12:30

12:30

13:00

13:00

13:30

13:30

14:00

14:00

14:30

14:30

15:00

15:00

15:30

15:30

16:00

16:00

16:30

16:30

17:00

17:00

17:30

17:30

18:00

18:00

18:30

18:30

19:00

19:00

A

B

C

D

E

F

G

305

Tuesday

12:00

Scientific Sessions 10:30 - 12:00

Room A

SS 1810 Tumors and marrow disease Moderators: E.E. Drakonaki; Iraklion/GR A.R. Mester; Budapest/HU

10:30

Role of single shot echoplanar diffusion weighted MR imaging in differentiation of malignant spinal tumors from benign lesions A.A.A. Abdel Razek1, R. Warda2, A. Elsherif2, K. Abdel Hamid2; 1Mansoura/EG, ²Gedda/SA ([email protected]) Purpose: To evaluate the role of single shot echoplanar diffusion weighted MR imaging in differentiation of malignant spinal tumors from benign lesions. Methods and Materials: Prospective study was conducted on 67 consecutive patients (47M, 30 F aged 21-74 ys: mean 47 ys) with spinal lesions. They underwent diffusion weighted MR imaging using a single shot echo planar imaging with a bfactor of 0, 250 & 500 sec/mm2. The scanning parameters were: TR=10000 ms, TE=108 ms, NEX=8-16, bandwidth=125 kHz, slice thickness=4 mm. Apparent diffusion coefﬁcient (ADC) map was reconstructed. The ADC value of the spinal lesion was calculated and correlated with surgical ﬁndings or biopsy. Results: Adequate ADC maps were obtained in 63 patients. The mean ADC value was 1.02p0.13X10-3 mm2/sec in metastasis, 0.97p0.12X10-3 mm2/sec in lymphoma, 0.86p0.07X10-3 mm2/sec in benign tumors, 1.47p0.07X10-3 mm2/ sec in pyogenic infection, 1.59p0.08X10-3 mm2/sec in tuberculous infection and 1.61p0.11X10-3 mm2/sec in traumatic and osteoporotic fracture. There was a statistically signiﬁcant difference in ADC values between malignant spinal tumors and benign lesions (p 0.021). When apparent diffusion coefﬁcient value of 1.30X10-3 mm2/sec was used as a threshold value for differentiating of malignant tumors from benign spinal lesions, the best result was obtained with an accuracy of 88%, sensitivity of 87%, speciﬁcity of 92%, positive predictive value of 95% and negative predictive value of 81%. Conclusion: Single shot echoplanar diffusion weighted MR imaging is a rapid non invasive imaging modality that provides valuable information for differentiating malignant spinal tumors from benign lesions.

B-702

10:39

Comparative study of whole body diffusion weighted MR imaging and bone scintigraphy in detection of bone metastases S. Li, H.D. Xue, F. Sun, Z.Y. Jin; Beijing/CN ([email protected]) Purpose: To evaluate the diagnostic value of whole body diffusion weighted MR imaging (WB-DWI) for detecting of bone metastases in comparison to bone scintigraphy. Methods and Materials: 44 patients (25 males, 19 females, aged 7-80 years) with malignant tumors (10 lymphoma, 13 lung cancer, three gastric cancer, two breast cancer, two leukemia, ﬁve hepatic cancer, four renal cancer, two colon cancer, one prostate cancer, two cervical cancer) were enrolled. All the patients underwent WB-DWI and bone scintigraphy within one week. WB-DWI was performed by using short TI inversion recovery diffusion weighted EPI sequence with free breathing and build-in body coil. Total scan time was 30 minutes for ﬁve stations covering from head to thigh. High resolution 3D MIP images of WB-DWI were obtained and displayed by inverted gray scales. The skeletal system was divided into seven regions (skull, sternum and clavicle, scapula, spine, ribs, pelvis, limbs). Results: In 23 cases WB-DWI identiﬁed 200 bone metastases, while bone scintigraphy demonstrated 188 lesions. There was no statistical difference between the two methods (p 0.05). Compared with bone scan, the regions missed by WB-DWI were mainly located in skull (n=6) and limbs (n=4). Most of them were osteogenic metastases. WB-DWI was more sensitive in the detection of metastases to the spine (n=9), pelvis (n=8) and ribs (n=5), which were conﬁrmed by axial T2w images. Besides, WB-DWI revealed more metastases to the lymph lodes and extraskeletal organs. Conclusion: WB-DWI has high accordance with skeletal scintigraphy in detecting bone metastases, and the two modalities are complementary to each other.

306

A

B

10:48

Bone disease in multiple myeloma patients: 3 T whole body MRI vs radiological skeletal survey. Initial clinical experience E. Squillaci, S. Mancino, C. Cicciò, D. Bartolucci, G. Manenti, G. Simonetti; Rome/IT ([email protected])

Musculoskeletal

B-701

B-703

C

D

E

F

G

Purpose: To compare the diagnostic accuracy of Whole-Body MR Imaging (WBMRI) vs conventional Radiological Skeletal Survey (RSS) in patients with Multiple Myeloma (MM) in the evaluation of bone disease, using a 3.0 Tesla whole-body scanner. Methods and Materials: Between March and September 2007, 15 patients (10 male, ﬁve female, mean age 55 years) with progressive and symptomatic MM underwent 3 T WBMRI (Philips Achieva, Best, The Netherlands), using a rolling table platform, including head, neck, thorax, abdomen, pelvis and upper and lower extremities. With seven-station approach, three different unenhanced image sets were acquired in the coronal plane, T1W TFE, T2W-TFE and T2W-STIR. In addition, the spine was imaged in the sagittal plane. All patients were staged according to staging system of Durie and Salomon (three patients stage I, eight patients stage II, four patients stage III), using RSS as gold standard in assessment of bone disease. MRI ﬁndings were blinded compared with those of previous RSS to detect bone inﬁltration. Results: WBMRI revealed excellent correlation with RSS in detection of bone lesions. A total of 24 bone lesions were identiﬁed on RSS and 21 lesions were detected on the same sites by WBMRI. The regions missed by WBMRI were located in the rib cage (n=2) and in the skull (n=1). WBMRI allowed to identify ﬁve additional bone lesions in spine, pelvis, and femur. Conclusion: WBMRI is a speciﬁc and sensitive method for detecting bone marrow inﬁltration, proving to be an alternative to RSS for staging in MM patients.

B-705

10:57

A prospective analysis of CT density measurements of bone metastases after treatment with zoledronic acid P. Dell’Aia, C.C. Quattrocchi, S. Piciucchi, D. Santini, R. Grasso, B. Beomonte Zobel; Rome/IT ([email protected]) Purpose: To prospectively determine CT density changes of bone metastases, before and after intravenous zoledronic acid for a maximum period of 12 months. Methods and Materials: Twenty-three consecutive patients presented with bone metastases and underwent therapy with zoledronic acid since December 2004. All patients underwent CT of the chest, abdomen and pelvis. Bone density, measured in HU, was determined by segmenting lesions on the same anatomical area of the metastasis sites in the axial images of the sequential series of CT examinations. The effects of zoledronic acid were evaluated by calculating absolute and relative increases of bone density. Results: The patients presented with multiple metastases in 65% of the cases. When compared to the baseline, all groups demonstrated a signiﬁcant increase in bone density which was signiﬁcantly (p 0.01) correlated with the number of zoledronic acid administrations. There was an increased bone density of at least 100% in 57%, and an increase of at least 50% in 87% of the patients. This increase was signiﬁcant in both lytic and sclerotic metastases after three months of therapy. No signiﬁcant bone density difference was found in normal-appearing bone. Conclusion: Bone density, as measured with CT, increases at the metastatic sites after zoledronic acid treatment, regardless of the type of metastasis but contrary to apparently normal bone.

B-706

11:06

Application study of 3 T MR functional imaging in patients with musculoskeletal tumors Z. Qi, C. Li, D. Yu, K. Zhang; Jinan City/CN ([email protected]) Purpose: To assess the usefulness of dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI), 31P- and 1H-spectroscopy (MRS), and diffusionweighted imaging (DWI) in distinguishing benign from malignant musculoskeletal tumors (MT). Methods and Materials: 35 patients with MT were examined with DCEMRI, DWI and 31P-MRS, and eight cases with 1H-MRS. The signal intensity of tumor shown in the time-signal intensity curve (TIC), maximum slope of increase (MSI), maximum slope of decrease (MSD), positive enhancement integral (PEI), signal enhancement ratio (SER), time to peak (Tpeak), the ratios of the other metabolites toB-ATP, NTP and Pcr, intercellular pH, 1H-MRS parameters, apparent diffusion coefﬁcient (ADC) and exponential apparent diffusion cofﬁcient (EADC) of the lesions were acquired. ANOVA was used as the statistical method for ADC and EADC. Results: Most malignant lesions were type I (rapid progression and gradual drain-

Scientific Sessions age) or II (rapid progression and stable or slight increase of signal intensity) of TICs. Using types I and II as the standard of malignancy, the diagnostic sensitivity, speciﬁcity and accuracy were 92.00, 70.59 and 83.33%, respectively. MSI and MSD of eight metastatic tumors were markedly higher than those of benign ones, while ADC was opposite. The ratios of Pcr/PME and PME/NTP between tumor groups and contrast, the TIC types, Tpeak, PME/B-ATP, PME/NTP, PME, PDE, Pcr, Pi and choline level between benign and malignant tumors were signiﬁcantly different. There were signiﬁcant differences of ADC between metastatic and benign tumors. Conclusion: The TIC plays an important role in differentiating benign from malignant MT, and the in vivo MRS provides the energy metabolite characteristics. DWI could show lesions sensitively.

B-707

11:15

Preliminary study of 3 T 1H MR spectroscopy in bone and soft tissue tumors Z. Qi, C. Li, Z. Li, K. Zhang, Q. Wang; Jinan City/CN ([email protected]) Purpose: To assess the usefulness of 1H-MRS in diagnosing bone and soft tissue tumors and distinguishing benign from malignant tumors. Methods and Materials: 56 patients with bone and soft tissue tumors proved clinically and pathologically were examined with 1H-MRS. 1H-MRS was performed to study malignant and benign musculoskeletal tumors, and normal muscle adjacent to lesions. The peak height of choline compounds (Cho), the creatine were observed, and metabolite ratios of Cho/Cr were manually calculated. Then according to the results tumors were judged benign or malignant. Kappa test was used to analyze all data. Results: 1H-MRS spectra style was different between bone and soft tissue tumors and normal muscle, and between benign and malignant tumors. Choline level and Cho/Cr in malignant tumor was markedly higher than that in benign tumors. The sensitivity, speciﬁcity, positive predictive value, negative predictive value and accuracy rate of the 1H-MRS was 94.44 (34/36), 83.33 (15/18), 91.89 (34/37), 88.29 (15/17) and 90.74% (49/54), respectively. The MRS results and the histopathology inspection conclusions had very good uniformity. The K value was 0.76p0.10 (p 0.01). Conclusion: The increase of Cho level and Cho/Cr measured by 1H-MRS is related with the bone and soft tissue malignant tumor, and they will play a vital role in the diagnosis and differential diagnosis of bone and soft tissue tumors.

B-708

11:24

Dynamic contrast enhanced MRI of spindle cell sarcomas: Counterintuitive relationship of perfusion to tumour necrosis A.P. Toms1, L.M. White2, R. Kandel2, R.R. Bleakney2, M. Noseworthy3, M.E. Blackstein2, J. Wunder2, L. Shepstone1; 1Norwich/UK, 2Toronto, ON/CA, 3 Hamilton, ON/CA ([email protected])

11:33

The relationship between MRI features and clinical outcome of agressive ﬁbromatosis R. Arkun, S. Dogan, M. Argin, T. Akalin, D. Sabah, Y. Bolukbasi; Izmir/TR ([email protected]) Purpose: Our aim is to describe MR imaging ﬁndings of agressive ﬁbromatosis and to investigate the relationship between MRI features and clinical outcome of the lesion in both both primary (P) and recurrent (R) cases. Methods and Materials: The MR images of 42 cases (29P, 13R) pathologically proven aggressive ﬁbromatosis were reviewed retrospectively. A total of 86 scans of 42 cases were revieved. The size, margin, local invasion, signal characteristics and both dynamic and static contrast enhancement pattern of tumor were evaluated and correlated with pathological ﬁndings in 42 cases and clinical outcome in 30 cases. Results: The tumor size was between 0.8-15 cm in both groups. Fifteen patients (14P,1R) showed well deﬁned and 27 patients (15P,12R) showed ill deﬁned margins (p 0.05). In follow-up, reccurent rate was 21% in patients with well deﬁned and 56% with ill deﬁned margins (p 0.05). Recurrence rate was 33% in primary and 55 in secondary cases (p 0.05). Local invasion was detected 48% in primary and 62% in recurrent cases. Heterogenous signal changes were seen 61% in T1-W and 81% in T2-W images. On T1-W images, signal intensity (SI) was hypointense in 22 (15P, 7R), isointense in 17 (12P, 5R) and hyperintense in three (2P, 1R) cases compared to muscle. On T2-W images, hyperintense SI was seen in 27 (18P, 9R) patients compared to subcutaneous fat tissue. Signal void band like area was detected in 11 (8P, 3R) patients. Contrast enhancement was seen in 33 patients. Conclusion: Aggressive ﬁbromatosis has MR features similar to other soft tissue tumors. There is no correlation between MR imaging ﬁndings and clinical outcome of patients.

B-710

11:42

Morphology, metabolism, microcirculation, and strength of skeletal muscles in cancer-related cachexia M.-A. Weber1, H. Krakowski-Roosen1, L. Schröder1, R. Kinscherf2, M. Krix1, A. Kopp-Schneider1, M. Essig1, P. Bachert1, H.-U. Kauczor1, W. Hildebrandt1; 1 Heidelberg/DE, 2Mannheim/DE ([email protected]) Purpose: To assess morphology, metabolism, and microcirculation in skeletal muscles of patients with cancer-related cachexia and to compare these data with matched healthy volunteers. Methods and Materials: In 19 patients with cancer-induced cachexia and 19 age-, gender-, and body-height-matched healthy volunteers body composition and aerobic capacity (VO2max) were analyzed. Skeletal muscle ﬁber size and capillarization were evaluated in biopsies of the vastus lateralis muscle. The cross-sectional area (CSA) of the quadriceps femoris muscle was measured by magnetic resonance imaging as well as its isokinetic and isometric forces. The energy and lipid metabolism of the vastus lateralis muscle was quantiﬁed by 31P and 1H NMR spectroscopy and parameters of its microcirculation by contrast-enhanced ultrasonography (CEUS). Results: Morphologic parameters were about 30% lower in cachexia than in volunteers (body mass index: 20p3 vs 27p4 kg m-2, CSA: 45p13 vs 67p14 cm2, total ﬁber size: 2854p1112 vs 4181p1461 μm2). VO2max was reduced in cachexia (23p9 vs 32p7 ml min-1 kg-1, p=0.03), whereas histologically determined capillary density and microcirculation in vivo were not signiﬁcantly different. Concentrations of muscular energy metabolites, pH, and intramyocellular lipids were not signiﬁcantly different in both groups. Absolute strength of quadriceps muscle was reduced in cachexia (isometric: 107p40 vs 160p40 Nm, isokinetic: 101p46 vs 167p50 Nm; p=0.03), but identical when normalized on CSA (isometric: 2.4p0.5 vs 2.4p0.4 Nm cm-2, isokinetic: 2.2p0.4 vs 2.5p0.5 Nm cm-2). Conclusion: Cancer-related cachexia is associated with a loss of muscle volume but not of functionality, which can be a rationale for muscle training.

Tuesday

Purpose: The aim of this study was to determine whether changes in pharmacokinetic modeling of dynamic contrast-enhanced MRI (DCE-MRI), during chemotherapy for spindle cell sarcomas, correlated with histological measures of total tumor necrosis. Methods and Materials: Twelve patients with appendicular spindle cell sarcomas were included in the study. Each patient had DCE-MRI before, and after completion, of pre-operative chemotherapy. The mean arterial slope (A), endothelial permeability coefﬁcient (Ktrans) and extravascular extracellular volume (Ve) were derived from each dataset using a modiﬁed 2-compartment pharmacokinetic model (Tofts). Total tumour necrosis rates were compared with changes in A, Ktrans and Ve. Results: Six patients had total tumour necrosis of = 90%, and 6 had a measure of 90%. The median percentage changes in A, Ktrans and Ve for the = 90% necrosis group were -52.5% (-83 to 6), -66% (-82 to 26) and 23.5% (-26 to 40) respectively. For the 90% necrosis group A= -35% (-75 to 132), Ktrans= -53 (-66 to 149) and Ve= -14.5% (-42 to 40). One patient with 90% necrosis had increase in all three measures. Comparison of the two groups generated P values of 0.699 for A, 0.18 for Ktrans and 0.31 for Ve. Conclusion: There was no statistically signiﬁcant correlation between changes in A, Ktrans or Ve with total tumour necrosis in this study. An individual positive correlation of perfusion with necrosis suggests that neoangiogenesis caused by normal tissue responses to necrosis, rather than driven by tumour factors, may confound expected perfusion patterns in primary bone tumours.

B-709

A

B

C

D

E

F

G

307

Scientific Sessions

Abdominal Viscera (Solid Organs)

Conclusion: Gd-BOPTA enhanced MRI can depict the hemodynamic characteristics and reﬂect the functional status of HCC lesions, thus contributing greatly to the diagnosis of HCC.

SS 1801a

B-713

10:30 - 12:00

Room B

HCC, still tantalising and bewildering Moderators: M.-F. Bellin; Villejuif/FR L. Crocetti; Pisa/IT

B-711

10:30

Gadoxetic acid-enhanced MR imaging of hepatocellular carcinoma (HCC) in patients with cirrhosis correlated to histological grading: Primary assessment M. Di Martino, D. Marin, C. Catalano, A. Guerrisi, F. Galati, M. Baski, S. Brozetti, R. Passariello; Rome/IT ([email protected]) Purpose: To investigate whether gadoxetic acid-enhanced MR imaging can be useful in assessing histological grades of hepatocellular carcinoma (HCC). Methods and Materials: Informed consent and ethical approval were obtained. Ten consecutive patients (eight men, two women; mean age, 63 years) with 19 histologically proven HCC nodules underwent gadoxetic acid-enhanced MR imaging and subsequent surgery within 30 days. MR studies were performed at 1.5 T scanner (Avanto, Siemens) by using sensitivity encoding (SENSE) technique. MR protocol included: breath-hold T2-weighted images and volumetric three-dimensional gadoxetic acid-enhanced T1-weighted GRE MR images acquired in the arterial (25 s), portal venous (60 s), equilibrium (180 s), and hepatobiliary phases (20 min). One radiologist performed operator-deﬁned region-of-interest (ROI) measurements of each nodule, the hepatic parenchyma, and the background noise on the hepatobiliary MR images. The lesion-liver contrast-to-noise ratio (CNR) was calculated (SIlesion-SIliver)/SDnoise).The relationship between the CNRs of each nodule and the histological grading according to the World Health Organization grading system was evaluated with the Spearman correlation analysis. Results: Of the 19 HCC lesions, ﬁve (26%) were well-differentiated, eight (42%) moderately-differentiated, and six (31%) poorly-differentiated at pathologic analysis. The lesion-liver CNRs on the hepatobiliary phase MR images were 5p12.1 (meanpstandard deviation) for well-differentiated, -83.2p8.9 for moderatelydifferentiated, and -172.8p7.5 for poorly-differentiated lesions. With the Spearman analysis, strongly positive correlation was found between HCC signal intensity and histological grades (r, 0.85; P 0.02). Conclusion: Gadoxetic acid-enhanced MR imaging is useful to distinguish moderately/poorly differentiated HCCs from well-differentiated HCCs. The lower signal intensities on MR images were related to a lower degree of histologic differentiation.

B-712

10:39

Value of Gd-BOPTA enhanced MR imaging in diagnosing hepatocellular carcinoma with correlation of pathology B. Song, Y. Li, J. Xu, B. Wu, Y. Yang; Chengdu/CN ([email protected]) Purpose: To investigate the value of a double-phasic MR contrast agent (Gadobenate Dimeglumine,Gd-BOPTA) in the diagnosis of hepatocellular carcinoma (HCC) with correlation of pathology. Methods and Materials: MRI scan was performed on 33 patients suspected of having focal liver lesions by CT or ultrasonography. The MRI scan protocol included routine axial T1WI, T2WI, coronal T2WI and 3D-MRCP, followed by Gd-BOPTA enhanced 3D gradient-echo T1W sequence acquiring dynamic data at 15 s, 55 s, 90 s,1 min, and delayed acquisitions at 10 min, 40 min and 80 min. Surgical specimens were collected for pathological analysis. Results: Surgery conﬁrmed the diagnosis of HCC in 21 patients with 27 lesions. The sensitivity, speciﬁcity and accuracy of MR imaging with combined phases (including hemodynamic phase and hepatobiliary phase) were 92.9, 93.3, and 92.9%, respectively. In the hemodynamic phase, 22 lesions showed early enhancement in arterial phase, fast “washout” in portal venous phase, and hypointense in delayed phase. Four lesions showed slight enhancement, while one lesion showed no enhancement in all hemodynamic phases. In the hepatobiliary phase, 23 lesions showed homogeneous hypointensity compared with the enhanced liver background, but four lesions (three well-differentiated and one moderately differentiated) had patchy or dot-like slightly hyperintense areas within the lesions. Pathologically cholestatic changes of tumor cells were observed and bile-containing small ducts were present.

308

A

B

C

D

E

F

G

10:48

Automated RECIST and WHO criteria measurements and volumetry of solid liver tumors in multi-slice CT V. Yaghmai1, S. Berggruen1, B. Atassi1, M. Rezvani1, J. Huang1, H. Soud2, P. Nikolaidis1, R. Salem1; 1Chicago, IL/US, 2Malvern, PA/US ([email protected]) Purpose: To evaluate the feasibility of automatic segmentation and measurement of hepatic lesions. The feasibility of automated volume calculation of hepatic tumors was also evaluated. Methods and Materials: Seventy-four solid hepatic masses on multislice CT were segmented using prototype automated software (Siemens Medical Solutions, Forchheim, GER). The radiologists used a 5-point scale to rate the quality of automated segmentation: 5:excellent-1:unable to segment. The software automatically generated RECIST and WHO measurements. These measurements were validated by phantoms.The radiologists also measured each lesion manually. Pearson correlation coefﬁcients (r) for automated and manual measurements were calculated. The software also calculated the volume of each liver lesion. Correlation coefﬁcients between volume and RECIST or WHO criteria were also calculated. Results: There is strong correlation between automated and manual measurements for either RECIST or WHO measurements (r=0.976, p 0.0001 and r=0.983, p 0.0001, respectively). Both automated and manual RECIST measurements are strongly correlated with WHO measurements using similar measurement techniques (r=0.959, p 0.0001 and r=0.968, p 0.0001, respectively). The software was able to automatically segment and measure the volumes of all hepatic lesions with average segmentation score of 4.6 (SD=0.79). Strong correlations between RECIST or WHO criteria and volumetric measurements were observed (r=0.886, p 0.001 and r=0.955, p 0.001, respectively). Conclusion: Accurate automated segmentation of solid liver masses is feasible. There is strong correlation between automated and manual RECIST and WHO measurements. Automated volumetric measurement of the solid liver lesions is also feasible.

B-714

10:57

Probabilistic separation of iodine: Virtually non-enhanced CT obtained with single source dual energy MDCT J. Sosna1, L. Goshen2, S. Amin Spector2, R. Carmi2, E. Libson1; 1Jerusalem/IL, 2 Haifa/IL ([email protected]) Purpose: Dual-energy CT may be obtained with a single-source, dual-detection system. Our purpose was to generate virtually non-enhanced CT images from enhanced clinical CT studies and to compare the image quality to regular nonenhanced CT. Methods and Materials: Feasibility study was performed using a dual-energy MDCT prototype (Philips Medical Systems, Cleveland, OH) in accordance with our IRB requirements. Data was obtained from 12 patients (ages 45-76 years) who underwent bi-phasic MDCT studies including a non-enhanced and an iodineenhanced phases. Study parameters were as follows: 2 mm slice thickness, 1 mm increment, 140 kVp, and 250 mAs. Spectral separation analysis was performed on the enhanced phase using a probabilistic mixture model and virtually non-enhanced CT images were generated. The virtual and regular non-enhanced CT images were compared by two experienced radiologists in consensus for image quality, noise, the visibility of calciﬁed areas and artifacts using a predetermined scale. Results: Virtually non-enhanced images could be generated for all patients and were of diagnostic quality. Image quality and noise were graded similar in 83% (n=10) and 75% (n=9) of patients, respectively (P=NS). Calciﬁed areas had the same conﬁguration in 66% (n=8) with mild reduction and marked reduction in calciﬁcation size in three (25%) and one (8.5%) patient, respectively. Mild streak like artifacts were seen in two patients. Conclusion: Virtually non-enhanced images can be obtained with dual-energy CT with similar image quality as regular non-enhanced images and may obviate the need for regular non-enhanced phase in multi-phasic MDCT.

Scientific Sessions B-715

11:06

Quantitative color mapping of the arterial enhancement fraction of the liver for detection of hepatocellular carcinoma in cirrhotic liver K. Kim1, J. Lee1, E. Klotz2, J. Lee1, S. Hong1, W. Kim1, S. Kim1; 1Seoul/KR, 2 Forchheim/DE ([email protected]) Purpose: To apply quantitative color mapping of the arterial enhancement fraction (AEF) to hepatocellular carcinoma (HCC) in liver cirrhosis (LC), and to deﬁne its usefulness. Methods and Materials: In 65 LC patients with 108 HCCs, who had multiphasic MDCT scans, quantitative color maps of the AEF (the ratio of attenuation increment during the arterial phase to attenuation increment during the portal phase) of the whole liver were created using a software (Hepacare) from the images of routine multiphasic CT. AEFs of non-neoplastic liver parenchyma and HCC were measured and analysed regarding those relationships with severity of LC, histologic grade and size of HCC. The diagnostic value of color mapping of AEF was evaluated with retrospective reviewing of the images by three radiologists. Results: The mean AEF of HCC (76.5p13.7) was signiﬁcantly higher than that of liver parenchyma (24.6p12.8) (p 0.001). The mean AEF of non-neoplastic liver parenchyma was signiﬁcantly increased as the patient's Child-Pugh class was increased (p 0.05), while no signiﬁcant difference in mean AEF of HCCs according to their size and histologic grade was noted. We set the cut-off value of AEF as 40 for diagnosis of HCC, which gives sensitivity 99.1% and speciﬁcity 93.8%. The mean sensitivity for HCC detection was increased from 77.5% with multiphasic CT only to 87.6% with CT and AEF color mapping, which was mainly due to increase in detection of small HCC. Conclusion: Quantitative color mapping of AEF can provide quantiﬁed hemodynamic information of HCC and liver without adding risk of radiation hazard, and increase the sensitivity for detection of HCC.

B-716

11:15

Hepatocellular carcinoma in cirrhotic liver disease: Functional CT with perfusion imaging for assessment of tumor angiogenesis D. Ippolito1, F. Invernizzi1, M. Pozzi1, L. Antolini1, S. Sironi1, F. Fazio2; 1Monza/IT, 2 Milan/IT ([email protected]) Purpose: To determine the value of functional computed tomography (CT) with perfusion imaging in quantitative assessment of tumor-related angiogenesis in cirrhotic patients with hepatocellular carcinoma (HCC). Methods and Materials: Study group consisted of 73 patients who met the following inclusion criteria: a) Child-Pugh class A or B liver cirrhosis; b)presence of single lesion suspected as HCC at ultrasound examination; c)maximum lesion diameter between 1 and 3 cm. All patients were examined with triple-phase MDCT and subsequently with CT-perfusion technique, consisting of acquisition of eight dynamic slices/scan per 40 scans after i.v. administration of 50 ml of iodinated contrast with a ﬂow rate of 6 ml/sec. Using a dedicated software (CT-perfusion Philips 2.0), ﬁve perfusion parameters were measured for both HCC and cirrhotic parenchyma: hepatic perfusion (HP); blood volume (BV); hepatic perfusion index (HPI); arterial perfusion (AP); and time to peak (TTP). Perfusion parameters were described with quartile values of their distribution, and statistically analyzed with Univariate paired Wilcoxon signed rank test. Results: HCC diagnosis was made in 35 out of 73 patients in accordance with the European Association for the Study of the Liver (EASL) criteria. The values of perfusion parameters measured within tumor tissue were: HP (mL/sec/100 g): median=47.0 (1 stqt=36.0, 3 stqt=61.4); BV (mL/100 mg): median=24.0 (1 stqt=18.7, 3 stqt=29.3); HPI (%): median=78.4 (1 stqt=62.9, 3 stqt=100); AP (mL/min): median=45.9 (1 stqt=39.0, 3 stqt=60.1); TTP (sec): median=18.7 (1 stqt =16. 3, 3 stqt =26.5). HP, BV, HPI, and AP were higher (p 0.001), while TTP was lower (p=0.001). Conclusion: In cirrhotic patients with HCC, functional CT with perfusion imaging can provide quantitative information about tumor-related angiogenesis.

B-717

11:24

Purpose: To compare the diagnostic capability of MDCT and MRI for detection of tumor nodules and the effect on patient management. Methods and Materials: 28 patients (25 male, 3 female, age 67p10.79) with biopsy proven HCC were investigated with 64-row MDCT (collimation 0.6 mm, slice 1 mm native, arterial and portal phase, 120 ml Imeron, 4 ml/s, delay by bolus trigger) and MRI (T1fs ﬂ2d TE/TR2.72/129 ms, T2tse TE/TR102/4000 ms, 5-phase dynamic contrast-enhanced T1fs ﬂ3d TE/TR1.56/4.6, 0.1 mmol/kgBW gadolinium-DTPA, slice

B-718

11:33

Comparison of enhancement with iomeprol-400 and iodixanol-320 in patients undergoing contrast-enhanced multidetector CT of the liver L. Romano1, C. Catalano2, L. Grazioli3, L. Bonomo2, J.-R. Xu4, K.-M. Chen4; 1 Naples/IT, 2Rome/IT, 3Brescia/IT, 4Shanghai/CN ([email protected]) Purpose: To compare contrast enhancement with iomeprol (400 mg I/mL) and iodixanol (320 mg I/mL) in patients undergoing contrast-enhanced liver multidetector CT (MDCT). Methods and Materials: 183 patients received equi-iodine doses (40 g I) of iomeprol-400 or iodixanol-320 IV at 4 mL/sec. Liver MDCT was performed using scanners with at least four detector rows. Two off-site, independent, blinded readers assessed images at the abdominal aorta, inferior vena cava (IVC), portal vein, and liver parenchyma. Descriptive statistics were used to summarize the contrast density (HU) measurements for the two study groups. The mean contrast densities achieved in each of the four regions of interest were compared and 95% conﬁdence intervals estimated. Results: Ninety-one patients received iomeprol-400 while 92 received iodixanol320. The two study groups were comparable with regard to sex, age, weight, and race. Iomeprol-400 resulted in signiﬁcantly greater arterial phase enhancement of the abdominal aorta compared to iodixanol-320 (Reader 1: 339.7 vs 292.6 HU, 95% CI [19.4, 65.5], p=0.0004; Reader 2: 327.9 vs 293.0 HU, 95% CI [6.6, 54.3], p=0.01). Iomeprol-400 also led to greater enhancement of liver parenchyma during the portal-venous phase (signiﬁcantly greater for Reader 1 [p=0.04] and of borderline signiﬁcance for Reader 2 [p=0.05]). No signiﬁcant difference was noted between the two study groups regarding enhancement of the IVC or portal vein. Conclusion: Iomeprol-400 provides signiﬁcantly greater enhancement in the arterial phase and improved enhancement of hepatic parenchyma in the portal-venous phase compared to iodixanol-320 at the same iodine dose in patients undergoing MDCT of the liver.

B-719

11:42

The combined assessment of hepatocellular nodule vascularity at 64row multidetector CT and contrast-enhanced US: Analysis of diagnostic performance E. Quaia1, V. Alaimo2, T.V. Bartolotta2, M.A. Cova1; 1Trieste/IT, 2Palermo/IT ([email protected]) Purpose: To assess the diagnostic value of 64-row multidetector-row CT (MDCT) combined to contrast-enhanced US (CEUS) in the assessment of hepatocellular nodule vascularity in patients with liver cirrhosis. Methods and Materials: 38 cirrhotic patients (26 male, 12 female; mean age p SD, 71 p 7 years) with 55 biopsy-proven hepatocellular nodules scanned both by 64row MDCT and CEUS were prospectively recruited. Multiphasic MDCT consisted of unenhanced scan, followed by arterial, portal venous, and delayed equilibrium phase (iodine 400 mg/mL at 5 mL/sec; 50 mL of saline chaser; bolus tracking with one ROI on superior abdominal aorta with 150 HU threshold). CEUS was performed after sulfur hexaﬂuoride-ﬁlled microbubble injection and each nodule was scanned at arterial (10-40 secs) and extended portal venous phase (from 45 secs after microbubble injection up to microbubble disappearance). The nodule vascularity and enhancement pattern at MDCT (arterial and delayed equilibrium phase), CEUS (arterial and extended portal venous phase) and combined MDCT/CEUS were evaluated side-by-side by 2 independent blinded readers who classiﬁed nodules as benign or malignant according to reference diagnostic criteria. Results: Final diagnoses comprised 44 hepatocellular carcinomas, 7 macroregenerative and 3 dysplastic nodules, and 1 necrotic nodule. The separate assessment of

A

B

C

D

E

F

G

309

Tuesday

MRI versus 64-row MDCT for diagnosis of HCC R. Kloeckner, G. Otto, S.C.A. Herber, K.F. Kreitner, C. Dueber, M.B. Pitton; Mainz/DE ([email protected])

4 mm). Consensus reading of both modalities (280 datasets) was used as reference and was positive if one or both methods were positive for hypervascularized nodules. Tumor nodules were analysed with respect to number, size and location. Results: 162 tumor nodules were detected by consensus reading. MRI detected signiﬁcantly more tumor nodules (159 versus 123, p 0.001) compared to MDCT, with best sensitivity for early arterial phase MRI. Sensitivity for tumor detection was 0.98 for MRI compared to 0.76 for MDCT. False negative CT ﬁndings included nodules b5 mm (n=5), b10 mm (n=17), b15 mm (n=12), b20 mm (n=4) and one nodule 20 mm. MRI missed two nodules b10 mm and one nodule b15 mm. Nodules demonstrated greater diameters in MRI compared to CT (29.2p25.1 mm, range 5-140 mm versus 24.1p22.7 mm, range 4-129 mm, p 0.001). 13 CT negative nodules in seven patients were located in the left lobe; ﬁnding them could have changed therapeutic strategy. Conclusion: Contrast-enhanced MRI is superior compared to 64-row MDCT for detection and sizing of HCC tumor nodules. Stratiﬁcation of HCC patients to interventional or operative treatment should be based on a dedicated MRI protocol.

Scientific Sessions MDCT and CEUS scans provided an overall diagnostic accuracy of 86/84% (reader 1/reader 2) and 87/82%, while combined MDCT/CEUS improved the diagnostic accuracy up to 96/98% (P 0.05, McNemar test). Conclusion: The combined assessment of hepatocellular nodule vascularity at MDCT and CEUS improved the diagnostic performance in the malignancy diagnosis in patients with liver cirrhosis.

B-720

11:51

Role of fusion imaging technique by a combined CT/MRI ultrasound navigation system in the diagnosis of small HCC ( 20 mm) on cirrhosis G. Vidili, F. Piscaglia, A. Bazzocchi, G. Zironi, A.G. Tewelde, M.G. Mirarchi, S. Costantini, R. Righini, R. Golﬁeri, L. Bolondi; Bologna/IT ([email protected]) Purpose: Aim of the present study was to test the potential of the ultrasound (US)based Virtual Navigator® (VN) system (Esaote, Genova, Italy) in the identiﬁcation and characterization of small ( 2 cm) lesions in cirrhosis detected by ce (contrastenhanced)-CT and ce-MR imaging and suspected to be hepatocellular carcinoma (HCC), but not visualized by a previous standard US scan. Methods and Materials: Fourteen cirrhotic patients (9M; 5 F), mean age 69.6p6.2 (SD) years, in whom ce-CT or ce-MR detected 17 new lesions (13.6p3.8 mm) suspected to be HCC, but not identiﬁed by US (performed blindly to CT/MR) were submitted to ce-US (CEUS) focused to the target zone identiﬁed by the navigation system, which displays simultaneously the CT/MR images reconstructed in 3D and real-time ultrasonography. A ﬁnal diagnosis was established with coincidental ﬁndings according to the international guidelines or by follow-up in discordant cases. Results: A ﬁnal diagnosis of HCC was reached in 15/17 cases. VN identiﬁed and showed typical CEUS pattern in 12/15 HCC (80%) (four identiﬁed at conventional US with VN and eight only after CEUS); 3/12 lesions (20%) could not be identiﬁed with VN. The 2/17 remaining lesions were not visualized by VN and ﬁnally judged as CT/MR false positives, as they disappeared also at CT/MR during a 12-month follow-up. Conclusion: Our preliminary results show that VN improves the potential of US in identifying and characterizing HCC in cirrhosis, permitting achievement of diagnosis according to non invasive criteria and opening the way for potential percutaneous ablation strategies.

10:30 - 12:00

Room C

GI Tract

SS 1801b Small bowel and appendix Moderators: G.J. Antes; Kempten/DE V. Maniatis; Athens/GR

B-721

10:30

Improved diagnosis of Crohn’s disease by peristaltic motion assessment: A comparison of standardized MRI with and without cine TrueFISP sequence C. Waldherr1, J.M. Fröhlich2, C. Stoupis3, S.M. Ertürk4, M.A. Patak1; 1Berne/CH, 2 Winterthur/CH, 3Männedorf/CH, 4Istanbul/TR ([email protected]) Purpose: Peristaltic motion evaluation in MR enterography (MRE) of the small bowel for Crohn’s disease (CD) has not been established. The aim was to evaluate the diagnostic accuracy of CD-MR by adding cine TrueFISP (cMRE) sequences for the assessment of CD. Methods and Materials: 40 patients with histologically proven CD (m:22, f:18) underwent standardized MRE of the small bowel (ingestion: 1000 ml of 3% Mannitol). Imaging was done by TrueFISP (cor & trans), T2w-HASTE (cor & trans) and coronal T1w-FLASH-3D 20, 60 and 90 sec after iv-contrast. Cine MRE were acquired covering the entire small bowel using cor plane 2D cine TrueFISP sequences with a slice repetition time of 500 msec. Both MRE and cMRE were evaluated by two readers for image quality, bowel distension and pathologic ﬁndings. Results: MRE image quality was scored as good in 70/78/78% (upper-/lower abdomen/terminal ileum). For cMRE, distension and motion were scored as good in 80/83% (upper abdomen), in 90/88% (lower abdomen) and in 85/90% (terminal ileum). Number of ﬁndings of cMRE compared to MRE: wall thickening 31/25 (p=0.01), mural or extramural masses 1/0, wall layering 17/12 (p=0.03), mucosal ulcers 5/2 (p=0.09), abscesses 1/0, enlarged lymph nodes 21/12 (p=0.005), ﬁstulae 0/1, comb sign 22/16 (p=0.006) and stenoses 24/19 (p=0.03). Overall, cMRE detected additionally 12 lesions in these patients, cMRE 44 lesions, MRE 32 lesions (p=0.008).

310

A

B

C

D

E

F

G

Conclusion: The inclusion of small bowel peristaltic motion evaluation in MR imaging of CD signiﬁcantly increases the lesion detection rate for CD related pathologic ﬁndings compared to standard MRE.

B-722

10:39

Sliding multislice: A novel adjunct to hydro-MRI for the evaluation of patients with recurrent Crohn’s disease G. Pache, A.-O. Schaefer, M. Langer; Freiburg/DE ([email protected]) Purpose: Evaluation of sliding multislice (SMS), a novel MRI technique for axial moving table acquisitions, as part of Hydro-MRI for the diagnostic work-up of patients with recurrent Crohn’s disease. Methods and Materials: 32 patients with known Crohn’s disease underwent Hydro-MRI at 1.5 Tesla. The imaging protocol consisted of axial and coronal HASTE, contrastenhanced coronal VIBE and axial T1w FLASH-2D in standard multistage and SMS techniques. Two radiologists analysed MRI data for lesion site, activity and complications of Crohn’s disease. Results in standard multistage FLASH-2D were compared with those of SMS FLASH-2D and correlated with Crohn’s disease activity index (CDAI). Results: Compared with the standard multistage technique, SMS reduces total examination time by 4.5 minutes thereby exhibiting no loss in image quality. MRI revealed at least one lesion site in 93% and extraluminal disease complications in 47% of the patients. Interobserver agreement ranged from 0.672 to 1.0 (K value) for all parameters evaluated. Regarding lesion detectability, excellent correlation between standard multistage and SMS FLASH-2D was found. No signiﬁcant correlation was found between bowel wall enhancement and the CDAI (R=0.09, p=0.63). Conclusion: In our opinion, SMS-based Hydro-MRI has the potential for a ﬁrst line imaging modality for patients with recurrent Crohn’s disease.

B-723

10:48

Comparison of True-FISP-MRI before and after i.v.-contrast administration in patients with inﬂammatory bowel disease (IBD): Is postcontrast TrueFISP imaging essential? M. Hellmann, G. Krombach, N. Krämer, M. Distelmaier, R. Günther, C. Hohl; Aachen/DE ([email protected]) Purpose: To evaluate the basic necessity of i.v.-contrast administration in True-FISP imaging of patients with IBD. Methods and Materials: 40 patients, aged 13-66 years, with suspected or known IBD underwent abdominal MRI on a 1.5 T-scanner employing a 32-channel receivercoil (Achieva 1.5, Philips, Best, Netherlands). Bowel preparation was performed by oral administration of 1 l mannitol-solution and an additional mannitol-enema. The MR-protocol included True-FISP-sequences (balanced-FFE, TR=5.2 ms, TE=2.6 ms, FA=80°) with a spatial resolution of 0.88x0.88x5.0 mm before and after i.v.-contrast administration (Gd-DTPA; 0.1 mmol/kg) in coronal and axial planes. Each imaging sequence was acquired in a single breathhold using sensitivity encoding based parallel imaging (SENSE) with an acceleration factor R=5. Image quality of pre- and postcontrast True-FISP-sequences was evaluated by signal-to-noise (SNR) measurements. The diagnostic impact was estimated by contrast-to-noise (CNR) measurements between pathologic bowel-wall, healthy bowel-wall and mesenteric fat. Statistical signiﬁcance was tested using the t-test (p 0.05). Results: 15 of 40 patients had inﬂammatory bowel-wall changes. After contrast injection, the mean CNR between pathologic and healthy bowel-wall increased from 1.57 to 13.93 whereas mean CNR between pathologic bowel-wall and mesenteric fat decreased from 17.45 to 5.4. These changes were statistically signiﬁcant. SNR showed no signiﬁcant difference after contrast administration. Conclusion: Postcontrast True-FISP helps to determine between pathologic and healthy bowel-walls. However, the contrast between bowel-wall and mesenteric fat is impaired after contrast administration. To avoid loss of diagnostic information, True-FISP should be performed before and after i.v.-contrast.

B-724

10:57

Gadolinium enhancement of mesenteric lymph nodes in different Crohn disease subtypes S. Gourtsoyianni, N. Papanikolaou, J. Grammatikakis, L. Bourikas, M. Roussomoustakaki, N. Gourtsoyiannis; Iraklion/GR ([email protected]) Purpose: To measure contrast enhancement ratio (CER) of mesenteric lymph nodes and adjacent vessels in patients with different Crohn disease clinical subtypes. Methods and Materials: Twenty patients, categorized clinically into three different Crohn disease subgroups, underwent MR enteroclysis examination. CER between the signal intensity of each mesenteric lymph node and nearby vessels

Scientific Sessions was calculated for all patients on FLASH images, acquired 75 seconds after intravenous administration of gadolinium. A one-way analysis of variance statistical test was applied to investigate for any signiﬁcant differences of CER between different subtypes. Results: A total number of 99 mesenteric lymph nodes were assessed, 62 in patients with active inﬂammatory (AI) disease, 30 in patients with ﬁbrostenotic (FS) and seven in patients with ﬁstulizing/perforating (FP) disease process. No correlation between the geometrical area and CER was found. Mesenteric lymph nodes of patients with AI disease presented with the highest CER (mean: 0.803, sd: 0.12), followed by mesenteric lymph nodes of patients with FP subtype (mean: 0.717, sd: 0.1) and FS disease (mean: 0.63, sd: 0.1). The differences in mean values of CER between mesenteric lymph nodes of AI and FS subtypes were statistically signiﬁcant (p 0.0001), while mesenteric lymph nodes of the FP subtype did not present any statistically signiﬁcant differences in CER either with that of AI or FS subtypes. Conclusion: Contrast enhancement ratio of mesenteric lymph nodes may vary across different subtypes of Crohn disease. Such differences may be used in clinical practice.

B-725

11:06

Assessment of the Crohn disease activity using intravenous application of the blood-pool contrast agent gadofosveset during MR enterography J. Ferda, J. Kozeluhova, B. Kreuzberg, V. Treska; Plzen/CZ ([email protected]) Purpose: To evaluate the feasibility of the MR enterography with intravenous application of the blood-pool agent gadofosveset in differentiation of Crohn disease activity. Methods and Materials: 50 patients (28 males, 22 females, mean age 34.5 y) with conﬁrmed Crohn disease underwent MR enterography after oral preparation using 2000 ml of 2.5% mannitol. Examinations were performed on 1.5 T MRI suite Avanto after written consent; study was approved by the local ethics committee. Imaging protocol contained T2 weighted balanced echo sequences (trueFISP) and 3D gradient echo T1 weighted sequences (FLASH - volume interpolated breath hold examination - VIBE). VIBE sequences were performed in coronal orientation after bolus application of the gadofosveset (0.12 ml/kg) in dynamic phase - in arterial, early-portal and late portal phases. VIBE in transversal orientation followed dynamic phase; last data set was acquired with delay of 10 minutes after actual beginning of the bolus application. The ﬁnding of the late stratiﬁed enhancement of the inner bowel wall layer was compared with the Crohn disease activity index (CDAI), and sensitivity and speciﬁcity were calculated. CDAI value more than 120 was present in 28 patients. Results: The sensitivity of the stratiﬁed enhancement still detected after 10 minutes for detection of the active Crohn disease was 92.9% (26/28), speciﬁcity 95.5% (21/22). Conclusion: The enhancement of the inner bowel wall layer delayed more than 10 minutes after gadofosveset application is a valuable sign of Crohn disease activity. Study was supported by the research project MSM-0021620819 of the Czech government.

B-726

11:15

Contrast-enhanced wideband harmonic imaging ultrasound and contrast media “blood-pool”: Quantitative measurement of inﬂammatory activity in Crohn’s disease using parametric imaging A. De Franco, A. Di Veronica, L. Guidi, A. Armuzzi, I. Roberto, I. De Vitis, G. Fedeli, F. Danza, E. Bock; Rome/IT ([email protected])

B-727

11:24

Magnetic resonance enteroclysis compared with conventional enteroclysis and computed tomography enteroclysis: A critically appraised topic R. Ryan, I.S.E. Heaslip, D.E. Malone; Dublin/IE Purpose: Recent advances in CT and MR technology, particularly the advent of multidetector CT (MDCT), the advent of rapidly changing gradients in industry standard MRI scanners enabling ultrafast sequences, have led to an expansion in the role of cross sectional imaging in the investigation of small bowel disorders. The purpose of this study was to conduct an evidence-based review of MR enteroclysis (MRE) and how it performs in comparison to CT enteroclysis (CTE) and the gold standard of conventional enteroclysis (CE) for diagnosis of small bowel Crohn’s disease and small bowel neoplasia. Methods and Materials: We used the standard 5-step evidence-based medicine method of ask, search, appraise, apply and evaluate. We performed a Pubmed search based on a standard PICO (Patient, Intervention, Comparison, Outcome) question using MESH terms. Each study was appraised and ranked as per the Oxford CEBM levels of evidence. Results: We found three relevant level 1B studies, and one level 3B study. No studies evaluating MRE in small bowel neoplasia were found. MRE does not perform as well as CE in evaluation of ﬁne mucosal detail, but the additional extraluminal detail and absence of ionising radiation enhances its overall performance. Conclusion: It was not possible to establish the relative diagnostic performances of MRE and CTE from existing literature. CTE does involve patient irradiation. For patients in whom jejunal intubation and enteroclysis are considered to evaluate the small bowel, MRE should be considered the ﬁrst-line investigation, local resources and expertise permitting.

B-728

11:33

MR enteroclysis in the evaluation of the small bowel neoplasms G. Masselli, E. Casciani, E. Polettini, L. Bertini, G. Gualdi; Rome/IT ([email protected]) Purpose: The role of MR in small bowel neoplasms has not been fully established in previous studies, due to the limited experience with these rare tumors. Aim of this study was to evaluate the usefulness of MR Enteroclysis (MRE) in the diagnosis of small bowel neoplasms. Methods and Materials: 110 patients (63 males and 47 females), who were suspected of having palpable abdominal mass (n=31), low grade small bowel obstruction (n=37), unexplained gastrointestinal bleeding (n=26), refractory celiac sprue (n=16) prospectively underwent MRE. MRE were performed on 1.5 T magnet by injection of 1.5-2 l of polyethylglycol (PEG) solution via a nasojejunal catheter. MRE ﬁndings were analyzed by two radiologists, working in consensus, and were compared with the results of endoscopy, enteroscopy, videocapsule endoscopy, surgery and clinical follow-up. Results: MRE was well tolerated in all 110 patients. In 12/110 patients (11%) that refused nasojejunal intubation MRE was performed with oral administration of PEG solution. MRE allowed the diagnosis small bowel neoplasms (n=19), small bowel lymphoma complicating celiac disease (n=5), IBD (n=27). Negative examination included 59 patients. Tumors ranged in diameter from 0.5 cm to 9 cm (mean range 4 cm). Sensitivity, speciﬁcity, accuracy, positive and negative predictive values of MRE were 94, 90, 92, 89, and 94%, respectively. Conclusion: MRE allows visualization of luminal, parietal and mesenteric abnormalities and with functional, morphologic, and vascular informations is an accurate examination for the evaluation of small bowel neoplasms.

B-729

11:42

MSCT-enteroclysis (MSCTe) in the diagnosis of bowel endometriosis E. Biscaldi, S. Ferrero, V. Remorgida, G.A. Rollandi; Genoa/IT ([email protected]) Purpose: To prospectively study the effectiveness of MSCTe in bowel endometriosis diagnosis over a three-year period. Methods and Materials: 365 women with clinical suspicion of bowel endometriosis

A

B

C

D

E

F

G

311

Tuesday

Purpose: Evaluate patients with ileal Crohn’s Disease (CD) using contrastenhanced wideband harmonic imaging US (SonoVue® - SV) by means of HDIQLab software elaboration to assess the correlation between US data and clinical/ laboratory activity indexes of disease. Methods and Materials: After informed consent, thirty-six patients (M:24, F:12, age range 17-58 years) with ileal CD were studied by digital US scanner Philips iU2 with dedicated software for low-mechnical-index studies and 4-8 MHz linear-array probe. According to Vienna classiﬁcation, 10 patients had inﬂammatory, 13 stricturing and 13 ﬁstulizing CD. Forty-ﬁve US examinations after SV i.v. administration (4, 8 ml) were performed. US data obtained allowed drawing time-intensity curves through software-assisted HDI-QLab and to obtain following parameters: contrast time of arrival (TA), time to peak (TTP), wash-in, maximum peak intensity, echo-media, echo-median, echo-standard deviation, area under the curve (AUC), wash-out (T/2). These quantitative data were correlated with CDAI, biological activity, C-reactive protein (CRP), erythrocyte sedimentation rate (ESR), white blood cells count (WBC), alpha-1 acid glycoprotein and ﬁbrinogen (student’s t-test, signiﬁcant p 0.05). Results: PCR correlated with AUC (p=0.028), TA (p=0.018), TTP (p=0.011), wash-in

(p=0.05); alpha-1 acid glycoprotein correlated with TTP (p=0.034), ﬁbrinogen and peak velocity (p=0.04). Active disease showed values of maximum intensity peak signiﬁcantly (p=0.038) higher than non-active disease. Conclusion: Signiﬁcant correlations were found between quantitative HDI-QLab software elaborated contrast-enhanced wideband harmonic imaging US and clinical and inﬂammatory parameters in CD. Quantitative contrast-enhanced US could improve the reliability of conventional US to detect active bowel wall inﬂammation in CD.

Scientific Sessions underwent both MSCTe and surgery. MSCTe (16 rows) was performed after intestinal cleansing, during intestinal hypotonization.The X-ray dose was optimised reducing the exposure of 20%, compared with standard studies. Multiplanar reconstructions (MPRs) were performed. Results: We detected bowel endometriotic nodules in 224 (61.4%) women; in 140 (38.4%) the colon was normal, one had Crohn’s disease. Endometriotic nodules are rounded, solid, inhomogeneously enhanced; they inﬁltrate the bowel wall from outside, a distinctive sign in differential diagnosis versus adenocarcinoma.The deeper nodules inﬁltrated the serosa in 55 (24.6%) patients, the muscularis propria in 117 (52.2%), the submucosa in 39 (17.4%), the mucosa in 13 (5.8%). 24 (10.7%) cases of ileal endometriosis were detected. Surgery and histology conﬁrmed the bowel endometriosis in all the patients identiﬁed by MSCTe. In 12, MSCTe did not identify serosal nodules. Sensibility, speciﬁcity, positive predictive value, negative predictive value for diagnosing bowel endometriosis were 94.9, 100, 100, 92.2%. MDCTe identiﬁed totally 173 nodules; in three cases the presence of nodules was not surgically conﬁrmed, these false positives were determined by sigmoid distortion due to post-surgical adhesions.17 additional serosal nodules, not detected at MSCTe, were surgically excised. Among the nodules correctly identiﬁed at MSCTe, the depth of bowel inﬁltration was correctly estimated in 149 cases (87.6%; 95% C.I. 81.7-92.2%); in 9.4% MSCTe overestimated the submucosal inﬁltration. MSCTe does not detect the nodules 7 mm. Conclusion: MSCTe is effective in diagnosing bowel endometriosis as a preoperative planning; the X-ray dose is optimizable.

B-732

B-730

11:51

Purpose: Diagnosing acute appendicitis using ultrasonography (US) is quite difﬁcult even if the appendix is visible. We wanted to determine which of the US parameters described in the literature is the most effective in the diagnostics of appendicitis in a series of large patient numbers and high frequency of visualized appendix. Methods and Materials: 838 (536 female, 302 male, mean 36.5+18.1 years) patients underwent US due to suspected appendicitis. In 580 (69%) the appendix was visible. The following criteria of appendicitis were compared with the ﬁnal diagnosis using SPSS statistical analysis: anterio-posterior diameter under compression (OD) 6 mm, maximal wall thickness (MWT) 3 mm, hypervascularisation (CD) (subjective 0-3 scale with standardised PRF, gain), absence of air in the lumen (AIR-), non-compressibility (nonCOMP), round cross-section (ROUND), inﬁltration of fatty tissue around (FAT) and pain just at the location of the visible appendix (PAIN). Histology was the ﬁnal diagnosis in operated cases and patients without surgery were considered negative if still not operated after a 2-week follow-up. Results: Of the 838 patients, 344 were operated and had histology. The following criteria reached the highest diagnostic values: sensitivity OD (97.4%), speciﬁcity CD (91.6%), positive prediction OD (83.1%), negative prediction OD (96.9%), accuracy OD (88.7%), although the differences were not statistically signiﬁcant between the ﬁrst three positions. Conclusion: In a clinical setting OD seems to be the most effective criterion in both proving and excluding appendicitis sonographically. To avoid false positives, adding CD can help.

10:30 - 12:00

Room E1

Cardiac

SS 1803a Coronary artery imaging Moderators: F. Pugliese; Rotterdam/NL H.-J. Smith; Oslo/NO

B-731

10:30

NIMIS-CAD study: Multicenter intraindividually controlled comparison study to evaluate the NPV of 16 and 64 slice MDCT imaging in patients scheduled for coronary angiography R. Marano1, F. De Cobelli2, C. Becker3, C. Herzog4, I. Floriani2, T. Nimiscad Study Group2, A. Del Maschio2, L. Bonomo1; 1Rome/IT, 2Milan/IT, 3 Munich/DE, 4Frankfurt/DE ([email protected]) Purpose: To assess the diagnostic performance of MDCT in patients enrolled for elective X-ray coronary-angiography (CA) in a prospective Italian multicenter study by using different scanner manufacturers.

A

B

C

D

E

F

G

10:39

Learning curve in dual-source CT coronary angiography in high risk patients T. Adla, J. Neuwirth, J. Veselka, V. Suchánek, M. Polovincák, R. Duchonová, L. Martinkovicová; Prague/CZ ([email protected])

Comparison of ultrasonographic signs of appendicitis in visible appendices Z. Tarján, A. Stefan, L. Lukács; Budapest/HU ([email protected])

312

Methods and Materials: Twenty participating sites prospectively enrolled 367 patients from July 2004 to June 2006, with a sample size (350) according to the targeted number. In every institution an MDCT scan was performed within two weeks before CA using a predeﬁned state-of-art scan-protocol for each type of 16- and 64-slice MDCT scanner, during iv injection of cm (Iopromide 370; BayerSchering-Pharma AG, Berlin, Germany) and characterized by the thinnest slice collimation (0.5-0.75 mm) and temporal resolution (single-segment algorithm) of 165-200 msec. MDCT ﬁndings were compared with quantitative CA. Patient with heart-rate 70 bpm received a ß-receptor-blocker iv or orally. Sensitivity, speciﬁcity, positive predictive value (PPV), negative predictive value (NPV) and diagnostic accuracy (DA) for detection of stenosis r50% were evaluated on per-patient and per-segment basis. An off-site per-patient analysis was also performed by two blinded readers. Results: Per-patient in-site analysis was conducted on 327 patients (23 with at least one coronary segment not assessable). Sensitivity, speciﬁcity, PPV, NPV, and DA of MDCT for detection of stenosis r50% were 94, 88, 91, 91, and 91%, respectively for in-site analysis, and 73, 93, 93, 73, and 82% for off-site evaluation. In per-segment (4628) in-site analysis the sensitivity, speciﬁcity, PPV, NPV, and DA were 70, 96, 72, 96, and 93%, respectively. Conclusion: This national multicenter study conﬁrms the capability of MDCT to exclude signiﬁcant CAD, even if conducted by using different scanner manufacturers.

Purpose: To evaluate progress of diagnostic accuracy for assessment signiﬁcant coronary artery disease at radiological department with a minimal previous experience with evaluation of cardiac computed tomography and newly installed dual source CT (DSCT). Methods and Materials: 134 patients (87 males, mean age 63.9p11.7 years) underwent CT coronary angiography (CTCA). The subgroup of 56 patients (40 males, mean age 65.9p7.9 years) underwent invasive coronary angiography (ICA) in following 1-2 days. Patients after coronary revascularisation were excluded. Imaging methods were assessed for signiﬁcant stenosis (narrowing 50%), CTCA by two radiologists and ICA by one cardiologist. Segments with a diameter 1.5 mm were excluded. Sensitivity (SE), speciﬁcity (SP), positive predictive value (PPV) and negative predictive value (NPV) were calculated for three equal learning periods (1-45, 46-90, 91-134 patients). Results: Mean heart rate during DSCT was 64.4p13.2. 659 coronary segments had diameter 1.5 mm, no segment has been considered as inaccessible. Mean calcium score was 557p672. SE, SP, PPV and NPV for the ﬁrst period were 40.0, 95.1, 66.7, 86.5%, for the second, 80.5, 96.0, 76.7, 96.8%, and 84.4, 96.4, 82.6, 96.8% for the third. Conclusion: The study suggests a learning curve of diagnostic accuracy of CTCA. The numbers indicate that sufﬁcient accuracy per segment was reached after 90 CTCA, but SE and PPV were growing substantially till the end of study.

B-733

10:48

Blood pool contrast agent for coronary MR angiography: Utility of ﬁrst pass and steady state acquisitions I. Carbone, E. Algeri, H. Grazhdani, D. Cannata, F. Ciolina, N. Galea, C. Catalano, R. Passariello; Rome/IT ([email protected]) Purpose: MR angiography is becoming a promising strategy for non invasive evaluation of coronary arteries, thanks to improvements in terms of spatial and temporal resolution. Blood-pool contrast agents are now available for cardiovascular applications. The aim of our study has been to evaluate and compare image quality of 3D coronary MR angiography performed with and without the intravenous (iv) injection of a blood pool contrast agent. Methods and Materials: Sixteen healthy volunteers underwent standard freebreathing 3D coronary MR-angiography, without any contrast media administration; after iv injection of 0.1 mMol/Kg of blood-pool contrast agent (Vasovist, Schering), ﬁrst-pass breath-hold acquisition was performed, followed by a steady-state acquisition. Quality of the three different exams was determined by two observers considering coronary segments visualized, vessel sharpness, signal intensity measured in each vessel, signal to noise ratio (SNR), visualization of coronary veins. An inter-individual/intra-individual comparison of image quality of the three sequences was performed. Interobservers’ agreement was calculated using kappa statistics.

Scientific Sessions Results: Signal intensity and vessel sharpness were signiﬁcantly higher at ﬁrstpass images (p 0.005). CNR and SNR were higher in sequences acquired after contrast media administration, compared with the precontrast acquisition (p 0.005). The number of segments identiﬁed was similar for the three groups. Coronary veins were visualized in a higher number of exams performed during steady state (n=13/15) compared with images acquired with no contrast (n=4/15) or at ﬁrst-pass (n=2/15). Conclusion: High resolution ﬁrst-pass coronary MR angiography with Vasovist showed an increased quality of images compared to steady-state and pre-contrast images.

B-734

10:57

The relationship between the angle of the left main bifurcation and the formation and distribution of atheroma in the left main bifurcation by 64-slice spiral CT X.-X. Tang1, J. Yuan1, K.-M. Chen2; 1Taiyuan/CN, 2Shanghai/CN ([email protected]) Purpose: To explore formation and distribution of the plaque of the left main bifurcation and its relationship with the bifurcation angle using high-resolution 64-slice CT coronary angiography (64SCTA). Methods and Materials: Fifty-ﬁve patients were evaluated using a 64SCT scanner. The mean age was 53p11.3 years. MIP, 3D volume rendering techniques and 3D reconstruction of the entire coronary tree were used to demonstrate the whole coronary artery. MIP of the coronary artery was used to calculate the angle of the bifurcation using a caliper tool provided by customized software, to calculate the distance between intersection of the LMCA bifurcation and vertical line of middle point in atherosclerotic plaques. Two readers independently assessed image quality. Results: Of the 29 patients with a normal LAD, 21 (72%) had a bifurcation angle 78.3o, whereas the 69% of the patients with any LAD disease had an angle 78.3o (p=0.002). Of the 43 patients with a normal LCX, 25 (58%) had a bifurcation angle 78.3o,whereas the 67% of the patients with any LCX disease had an angle 78.3o (p=0.128). Of the 23 patients with a LAD disease, the angle of the left main bifurcation and the distance between plaques and the bifurcation are negatively correlated (r=-0.515; p=0.012). Conclusion: As a potential factor, the left main bifurcation angle affects ﬂuid dynamics in the formation and distribution of atherosclerotic plaques within the left main bifurcation.

B-735

11:06

64-slice computed tomography coronary angiography: Diagnostic accuracy in the “real world” F. Cademartiri1, E. Maffei2, A. Palumbo2, M. Fusaro2, F. Notarangelo2, A. Aldrovandi2; 1Rotterdam/NL, 2Parma/IT (ﬁ[email protected])

11:15

Detection of coronary artery stenosis with 3 Tesla coronary MR angiography L. Job, F. Bonnet, C. Hoeffel, S. Perotin, A. Coupteau, C. Marcus; Reims/FR ([email protected]) Purpose: We sought to determine the feasibility of 3 Tesla coronary MR angiography for detecting coronary artery stenosis. Methods and Materials: 25 patients (15 male, 10 female, age range 30-82 years; mean 57.9) were prospectively imaged with both X-ray coronary angiography and 3 Tesla coronary MR angiography. Patients had either a history of infarct, angina, positive stress exercise, or risk factors for coronary artery disease. MRA was performed on a 3 T MR unit (Achieva, Philips) using a 6-element cardiac phased-array coil and parallel imaging. The coronary images were acquired using 3D sequences during free breathing and prospective respiratory navigator-gating. MIP reconstructions were performed. A subjective analysis and grading of the image quality of vessel lumen was performed by two experienced investigators, independently. Images were also evaluated for the detection of stenosis. The RCA was divided into three segments and an arbitrary division was used for the other arteries. Cohen’s kappa test was used to calculate interobserver’s agreement. Results: 184 segments with 37 stenoses were analyzed. Interpretation was feasible for 96 segments. Sensitivity of MRI for the detection of stenoses was 52 and 73%, and speciﬁcity 87% for each observer. Kappa was 0.64, corresponding to a good interobserver agreement. Conclusion: 3 Tesla coronary MR angiography is feasible. However, further technical advances are required to improve the sensitivity of this modality.

B-737

11:24

Comparison of 16 and 64 multi-detector row spiral computed tomography for noninvasive detection and evaluation of coronary atherosclerotic plaques B. Lu, W. Wu, S. Qiao, S. Jiang; Beijing/CN ([email protected]) Purpose: To evaluate and compare the ability and accuracy of 16- and 64-multidetector row spiral computed tomography (MDCT) in detecting coronary atherosclerotic plaques. Methods and Materials: Contrast-enhanced coronary artery scans were performed on 16-row (in 30 patients) and 64-row (in 18 patients) MDCT scanner (Lightspeed Ultra-16 and VCT, GE, USA). All of these 48 patients had undergone conventional coronary angiography and intracoronary ultrasound (ICUS). A total of 119 coronary segments were studied. MDCT datasets were blindly evaluated for the presence, areas of plaques and plaque compositions in comparison with ICUS results on a segmental basis. Results: Ninety-four and 25 coronary artery segments were imaged and analyzed on 16-row and 64-row MDCT, respectively. For the detection of segments with any plaque, 16-row and 64-row MDCT had a sensitivity of 82% (46/56) and 96% (22/23), respectively (p 0.05), and speciﬁcity of 89% (34/38) and 100% (2/2), respectively (p 0.05). The mean plaque areas measured by ICUS and 16-row MDCT were 5.3 mm² and 5.6 mm² (r=0.58, p 0.01). The mean plaque areas measured by ICUS and 64-row MDCT were 7.8 mm² and 9.2 mm² (r=0.78, p 0.01). Conclusion: The results indicated that MDCT had potential to detect coronary atherosclerotic plaque with higher sensitivity and speciﬁcity especially on 64-row MDCT. Measurements of plaque areas derived by MDCT correlated well with ICUS. MDCT as a noinvasive method might become an important tool for risk stratiﬁcation of coronary heart disease.

B-738

11:33

Feasibility of 64-slice cardiac CT in obese patients: A retrospective analysis of assessable arteries and segments in 567 patients R. Salgado, B. Shivalkar, P.M. Parizel, C. Vrints; Edegem/BE ([email protected]) Purpose: To study the feasibility of CT angiography of the coronary arteries in obese patients. Methods and Materials: The results of 567 patients with a BMIr30, referred for CT angiography of the coronary arteries, were retrospectively reviewed. All patients were scanned with a 64-slice CT. Three subgroups were deﬁned: Obesity class I (BMI 30-34, n: 417, 6328 coronary segments), obesity class II (BMI 35-39, n: 104, 1573 segments) and morbid obesity (BMIr40, n: 46, 696 segments), with a total of 8597 coronary segments for analysis. Results: The average percentage of assessable segments per patient was 94.1 (92.7-95.5) in the class I group, 93.0 (89.4-94.0) in the class II group, and 85.8 (82.6-89.1) in the morbid obesity group. On a per coronary artery analysis, the

A

B

C

D

E

F

G

313

Tuesday

Purpose: To evaluate diagnostic accuracy of 64-slice CT coronary angiography (CT-CA) for the detection of signiﬁcant coronary artery stenosis in the real clinical world. Methods and Materials: From the database of CT-CA of our Institution we enrolled 145 patients (92 males, 52 females, mean age 63.4p10.2 years) with suspected coronary artery disease. All patients presented with atypical or typical chest pain and underwent contrast enhanced (Iomeprol 400, Bracco) CT-CA (Sensation 64, Siemens) and conventional coronary angiography (CAG). The reports of CT-CA and CAG were used to evaluate diagnostic accuracy using the threshold for signiﬁcant stenosis at r50% for CT-CA and r70% for CAG. Results: Eleven patients were excluded from the analysis because of a non diagnostic quality of CT-CA. The prevalence of disease demostrated at CAG was 63% (84/134). Sensitivity, speciﬁcity, positive and negative predictive values for CT-CA on a per-segment, per-vessel, and per-patient basis were 75.6, 85.1, 97.6; 8.9, 81.8, 58.0%; 48.2, 68.1, 79.6%; and 95.7, 92.3, 93.5%; respectively. Two patients of 134 eligible resulted in false negative. The heart rate did not show signiﬁcant inﬂuence on diagnostic accuracy, while the absence or the presence of low coronary calciﬁcations improved diagnostic accuracy. The positive and negative likelihood ratios at per-patient level were 2.32 and 0.041, respectively. Conclusion: CT-CA in the real clinical world shows a diagnostic performance lower than in previously reported validation studies. The excellent negative predictive value and negative likelihood ratio make CT-CA a non invasive gold standard for the exclusion of signiﬁcant coronary artery disease.

B-736

Scientific Sessions largest decline in assessability was noted in the LAD (95.5-82.6%), followed by the LCX (94.7-85.9%). On a per coronary segment analysis, the largest decline in assessability was noted in the proximal LAD (97.3-80.4%), followed by the distal LAD (94.7-80.4%). Conclusion: A BMIr40 signiﬁcantly decreases the number of assessable segments, with the LAD and LCX being the most affected coronary arteries, and the proximal and distal LAD the most affected segments. However, no signiﬁcant difference was noted between class I and II obesity patients, with an overall excellent percentage of per patient segment assessability of 93-94.

B-739

11:42

Whole-heart coronary MRA at 3.0 Tesla: A comparison study of MRA and CTA in patients with coronary stenosis Q. Yang1, K. Li1, D. Li2, R. Merges1, J. Chen1, J. An1; 1Beijing/CN, 2Chicago, IL/US ([email protected]) Purpose: To compare the diagnostic accuracy of whole-heart coronary MRA at 3.0 Tesla with 64-slice MDCT for detecting coronary stenosis. Methods and Materials: Twelve patients with suspected coronary artery disease underwent a 64-slice MDCT coronary angiography (Sensation Cardiac 64, Siemens AG, Germany). All the12 patients also underwent an MR examination within three days with a 3.0 T MR whole-body scanner (Trio, Siemens AG, Germany). A navigator-gated, ECG-triggered, segmented FLASH sequence was employed for whole-heart coronary MRA. 56 transverse slices were acquired with 1.3 mm slice thickness (interpolated to 112 slices of 0.65 mm thickness). The in-plane spatial resolution was 1.3×1.3 mm (interpolated to 0.65×0.65 mm2). Other imaging parameters included: TR/TE=3.0/1.4 mm; ﬂip angle=15°; readout bandwidth=610 Hz/pixel; voxelsize=1.3×1.3×1.3 mm. Contrast-agent (0.2 mmol/kg body weight, Multihance, Bracco Imaging SpA, Italy) was intravenously administered at the rate of 0.3 ml/ sec using a Medrad power injector. The accuracy of 3.0 T CE-MRA in detecting coronary stenosis was evaluated by two experienced radiologists in consensus without knowledge of the ﬁndings on MDCT. Results: Fifteen plaques with signiﬁcant stenosis ( 50%) were detected on 64-slice MDCT. Thirteen plaques with signiﬁcant stenosis ( 50%) were correctly detected by CE-MRA compared to MDCT. Two plaques were not detected with MR. There are 31 accessible vessels, and 5 non accessible vessels on CE-MRA. Conclusion: Whole-heart contrast-enhanced coronary MRA showed good diagnostic accuracy for detection of signiﬁcant stenosis. Further optimization of the contrast injection scheme and improvements in spatial resolution and imaging speed are still required for clinical applications.

B-740

11:51

Is 64-slice computed tomography coronary angiography accurate and safe in asymptomatic, intermediate-to-high Framingham risk patients to exclude coronary artery disease? A prospective study G. Feuchtner, D. Jodocy, D. Junker, A. Klauser, W. Dichtl, O. Pachinger, P. Soegner, D. zur Nedden, G. Friedrich; Innsbruck/AT ([email protected]) Purpose: To assess the accuracy and safety of 64-slice computed tomography (CT) coronary angiography prospectively in asymptomatic patients with clinically suspected coronary artery disease (CAD) in a routine clinical setting. Methods and Materials: 102 asymptomatic patients with an intermediate (62/102; 61%) or high-risk Framingham risk proﬁle (40/102; 39%) and an inconclusive, borderline or mild-pathologic treadmill ECG-stress test were examined prospectively with 64-slice CT (Sensation 64TM, Siemens) (64x0.6 mm; 0.33 s; pitch 0.2; 120 kV/700-900 mAs, eff.sl. 0.75 mm/inc. 0.4/B 25 f; retrospective ECG-gating; 80-100 ml iodixanol (Visipaque 320TM,GE) or iomeprol (Iomeron 400TM, Bracco), 5-6 ml/s ﬂow rate, bolus-tracking. A 9 month-follow-up (phone-call) was performed. Invasive coronary angiography (IA) was performed if indicated. Results: Coronary plaques were found in 78% (80/102). 2.9% (3/102) of patients had exclusively non-calcifying plaques. 59% (60/102) had non-obstructive plaques. The prevalence of 50% stenosis conﬁrmed by IA was 18% (19/102). The diagnostic accuracy of CT to detect coronary stenosis 50% per coronary segment was: sensitivity 91% (30/33), speciﬁcity 95% (338/357), PPV 61% (30/49); NPV 99% (338/342). After nine months, none of the patients had a cardiac event or onset of symptoms [The accuracy to diagnose "signiﬁcant CAD" (either 50% stenosis or cardiac event) per patient was: sensitivity 100% (19/19), speciﬁcity 93% (77/83), PPV 76% (19/25), NPV 100% (77/77)]. Conclusion: 64-slice CT coronary angiography used as a second non-invasive test in asymptomatic, intermediate-to-high Framingham risk individuals is safe and accurate to rule out signiﬁcant CAD. 64-slice CT revealed a high percentage of patients with non-obstructive coronary plaques, who would beneﬁt from speciﬁc medication.

314

A

B

C

D

E

F

G

10:30 - 12:00

Room F1

Genitourinary

SS 1807 Prostate imaging: Diagnosis, therapy and follow-up Moderators: N.A. Al-Nakshabandi; Riyadh/SA J.C. Vilanova; Girona/ES

B-741

10:30

Comparison of preoperative prostate cancer staging performance between 3 T endorectal MR imaging and real-time transrectal ultrasound S.W.T.P. Heijmink, T. Hambrock, T.W.J. Scheenen, C.A. Hulsbergen-van de Kaa, B.C. Knipscheer, J.A. Witjes, J.O. Barentsz; Nijmegen/NL ([email protected]) Purpose: The purpose of this study was to compare the local staging accuracy of real-time gray-scale transrectal ultrasound (TRUS) and T2-weighted 3 T endorectal (ERC) MRI with whole-mount section histopathology. Methods and Materials: After written informed consent, 25 consecutive patients with biopsy-proven and clinically localized prostate cancer underwent 3 T ERC MRI prior to radical prostatectomy. Three-directional T2-weighted imaging was obtained (TR/TE 5000/153 ms; voxel size: 0.26x0.26x2.50 mm3). Subsequently, real-time axial gray-scale TRUS (Aplio system, Toshiba) movie clips were recorded. Radiologists A and B, with different levels of experience, independently read all imaging sets and scored the local disease stage on a 5-point probability scale. Whole-mount section histopathology was the standard of reference. Areas under the ROC curve (AUC) were determined and diagnostic performance parameters were calculated. p 0.05 was statistically signiﬁcant. Results: Nine patients had locally advanced disease. For the experienced radiologist A, the AUC for TRUS and MRI were 0.69 and 0.96, respectively (difference, p 0.05). For the less experienced radiologist B, the AUC for TRUS and MRI were 0.65 and 0.83, respectively (p 0.05). Sensitivity for detecting locally advanced disease increased signiﬁcantly for radiologist A from 22% (2/9) with TRUS to 89% (8/9) with MRI (p 0.05). Also for radiologist B, sensitivity increased signiﬁcantly from 0% (0/9) to 67% (6/9), respectively (p 0.05). Speciﬁcity was high for both MRI (88%) and TRUS (94-100%). Conclusion: 3 T ERC MRI had a signiﬁcantly higher local staging AUC compared with TRUS for the experienced reader and signiﬁcantly improved staging sensitivity for both readers. Both TRUS and MRI achieved high speciﬁcity.

B-742

10:39

Value of endorectal coil T2-weighted MRI combined with 3D-CSI spectroscopy in the diagnosis of prostate cancer in patients with elevated PSA G.M. Villeirs, W. Oosterlinck, V. Fonteyne, G.O. De Meerleer; Gent/BE ([email protected]) Purpose: To evaluate combined magnetic resonance imaging and spectroscopy (MRI/MRS) in the detection of prostate cancer. Methods and Materials: Between March 2002 and August 2007, 677 patients were examined with 4-mm transverse, coronal and sagital fast-T2-weighted images and 3D-CSI MRS on a 1.5 T scanner. Follow-up data (pathology report or at least 12 months clinical follow-up) were available for 366 patients. Results: The sensitivity, speciﬁcity, and accuracy were 55.3, 94.2, and 69.9% for MRI alone, 61.4, 94.9, and 74.0% for MRS alone, and 70.2, 92.0, and 78.4% for combined MRI/MRS, respectively. Corresponding Az-values (area under ROC curve) were 0.79, 0.80, and 0.85, respectively. There was a moderate-strong correlation between Gleason score and the sensitivity of MRI, MRS and MRI/MRS (corresponding Pearson correlation coefﬁcients of 0.78, 0.74, and 0.77, respectively). The sensitivity for Gleason 3+4 or lower grade tumors was 65.1% and for Gleason 4+3 or higher grade tumors 92.9%. The negative predictive value for the latter tumors was 97.7%. Conclusion: MRS is a very valuable adjunct to MRI in the assessment of patients with elevated PSA, resulting in a 15% increase of sensitivity for a comparable speciﬁcity. A negative MRI/MRS virtually excludes the presence of a Gleason 4+3 or higher grade tumor.

Scientific Sessions B-743

10:48

Differentiation of normal prostate tissue, chronic prostatitis, and prostate cancer by quantitative perfusion analysis using a dynamic contrastenhanced inversion-prepared dual-contrast gradient echo sequence T. Franiel, L. Lüdemann, B. Rudolph, H. Rehbein, A. Staack, M. Taupitz, D. Prochnow, D. Beyersdorff; Berlin/DE ([email protected]) Purpose: To quantify independent pharmacokinetic parameters for differentiation of prostate pathology. Methods and Materials: The new dynamic contrast-enhanced, inversion-prepared dual-contrast gradient echo sequence (T1/T2*-weighted, 1.65 s temporal resolution) was acquired in 27 patients with biopsy-proven prostate cancer using a combined endorectal/body phased-array coil at 1.5 Tesla. Perfusion, blood volume, mean transit time, delay, and dispersion were calculated using a sequential 3-compartment model. Twenty-three patients underwent prostatectomy. For histological correlation a pathologist mapped areas of normal prostate tissue, chronic prostatitis and prostate cancer (total of 63 areas) on histological sections corresponding to the MRI plane. Results: Histological correlation showed higher perfusion for low-grade cancer (Gleason scoreb6) versus normal prostate tissue (1.01 vs 0.26 ml/cm3*min, p=0.050) whereas high-grade cancer (Gleason scorer7) showed higher perfusion (1.21 vs 0.26 ml/cm3*min, p 0.001), higher blood volume (1.44 vs 0.95%, p=0.005), shorter mean transit time (3.55 vs 4.40 s, p=0.019), shorter delay (10.15 vs 13.36 s, p=0.015) and smaller dispersion (8.56 vs 12.11 s, p=0.020). High-grade cancer showed higher perfusion than chronic prostatitis (1.21 vs 0.90 ml/cm3*min, p=0.041). Compared to normal prostate tissue, chronic prostatitis showed higher perfusion (0.90 vs 0.26 ml/cm3*min, p=0.006), higher blood volume (1.53 vs 0.95%, p=0.046), shorter delay (11.42 vs 13.36 s, p=0.015) and smaller dispersion (10.49 vs 12.11 s, p=0.020). There were no statistically signiﬁcant differences between low-grade and high-grade cancer or between low-grade cancer and chronic prostatitis. Conclusion: The pharmacokinetic parameters investigated, especially perfusion, allow statistically signiﬁcant in situ differentiation of normal prostate tissue from cancer and chronic prostatitis and of high-grade cancer from chronic prostatitis with unprecedented accuracy.

B-744

10:57

Correlation of diffusion-weighted MR imaging (DWI) and 11Cholin PET-CT in the differentiation of benign and metastatic pelvic and inguinal lymph nodes in patients with prostate cancer: Preliminary experience M. Eiber, K. Holzapfel, C. Ganter, M. Souvatzoglou, E.J. Rummeny, J. Gaa; Munich/DE ([email protected]) Purpose: To determine whether diffusion-weighted MR imaging (DWI) using a high resolution echo-planar imaging (SSEPI) sequence is helpful for differentiating between benign and metastatic pelvic and inguinal lymph nodes in patients with prostate cancer. Methods and Materials: 17 patients with prostate cancer and former imaging by 11 Cholin PET-CT underwent MR imaging at 1.5 T (Avanto, Siemens) using a body phased array coil. DWI was performed by a non breath-hold SSEPI sequence with b values of 50, 300 and 600 sec/mm2 (TR=d, TE=69 ms, matrix 196x144, slice thickness 5 mm, GRAPPA 2). Pelvic and inguinal lymph nodes with a diameter of at least 10 mm were selected in T2- and SSEPI sequences and Apparent Diffusion Coefﬁcients (ADCs) were calculated for each lesion. Results of prior 11Cholin PET-CT were taken as gold standard. In total, 15 11Cholin PET-CT positive and 23 negative lymph nodes were evaluated. Mann-Whitney test was used for statistical analysis. Results: The mean ADC values (10-3 mm2/sec) were 1.13p0.22 for 11Cholin PET-CT positive lymph nodes and 1.83p0.38 for 11Cholin PET-CT negative lymph nodes. Statistical analysis showed a highly signiﬁcant difference between ADC values of metastatic and tumor-free lymph nodes. Conclusion: Calculation of ADC value seems to have the potential to differentiate between malignant and benign lymph nodes in patients with prostate cancer. Nevertheless, further correlation with histopathologic ﬁndings is necessary to conﬁrm these preliminary ﬁndings.

11:06

Effect of hormonal therapy on dynamic CAD-based analysis of prostate cancer prior to iMRI-therapy A. Malich, O. Mikulik, A. Kott, S. Mikulik, D. Hentrich, W. Oehler; Nordhausen/DE ([email protected]) Purpose: To analyze inﬂuence of antiandrogen hormonal therapy on dynamic prostate MRI, permeability and tumor volume calculation using CAD. Methods and Materials: All prostatic MRI performed on behalf of our radiooncologic department during 2006 were analyzed (n=112). Eighty-one adenocarcinoma, two

B-746

11:15

Value of dynamic contrast-enhanced MR imaging in detecting local tumor progression after HIFU ablation of localized prostate cancer C. Kim, B. Park, H. Choi, H. Lee; Seoul/KR ([email protected]) Purpose: To evaluate the diagnostic performance of dynamic contrast-enhanced MR imaging (DCEI) in detecting local tumor progression after high-intensity focused ultrasound (HIFU) ablation of localized prostate cancer. Methods and Materials: Twenty-three patients with increased prostate-speciﬁc antigen levels after HIFU ablation were included in this study. All MR examinations were performed using T2-weighted (T2WI) and DCEI, followed by transrectal ultrasound-guided biopsy. The MRI and biopsy results were correlated in six prostate sectors. Residual or recurrent prostate cancer after treatment was deﬁned as local tumor progression if biopsy results showed any cancer foci. Two independent readers interpreted the MR images. Results: Biopsy cores were obtained in 138 sectors. Of them, 47 (34%) prostate sectors were positive for cancer in 16 patients. For the detection of local tumor progression of prostate cancer, the sensitivity of DCEI and T2WI were 70 and 43% for reader 1 (p 0.001) versus 87 and 64% for reader 2 (p 0.001), respectively, whereas the speciﬁcity of DCEI and T2WI were 60 and 74% for reader 1 (p 0.001) versus 48 and 59% (p=0.002) for reader 2, respectively. The accuracy of DCEI and T2WI in both readers was similar. For the detection of local tumor progression of prostate cancer, DCEI and T2WI showed moderate (k=0.52) and fair agreement (k=0.21), respectively. Conclusion: For detecting local tumor progression of prostate cancer, DCEI was more sensitive than T2WI, with less interobserver variability than T2WI. However, T2WI was more speciﬁc than DCEI.

B-747

11:24

MR imaging in the prediction of the probability of biochemical recurrence in patients undergoing radical prostatectomy M.H. Fuchsjäger1, A. Shukla-Dave1, H. Hricak1, L. Wang1, J.H. Donohue1, P.T. Scardino1, J.A. Eastham1, M.W. Kattan2; 1New York, NY/US, 2Cleveland, OH/US ([email protected]) Purpose: To investigate whether the addition of endorectal magnetic resonance imaging (erMRI) data improves the accuracy of clinical nomograms for the prediction of freedom from biochemical recurrence after radical prostatectomy (RP). Methods and Materials: We conducted a retrospective study of 610 consecutive patients (median age, 60 years, range 36-87) with biopsy-proven prostate cancer who underwent erMRI before RP with whole-mount step-section pathology between January 2000 and October 2004. Preoperative erMRI ﬁndings were scored based on the TNM staging system as follows: 1, no tumor seen; 2, tumor seen, no extracapsular extension (ECE); 3, tumor seen, ECE cannot be ruled out; 4, unilateral ECE; 5, bilateral ECE; 6, seminal vesicle invasion; 7, lymph node metastasis. erMRI scores were added to the Kattan (1998) and Stephenson (2006) clinical preoperative nomograms for predicting biochemical recurrence at ﬁve and 10 years, respectively. The predictive accuracy of erMRI was quantiﬁed as the differences in bootstrapcorrected concordance indices of the models with and without erMRI. Results: As of August 2007, 64 (10.5%) patients had experienced biochemical recurrence. erMRI scores were associated with biochemical recurrence (p 0.0001) with a hazard ratio of 1.67. Actual recurrence rates by erMRI score were: 1, 0%; 2, 4.5%; 3, 9%; 4, 24.1%; 5, 33.3%; 6, 69.2%; 7, 100%. With the addition of erMRI data, the concordance index for the 10-year nomogram model increased from 0.849

A

B

C

D

E

F

G

315

Tuesday

B-745

other malignancies, two benign hyperplasia, 15 discrepant cases (MR suspicious, biopsy up to now benign or borderline, PSA still increasing) and six unsuspicious cases after surgical prostatectomy could be analyzed, whereas in six cases technical pitfalls occured during CAD-application (CADSciences, USA) not allowing a useful analysis. Permeability, extracellular volume fraction, dynamic curve types, volume calculation, PSA-values and hormonal treatment prior or during MRi were recognized and embedded in the analysis of the data. Results: Permeability of treated/untreated tumor was: perm25: 2.4/2.0; perm50: 3.7/3.5; perm75: 6.1/6.8. Mean wash in ratio ( 100% initially) with/without hormonal treatment was: 54.9/32.7%, wash out: 41.6/51.2%. Correlations of tumor volume and PSA-value were signiﬁcantly positive for untreated cases (r=0.50); treated cases r=0.08. Mean permeability values were: 3.74/3.39. Conclusion: Presurgical/preradiooncological hormonal treatment of prostate cancer signiﬁcantly inﬂuences CAD-based analysis of contrast uptake, permeability, detectable tumor-volume. Hormonal treatment seems to induce depression of less aggressive tumor parts initially causing a reduced visible tumor volume. Due to this preselection, permeability of still enhancing tumor parts is more aggressive, reﬂected by a higher permeability. Radiologists have to be aware of hormonal therapy as one of the major inﬂuencing parameters in order to avoid misinterpretation of prostate-MRI.

Scientific Sessions to 0.875, and the concordance index for the 5-year nomogram model increased from 0.836 to 0.864. Conclusion: The addition of erMRI improves clinical preoperative predictions of biochemical recurrence.

B-748

11:33

3D multiplanar TRUS with power Doppler in detecting prostatic fossa recurrences following radical prostatectomy V. Gazhonova, V. Solovyev, A. Zubarev; Moscow/RU Purpose: To identify the US features of local tumor progression by means of 3D TRUS with PD in patients with unequivocal (rising or persistent) PSA levels after radical prostatectomy (RP) for prostate cancer. Method and Materials: Fourteen patients with increased PSA levels following RP underwent 3D TRUS study for identifying local tumor recurrence. Mean duration after RP was 38 months (range 6-10 years). MRI and target biopsies was performed in all cases. 3D TRUS images were evaluated for the presence of focal lesions, residual tissue volume and vascularity degree (VD). Results: Nine of 14 pts had positive TRUS for residual tumor tisue. Cancer was detected in eight of 14 pts (2-T2bNoMo, 4-T3aNoMo, 2-T3bNoMo). 3D US coronal scans were superior to conventional 2D images for interpretation of tumor recurrence and target biopsy. The mean volume of identiﬁed tissue in the prostatic fossa was higher in the cases with tumor recurrence than in normal cases (p 0.05). There were very signiﬁcant differences for VD values between local recurrence areas than in non-cancerous zone (p 0.001). The overall accuracy, sensitivity and speciﬁcity of tumor recurrence detection after RP in patients with elevated PSA levels using 3D TRUS in were 93%, 89%, 83%, respectively. Conclusion: 3D TRUS with PD analysis appears to be clinically feasible with added anatomic information from the coronal plane for better detection of the recurrent tumor. Increase in the vascularity degree in the prostatic fossa could be used as criteria for diagnosis of tumor recurrence.

B-749

11:42

The role of diffusion tensor imaging with 3 Telsa in the diagnosis of prostate cancer B. Gurses, N. Kabakci, Z. Firat, T. Alp, A. Bayram, M.A. Ulug, M.M. Cihangiroglu, I. Kovanlikaya; Istanbul/TR ([email protected]) Purpose: The sensitivity and speciﬁcity of MRI in the diagnosis of prostatic carcinoma are limited. Recently, the role of diffusion and diffusion tensor imaging (DTI) are investigated as adjunctive methods to increase the diagnostic role of MRI. The aim of this study was to ﬁnd out the possible differences between DTI parameters of normal and cancerous prostate tissue. Methods and Materials: 18 patients with a high suspicion of prostate cancer with clinical and laboratory ﬁndings had undergone MR imaging with 3 Tesla system (Intera, Achieva, Philips). For each patient, axial T2 TSE, DTI with ss-EPI were obtained. TRUS guided biopsy was performed after the MRI examination. 15 patients were proven to have prostatic carcinoma. FA and ADC values of these patients were calculated for each biopsy zone. The parameters for normal and cancerous tissue were compared using t test as the statistical analysis. Results: The mean FA value of the tumor tissue (0.29p0.07) is found to be signiﬁcantly higher (p 0.001) when compared with the mean FA of the normal tissue (0.21p0.03). The mean ADC value of the tumor tissue (1.31p0.38) was found to be signiﬁcantly lower when compared with the mean ADC of normal tissue (1.50p0.20). Conclusion: The DTI parameters obtained from normal and tumor tissue differ signiﬁcantly. DTI technique can be used as an adjunctive method to conventional MRI for prostate cancer.

B-750

11:51

Prospective study on MR-guided galvanotherapy of prostate cancer: Initial results T.J. Vogl1, H.P. Mayer2, S. Zangos1, F.B. Mayer1; 1Frankfurt a. Main/DE, 2 Regensburg/DE ([email protected]) Purpose: To evaluate prospectively the safety and effectiveness of MR-guided galvanotherapy (MR-GT) for the treatment of prostate cancer. Methods and Materials: Sixty patients (mean: 59.3 years) with histologically proven prostate cancer were prospectively treated with MR-GT. After transgluteal puncturing of the prostate under local anesthesia, two MR-compatible electrodes were positioned under MR guidance in the periphery of the right and left lobes of the prostate with direct tumor contact. The patients were treated three times in 1-week intervals, and a total of 350 Coulomb was applied to the localized cancer in the prostate gland. Follow-up using laboratory data (PSA) and endorectal MRI was performed at three-, six- and twelve-month intervals after the procedure.

316

A

B

C

D

E

F

G

Results: All patients tolerated the MR-GT well without any major side effects or complications. Twenty-ﬁve patients had some reversible urinating problems after one treatment and 10 patients reported temporary parasthetic feeling in one leg. Tumor size decreased from a median of 2.5 cm³ to 1.02 cm³ which corresponded to a signiﬁcant (p 0.01) reduction of 59.2%. Five patients (8.3%) had a complete remission and 32 (53.3%) had a partial remission at follow-up MR evaluation 12 months post therapy. In 19 of the patients (31.6%) stable disease was achieved. Four patients (6.0%) presented progressive ﬁndings. The median PSA values decreased in the twelve-month control period from PSA of 7.8 ng/ml to 2.2 ng/ml (p 0.01). Conclusion: MR-GT is a safe procedure and may result in local control of prostatic carcinoma with a concomitant reduction in the PSA value.

10:30 - 12:00

Room F2

Breast

SS 1802 MR: Role in patient management Moderators: H. Dobson; Glasgow/UK R.A. Kubik-Huch; Baden/CH

B-751

10:30

Correlation between dynamic contrast enhanced MRI (DCE-MRI) morphological and dynamic features and histopathological prognostic factors in invasive breast cancer V. Girardi1, G. Carbognin1, A. Russo1, M. Tonegutti2, F. Bonetti1, R. Pozzi Mucelli1; 1 Verona/IT, 2Peschiera/IT ([email protected]) Purpose: To evaluate whether dynamic DCE-MRI could provide histopathological prognostic factors in invasive breast cancer. Methods and Material: 54 women with surgery proven invasive breast cancer underwent preoperative DCE-MRI. DCE-MRI ﬁndings, according to the scoring system described by Fischer in 1999, included tumor shape, margins, internal enhancement, signal intensity increase, signal intensity course, Fischer’s score. In addition, the Fischer’s score was divided into two subgroups: the morphological score 0-4 (resulting from the sum of point of morphological criteria) and the dynamics score 0-4 (resulting from the sum of point of kinetics criteria). The evaluated histopathological prognostic factors included histological type, tumor grading, DCIS associated component, lymphnodes metastasis and ER, PgR, Ki67 and HER2 expression. The Fisher’s exact test was used to correlate DCE-MR imaging parameters and histopathological ﬁndings. Results: Shape and margins of tumors have no signiﬁcant correlation with prognostic factors. Extremely signiﬁcant correlation (p 0.0001) was identiﬁed between internal enhancement with heterogenous, ring pattern and inﬁltrative ductal carcinoma type. Signiﬁcant correlation (p 0.02) was identiﬁed between rapid signal intensity increase and positive expression of HER2. Signiﬁcant correlation (p 0.04) was found between dynamic Fischer's score 3 and histopathological grade 3. Not quite signiﬁcant correlation was found between signal intensity course and Ki67 expression (p 0.07), Fischer's score and HER2 expression (p 0.08), dynamic Fischer's score and ER (p 0.06) and PgR expression (p 0.08). Conclusion: Based on our study, the DCE-MRI ﬁndings which correlate with histopathological prognostic factors are the internal tumor enhancement and the dynamic Fischer’s score.

B-752

10:39

The value of functional MRI methods in diagnosis of breast carcinoma and assessment of early response to neoadjuvant chemotherapy X.X.H. Wang, W.J. Peng, K. Shen, F. Sun; Shanghai/CN ([email protected]) Purpose: To assess the value of proton Magnetic Resonance Spectrum (1H-MRS), diffusion-weighted imaging (DWI) and MR ﬁrst pass Perfusion Imaging (PWI) in diagnosis of breast carcinoma and assessment of early response to neoadjuvant chemotherapy. Methods and Materials: The study included 19 patients with breast carcinoma (18 cases of invasive ductal carcinoma, one mucinous carcinoma) undergoing neoadjuvant chemotherapy. The results of 1H-MRS, DWI, PWI were measured before and after two cycles of treatment. Each measure was assessed to predict the curative effect. All cases were proved by biopsy. Results: (1) DWI (b=1000) can detect breast carcinoma (17/19), the signal intensity of lesions was high on DWI map, and the mean ADC value (1.12×10-3p0.03 mm2/

Scientific Sessions sec) was lower than normal tissue. After two cycles of treatment, the mean ADC value of 17 cases that were partially relieved was signiﬁcantly increased (mean: 1.29×10-3p0.05 mm2/sec), with decreased signal intensity and lesion size. (2) The technical successful rate for 1HMRS was 89.5% (17/19), sensitivity was 64.7% (11/17). The choline-containing compounds (Cho) can be detected. After two cycles of treatment, the peak amplitude of composite choline in 10/11 partially relieved cases decreased or vanished. (3) The sensitivity of PWI in detecting breast carcinoma was 90.9% (10/11), time-intensity curve showed sharp drop of MR signal in the perfusion phase compared with pre-contrast baseline. The mean intensity decreasing ratio (minimum intensity divided by baseline) decreased after therapy. Conclusion: 1HMRS, DWI, PWI can detect breast carcinoma, and provide a sensitive assessment of treatment efﬁcacy by measuring changes of physiological functions and biochemistry.

B-753

10:48

Magnetic resonance imaging (MRI) in predicting disease-free survival in patients treated with primary chemotherapy for locally advanced breast cancer M. Sentís Crivellé1, M. Seguí Palmer1, L.A. Fernandez Morales1, E. Dalmau1, C. Santos1, S. Ganau Macias1, S. Fernandez1, E. Pous1, J. Font1, A. Arcusa2, E. Saigi Grai1; 1Sabadell/ES, 2Terrassa/ES ([email protected]) Purpose: To evaluate the value of breast magnetic resonance imaging (MRI) in predicting disease-free survival (DFS) in patients with locally advanced breast cancer (LABC) treated with primary chemotherapy (PC). Methods and Materials: Patients with LABC treated with PC including anthracyclines and taxanes, who preoperatively underwent breast MRI examination, were analysed. The relation between the response to PC using breast MRI and disease-free survival was investigated. Results: From June 1998 to March 2005, a total of 110 patients were treated with PC. Ninety-eight patients (89.1%) were evaluated with breast MRI at diagnosis and before surgical treatment. There were 19 patients (17.3%) who obtained a pathological complete response (pCR) to PC. When we evaluated the response to PC using breast MRI, there were 30.6% of patients (30/98) who obtained complete response (CR). Eleven patients (36.7%) with CR by MRI (CR-MRI group) had a pCR, whereas four patients (5.9%) without CR by MRI (not CR-MRI group) had a pCR (p 0.01). With a median follow-up time of 46.9 months, two patients (6.7%) in the CR-MRI group and 19 patients (27.9%) in the not CR-MRI group recurred. The estimated DFS at three and ﬁve years was 92.6% for patients in the CR-MRI group, while in the not CR-MRI group the estimated DFS was 80.6% at three years and 69.1% at ﬁve years (p=0.0491). Conclusion: In this study, breast MRI is useful in predicting DFS in patients with LABC treated with primary chemotherapy, probably because the response evaluation using MRI could be a surrogate marker of pathological response.

B-754

10:57

MRI of the breast improves detection of invasive cancer, preinvasive cancer and premalignant lesions during surveillance of women with a high risk for breast cancer C.C. Riedl, L. Ponhold, C. Rainer, S. Jaromi, D. Flöry, M. Fuchsjäger, T. Wagner, T.H. Helbich; Vienna/AT ([email protected])

11:06

Changes from breast conserving surgery to mastectomy in breast cancer patients based on preoperative breast MRI: Reason, frequencies and results S.-H. Mun, E.-Y. Ko, J.-H. Shin, S.-J. Kim, B.-K. Han; Seoul/KR Purpose: To evaluate the causes and frequencies of changing surgical plan from breast conserving surgery (BCS) to mastectomy after preoperative breast MRI in breast cancer patients with candidates for BCS based on conventional images, and to correlate MR ﬁnding with the histological results. Methods and Materials: Between January 2004 and December 2006, 528 consecutive patients with breast cancer underwent preoperative breast MRI. Of them, 357 patients including three bilateral cancer patients were candidates for BCS based on conventional images before MRI. We evaluated the MR ﬁndings causing changes of surgical plan from BCS to mastectomy, and frequency of each MR ﬁnding. We also reviewed the pathology of mastectomy specimen, and correlated MR ﬁndings with the histologic results. Results: In 29/357 patients (7.3%) surgical management was altered from BCS to mastectomy according to following MR ﬁndings: substantially larger extent or size of lesions (n=4), more tumor extension toward the nipple or subareolar area making BCS difﬁcult (n=9), multifocal or multicentric lesions that were not demonstrated on mammography or US (n=12). In one patient, both more tumor extension toward subareolar area and substantially larger extent were noted on MRI. On pathologic reviews, false positive rates of each ﬁnding were 0, 55.6%, and 25% respectively. Conclusion: Breast MRI is sensitive for assessment of tumor extent and multiplicity preoperatively. However, false positive rate is relatively high in assessment of subareolar tumor extension, and various benign lesions can mimic multiple malignancies. Therefore, correlation with conventional images and preoperative MR-guided biopsy are recommended.

B-756

11:15

Contrast-enhanced magnetic resonance mammography: Does it affect surgical decision-making in patients with breast cancer? S. Padula, F. Pediconi, A. Roselli, V. Dominelli, S. Cagioli, C. Catalano, R. Passariello; Rome/IT ([email protected]) Purpose: Diagnostic imaging in women with suspected breast cancer should accurately detect and diagnose malignant tumors and facilitate the correct therapy choice. Contrast-enhanced MR mammography (CE-MRM) is potentially the imaging modality of choice for accurate patient management decisions. Methods and Materials: 164 women with suspected breast cancer based on clinical examination, conventional mammography and/or ultrasound each underwent preoperative bilateral CE-MRM using an axial 3D dynamic T1-weighted gradientecho sequence and 0.1 mmol/kg gadobenate dimeglumine (MultiHance®; Bracco) as contrast agent. Images were evaluated by two readers in consensus. Histological evaluation of detected lesions was performed on samples from core biopsy or surgery. Determinations were made of the sensitivity and accuracy of CE-MRM compared to mammography/ultrasound for the detection of malignant lesions and of the impact of CE-MRM on surgical decision-making. Results: Conventional mammography/ultrasound detected 175 lesions overall. CE-MRM revealed 51 additional lesions in 34/164 patients; multifocal and multicentric cancer was detected in seven and four additional patients, respectively, contralateral foci in 21 additional patients and pectoral muscle inﬁltration in two additional patients. CE-MRM also conﬁrmed the absence or benignity of three and one lesions suspected of malignancy on mammography/ultrasound. The sensitivity and accuracy for malignant lesion detection and identiﬁcation was 100 and 93.4%, respectively, for CE-MRM compared to 77.3 and 72.1% for mammography/ultrasound, respectively. Patient management was altered for 32/164 (19.5%) patients as a result of CE-MRM. Conclusion: CE-MRM positively impacts patient management decisions and should be performed in all women with suspected breast cancer based on clinical examination, mammography and/or ultrasound.

B-757

11:24

The accuracy of MRI in evaluating patients for residual breast cancer after lumpectomy O. Golan, F. Sperber, A. Shalmon, A. Gat; Tel Aviv/IL ([email protected]) Purpose: Residual tumor after excisional biopsy has been reported in 32-63% of cases with breast cancer. Positive margins are associated with increased long term risk of cancer recurrence in the ipsilateral breast. To assess the diagnostic accuracy of MRI in evaluating presence and extent of residual disease in patients with positive or close margins after lumpectomy.

A

B

C

D

E

F

G

317

Tuesday

Purpose: To compare the values of mammography, ultrasound, and breast magnetic resonance imaging (MRI) for high risk breast cancer screening. Methods and Materials: In a prospective comparison study, 327 women at high risk for breast cancer received 672 complete annual screening examinations, consisting of mammography, ultrasound, and MRI, at a single tertiary care breast center. Sensitivity and speciﬁcity of each modality were based on the histopathological evaluation of suspicious ﬁndings from all modalities plus the detected interval cancers. Results: Of the 28 detected cancers, 14 were detected by mammography, 12 by ultrasound, and 24 by MRI, which resulted in sensitivities of 50, 42.9, and 85.7%, respectively (p 0.01). MRI detected not only signiﬁcantly more invasive but also signiﬁcantly more preinvasive cancers. Five of 11 DCIS (54%) were detected only by MRI and one (9%) was missed by MRI. In addition, MRI was superior in the diagnosis of atypical ductal hyperplasias, lesions considered to be of premalignant character. Of a total of 35 atypical ductal hyperplasias, nine (26%) were detected by mammography, two (6%) with ultrasound, and 32 (91%) with MRI (p 0.01); 25 (64%) were detected only by MRI. Mammography, ultrasound, and MRI led to 25, 26, and 101 false-positive ﬁndings, which resulted in speciﬁcities of 98, 98, and 92%, respectively (p 0.05). Conclusion: MRI of the breast not only improves the detection of invasive cancer but also of preinvasive and premalignant lesions in a high-risk population, and should therefore become an integral part of breast cancer surveillance in these patients.

B-755

Scientific Sessions Methods and Materials: Forty-ﬁve patients who had undergone exicional biopsy with positive or close resection margins underwent MRI for evaluation of residual breast cancer and surgical planning. Dynamic and morphologic enhancement features were used for lesion characterization. Imaging ﬁndings were correlated with histopathology results of the relumpectomy. Sensitivity, speciﬁcity, positive predictive value and negative predictive value were calculated. Results: The sensitivity and speciﬁcity of MRI for detecting residual disease were 67 and 75%, respectively. Positive predictive value and negative predictive value were 94 and 27%. respectively. In 10% of patients MRI identiﬁed additional suspicious lesions distant from the biopsy site. Conclusion: MRI has recently become the imaging modality of choice for evaluating residual disease. Overlap in the appearance of benign and malignant lesions limits MRI ability to evaluate residual cancer. MRI depicts additional suspicious ﬁndings which may alter the surgical management. In the future we plan on adding spectroscopy to the MR examinations of similar cases, which may improve these results.

B-758

11:33

Role of breast MRI in patients with unilateral nipple discharge: Preliminary study L. Ballesio, S. Savelli, C. Maggi, I. D’Ambrosio, L.M. Porﬁri, C. De Felice, L. Manganaro; Rome/IT ([email protected]) Purpose: Role of breast Magnetic Resonance Imaging (MRI) in patients with unilateral nipple haematic or serohematic discharge. Methods and Materials: 42 patients with haematic or serohematic nipple discharge, negative mammographic ﬁndings (37/42 cases), underwent MRI to evaluate breast ducts. Ultrasonography (US) negative in 15 patients, identiﬁed 16 ductal ectasia (seven simple, nine with solid intraductal echoes), 11 solid masses; galactography negative in four and positive in seven. 16 women were followed-up clinically, with US, cytologic evaluation of nipple discharge (6-12 months), 23 underwent corebiopsy (followed by excisional biopsy), three excisional biopsy. Results: MRI correlated with histopathologic ﬁndings: ﬁve papillomatosis were identiﬁed as a patchy, homogeneous enhancing area, 16 intraductal papillomas as areas with well-deﬁned margins and II type intensity/time curve (BI-RADS3), two atypical ductal hyperplasia as diffuse nodular enhancement. Three patients found to have enhancing areas with ill-deﬁned margins (BI-RADS4) at MRI; two micro-papillary DCIS and a papillary carcinoma was identiﬁed on excisional biopsy. Negative follow-up: 15 patients, MRI were interpreted as normal (BIRADS1) and one with a late patchy enhancing (BIRADS2). Conclusion: Breast MRI can be considered an alternative to galactography for the study of ducts based on high sensibility in the diagnosis of intraductal lesions in patients with nipple discharge.

B-759

11:42

The role of real-time virtual sonography in detecting enhancing incidental lesion on contrast-enhanced MR imaging of the breast S. Nakano1, K. Yorozuya1, M. Takasugi1, Y. Mouri1, T. Fukutomi1, K. Ohno1, O. Arakawa2, T. Mitake2; 1Aichi/JP, 2Chiba/JP ([email protected]) Purpose: The aim of this study was to assess the ability of real-time virtual sonography (RVS) to identify enhancing incidental lesions (EIL) detected on contrastenhanced MR imaging (MRI) in patients undergoing evaluation for breast cancer. Methods and Materials: Between September 2006 and August 2007, MRI was performed on 51 women prior to undergoing breast conservation in our hospital. All patients were examined using mammography, sonography, and RVS, enabling both breast sonography and contrast-enhanced MRI cutaway images of the same site to be displayed in real time. Breast MR images were obtained on a 1.5-T imager in the supine position using a ﬂexible body surface coil. Results: The overall sensitivity for detecting primary breast tumors was 71% (36/51) for mammography, 88% (45/51) for sonography, 94% (48/51) for MRI. As three lesions were not visualized on MRI, RVS was performed in 48 enhancing lesions. A virtual MPR image of the target tumor was displayed under good condition correspondence with the sonography image in all patients. EIL were found in 23% (11/48), which were not expected from the previous conventional techniques. Of these, 55% (6/11) of EIL could be identiﬁed only on repeated sonography, but 91% (10/11) of EIL could be identiﬁed easily using the RVS system. Histologically, 20% (2/10) of EIL were invasive ductal carcinoma. Conclusion: The RVS system could project enhanced MRI information onto a body surface correctly while checking sonography from images without large-scale equipment or radiation exposure. The present results suggest that the RVS system offers excellent accuracy for identiﬁcation of EIL.

318

A

B

C

D

E

F

G

B-760

11:51

Differentiation of recurrent breast cancer from post treatment changes with diffusion weighted MR imaging A.A.A. Abdel Razek1, N. Abdel Razek2, A. Denewar1, N. Nada1; 1Mansoura/EG, 2 Cairo/EG ([email protected]) Purpose: To evaluate the role of diffusion weighted echoplanar MR imaging in differentiation of recurrent breast cancer from post treatment changes. Methods and Materials: Prospective study was conducted on 47 consecutive female patients (aged 47-64 ys: mean 53 ys) with breast cancer after surgery and complete course of neoadjuvant chemotherapy. All patients were examined at 1.5 Tesla using dedicated bilateral breast coil. They underwent dynamic contrast enhanced MR imaging and diffusion weighted MR imaging of the breast using a single shot echo planar imaging with a b-factor of 0.200 & 400 sec/mm2. The scanning parameters were: TR=10000 ms, TE=108 ms, NEX=2, bandwidth=125 kHz, FOV=20-25 cm, slice thickness=4 mm. Apparent diffusion coefﬁcient (ADC) maps were reconstructed. The ADC value of the breast lesion was calculated and correlated with biopsy results. Results: The mean ADC value of residual breast cancer was 1.22p0.13X10-3 mm2/ sec and of post treatment changes was 1.77p0.22X10-3 mm2/sec. There was a statistically signiﬁcant difference in mean ADC values between recurrent breast tumors and post treatment changes (p 0.003). When apparent diffusion coefﬁcient value of 1.6X10-3 mm2/sec was used as a threshold value for differentiating of residual breast cancer from post treatment changes, the best result was obtained with an accuracy of 88%, sensitivity 91%, speciﬁcity 92%, positive predictive value 92% and negative predictive value of 83%. Conclusion: Apparent diffusion coefﬁcient value is a new non-invasive imaging parameter that can be used for differentiation of recurrent breast cancer from post treatment changes.

10:30 - 12:00

Room G/H

Head and Neck

SS 1808 Skull base and temporal bone Moderators: R.B.J. de Bondt; Maastricht/NL S. Kösling; Halle a.d. Saale/DE

B-761

10:30

Preoperative assessment of cochlear implant patients using MSCT and MRI H. Ma1, P. Han1, J. Xiao2, R. Merges2, W.-J. Kong1, G.-S. Sheng1, H. Shi1; 1 Wuhan/CN, 2Shanghai/CN Purpose: To investigate the value of MSCT and MRI imaging in assessment of preoperative cochlear implant patients. Methods and Materials: 43 cochlear implant candidates were involved in this study. All candidates were examined using MSCT and MRI. The axial imaging was performed on a SOMATOM Sensation 16 CT scanner (Siemens, Germany) using the following parameters:120 kv, 100 mAs, 0.75 mm collimation, 1 mm reconstruction increment.The axial images of interested ears were reconstructed with 0.1 mm reconstruction increment and a FOV of 50 mm. 2D and 3D reconstruction were done on the workstation. All candidates were performed on a MAGNETOM VISION 1.5 T (Siemens,Germany) scanner with 3D T2-weighted CISS sequence. Results: Both MSCT and MRI showed 20 normal inner ears in 43 patients. 23 patients with 42 ears were abnormal, which involved 20 patients with 39 inner ear malformations and 3 patients with labyrinthitis ossiﬁcans in 3 ears. The malformations were: Michel deformity (one ear), common cavity deformity (3 ears), incomplete partition type I (3 ears), incomplete partition type II (Mondini deformity) (5 ears), vestibular and semicircular canal malformations (13 ears), pure dilated vestibular aqueduct (14 ears). Five ears with cochlear nerve deﬁciency and two ears with cochlear nerve hypoplasia were detected by MRI. 26 ears were detected with mastoid pneumatization type I by MSCT, 36 ears with type II, and 24 ears with type III. Conclusion: MSCT and MRI are important modalities to evaluate the cochlear implant candidate and can help in selecting the patient, forecasting the operation and formulating the appropriate operation scheme. MSCT plays a central role in diagnosing bony abnormalities. MRI should be used to study cochlear nerves.

Scientific Sessions B-762

10:39

Evaluation of four MSCT systems in postoperative imaging of a cochlear implant B.M. Verbist, R.M.S. Joemai, W.M. Teeuwisse, J.H.M. Frijns, J. Geleijns; Leiden/NL ([email protected]) Purpose: To evaluate visualization of a Clarion CII HiFocus I cochlear implant (CI) and accuracy of measurements of electrode positions for acquisitions with 64-slice CT scanners of four major CT systems (Toshiba Aquilion-64, Philips Brilliance-64, GE LightSpeed-64, Siemens Sensation-64). Methods and Materials: An implanted human cadaver temporal bone, a polymethylmethacrylate (PMMA) phantom containing a CI and a point spread function (PSF) phantom were scanned. In the human cadaver temporal bone the visibility of cochlear structures and electrode array were assessed using a visual analogue scale (VAS). Distinction of individual electrode contacts was evaluated using a quantitative score. Quantitative assessment of electrode positions was achieved with the PMMA phantom by measurement of the displacement of the center of an electrode relative to its known position. In addition, PSF was measured to evaluate spatial resolution performance of the CT scanners. Results: VAS score and qualitative score were lower for Brilliance-64 and LightSpeed-64 compared to Aquilion-64 and Sensation-64. Displacement of electrode contacts ranged 0.05-0.14 mm on Acquilion-64, 0.07-0.16 mm on Brilliance-64, 0.07-0.61 mm on Light Speed-64 and 0.03-0.13 mm on Sensation-64. PSF measurements show an in-plane and longitudinal resolution varying from 0.48 to 0.68 mm and from 0.70 to 0.98 mm, respectively, over the four scanners. Conclusion: All scanners meet the technical requirements to acquire adequate postoperative data in cochlear implant patients with a HiRes 90k with HiFocus 1J electrode. The assessment of visibility of the CI and the quantitative assessment of electrode contact positions, however, varies between scanners.

B-763

10:48

CT ﬁndings of ﬁrst branchial anomalies K. Lee, H.-J. Kim, S. Kim, P. Jeon, K. Kim, Y. Yim; Seoul/KR ([email protected]) Purpose: To evaluate CT ﬁndings of ﬁrst branchial anomalies (FBAs) and the difference between histologic type 1 and 2 FBAs. Methods and Materials: Forty seven patients (M:F=26:21, mean age: 19 years, range: 1-47 years) were pathologically proved as FBAs. Of the 16 who underwent CT, ﬁve were type 1 and six were type 2. The remaining ﬁve were unclassiﬁed. We analyzed CT images on location, size, shape, internal architecture, associated infection/inﬂammation, and relation to external auditory canal (EAC). Results: Sixteen FBAs showed unilocular (n=8) or multilocular (n=2) cyst, solid nodule (n=4), sinus (n=1) or ﬁstula (n=1), located in intraparotid (n=5), extraparotid (n=7), or both intra- and extraparotid (n=4) regions. Twelve were accompanied by infection or inﬂammation of adjacent skin or parotid gland. Twelve showed intimate relationship with the EAC including broad attachment (n=9) and sinus or ﬁstula (n=3). Five cases of type 1 FBAs showed unilocular cyst (n=4). solid nodule (n=1), located in intraparotid (n=2), extraparotid (n=3) region. Two had broad attachment to EAC. No case had sinus or ﬁstula. In contrast, six cases of type 2 FBAs showed unilocular (n=3), multilocular (n=1) cyst, solid nodule (n=1), and sinus tract (n=1), located in intraparotid (n=2), extraparotid (n=3), or both intra- and extraparotid (n=1) regions. All six lesions were broadly attached to EAC and one showed a sinus tract with cartilaginous EAC. Conclusion: CT can demonstrate detailed anatomy and morphology of FBAs in preoperative settings. Compared with type 1 FBAs, somewhat more complicated ﬁndings might be demonstrated in type 2 FBAs.

B-764

10:57

Congenital deformities of the ear V.K. Katsaros, N. Marangos, C. Drossos; Athens/GR ([email protected])

B-765

11:06

Imaging of intralabyrinthine schwannomas: A retrospective study of 52 cases with emphasis on lesion growth A. Tieleman1, J.W. Casselman1, T. Somers2, J. Delanote1, J. Ghekiere1, R. Kuhweide1, B. De Foer2, E.F. Offeciers2; 1Bruges/BE, 2Antwerp/BE ([email protected]) Purpose: Assess clinical signs, prevalence, MR characteristics, location, clinical management and especially growth potential/patterns of intralabyrinthine schwannomas (ILS) in the largest series reported until now. Methods and Materials: Clinical presentation, exact intralabyrinthine lesion localisation, MR characteristics, lesion growth and clinical management were retrospectively reviewed in all patients diagnosed with an ILS between February 1991 and August 2007. Results: Fifty-two patients were included. They most frequently presented with progressive hearing loss and non-pulsatile tinnitus. Initially, ILS most frequently occur in the cochlea (81%). They have a spontaneous hyperintense signal on unenhanced T1-weighted images, enhance strongly after gadolinium administration, are sharply circumscribed and have a hypointense signal on thin heavily T2-weighted 3D-images. The scala tympani is more frequently or more extensively involved than the scala vestibuli. Follow-up MR, available in 27 patients, showed growth in 59% of them. Growth was seen from the scala tympani into the scala vestibuli and from the perilymph space in the scala vestibuli to the perilymphatic space around the saccule and vice versa. Twelve ILS were resected. Conclusion: ILS are more frequent than previously thought and grow in more than 50% of the cases. They are most frequently found intracochlear, often anteriorly in the area between the basal and second turn. Cochlear ILS most often originate in the scala tympani and only later will grow into the scala vestibuli. Growth can occur from the cochlea into the vestibule or vice versa through the anatomical open connection between the perilymphatic spaces in the scala vestibuli and saccule.

B-766

11:15

MRI of the inner ear at 1.5 T and 3 T G.A. Krombach, T. Plum, M. Westhofen, C. Hohl, R.W. Günther; Aachen/DE ([email protected]) Purpose: To evaluate the potential for MR imaging of the inner ear at 3 T compared to 1.5 T. Methods and Materials: In 20 healthy subjects and 15 patients a T2-TSE was carried out at 3 T (TR/TE 2500/200 ms, matrix 368x512, FOV 160 mm², imaging time 5.42 min). At 1.5 T true-FISP (TR/TE 9.8/4.2 ms, 80°, matrix 512x512, FOV 230 mm², imaging time 4.19 min) and TSE images (TR/TE 2000/500 ms, matrix 128x128, FOV 90 mm², imaging time 5.46 min) were obtained. Two radiologists evaluated visualization of anatomical structures and pathologies using a 4-point non-parametric scale. Results: Small structures such as modiolus, osseous spiral laminal, facial, vestibulocochlear, cochlear, superior and inferior vestibular nerve were signiﬁcantly better delineated at 3 T images (p 0.001). There was a small difference for delineation of large structures with high signal intensity (vestibule, semicircular canals). Pathological ﬁndings were encountered in seven cases. Recognizability of pathologies was similar at 1.5 T and 3 T, while pathologies were delineated with higher image quality at 3 T. Conclusion: Imaging of the temporal bone at 3 T allows for delineation of anatomical details, with short imaging times. Compared to TSE or true-FISP images obtained at 1.5 T, more anatomical details can be assessed due to differences in signal to noise ratio. This may help in the assessment of pathologies.

B-767

11:24

The facial nerve in the petrous bone: Improvement of visualisation using 3 T MRI H.P. Burmeister, F. Hause, P. Schmidt, O. Guntinas-Lichius, J. Sedlacik, H.-J. Mentzel, W.A. Kaiser; Jena/DE ([email protected]) Purpose: The aim of this study was to assess the improvement of visualisation with focus on the intrapetrous facial nerve using 3 T-MRI-imaging and to compare it to 1.5 T-MRI-imaging regarding the signal intensity-to-noise-ratio (SNR) and the identiﬁability of the branches leaving the nerves’ trunk.

A

B

C

D

E

F

G

319

Tuesday

Purpose: To present our experience in congenital deformities of the ear and to propose a new classiﬁcation scheme for them. Methods and Materials: In the last 25 years, cochlear implant candidates (1200 children) were examined by CT, MRI or both methods. CT included high resolution technique with slice thickness 0.5-1 mm. MRI of the IAC included high resolution T1 weighted, before and after intravenous administration of Gadolinium chelates and T2 weighted, as well as CISS 3D images. Results: As high resolution computed tomography of the temporal bones and occasionally magnetic resonance imaging are used in the evaluation of cochlear implant candidates with hereditary hearing loss, several ear malformations have been observed that cannot be classiﬁed as Mondini dysplasias. Thus, a new classiﬁcation of ear deformities is necessary, based on arrested or altered inner ear

embryogenesis. We found 108 cases of various temporal bone malformations not necessarily included in previous classiﬁcations but with impact in cochlear implant surgery or outcome. Conclusion: Cochlear implant is contraindicated on some of these deformities; on others, electrode insertion might be a challenge and in some cases, poorer auditory performance with the cochlear implant may occur.

Scientific Sessions Methods and Materials: Twenty healthy subjects underwent T1- and T2-weighted MRI imaging of the temporal bone. Scanning parameters were adapted to the nervous and soft tissue that the fallopian canal contains. Imaging was performed at 3 T by using an 8-channel head coil and an 8-channel multifunctional coil and at 1.5 T using head coil. SNRs for the geniculate ganglion and the pons were measured. The ﬁndings of all three groups were compared to each other. The identiﬁability of the nerves’ course and the branches was assessed. Results: Statistical testing showed that the SNRs achieved by the head coil at 3 T were superior to those using the multifunctional coil at 3 T and the head coil at 1.5 T (p 0.001). SNR values gained by the multifunctional coil at 3 T were only superior to those of the geniculate ganglion at 1.5 T. In all subjects the course of the intrapetrous facial nerve was discernable by T1- and T2-weighted MRI imaging, but only at 3 T the smallest human nerves, like the tympanic chorda and the stapedial nerve, are discernable. Conclusion: The study shows that in spite of gaining susceptibility artifacts, 3 TMRI enables a much more valid depiction of the intrapetrous facial nerve and even some of its smallest branches.

B-768

11:33

Superior semicircular canal dehiscence syndrome: CT-MPR of the superior semicircular canal improves diagnostic accuracy R.T. Geracimos, C.M. Silverberg, A. Iaia, B. Kung, G. Sparcia, K. Karcher, M. Teixido; Newark, DE/US ([email protected]) Purpose: To determine whether multidetector CT (MDCT) with multiplanar reformations (MPR) perpendicular and parallel to superior semicircular canal (SSC) improves diagnostic accuracy for detection of SSC dehiscence in patients with superior semicircular canal dehiscence syndrome. Methods and Materials: Inclusion criteria for SCD syndrome consisted of a constellation of clinical signs, symptoms, imaging, and laboratory testing including VEMP and audiometry. Sixteen patients with SCD syndrome were evaluated with MDCT of the temporal bones, along with 16 normal controls. MDCT temporal bone examinations containing 0.8-1.0 mm coronal and axial images were performed. Additionally, a second image data set was generated for each subject which included the original axial and coronal images through the temporal bones as well as 0.4-0.5 mm parallel and perpendicular reformatted images through the SSC. CT scanners employed ranged from 4 to 64 slices. Findings were evaluated by single blinded CAQ neuroradiologist. Results: SCD was diagnosed in 12/16 (SCD+ data set) patients using coronal and axial images only. When the same data set was reviewed with the addition of MPR images, 14/16 were diagnosed with SCD (SCD+ data set). Control group examinations (SCD- data set) were diagnosed as negative 32/32. Sensitivity was 75% (axial & coronal) and 87% (axial, coronal, and MPR). The speciﬁcity was 100%. The additon of MPR imaging through the SSC improved diagnostic accuracy from 87% (Ax/Cor images) to 94% (Ax/Cor + MPR). Conclusion: MDCT temporal bone imaging with MPR through SSC improves detection of SCD in patients suspected of having superior semicircular canal dehiscence syndrome.

B-769

11:42

Traumatic ossicular chain separation: Sliding-thin-slab maximum intensity projections for diagnosis B. Zhao, J. He, S. Liu, T. Ma; Jinan/CN ([email protected]) Purpose: To retrospectively determine, by using MSCT, whether additional interpretation of sliding-thin-slab (S-T-S) maximum intensity projections (MIP) reformations images improves diagnostic accuracy when compared with the diagnostic interpretation of conventional S-T-S multiplanar reformation (MPR) CT images for traumatic ossicular chain separation. Methods and Materials: Twenty-nine patients with traumatic ossicular chain separation who underwent temporal bone MSCT were retrospectively identiﬁed from electronic medical records. An additional 29 control subjects who underwent temporal bone MSCT for other reasons, were retrospectively selected from the same period. Two neuroradiologists independently reviewed the 116 temporal bones twice. One review was restricted to MPR (“MPR only”) images. The other review used MIP images and MPR (“all reformations”) images. The observers were blinded to clinical history, and the two reviews took place four weeks apart to avoid recall bias. The chi-square test was performed for diagnostic accuracy between MPR images and all reformation images. Cohen’s kappa statistics were used to evaluate interobserver variability. Results: With MPR and “all reformations” images, Observer 1 diagnosed traumatic ossicular chain separation in 19 (16.4%) and 28 (24.1%) of 116 temporal bones, respectively (p 0.05); Observer 2 diagnosed traumatic ossicular chain separation in 20 (17.2.0%) and 27 (23.3%) of 116 temporal bones, respectively (p 0.05). The K values revealed high inter-observer agreement, independent of the imaging

320

A

B

C

D

E

F

G

modality ("MPR only" or "all reformations" images). Conculsion: Interpreting MIP images additionally provided more accurate diagnoses than interpreting MPR images alone for diagnosis of traumatic ossicular chain separation.

B-770

11:51

Detailed imaging of the jugular foramen and its content at 3 Tesla MRI J. Linn, F. Schwarz, H. Brückmann, I. Yousry; Munich/DE ([email protected]) Purpose: To determine the value of 3 Tesla MRI in the visualization of the detailed anatomy of the jugular foramen (JF) using a contrast-enhanced 3D-FIESTA (fast imaging employing steady state acquisition) sequence. Methods and Materials: 25 patients without pathologies affecting the JF were included. Imaging was performed on a 3 Tesla scanner (Signa, General Electric) using a gadolinium-enhanced 3D-FIESTA sequence. The data were analysed collaboratively by two readers. It was recorded if the glossopharyngeal (CNIX), the vagus nerves (CNX), and their ganglia could be identiﬁed in the JF. Furthermore, landmarks for their identiﬁcation were described. Measurements were performed for each side separately. Results: CNIX could be identiﬁed in 91.7 and 88%, CNX in 87.5 and 100% (on the right and left sides, respectively). The superior ganglion of CNIX was dicernible in 91.6 and 88%, the ganglion of CNX in 95.8 and 100% (on the right and left sides, respectively). The inferior petrosal sinus, the hypoglossal canal, and the external opening of the cochlear aqueduct were found to be useful landmarks for these structures and could all be identiﬁed in 100% on both sides. Conclusion: The contrast-enhanced 3D-FIESTA sequence performed on a 3 T scanner is an adequate tool for the vizualisation of the complex anatomy of the JF. Thus, high ﬁeld MR imaging might be helpful in the formulation of differential diagnosis and in presurgical planning regarding pathologies involving the JF.

10:30 - 12:00

Room I

Vascular

SS 1815 Therapy, evaluation and outcome Moderators: M. Heller; Kiel/DE P. Vilela; Almada/PT

B-771

10:30

Endovascular revascularization for limb salvage in diabetic patients L. Di Vito, S. Fabiano, R. Gandini, M. Stefanini, M. Messina, C. Reale, E. Pampana, G. Simonetti; Rome/IT ([email protected]) Purpose: To evaluate short- and long-term results in diabetic patients who underwent endovascular revascularization for limb salvage. Methods and Materials: From June 2001 to June 2007, 635 patients (average age 69.5 years), all diabetics, were treated with different endovascular techniques to obtain a direct ﬂow to the foot and to avoid major amputation. All the lesions were staged with Texas classiﬁcation: 65 patients (10.25%) stage IIB, 80 (12.6%) IIC, 109 (17.16%) IID, 16 (2.52%) IIIA, 115 (18.11%) IIIB, 129 (20.31%) IIIC and 121 (19.05%) IIID. The only contraindication is the lack of visualization of the vessel at the plantar arch and/or the pedal artery. All patients were previously assessed with clinical evaluation, TcPO2/TcPCO2 measurement and angiographic study (magnetic resonance or CT). The patients had several comorbidities (coronary arteries disease, hypertension, nephropathy) and were unsuitable for surgery. Results: Post-intervention evaluation included measurement of TcPO2/TcPCO2, clinical evaluation and color-duplex ultrasound. Technical success rate was 95.6% with only 4.4% failure. 1.9% of patients despite technical success underwent amputation for presence of severe osteomyelitis and alterated microcirculation. 10.7% of patients underwent re-intervention, because of unhealed wounds or decrease of TcPO2, and the success rate was 8.34%. The mean follow-up time was 24 months. The percentages of limb salvage and major amputation were 91.34 and 8.66, respectively. Conclusion: Endovascular revascularization shows a high technical success rate with an elevated success rate of re-intervention and should be considered as the primary preferred therapeutic option in lower limb salvage for diabetic patients.

Scientific Sessions B-772

10:39

Clinical validity of ECG-gated CT angiography for prediction of patient outcome after coronary artery bypass graft surgery P. Suranyi, B.J. Williams, S.A. Bagg, A.R. Smith, P. Huang, J.B. Selby, P. Costello, U.J. Schoepf; Charleston, SC/US ([email protected]) Purpose: To assess the clinical validity of ﬁndings at contrast enhanced ECG-gated CT angiography (CTA) performed early after coronary artery bypass graft (CABG) surgery to predict outcome in asymptomatic patients. Methods and Materials: 118 asymptomatic patients underwent ECG-gated CTA three months after CABG surgery. CTA was acquired using 16-slice CT (n=55) or 64-slice CT (n=63). CTA studies were reviewed for signs of graft failure (occlusion, stenosis, plaque formation, surgical complications). Systematic follow-up was performed 12 months after the CT scan using medical records and phone-calls to the referring physician and the patients to record repeat catheterization and subsequent cardiac events (recurrent angina, repeat revascularization, cardiac death). Data were analyzed using chi-square and non-parametric Wilcoxon testing. Results: The 118 patients received 251 grafts. Thirty-two patients showed signs of graft failure on CTA three months after surgery. Seven of those (ﬁve occlusions, two critical stenoses) were immediately intervened on based on CTA ﬁndings and excluded from further analysis. Eighty-six patients showed unremarkable grafts. During 12 months follow-up, 57 repeat catheterizations were performed in the group of patients with unremarkable grafts versus 67 in patients with CTA signs of graft failure (p=0.0042). Fifty percent of patients with signs of graft failure on CTA had a cardiac event during follow-up (81% speciﬁcity) compared with only 19% of patients with unremarkable grafts (78% NPV; p=0.0032). Conclusion: Even in the absence of symptoms, individuals with signs of graft failure on routine follow-up CTA early after CABG surgery may beneﬁt from closer monitoring and more aggressive treatment.

B-773

10:48

Juxtarenal and thoracoabdominal aneurysms: Successful treatment with fenestrated and branched stent grafts C.D. Bicknell, N.J.W. Cheshire, P. Bourke, C.V. Riga, J.H.N. Wolfe, R.G.J. Gibbs, M.P. Jenkins, M. Hamady; London/UK ([email protected]) Purpose: The advent of fenestrated custom-made stents has allowed the endovascular treatment of short-necked infrarenal abdominal aortic aneurysms (IRAAA) and suprarenal aneurysms (TAAA). We describe our early experience to treat aneurysmal disease at these levels. Methods and Materials: Twenty-two patients underwent fenestrated/branched stent grafting (2005-2007). Short neck ( 1 cm) IRAAAs (n=17), type 1 (n=2), type 3 (n=2) and type 4 (n=1) TAAAs were stented using covered stents or branches from the main body to patent visceral vessels. CT surveillance was performed at 1, 3 and 12 months. Results: Technical success for cannulation and stenting of the target vessels was 98%. In-hospital mortality was 4.5% (one patient died from bleeding from an infected iliac conduit). Six patients had major complications following successful grafting, including paraplegia in one (type 3 TAAA), which resolved, and one patient who underwent open repair after failure to cannulate the renal fenestration. Median length of stay was nine (8.5-11.5) days. At a median follow-up of 8.5 months (IQR 3.25-14.5), CT conﬁrmed all target vessels remain patent. Three type 2 endoleaks, treated conservatively, and two type 3 endoleaks, successfully treated percutaneously, were detected. One patient died after rupture 16 months after endograft placement following a type 3 endoleak while waiting for intervention. Conclusion: In selected patients, fenestrated/branched stent grafts are safe and effective alternative to surgery for short-necked IRAAAs and TAAAs. The long-term durability needs to be proven.

B-774

10:57

Purpose: To compare the accuracy of color duplex ultrasound (CDUS) and computed tomography (CT) in assessing aortic aneurysm diameters and detecting endoleaks (EL) after endovascular aortic aneurysm repair (EVAR). Methods and Materials: From November 1998 to January 2007, 196 patients (M/F: 191/5, mean age 72 years) underwent EVAR. Patients were followed by CT and CDUS. Annual paired CT and CDUS examinations were reviewed to assess agreement in measuring maximum aneurysm axial diameter (Dmax) and in identifying signiﬁcant (p10%) modiﬁcations in Dmax; moreover, using CT as the gold standard, diagnostic accuracy of CDUS in detecting endoleaks was assessed.

B-775

11:06

The role of contrast-enhanced ultrasonography (CE-US) in the detection of endoleaks in patients who underwent endovascular abdominal aortic aneurysm repair (EVAR) R. Iezzi, R. Basilico, V. Pasqua, M. Santoro, A.R. Cotroneo, M.L. Storto; Chieti/IT ([email protected]) Purpose: To assess the clinical value and potential impact of real time contrastenhanced ultrasonography in the detection of endoleaks in comparison with unenhanced ultrasound and CT-angiography. Methods and Materials: Eighty-four patients with unruptured abdominal aortic aneurysm treated with endovascular repair (EVAR) who underwent CT-angiography follow-up were enrolled in the study. In the same day, CT-angiography (CTA) (4x1 mm collimation, 1.25 mm slice width), unenhanced ultrasound (US) and contrastenhanced ultrasonography (CE-US) were performed in all patients. CE-US was performed after intravenous bolus injection of 1.2/2.4 mL of a second-generation contrast agent (Sonovue, Bracco, Italy) with continuous low-mechanical index (0.01-0.04) real-time tissue harmonic imaging. US and CE-US exams were assessed in two different sessions by two independent experienced readers for the presence of endoleaks, according to a ﬁve-point conﬁdence scale. The standard of reference was represented by the consensus reading of CTA, US and CE-US performed by two experienced radiologists not involved in the image analysis. Sensitivity, speciﬁcity and diagnostic accuracy of each reading session were compared. Results: Sensitivity, speciﬁcity and diagnostic accuracy of CE-US were signiﬁcantly higher (p 0.05) than those obtained with US (97.5, 83, and 89.3% versus 62.5, 63.6, and 63.1%). No adverse events were recorded during the study. Conclusion: CE-US is a non-invasive, reliable, and fast tool for diagnosing endoleaks in patients who underwent EVAR, providing results higher than US, potentially avoiding the need for CT-angiography.

B-776

11:15

New frontiers of vascular imaging - dynamic MRI derived models for morphologic and functional evaluation of thoracic aorta: Clinical application in EVAR patients M. Midulla1, R. Moreno1, L. Veunac1, F. Nicoud2, B. Marcheix1, V. Chabbert1, F. Joffre1, H. Rousseau1; 1Toulouse/FR, 2Montpellier/FR ([email protected]) Purpose: So far there is no imaging method which is routinely performed for a morphologic and functional study of the thoracic aorta. With this study we want to report the ﬁrst clinical application of a new imaging method coupling dynamic MRI and Computational Fluid Dynamics (CFD) in pre and postoperative evaluation of patients candidated and subjected to endovascular aortic repair (EVAR) of thoracic aorta. Methods and Materials: From November 2006 to October 2007, 14 patients (13 male, one female) subjected to EVAR for different thoracic aortic diseases (four TAA, one PU, one IMH, seven dissections and one ATAR) have been evaluated by a dedicated MR protocol (morphological BTFE T1 sequences followed by Q-ﬂow at different levels). The images obtained were re-elaborated by a method focused to build patient-speciﬁc geometric data and boundary conditions for unsteady CFD runs with variable meshes valid over the cardiac cycle. Results: Fourteen of 16 examinations (87.5%) were suitable for the subsequent elaboration by the CFD method. Anatomical dynamic models were obtained after the processing phase. They provided detailed information on wall displacement (moving meshes), ﬂow turbulence, ﬂow velocity, and parietal stresses (wall shear stress and relative parietal tension). Conclusion: This method is a promising, feasible and reliable technique of new imaging of the thoracic aorta. It will be an exciting challenge for next experiment-

A

B

C

D

E

F

G

321

Tuesday

Color-duplex ultrasonography surveillance after endovascular repair of abdominal aortic aneurysm I. Bargellini, P. Torri, F. Guidocci, P. Petruzzi, R. Cioni, A. Cicorelli, C. Vignali, C. Bartolozzi; Pisa/IT ([email protected])

Results: Mean follow-up was 3.5p2.4 years (range 0-8.5). The 5-year cumulative survival was 75.7%, with only one (0.5%) case of AAA-related death. The 5-year cumulative incidence of EL was 44.3% (70 patients), with a higher incidence in the initial six months follow-up. 634 paired CT and CDUS annual examinations were collected in 171 patients. A high agreement (k=0.96) was observed between examinations in measuring Dmax, with a mean difference between CDUS and CT of -2.5 mm. A good agreement (k=0.67) was also observed in identifying signiﬁcant Dmax changes compared to the preprocedural value. Versus CT, sensitivity, speciﬁcity, positive, and negative predictive values of CDUS for detecting endoleaks were 44.5, 97.6, 77.6 and 90.3%, respectively. Conclusion: EVAR follow-up costs can be reduced by annually monitoring aneurysm Dmax modiﬁcations with CDUS; keeping in mind the underestimation of Dmax at CDUS, CT should be reserved for patients with proven increased or stable Dmax.

Scientific Sessions ers to better comprehend the practical implications that the multiple results of this imaging technique can suggest and, on the other side, to identify new clinical application ﬁelds.

B-777

11:24

Multimodality evaluation of consecutive patients with abdominal aneurysm treated with endovascular graft: Color-Doppler-US vs low-MI CEUS as compared with 64-slice angio-CT and angio-MRI V. Cantisani, P. Ricci, A. Napoli, C. Marigliano, E. Marotta, C. Catalano, F. Fanelli, U. D`Ambrosio, R. Passariello; Rome/IT ([email protected]) Purpose: To evaluate the efﬁcacy of color-Doppler-US and low-MI CEUS in the assessment of endovascular graft treatment for abdominal aneurysm as compared with 64-slice-angio-CT and angio-MRI. Methods and Materials: From February 2006 to March 2007, 60 consecutive patients (35 M, 15 F - mean age: 63) treated with endovascular aortic graft underwent color-Doppler-US, low-MI CEUS, 64-slice-angio-CT, angio-MRI and angiography if re-treatment was indicated. Sensitivity and speciﬁcity of ultrasound examinations were compared with CT and MRI as the reference standards or when available with the angiography. McNemar test was then calculated. Results: Six true endoleaks (type II: 5, Type III: 1) were identiﬁed (10%) by the FU or by angiography. Sensitivity and speciﬁcity of color-Doppler-US, CEUS, angio-CT, and angio-MRI were: 50, 100, 83, 100% and 60, 80, 80, 80%, respectively. CEUS was signiﬁcantly more accurate than US (p 0.001) and highly comparable to angio-CT and angio-MRI. Consequences to treatment occurred in three patients (50%). Conclusion: CEUS seems to be a feasible tool in the long-term surveillance of endovascular aortic aneurysm repairs as it may identify endoleaks missed at other imaging techniques. However, further investigation is mandatory.

B-778

11:33

Device-speciﬁc outcome in terms of aneurysm shrinkage and growth after endovascular abdominal aortic aneurysm repair (EVAR): A comparison between suprarenal and infrarenal ﬁxation stent-grafts R. Iezzi, D. Giancristofaro, V. Pasqua, D. Gabrielli, A.R. Cotroneo, M.L. Storto; Chieti/IT ([email protected]) Purpose: To determine whether aneurysm shrinkage and growth after endovascular aneurysms repair (EVAR) are inﬂuenced by stent-graft implants. Methods and Materials: 198 patients who had undergone EVAR with a CT follow-up of at least two years were included. AneuRX (n=23), Excluder (n=63), Talent (n=88), Zenith (n=24) devices were implanted. Patients were grouped according to the level of ﬁxation: suprarenal (Group A) and infrarenal (Group B). CT angiograms at 12 and 24 months were reviewed to evaluate aneurysm shrinkage or growth, using the pre-procedural CT images as reference scan. Aneurysm shrinkage or growth was deﬁned as decrease or increase of 5 mm or more in the minor dimension of the sac. The two groups were also compared on the basis of the pre-procedural diameter (subgroups: b6 cm and 6 cm). Mann-Whitney U test and paired t-test were used. Results: At 12 and 24 months, median diameter decrease was 1.67p0.9 and 3.92p1.8 mm (Group A) and 2.16p1.1 and 4.49p1.6 mm (Group B), with a percentage change in sac diameter of 3.09 and 7.12% (Group A) and 4.13 and 9.8% (Group B), respectively (p 0.05). No signiﬁcant differences were found in terms of sac shrinkage (36 vs 48.5%) and growth (6 vs 4.4%) at 24 months. At 24 months, a signiﬁcant median diameter decrease and percentage change were only obtained in pre-treatment aneurysms of less than 6 cm in diameter (4.54 vs 2.69 mm and 9.45 vs 3.61%, p 0.05), without any device differences. Conclusion: Our study seems to demonstrate that aneurysm shrinkage/growth rates are not inﬂuenced by stent-graft implants, being only strictly related to pretreatment sac diameter.

B-779

11:42

Detection of vascular graft infection with dual-modality FDG-PET/CT: Initial results J. Votrubová, O. Belohlavek, M. Spacek, P. Sebesta; Prague/CZ ([email protected]) Purpose: Prosthetic graft infection is one of the most serious complications in vascular surgery. Morphological changes often fail in correct diagnosis of low-grade graft infection. The aim of this work was to deﬁne the role of FDG PET/CT in this situation. Methods and Materials: Forty-ﬁve patients (53 grafts) underwent PET/CT due to clinical suspicion of infection. PET parameters like visually assessed FDG uptake and ratio of maximal uptake in graft to reference region in aorta; CT parameters like ﬂuid collection, irritation at boundary and presence of pseudoaneurysm and

322

A

B

C

D

E

F

G

laboratory parameters like leucocytes and CRP were evaluated against surgical assessment at reoperation or follow-up. The duration of follow-up was, on average, 15 months (3-26). Diagnostic value was assessed by stepwise logistic regression and ROC analysis. Accuracy was expressed as sensitivity, speciﬁcity, negative and positive predictive values including 95% conﬁdence interval (CI). Results: Infection was present in 40 grafts and not present in 13 grafts. Visually assessed focal uptake represents the best discriminator (p 0.0001); the other parameters add no signiﬁcant information (p 0.1) according to the stepwise logistic regression. In six grafts there was no focal uptake; all were true negative. In 34 grafts there was intense focal uptake; all but one were true positive. It represents sensitivity 100% (CI: 87-100%), speciﬁcity 86% (CI: 42-99%), PPV 97% (CI: 93-100%), NPV 100% (CI: 52-100%). Conclusion: Visually assessed focal FDG uptake is the best parameter of PET/CT to diagnose infection of vascular graft. CT part of PET/CT is mandatory for exact localisation of FDG uptake.

B-780

11:51

Uterine artery pseudoaneurysms as an uncommon cause of delayed postpartum hemorrhage and their endovascular management V. Gupta, N. Khandelwal, M. Singhal, J.R. Bapuraj, A. Lal, L.K. Dhaliwal; Chandigarh/IN ([email protected]) Purpose: To highlight the importance of uterine artery pseudoaneurysms as etiological factor in delayed post partum hemorrhage (PPH) and their endovascular management. Methods and Materials: Eleven women with intractable delayed postpartum hemorrhage were referred for uterine artery embolization. Angiographic evaluation by selective angiography of uterine arteries was done in all patients. Selective embolization of uterine arteries was done. Results: Pseudoaneurysms as a cause of delayed postpartum hemorrhage was evident on angiography in ﬁve patients with active extravasation of contrast in two cases. After embolization of uterine arteries immediate cessation of bleeding was observed in all cases. No complications related to embolization were found. Conclusion: Uterine artery pseudoaneurysms are infrequently reported but an important cause for delayed postpartum hemorrhage. Selective emergency uterine artery embolization is an effective means of controlling hemorrhage.

10:30 - 12:00

Room K

Pediatric

SS 1812 Advances in CT and MRI Moderators: S. Chapman; Birmingham/UK R.R. van Rijn; Amsterdam/NL

B-781

10:30

The corpus callosum as a paradigmatic region for motor development: A study with diffusion tensor MR imaging (DTI) and transcranial magnetic stimulation (TMS) B. Ertl-Wagner, I. Koerte, R. Stahl, M. Bonfert, W. Flatz, C. Rummeny, M. Reiser, F. Heinen; Munich/DE ([email protected]) Purpose: In order to identify a paradigmatic region for motor development we combined MRI with diffusion tensor imaging (DTI) and Transcranial Magnetic Stimulation (TMS). The Corpus Callosum represents a key structure for bimanual motor development and is amenable to investigation by TMS and DTI. Methods and Materials: We examined six healthy pre-school aged children, ﬁve healthy adolescents and a control group of six adults with both TMS (motor threshold, ipsilateral silent period (ISP)) and DTI. Values for the apparent diffusion coefﬁcient (ADC), fractional anisotropy (FA) and relative anisotropy (RA) were calculated for the genu, corpus, isthmus and splenium. Results: The three study groups demonstrated an age-dependence of activated and resting motor threshold in TMS (p 0.001). Adolescents, adults and three pre-school children presented ISP with signiﬁcant, age-related differences concerning duration (p 0.12) and nAreaISP (p 0.006). No differences could be detected regarding ISP latency. RA and FA showed no difference for the genu, corpus and isthmus between adolescents and adults (p 0.05), but were signiﬁcantly lower for the splenium (p 0.001). RA and FA were, moreover, signiﬁcantly lower in the isthmus in the pre-school children compared to both adolescents and adults (p 0.001).

Scientific Sessions Conclusion: The results reﬂect the development of inhibitory cortical networks and corresponding axonal pathways via the corpus callosum. The corpus callosum may therefore serve as a paradigmatic region in the evaluation of motor development in children and adolescents and DTI promises to provide a deeper insight into developmental principles where connectivity precedes cortical networking.

B-782

10:39

Symmetrical CTT (central tegmental tract) hyperintensity lesion on MR imaging in children S. Yoshida1, K. Hayakawa1, A. Yamamoto1, N. Aida2, S. Okano1, H. Matsushita1, T. Kanda1, Y. Yamori1, N. Yoshida1, A. Hirota1; 1Kyoto/JP, 2Yokohama/JP ([email protected]) Purpose: Central tegmental tract (CTT) is mainly the extrapyramidal tract connecting the red nucleus and inferior olivary nucleus. There are only few case reports on CTT abnormalities on magnetic resonance imaging (MRI) in congenital metabolic disorders. Our purpose was to evaluate the frequency of CTT lesion and characteristics on MRI, and to correlate MR imaging with the clinical aspect. Methods and Materials: We reviewed MR imagings of 392 children (215 boys and 177 girls) retrospectively of age range 1-6 in two institutions. To evaluate symmetrical CTT hyperintensity lesion, we deﬁned the CTT lesion as bilateral symmetrical hyperintensity in tegmentum pontis on both T2-weighted images and diffusion-weighted images in more than two slices. We measured the apparent diffusion coefﬁcient (ADC) values of symmetrical CTT hyperintensity, and compared with normal CTT. Results: Among 392 children, CTT lesion was detected in 20 cases (5.1%). The mean ADC value of 20 children with CTT lesion was signiﬁcantly lower than that of normal CTT (p 0.001). On MR imaging, other than CTT lesion associated parenchymal lesion, none (n=6), periventricular leukomalacia (n=3), thin corpus callosum (n=3), ventricular dilatation (n=2), encephalopathy (n=2), and so forth were detected. Clinically, the frequency of cerebral palsy in children with CTT lesions was 30%, signiﬁcantly higher than that without CTT lesion (13%) (p 0.05). Conclusion: The CTT lesion was detected in 5.1% of all cases. The mean ADC value of CTT lesion was signiﬁcantly lower than that of normal CTT. Cerebral palsy was the most frequent clinical diagnosis.

B-783

10:48

Diffusion abnormalities of the optic nerve head in pediatric brain tumor patients A. Vossough, T. Feygin, L.T. Bilaniuk, R.A. Zimmerman; Philadelphia, PA/US ([email protected]) Purpose: To determine the prevalence and features of diffusion abnormalities of the optic nerve head in pediatric patients presenting with brain tumors. Methods and Materials: MRI examinations of 180 pediatric brain tumor patients at the time of initial imaging diagnosis were reviewed for the presence of diffusion abnormalities of the optic nerve head (ischemic papillitis). Other factors evaluated included tumor histology, tumor location, papilledema, and presence of hydrocephalus or herniation. A group of 50 pediatric patients presenting with indications and signs of increased intracranial pressure, but without brain tumors, were also evaluated as controls. All patients ranged in age from 6 months to 15 years. Results: Decreased diffusion in the optic nerve head was seen in 20% of pediatric brain tumor patients at the time of initial MR imaging diagnosis. Higher proportions were seen in patients with posterior fossa and sella/suprasellar tumors. There was a correlation with degree of hydrocephalus, but not with midline shift. There was a signiﬁcant but incomplete correlation with demonstration of optic nerve head swelling (papilledema) on MRI (r=0.71). Ischemic papillitis was present signiﬁcantly more commonly in tumor patients than in the control group presenting with other signs and indications of increased intracranial pressure but without tumor (p 0.05). Conclusion: Ischemic papillitis occurs in a signiﬁcant proportion of pediatric patients presenting with brain tumors. The exact pathogenesis and clinical implications of this ﬁnding should be further evaluated.

10:57

Cerebral perfusion CT in the evaluation of children with severe traumatic brain injury S. Binaghi, L. Racine-Parret, S. De Ribaupierre, F. Gudinchet; Lausanne/CH ([email protected]) Purpose: To assess the value of cerebral perfusion CT (PCT) in children with traumatic brain injury in predicting their consecutive clinical outcome. Methods and Materials: Twelve paediatric patients with acute traumatic brain injury underwent cerebral CT coupled with PCT during their admission at the emergency room (ER). PCT maps were reviewed for mean transit time (MTT), regional cerebral

B-785

11:06

Estimated risk of radiation-induced thyroid cancer from pediatric head MDCT examinations: Sequential versus spiral scanning M. Mazonakis, A. Tzedakis, J. Damilakis, N. Gourtsoyiannis; Iraklion/GR ([email protected]) Purpose: The thyroid gland of children is one of the most radiosensitive organs. This study aims to estimate thyroid dose and associated cancer risk from head MDCT examinations performed in sequential and spiral acquisition modes during childhood. Methods and Materials: The Monte Carlo N-particle transport code was employed to simulate sequential and spiral examinations of the brain, paranasal sinuses, orbits and inner ear on a 16-row MDCT scanner. Thyroid dose was calculated using mathematical phantoms representing the average individual as newborn and child of 1, 5, 10 and 15 years old. Monte Carlo calculations were compared with measurements performed on four physical humanoid phantoms and 18 children using thermoluminescent dosemeters. Results: Sequential scanning of the brain, sinuses, orbits and inner ear resulted in a thyroid dose of 2.1-2.8, 1.0-1.6, 0.5-0.9 and 0.6-0.9 mGy, respectively, depending upon the patient’s age. The corresponding dose from spiral examinations was increased by 1.2-3.7 times due to the effect of z-overscanning on the exposure to tissues beyond the imaged volume. Normalized data are presented to estimate thyroid dose from any head examination taking into account the patient’s age, scanned region, MDCT protocol parameters and acquisition mode. The mean difference between dosimetric measurements and Monte Carlo calculations was 9.7%. The excess lifetime risk for thyroid cancer induction from sequential and spiral scanning was below 21 and 65 per million patients, respectively. Conclusion: The use of sequential instead of spiral acquisition mode for pediatric MDCT studies of the head can considerably reduce thyroid dose and associated cancer risk.

B-786

11:15

How low can we go? Low dose 64 slice PET/CT protocols for paediatric patients: The importance of multidisciplinary teamwork L. Heiberg, C.E. de Lange, L.A. Refsum, W. Nordhøy, K. Johnsrud, H. Solheim, A.-E. Spiten, M. Sheikhaeri; Oslo/NO ([email protected]) Purpose: To minimize the CT dose necessary for PET interpretation in PET/CT examinations through collaboration between different disciplines. Methods and Materials: 20 consecutive paediatric examinations were performed with Siemens Biograph 64. The patient age range was 4-16 years. CT scan parameters, image quality, 18F-FDG-uptake and administered activity with estimation of effective doses were evaluated by two paediatric radiologists, three nuclear medicine specialists and one medical physicist by consensus. CT scan parameters for paediatric protocols provided by Siemens were replaced by weight based protocols, and adjusted step-by-step to further reduce dose. Results: kVp was reduced according to weight and mAs was reduced to the minimal values allowed by the scanner, maintaining sufﬁcient image quality for anatomic localization and attenuation correction for PET interpretation. Effective doses for CT decreased 5-10 times, ranging from 0.1 to 10.1 mSv, CTDI volume from 0.7 to 10.3 mGy. 18F-FDG administered activity was kept unchanged and estimated to 7-10 mSv, contributing up to 90% of the total effective dose. Conclusion: Following the ALARA principle, adjusting low dose paediatric CT protocols for PET/CT will reduce effective dose considerably while maintaining sufﬁcient image quality for PET interpretation. The administered 18F-FDG activity contributes the major part of effective dose and will be reduced in the future. Optimizing multimodality imaging protocols can only be done through good coordination and collaboration in multidisciplinary teams.

A

B

C

D

E

F

G

323

Tuesday

B-784

blood ﬂow (rCBF) and regional cerebral blood volume (rCBV) abnormalities. PCT results were compared to short- and mid-term clinical outcome. Results: Three patients with low Glasgow Coma Scale (GCS) (b8) and bad clinical outcome showed an increased MTT and decreased rCBV and rCBF. Five patients with low GCS and good clinical outcome showed an increased MTT without abnormalities of rCBV and rCBF. In patients with GCSr8 and good outcome, PCT maps were normal in two cases; transient PCT abnormalities were identiﬁed in one case with an embedded fracture of the skull and in one case with an epileptic seizure. Conclusion: Cerebral PCT can identify diffuse abnormalities of cerebral perfusion in children with traumatic brain injury showing a low initial GCS and a bad outcome. PCT can be a valuable tool to predict the severity of the prognosis of these patients as soon as they are evaluated by CT-scan during their admission at the ER.

Scientific Sessions B-787

11:24

Automatic exposure control in pediatric multi-detector CT examinations: Can radiation dose be substantially reduced? A. Papadakis, K. Perisinakis, J. Damilakis, N. Gourtsoyiannis; Iraklion/GR ([email protected]) Purpose: The purpose of this study was to investigate the potential of automatic exposure control (AEC) on dose reduction in pediatric patients undergoing multidetector CT examinations. Methods and Materials: Four physical anthropomorphic phantoms that simulate the average individual as neonate, 1-, 5-, and 10-year-old child were employed in the current study. Phantoms were scanned with a 16-row CT scanner using the AEC mechanism. This mechanism combines z-axis and angular (x-y) mA modulation. Percent dose reduction (%DR) factors were determined for standard helical scan protocols used for head and neck, thorax, thorax and abdomen, abdomen, abdomen and pelvis, pelvis, and whole body examinations. The effects of the direction of the scanogram acquisition and tube potential on %DR and image quality were investigated. Results: The mean %DR factors were 4.4 for the neonate, 24.1 for the 1-year-old, and 10.9 for the 5-year-old child. A mean dose increase of 2.1% was observed for the 10-year-old child. AEC did not affect image quality. Scanogram acquisition in the lateral as opposed to anterior-posterior direction and/or the use of a higher tube potential may result in a substantial increase in the dose reduction without impairing image quality. Conclusion: AEC results in a signiﬁcant dose reduction in children up to ﬁve years old. Scanogram acquisition in the lateral direction is required to achieve a substantial dose saving in older children. Careful selection of the scan parameters is required to optimize the efﬁciency of AEC in dose reduction while preserving image quality.

B-788

11:33

Which is the best imaging modality to capture bone erosion in JIA? M.B. Damasio, M. Valle, C. Malattia, F. Magnaguagno, C. Martinoli, S. Viola, A. Ravelli, A. Martini, P. Tomà, A. Pistorio; Genoa/IT ([email protected]) Purpose: To compare sensitivity of magnetic resonance imaging (MRI), sonography (US) and conventional X-ray (X-ray) in evaluating bone erosion in joints of patients (pts) with juvenile idiopathic arthritis (JIA) and to relate imaging results and clinical parameters. Methods and Materials: Forty-four JIA pts followed at our institution underwent wrist (26 pts) or hip (18 pts) radiographs, US and MRI in the same day. MRI, US and X-ray erosion score (E-score) scales were decided and E-score assigned independently by two experienced investigators, blinded to clinical ﬁndings. The relationship between imaging E-score and clinical data: number of limited joints (LOM); severity of joint involvement; Childhood Health Assessment Questionnaire (CHAQ) and Articular Juvenile Assessment Damage Index Score (JADI) were evaluated. Results: The inter-reader intraclass correlation coefﬁcients were excellent for both MRI (ICC wrist: 0.97; hip: 0.95) and X-ray (ICC wrist: 0.97; hip: 0.85). 81.8, 45.4, and 31.7% of the patients had erosions identiﬁed on MRI, X-ray and US, respectively. 16/44 patients (36.3%) had erosions identiﬁed only by MRI not on X-ray. All eight patients negative on MRI were also negative on X-ray. Spearman’s correlation between MRI and X-ray scores was satisfactory (rs=0.74). Wrist MRI and X-ray scores correlated with LOM score (rs=0.69 and rs=0.77); hip MRI and X-ray correlated with disease duration (rs=0.55) and JADI (rs=0.56), respectively. The correlations between MRI and X-ray and clinical indicators of disease activity were low, as expected. Conclusion: MRI appears to be the most sensitive modality to capture bone erosion compared to US and X-ray.

B-789

11:42

Purpose: To determine levels of satisfaction of UK radiologists regarding current training, service and support received in the ﬁeld of suspected physical abuse. To seek their opinion on how to meet professional development needs in this area. Methods and Materials: A list of UK radiology departments was obtained from the RCR. Researchers contacted these departments and spoke to a radiologist specialising or with an interest in paediatrics. Responses to speciﬁc questions were recorded using a web-based questionnaire - www.surveys.emnm.co.uk/gosh/index. php linked to a Microsoft Excel spreadsheet.

A

B

B-790

11:51

Value of diffusion-weighted imaging in routine MR scans using a fusion technique for visualization of malignant tumors in paediatric oncology patients S. Alibek, A. Cavallaro, W.A. Bautz, G. Staatz; Erlangen/DE ([email protected]) Purpose: To evaluate the value of diffusion-weighted imaging (DWI) for detection and delineation of malignant thoracic and abdominal tumors in children using fusion technique in routine MR scans. Methods and Materials: 10 paediatric patients with suspected malignant tumors were scanned at 1.5 T MR (Magnetom ESPREE) with routine contrast enhanced scan protocol. Additionally DWI was performed with standard scan parameters (B=0, 500, 1000; slice thickness 4 mm, transverse plane). Routine MR scans and DWI were read separately and combined by two experienced radiologists. Diagnosis was established in consensus. DWI images were evaluated regarding tumor detection, delineation, anatomical correctness and compared with routine T2 and T1w pre/post-contrast images. DWI and anatomical T1w precontrast images were fusioned. Results: 1. DWI showed signiﬁcantly high signal intensities in all patients for high B-values, matching with areas of contrast-uptaking lesions allowing correct detection of malignancy. 2. Conventional T2 and T1w images are superior to DWI for anatomical correctness. 3. Fusion of DWI images with T1w non-enhanced images allows correct anatomical delineation of tumor extent. 4. DWI images add value to routine MR scans enabling combined visualization of anatomy and functional MRI. Conclusion: 1. DWI is a stable functional imaging method. 2. Areas of high signal intensity in DWI at high B-values correlate accurately with contrast-enhancing malignant processes. 3. Deﬁciency of demonstrating exact anatomy of malignancies can be compensated by fusion of DWI and T1w images. 4. MRI without Gdadministration seems to be possible for tumor staging and has to be investigated in further studies.

10:30 - 12:00

Room L/M

Contrast Media

SS 1806a From research to practice Moderators: J.-M. Correas; Paris/FR D. Reddan; Galway/IE

B-791

10:30

Gadovist® 1.0M: Tolerance and clinical safety M. Forsting; Essen/DE ([email protected])

UK radiologists’ satisfaction with training and support in suspected physical abuse: A questionnaire survey A.C. Offiah1, A. Pervaiz1, C. Nwachukwu1, C. Walker2, C. Landes2; 1London/UK, 2 Liverpool/UK (ofﬁ[email protected])

324

Results: So far researchers have solicited the views of 60 departments, 39 of which (65%) were in a DGH. 33 (55%) have a dedicated paediatric radiologist. 31 (52%) have a named radiologist reporting child abuse cases. 10 (16%) refer all suspected abuse cases elsewhere for reporting. Three departments (5%) did not have ready access to a paediatric radiologist. 30 (50%) have double reporting. 39 (65%) receive b10 cases of suspected physical abuse per annum. 30 (50%) feel the service could be improved. The majority favour a) formal guidance for double reporting and brain imaging b) regular regional meetings c) the development of a national reference archive and d) formation of local/national "chambers" of expert witness paediatric radiologists. Conclusion: Findings should inform the design of future radiology child abuse training and professional development programmes. More robust training and support may attract an increasing number of radiologists to this difﬁcult and often controversial ﬁeld. Practice in other European countries should be ascertained.

C

D

E

F

G

Purpose: To determine the tolerance and clinical safety of Gadovist® 1.0M (Gadobutrol, Bayer Schering Pharma AG, Berlin, Germany) for CE-MRI, in the clinical routine setting in a large number of unselected patients. Methods and Materials: Six prospectively planned surveillance studies with gadobutrol were conducted. Individual patient data and occurrence of ADRs were collected utilizing a standardized questionnaire. Results: A total of 14,422 patients were enrolled. Body regions most frequently investigated by CE-MRI were head/neck/brain (54.3%) and spine (7.2%). The mean age of the patients was 53.7 years. 1.3% of the patients were less than 18 years old. The mean applied volume of Gadovist® was 11.96 ml. The contrast quality was assessed in two studies and was rated as very good/good in 94.5%. 78 of the 14,422 patients reported at least one ADR, corresponding to an overall incidence

Scientific Sessions rate of 0.55%. Two serious adverse drug reactions (SADR) were reported (rate 0.01%), of which one was severe anaphylactoid reaction requiring hospitalisation and medical treatment of the patient; the other also requiring hospitalisation was described as itching and swelling of the throat. The most frequently reported ADR was nausea, which occurred in 36 patients (0.252%; uncommon). Conclusion: Gadovist® 1.0M provides excellent contrast quality in CE-MRI. The contrast agent is very well tolerated and has an excellent safety proﬁle. The adverse drug reactions observed following the i.v. injection of Gadovist® 1.0M are well comparable to the published data of Gadolinium-DTPA and other 0.5M Gd-chelates.

B-792

10:39

Sinerem-enhanced intravenous MR lymphography for staging pelvic cancers: Results of a European multicenter phase III trial including 271 patients B. Hamm1, F. Caseiro-Alves2, M.-F. Bellin3, A. Padhani4, R. Passariello5, T.H. Helbich6, C. Roy7, J.O. Barentsz8, D. Hahn9, M. Lemort10, D. Portalez11, P. Marano5, A. Chiesa12, R. Aziza13, U. Mödder14; 1Berlin/DE, 2Coimbra/PT, 3Villejuif/FR, 4Northwood/UK, 5 Rome/IT, 6Vienna/AT, 7Strasbourg/FR, 8Nijmegen/NL, 9Würzburg/DE, 10Brussels/BE, 11 Saint-Jean/FR, 12Brescia/IT, 13Toulouse/FR, 14Düsseldorf/DE ([email protected]) Purpose: To assess the diagnostic accuracy of Sinerem®-enhanced versus unenhanced MRI for lymph node (LN) staging in patients with pelvic cancers. Methods and Materials: A total of 271 patients with cancer of the prostate (n=116), uterus (n=106), or bladder (n=49) with doubtful N staging were enrolled at 16 European centers. Standardized MRI was performed before and 24-36 hours after i.v. infusion of Sinerem® (Guerbet, France) at a dose of 2.6 mg Fe/kg. Blinded readings were performed off-site (three readers) using unenhanced images (pre), combined unenhanced and Sinerem®-enhanced images (pre&post) and Sinerem®enhanced images (post) as also on-site (pre versus pre&post). Based on LN size (pre), patterns of signal changes (pre&post), and signal intensities (post), LN were classiﬁed as nonmetastatic or metastatic. Histopathology served as gold standard. Evaluation was performed on a patient basis. Results: Sixty-eight patients (26.6%) were N+ at histology; 328 LN (mainly b10 mm) were individually correlated between MRI and histology. Sensitivity was higher for pre&post than pre alone in the off-site reading, highly signiﬁcantly so for two readers ($ 22%, p 0.001), as also on-site (83 vs 48%, p 0.01). NPV was increased for all readers (84-86% pre&post versus 79-83% pre), as also on-site (97 vs 46%). Post and pre&post readings yielded similar results. Conclusion: In nodal staging of patients with pelvic cancer Sinerem®-enhanced MRI signiﬁcantly improves diagnostic accuracy by higher sensitivity and NPV, and has a signiﬁcantly favorable inﬂuence on the therapeutic strategy. Precontrast MRI can be omitted.

B-793

10:48

3-D multiphasic low dose contrast enhanced MR angiography using TWIST on a 32-channel head coil at 3 Tesla: Comparison of a conventional gadolinium chelate with a high relaxivity agent F.L. Giesel1, V.M. Runge2, S. Baumann2, B. Corell1, E.L. Lowther2, J. Gendron2, H.-U. Kauczor1, M. Essig1; 1Heidelberg/DE, 2Temple, TX/US ([email protected])

10:57

Comparison of the contrast media Gd-DTPA and gadofosveset for absolute quantiﬁcation of myocardial perfusion using high-resolution MRI C.O. Ritter1, A. Wilke1, T. Wichmann2, M.J. Beer1, D. Hahn1, H. Köstler1; 1 Wuerzburg/DE, 2Rimpar/DE ([email protected]) Purpose: Aim of this study was to compare the contrast media (CM) Gd-DTPA and Gadofosveset for the absolute quantiﬁcation of myocardial perfusion using a high resolution GRAPPA-sequence. Methods and Materials: 10 healthy volunteers were examined twice at two different days with a ﬁrst-pass-perfusion examination at rest using the prebolus technique. A 1.5 T-scanner and a 32-channel heart-array coil with a saturationrecovery TrueFISP-GRAPPA-sequence (acceleration factor 3) and a high resolution of 1.8x1.8 mm were used. Manual delineation of the myocardial contours was performed on three slices. The myocardium was automatically divided into eight sectors. Absolute quantiﬁcation was performed after coil-, baseline- and contamination-correction. On the ﬁrst appointment, 1 cc/4 cc of the extracellular CM Gd-BOPTA (Multihance®) were administered, on the second date 1 cc/4 cc of the CM Gadofosveset (Vasovist®). Results: Using Gd-BOPTA, the perfusion value at rest was 0.62p0.25 cc/g/min (meanpstandard deviation), for Gadofosveset 0.66p0.25 cc/g/min. The Wilcoxon test showed no signiﬁcant difference (p=0.69). The apparent mean residence time in the myocardium was 20 s for the Gadofosveset and 24 s for Gd-BOPTA. Neither the measured SNR-ratio (p 0.05) nor the subjectively rated image contrast during manual segmentation of both examinations showed a signiﬁcant difference. Conclusion: The use of the CM Gadofosveset for the absolute quantiﬁcation of the myocardial perfusion is possible. The acquired perfusion values show no signiﬁcant difference from those determined with Gd-BOPTA. The use of Gadofosveset neither improves the SNR of the images nor the variation of the perfusion values. Gadofosveset does not show an intravasal concentration time course for ﬁrst-pass-perfusion studies.

B-795

11:06

Intraindividual comparison of Gadovist®, Omniscan®, Multihance® and Vasovist® for CE-MRA of the supra-aortic arteries B. Buerke1, T. Allkemper1, H. Kugel1, M. Voth2, C. Bremer1, S. Evers1, W.L. Heindel1, B. Tombach1; 1Muenster/DE, 2Berlin/DE ([email protected]) Purpose: In addition to hardware and software conﬁguration and contrast agent application strategy, the contrast agent itself is supposed to have a high impact on image quality of CE-MRA depending on its physicochemical properties (e.g. relaxivity, blood-pool-effect). The purpose of this intraindividual comparison study was to evaluate differences of four different contrast agents for CE-MRA of the supraaortic arteries. Methods and Material: In a total of 11 patients with headache symptoms four different contrast agents (1M Gadovist®; 0.5M Omniscan® each 0.1 mmol/kg bw; 0.5M Multihance®; 0.05 mmol/kg bw and Vasovist®; 0.03 mmol/kg bw; all 2 ml/s) were compared intraindividually for CE-MRA of the supraaortic arteries on a 1.5 T scanner (Gyroscan Intera®, Philips Medical Systems, The Netherlands). To avoid potential accumulation effects the time interval between two injections was 24 h. Qualitative analysis regarding image quality and contrast was performed by two radiologists in consensus. For quantitative analysis and intraindividual comparison SI and SD were measured to allow intraindividual comparison of SNR and CNR calculations. Results: Qualitative analysis documented a diagnostic image quality for all CEMRA. The ranking regarding image contrast showed slightly superior contrast levels for Gadovist® and Vasovist® as compared to Multihance® and Omniscan®. SNR and CNR (SNR/CNR: Gadovist® 18.8p7.4/15.6p5.9; Omniscan® 18.1p6.4/15.2p7.1; Multihance® 16.7p5.8/14.3p5.7; Vasovist® 20.2p6.4/16.2p5.8) showed no signiﬁcant differences between all contrast agents (p 0.05). Conclusion: All analysed contrast agents allowed a highly diagnostic image quality. No signiﬁcant differences in SNR/CNR between Vasovist®, Gadovist®, Multihance® and Omniscan® enhanced CE-MRA were proven. Subjective ranking of image quality was slightly superior for Gadovist® and Vasovist® in the investigated doses.

A

B

C

D

E

F

G

325

Tuesday

Purpose: With the recognition of nephrogenic systemic ﬁbrosis, gadolinium chelate administration in patients with renal failure has become a major concern, in particular accumulated dosage. The purpose of this investigation was to explore the possibility of low dose MR contrast enhanced angiography with high temporal and spatial resolution. Methods and Materials: Eleven male New Zealand white rabbits were evaluated with contrast enhanced MRA (TWIST: TR/TE=2.5/1.1, 1.2x1x1 mm3 voxel size, IPAT=3, TA=1.95 sec), examining the abdominal vascular structures at 3 Tesla using a prototype 32-channel head coil (Siemens Tim Trio). Two different contrast media (i) gadobenate dimeglumine and (ii) gadoteridol were injected in separate sessions ( 24 h) using a power injector (0.04 mmol/kg contrast dose). Quantitative analysis was performed by placing a circular region of interest (ROI) in the aorta at the level of the renal bifurcation and another in the adjacent tissue to assess contrast-to-noise ratio (CNR). Results: All 11 rabbits were evaluated successfully twice with multiphasic MRA using TWIST. Signiﬁcantly higher maximum signal intensity (~25-30%) was found in the aorta with the gadobenate dimeglumine, when compared to gadoteridol (p=0.001). The improved CNR also led to a better depiction of peripheral vessels while using almost 1/3 of the normal contrast dose. Conclusion: The utilization of higher ﬁeld strength, improved coil design, and higher relaxivity contrast media, combined with multiphasic MR angiography, permits high quality low dose MR angiography, representing a possible approach to address the NSF concern.

B-794

Scientific Sessions B-796

11:15

Contrast-enhanced MRA of the abdominal and peripheral vessels: Is there a need for blood pool agents? R. Iezzi1, S. Thurnher2, G. Soulez3, G. Schneider4, N. Shen5, M.A. Kirchin6, G. Pirovano5, A. Spinazzi5; 1Chieti/IT, 2Vienna/AT, 3Montreal, QC/CA, 4 Homburg a.d. Saar/DE, 5Princeton, NJ/US, 6Milan/IT ([email protected]) Purpose: To compare safety and diagnostic performance data from Phase III regulatory trials to evaluate gadobenate dimeglumine (Gd-BOPTA) and gadofosveset (MS-325) for renal and peripheral contrast-enhanced MR angiography (CE-MRA). Methods and Materials: Similar examination and blinded assessment methodology was utilized in all studies to determine the safety and diagnostic performance of the agents for detection of signiﬁcant ( 50%) steno-occlusive disease. Digital Subtraction Angiography (DSA) was used as the standard of truth. Diagnostic performance data (sensitivity, speciﬁcity, accuracy, predictive values [PVs], likelihood ratios [LRs]) from the regulatory programs and blinded reader agreement were compared (Chi-square test and Kappa statistics). Results: CE-MRA with Gd-BOPTA was more speciﬁc (92.4 vs 80.5%, p 0.0001) and accurate (83.6 vs 77.1%, p=0.022) than CE-MRA with MS-325 in the detection of signiﬁcant renal artery stenosis. The average sensitivity was higher for MS-325 (74.4 vs 67.3%, p=0.011) in peripheral vessels although Gd-BOPTA was more speciﬁc (93.0 vs 88.2%, p 0.0001) with no difference in accuracy (86.6 vs 86.3%, p=0.66). Positive PVs were higher (p 0.0001) for Gd-BOPTA in both territories. Pre- to post-test shifts in the probabilities of signiﬁcant disease were greater after Gd-BOPTA. Kappa values of 0.66 and 0.69 were obtained for Gd-BOPTA compared with between 0.32 and 0.42 for MS-325. The incidence of adverse events was higher after MS-325. Conclusion: The diagnostic performance of CE-MRA for the detection of signiﬁcant steno-occlusive disease in the renal or peripheral arteries may be similar with MS-325 and Gd-BOPTA although the rate of adverse events appears higher with MS-325.

B-797

11:24

Inﬂuence of the contrast agent iodine concentration on the visualization of coronary arteries using automated and manual segmentation methods in dual-source CT angiography A. Kuettner1, K. Anders1, S. Busch2, C. Becker2; 1Erlangen/DE, 2Munich/DE ([email protected]) Purpose: To compare iomeprol-300 (mg iodine/mL) and iomeprol-400 for coronary artery visualization in dual source CT angiography using automated and manual segmentation methods. Methods and Materials: Fifty-seven patients referred for coronary CTA received iomeprol-300 (n=29) or iomeprol-400 (n=28) at bodyweight adjusted volumes and ﬂow rates. Iodine ﬂow rates were 16.5-17.5 mg I s-1 kg-1 and 21.8-23.3 mg I s-1 kg-1, in the iomeprol-300 and -400 groups, respectively. Dual-source CT scanners were used with individual bolus arrival time determined using bolus-tracking software. The length and number of coronary segments were determined for the two study groups with automatic, semiautomatic, and manual segmentation methods. Mean contrast densities were recorded for the right and left ventricle and for the right, left main, left anterior descending, and left circumﬂex coronary arteries. WilcoxonMann-Whitney U tests were applied. Results: Using iomeprol-400 a signiﬁcantly (p=0.047) greater number of visualized coronary artery segments was noted for the automatic segmentation method but not for the manual or semiautomatic methods. A trend towards better deﬁnition of segment length was noted for iomeprol-400 for all three segmentation methods. Contrast density measurements were signiﬁcantly (p=0.019) higher for iomeprol-400 compared with iomeprol-300 for all arterial regions of interests. Using this bodyweight adjusted iodine ﬂux protocol, resulting contrast densities were independent of bodyweight. Conclusion: The higher iodine inﬂux with iomeprol-400 leads to improved visualization of the length and number of coronary artery segments and signiﬁcantly greater contrast enhancement. Beneﬁts are apparent for all segmentation methods and particularly for automatic segmentation.

B-798

11:33

The role of the oral contrast medium in 18 F-FDG-PET studies S.M. Shaikh; Hyderabad/IN Purpose: It is postulated that the use of computed tomography (CT) contrast enhancement agents have led to attenuation-induced artefacts on 18 F-ﬂuorodeoxyglucose positron emission tomography (18 F-FDG-PET/CT) systems. Some

326

A

B

C

D

E

F

G

centres may withhold the use of such agents. Whilst there is theoretical evidence to support the claim, the clinical relevance of these induced artefacts has not been widely established. Moreover, the potential beneﬁts of bowel enhancement on PET/ CT have yet to be formerly evaluated. We therefore prospectively examined PET/CT studies to assess whether the use of oral contrast medium induced clinically relevant artefacts and whether the use of these agents was diagnostically helpful. Methods and Materials: Over a two-month period, 18 F-FDG-PET/CT images were prospectively reviewed from 200 patients following Gastrografﬁn administration two hours prior to examination. Both a radiologist and a nuclear physician reviewed the images for contrast medium mediated clinically relevant artefacts. Artefacts were sought on the CT attenuation corrected images and were compared to the nonattenuated corrected images. The number of examinations where the oral contrast aided image interpretation was also noted. Results: There were no oral contrast medium induced clinically signiﬁcant artefacts. In 38/200 patients, oral contrast aided image interpretation (differentiation of mass/ node from bowel, discrimination of intestinal wall from lumen or by deﬁning the anatomy of a relevant site). Of the 38 patients, in 33 the anatomical site of interest was the abdomen/pelvis. Conclusion: The use of oral contrast medium in 18 F-FDG-PET studies should not be withheld as it improves image interpretation and does not produce clinically signiﬁcant artefacts.

B-799

11:42

Comparison of 68Ga-DOTA-Tyr (3)-Octreotide PET and diagnostic CT for treatment evaluation of radionuclide therapy M. Gabriel, A. Oberauer, D. Kendler, D. Putzer, R. Bale, R. Madleitner, C. Uprimny, P. Kovacs, I.J. Virgolini; Innsbruck/AT ([email protected]) Purpose: 68Ga-DOTA-Tyr (3)-Octreotide PET has proven usefulness in patients with neuroendocrine tumours. Radionuclide therapy (90Y-DOTA-TOC and/or 177LuDOTA-TATE) is one choice of treatment which also needs an accurate diagnostic modality for early evaluating treatment response. In our study PET and CT - using RECIST criteria - were prospectively compared. Methods and Materials: Thirty-two patients (22 male,10 female; age range 35-84 years) suffering from advanced neuroendocrine tumours were investigated prior to and after 3-7 cycles of radionuclide therapy. Six weeks after the last therapy cycle the follow-up control was performed. PET images were visually interpreted by two experienced nuclear medicine physicians. For comparison helical CT scans were obtained (GE HiSpeed CT/i Advantage) using Ultravist 370 contrast media. Results: In twenty patients PET and CT produced a comparable result (SD in 16, PR in 1, CR in 1, PD in 2 patients). Differences were observed in the remaining 12 patients without statistical signiﬁcance between both modalities (p = 0.15, McNemar test). In one patient diagnostic CT conﬁrmed pulmonary metastases not detected by PET like in two other cases with progressive liver disease. In nine patients PET provided additional information not obtained by CT. Five patients showed treatment response while in four patients tumour progression was observed, including two patients with progressive liver and two patients with progressive bone metastases. All available imaging techniques and course of disease were used as reference standard. Conclusion: PET and CT indicate to play a complementary role in the follow-up of patients treated with radionuclide therapy.

B-800

11:51

Highly iodinated intravenous contrast agent for PET/CT: Feasible? Useful? Advisable? T. Heusner, H. Kuehl, P. Veit-Haibach, M. Forsting, A. Bockisch, G. Antoch; Essen/DE ([email protected]) Purpose: Highly concentrated intravenous contrast materials (HCCM) have been found of beneﬁt in CT imaging. However, iv contrast may inﬂuence tracer quantiﬁcation and cause artifacts in PET/CT when using the CT data for PET attenuation correction. The aim of this study was to assess the feasibility of applying a HCCM (Imeron™ 400, Bracco Altana Pharma) containing 400 mg iodine/ml in PET/CT imaging and to compare HCCM to a lower concentrated contrast material containing 300 mg iodine/ml (LCCM, Imeron™ 300, Bracco Altana Pharma). Methods and Materials: In 60 oncological patients whole-body FDG-PET/CT was performed with HCCM (n=30) or LCCM (n=30). PET tracer uptake (SUVmax) and CT enhancement (Hounsﬁeld Units, HU) were quantiﬁed at 16 predeﬁned positions in different blood vessels and organ parenchyma. Potential PET artifacts were documented. Tumor delineation was evaluated qualitatively on CT according to a 4-point-scale. Differences between HCCM and LCCM were tested for signiﬁcance by Mann-Whitney-Wilcoxon-Test. Results: Compared with LCCM the use of HCCM did not have an effect on SUVmax (p 0.05) or on the number of PET artifacts (p 0.05). However, CT contrast

Scientific Sessions enhancement was signiﬁcantly better with HCCM in 12 of 16 positions (p 0.002) compared to LCCM. There was no difference in tumor delineation between HCCM and LCCM (p=0.464). Conclusion: Highly iodinated contrast materials did not cause any more PET artifacts or inaccuracies in PET tracer quantiﬁcation than LCCM. Therefore, HCCM may be considered an option in selected PET/CT indications where strong CT contrast enhancement is desired.

10:30 - 12:00

Room N/O

Neuro

SS 1811 Epilepsy and development issues Moderators: A. Cianfoni; Rome/IT J.M. García Santos; Murcia/ES

B-801

10:30

Incomplete inversion of the hippocampus in subjects without severe developmental anomalies: Is there a relationship to epilepsy? D. Bajic, E. Kumlien, P. Mattsson, S. Lundberg, C. Wang, O. Eeg-Olofsson, R. Raininko; Uppsala/SE ([email protected]) Purpose: Incomplete inversion of the hippocampus, an imperfect fetal development, has been described in patients with epilepsy or severe midline malformations. We have reported this condition in 19% of nonepileptic subjects without obvious developmental anomalies. Now we compare the frequency of the incomplete hippocampal inversion in populations with and without epilepsy. Methods and Materials: 296 patients, examined with brain MRI, were drawn from the regional epilepsy register. 74 of them were excluded because of a disease or condition affecting the temporal lobes. 150 subjects without epilepsy were used as controls: 34 healthy volunteers, 116 patients without obvious intracranial developmental anomalies or any condition which may affect the temporal lobes. The images were analysed without knowing clinical data. When clinical records were checked, 20 patients were excluded for incomplete data. The ﬁnal material consisted of 202 epilepsy patients and 150 controls. Results: 57/202 epilepsy patients (28%) had temporal lobe epilepsy (TLE). Incomplete hippocampal inversion was found in 18% of the controls and in 30% of the epilepsy patients. The difference was statistically signiﬁcant (p 0.05). Incomplete hipppocampal inversion was found in 19% of the TLE patients but in 32% of the patients with other types of epilepsy (p 0.05). Conclusion: Incomplete hippocampal inversion is more common in an epileptic than non-epileptic population but the frequency in TLE was the same as in the non-epileptic population. So this morphological variation seems to be related to deviations in parts other than temporal region of the brain.

B-802

10:39

Diffusion tensor imaging in lateralisation and localisation of epileptogenis focus in temporal lobe epilepsy with and without hippocampal sclerosis S. Gan, W. Handoko, S. Christensen, P.M. Desmond, T.J. O’Brien; Melbourne, Victoria/AU ([email protected])

B-803

10:48

Presurgical and postsurgical ﬁndings in refractory patients with anterior temporal lobe resection. Are they predictive factors of surgical outcome? N. Bargalló, L. Olondo, S. Capurro, S. Gonzalez, J. Rumià, M. Carreño, A. Donaire; Barcelona/ES ([email protected]) Purpose: To determine if the presurgical and postsurgical ﬁndings after anterior temporal lobe resection in refractory epilepsy patients are predictive factors of surgical outcome. Methods and Materials: Thirty-ﬁve refractory epilepsy patients with anterior temporal lobe resection were retrospectively revised. All patients had temporal lobe seizures and were previously EEG-video monitored. Magnetic resonance imaging was performed before and after surgery following a speciﬁc epilepsy protocol that included 1.5 mm slice coronal SPGR, 3 mm slice coronal FLAIR and T2WI centred in temporal lobe. Presurgical ﬁndings, the extent of resection, and postsurgical ﬁndings such as hemosiderin and malacia were evaluated. Annual clinical follow-up was performed using International League Against Epilepsy criteria. Results: Twenty-three patients had mesial sclerosis, ﬁve patients had dual lesions, three patients had tumour lesions and in four patients MRI was normal. All patients with tumour lesions had good outcome, while three of the four patients with normal MRI had seizures after surgery. Patients with mesial sclerosis had good outcome in 67% of the cases. In 22 patients, the anterior temporal lobe resection was complete though seven patients had seizures after surgery. In 12 patients, hyperintense hippocampus tail and/or body were present in the postsurgical studies although 58% of these patients were free of seizures after surgery. Conclusion: Presurgical ﬁndings in refractory epilepsy are predictive factors of good outcome; however, incomplete hippocampus resection was not an indicator of seizure recurrence after surgery.

B-804

10:57

Unverricht-Lundborg disease (ULD): Magnetic resonance imaging with voxel-based morphometry in nineteen patients and nineteen matched controls P. Koskenkorva, E. Niskanen, M. Könönen, E. Mervaala, R. Kälviäinen, R. Vanninen; Kuopio/FI (paivi.koskenkorva@kuh.ﬁ) Purpose: Unverricht-Lundborg disease (ULD), progressive myoclonic epilepsy type 1 (EPM1, OMIM254800) caused by mutations of the cystatin B (CSTB) gene, is a rare autosomal recessive neurodegenerative disorder characterized by age of onset from 6-16 years, stimulus-sensitive myoclonus, and tonic-clonic epileptic seizures. Thus far, imaging studies of ULD patients are scanty and include only small numbers of patients. The purpose of this study was to evaluate MRI characteristics of ULD patients. Methods and Materials: Nineteen patients with genetically veriﬁed ULD (four compound heterozygotes, 15 homozygotes) underwent MRI (1.5 Tesla, Siemens Avanto). MR imaging included T1- and T2-weighted spin-echo sequences, ﬂuid attenuated inversion recovery sequence (FLAIR), T1-weighted three-dimensional images, MR spectroscopy and diffusion tensor imaging. Motion correction software was used if necessary. T1-3D images were analyzed using optimized voxel-based morphometry (VBM) and statistical parametric mapping (SPM2 running under Matlab 6.5) and compared with matched controls (n=19). Results: Conventional images revealed no focal abnormalities. In VBM (p 0.05), ULD patients showed cortical gray matter atrophy in frontal premotor and supplementary motor areas, left cuneus and right occipital lobe. No differences were seen in cerebellum or brain stem. Conclusion: In spite of the stimulus-sensitive myoclonus, MR imaging was feasible in all patients. Detailed MR imaging together with neurophysiological evaluation may help to reveal the pathogenesis of ULD.

B-805

11:06

Electrical impedance tomography (EIT): A new imaging method A. Romsauerova1, L. Horesh2, A. McEwan1, D. Holder1; 1London/UK, 2Atlanta, GA/US ([email protected]) Purpose: Electrical Impedance Tomography (EIT) is a new medical imaging method in which tomographic impedance images are rapidly produced using arrays of electrodes placed around the body. EIT has the potential to distinguish between ischaemic and haemorrhagic strokes in a non-invasive, safe, cost effective and timely manner, in order to give an early diagnosis of ischaemic stroke and allow

A

B

C

D

E

F

G

327

Tuesday

Purpose: Changes in ADC and FA are reported in patients with temporal lobe epilepsy (TLE). The isotropic portion (p), anisotropic portion (q) and Euclidean magnitude (L) of the tensor allow more precise dissection of the tensor. This study aims to determine whether these DTI changes primarily reﬂect the presence of hippocampal sclerosis (HS) or the epileptogenic substrate independent of the nature of the underlying pathology by examining the DTI variables in the hippocampus and anterior temporal pole (ATP) in TLE patients with HS on MRI (HS+veTLE) versus those with no apparent structural abnormality (HS-veTLE). Methods and Materials: T1-weighted images of 18 HS+veTLE and 10 HS-veTLe patients were realigned to the diffusion weighted images. CSF partial voluming effects were removed. DTI and volumetric measurements were performed in the regions of interest. Results: Hippocampal ADC increased on the epileptogenic side in HS+veTLE (1043 vs 977 mm2/s; p 0.005) but not in HS-veTLE. Increased ADC (842 vs 808 mm2/s; p 0.05), p (1459 vs 1400 mm2/s; p 0.05) and decreased volume (3453 vs 4206 mm3; p 0.05) were found in the ipsilateral anterior temporal pole white matter in HS-veTLE but did not differ between the TLE subtypes. Conclusion: Increased hippocampal diffusivity in TLE likely reﬂects the structural changes of HS rather than epilepsy per se. Increased diffusivity in the ATP white

matter in HS-veTLE suggests the epileptogenic region is in neocortical ATP. DTI may provide further localisation information in these patients. p, q and L may be useful in future studies.

Scientific Sessions early thrombolysis. The purpose of this work was to assess the performance of the EIT system, the UCLH Mk2.5 developed at UCL, for imaging acute stroke. Methods and Materials: The performance of the EIT was assessed using computer simulation, animal studies and phantom studies that closely simulated acute stroke. Results: The computer simulation of acute stroke showed that impedance changes can be detected using our EIT system. The maximal scalp impedance changes for ischemic stroke ranged between 0.2 and 1.9%, whereas for haemorrhagic stroke they ranged from -0.7 to -6.1%. In the animal model of ischaemic stroke, four vessel cerebral ischaemia of 15 min caused an increase of impedance by 50-200% measured from the brain surface. Controlled phantom studies of acute stroke were coherent with the computer simulations and animal studies. Conclusion: These results demonstrated that our multi-frequency EIT system can detect and image changes across frequencies associated with ischaemic or haemorrhagic stroke, when compared with the normal brain. Therefore, we are now continuing with patient measurements on the ﬁrst pilot study of brain EIT in acute stroke at the National Hospital for Neurology and Neurosurgery and UCLH, London.

B-806

11:15

Impact of quality management for emergency teleneuroradiology-process management for the outcome of neurology patients: Results after six years experience P. Soegner, F. Koppelstaetter, B. Oefner, D. zur Nedden; Innsbruck/AT ([email protected]) Purpose: Does quality-management-controlled teleradiology processes have an impact in the outcome of ischemic patients? After the experience of six years with the ISO 9001:2001 certiﬁed teleradiology process we want to evaluate the difference between the beginning and six years afterwards. Methods and Materials: Used teleradiology process was ISO-9001:2001-certiﬁed since December 2001. Forty-eight radiology technologists and 34 radiologists from seven public hospitals were instructed on examination-standardization of different CT-scanners. Perpetual quality improvements were necessary during whole timeperiod due to changes of technical and human resources. Most important clinical process indicator is the overall time-period between onset and qualiﬁed treatment. We re-evaluated the process efﬁciency periodically (time-period between clinical request and delivery of legal ﬁnal report). Results: From 964 emergency teleneuroradiology-CCT 483 patients had positive radiology ﬁndings. 287 patients with only ischemia; 168 only bleeding; 18 intracerebral-tumors, 10 mixed ﬁndings. Forty-seven patients were transferred to the neurosurgery department of the MUI (Medical University of Innsbruck); 163 ischemic patients were treated with lyses therapy after exclusion of bleedings. The aim to serve the referring physicians in rural hospitals as quick as possible was increased from 84% (2001) to 94.6%. Decrease of technical errors from 24/a to 6/a; decrease of human errors from 28/a to 22/a, decrease of average time gap from 64 min/2001 to 44 min/2007. Conclusion: The time gap between onset of ischemic symptoms and start of the lyses therapy is only 180 minutes. Therefore the process quality of teleradiology has to be as good and fast as possible. Neuroradiology qualiﬁed diagnosis for lyses therapy could more often be offered by rural hospitals in cooperation with experts from the Tyrolean stroke unit.

B-807

11:24

Increased incidence of hypophosphatemia associated with serial triple-dose gadolinium administration among 75 participants in the BECOME MS trial L.J. Wolansky, T. Phatak, B. Peng, J. Kaplan, A.R. Jedynak, P. Baladandapani, M. Haghighi, J.H. Skurnick, S.D. Cook, D. Cadavid; Newark, NJ/US ([email protected]) Purpose: The “BECOME” study compares the efﬁcacy of Interferon beta-1b (Betaseron®) with glatiramer acetate (Copaxone®) in 75 patients with the relapsingremitting or clinically isolated forms of multiple sclerosis (MS) using 3 Tesla MRI with triple-dose gadolinium (3-dose Gd). As per protocol, 3-dose Gd (0.3 mmol/kg) Gadopentetate Dimeglumine (Magnevist®) was used monthly on up to 24 occasions per subject. Our purpose is to report on increased incidence of hypophosphatemia associated with increased cumulative triple-dose gadolinium. Methods and Materials: All subjects in the BECOME study underwent blood and urine examination monthly prior to each MRI. The incidence of abnormal laboratory values while undergoing monthly 3-dose Gd was compared to pre-3-dose Gd values. To control for biochemical abnormalities caused by Betaseron and Copaxone, the results were stratiﬁed by treatment group. Results: Analysis of 870 instances of laboratory testing revealed that the incidence of mild and moderate hypophosphatemia was increased to 15% of tests in patients

328

A

B

C

D

E

F

G

undergoing triple-dosing compared with 3% in baseline pre-study test results (p 0.001). This could not be ascribed to Betaseron or Copaxone. 12 patients had moderate (1-2 mg/dL) hypophosphatemia on at least one visit while four had more than one episode. Furthermore, in 11 of these patients with values through nine months, there was an increased incidence of mild or moderate hypophosphatemia with cumulative 3-doses over time (p=0.01). Conclusion: Serial administration of triple-dose gadolinium may be associated with an increased incidence of mild or moderate hypophosphatemia, which appears to be related to cumulative dose.

B-808

11:33

Postmortem magnetic resonance imaging in legal cases: Where can it help? E.H. Whitby, M. Paley, M. Cohen; Shefﬁeld/UK (e.whitby@shefﬁeld.ac.uk) Purpose: Post mortem magnetic resonance imaging (PMMR) has been proposed as a replacement for autopsy. Falx haemorrhage has been reported in cases of hypoxia. We have assessed the role of the PMMR in the coronial autopsy and have evaluated the presence or absence of falx haemorrhages in all cases. Methods and Materials: Thirty cases that were either Sudden Unexpected Death in Infancy (SUDI) or a perinatal death of unknown cause. PMMR was performed prior to the autopsy. The images and autopsy reports were reviewed retrospectively to assess the impact of each. The falx was reviewed on images and microscopically for the presence of heamorrhage. Results: Age range of one day to four years. Causes of death at PM were: 12 SUDI; ﬁve hypoxic ischaemic encephalopathies (HIE); ﬁve cardiopulmonary failures, four infectious diseases; one road trafﬁc accident; two birth trauma, one possible infanticide. Complete agreement in 21 (70%) cases; both PMMR and autopsy provided supplementary information in one (3.33%) case; PMMR provided additional information in two (6.66%) cases. Autopsy was superior in six (20%) cases. Histology was taken from the falx in 23/30 cases. These showed diffuse haemorrhage with evidence of hypoxia in 9/23 (four had small subdural haemorrhages, one HIE and one pulmonary hypertension) on both the autopsy and PMMR. Haemorrhage without hypoxia was seen in 9/23 cases (eight on histology only). No haemorrhage was seen in 5/23 cases. Conclusion: PMMR investigation is an important adjunct investigation to the conventional autopsy and provides a permanent comprehensive digital record for future reference.

B-810

11:42

Diffusion tensor imaging in meningoencephalitis: Anisotropy as an indicator of meningeal inﬂammation and its differential diagnosis R.H.N. Hanasoge, R.V.S.V. Vadapalli, U.D. Patil, U. Murgod, R. Gowda, S.K.V. Bommana; Bangalore/IN ([email protected]) Purpose: To determine if there is signiﬁcant anisotropy associated with meningeal inﬂammation and to analyze the FA (fractional anisotropy) values in different causes of meningoencepahlitic processes. Methods and Materials: Total of 18 patients (mean age 35 years) with clinically suspected and/or conﬁrmed meningoencephalitis were studied on 1.5 T GE HDX Twin scanner. Conventional MRI without and with contrast (T1, T2, FLAIR, DWI, 3D SPGR (BRAVO), MRS) and diffusion tensor imaging with 25 directions and b value of 1000 were performed with post processing on Functool platform. ROI (region of interest) was placed on the abnormal and normal appearing areas on FLAIR and post contrast images and FA values were quantiﬁed. Bacterial (n=5), viral (n=4), tuberculous (n=3), fungal (n=2), carcinomatous (n=1), chemical (n=1) and inconclusive (n=1) meningitis cases were analyzed. Nine age and sex matched controls were also scanned and mean FA values were quantiﬁed. Results: Increased anisotropy was noted in patient group compared to control group. Highest FA values were seen in fungal, tuberculous and some bacterial cases. The FA values were not signiﬁcant in viral, chemical and carcinomatous meningitis. Parenchymal involvement and presence of cerebritis or abscess showed deﬁnitive anisotropy differences. Conclusion: Anisotropy weighted imaging with quantiﬁcation of FA values could aid in identifying meningeal inﬂammation noninvasively and may indicate the possible cause.

Scientific Sessions 10:30 - 12:00

Room R1

Chest

SS 1804 Technical advances/Pulmonary hypertension and perfusion Moderators: M. Bellomi; Milan/IT C. Mueller-Mang; Vienna/AT

B-811

10:30

Standard dose and dose-reduced mobile direct radiography versus computed radiography for bedside chest imaging: Evaluation of image quality and reader agreement D.W. De Boo, N.J. Freling, E.E. Deurloo, H.W. Venema, D.A. Dongelmans, C. Schaefer-Prokop; Amsterdam/NL ([email protected]) Purpose: To assess image quality and potential for dose reduction of a mobile detector unit (DR) as compared to computer radiography (CR) for ICU chest radiographs. Methods and Materials: Three groups of age-, weight- and disease-matched ICU patients (n=50 each) underwent clinically indicated bedside chest radiography obtained either with CR (single read-out powder plates) or mobile DR (GOS-TFT detectors) at identical or 50% reduced dose (DR50%). Delineation of anatomic structures and monitoring materials, overall image quality and pathology were scored on a 3-point scale (3=best) by four readers of varying experience. In 12 patients, pairs of follow-up CR and DR images and in 15 patients, pairs of CR and DR50% images were available. In these pairs overall image quality was compared side-to-side. Results: Delineation of anatomy in the mediastinum (trachea, carina, retro-cardiac vessels and spine) was scored better with DR (2.1, p 0.05) or DR50% (2.0) than with CR (1.8). Monitoring materials were seen best with DR over DR50% and CR (2.5 vs 2.3 vs 2.3). In the side-to-side comparison of follow-up images in the same patient, overall image quality of DR and DR50% was rated better than that of CR in 96% (46/48) and 87% (52/60), respectively. Inter-observer agreement for assessment of pathology was generally low (K 0.4). Conclusion: Mobile DR units offer better image quality than CR for bedside chest radiography and allow for 50% dose reduction over CR without loss of relevant image quality. Inter-observer agreement is low and not improved by better image quality.

B-812

10:39

Radiation dose during nongated MDCT angiography of the chest: Comparison between single- and dual-source CT technology C. Santos Name, M. Rémy-Jardin, P. Dupuis, V. Deken, J.-B. Faivre, J. Rémy; Lille/FR ([email protected])

10:48

Prediction of perceptible artifacts in JPEG2000 compressed chest CT images using mathematical and perceptual quality metrics B. Kim1, K. Lee1, K. Kim1, R. Mantiuk2, S. Hahn3, T. Kim1, Y. Kim1; 1Seongnam-si, Gyeonggi-do/KR, 2Saarbrücken/DE, 3Seoul/KR ([email protected]) Purpose: To determine if peak signal-to-noise ratio (PSNR) and a perceptual quality metric (High-Dynamic Range Visual Difference Predictor, HDR-VDP) can predict the presence of perceptible artifacts in JPEG2000 compressed chest CT images. Methods and Materials: One-hundred chest CT images were compressed to 5:1, 8:1, 10:1, and 15:1. Five radiologists determined if the original and compressed images were identical (negative response) or different (positive response). Correlation between each metric result and the number of positively-responding readers was evaluated using the Spearman rank correlation test. Using the pooled readers’ responses as the reference standard, receiver-operating-characteristic (ROC) analysis was performed to determine the cutoff values balancing sensitivity and speciﬁcity and yielding 100% sensitivity in each metric. Using these cutoff values, visually lossless thresholds (VLTs) for the compression were estimated for the 100 original images and the estimation accuracy was compared (McNemar test) between the two metrics. Results: The correlation coefﬁcients were -0.918 and 0.925 for PSNR and HDRVDP, respectively. The areas under the ROC curves for the two metrics were 0.983 and 0.984, respectively (p=0.11). PSNR and HDR-VDP predicted the VLT accurately for 69 and 72% of the 100 images, respectively (p=0.68), at the cutoff values balancing sensitivity and speciﬁcity; for 43 and 47%, respectively (p=0.22), at the cutoff values reaching 100% sensitivity. Conclusion: Both metrics are promising in predicting the perceptible compression artifacts, and therefore can be potentially used to estimate the VLT.

B-814

10:57

The reliability of CT signs in mild and severe pulmonary hypertension A. Devaraj, A.U. Wells, M. Meister, T. Corte, J. Wort, D.M. Hansell; London/UK ([email protected]) Purpose: To compare the reliability of pulmonary artery signs on multi-detector CT (MDCT) in mild and severe pulmonary hypertension (PH). Methods and Materials: Eighty-two patients who had undergone thoracic MDCT and right heart catheterisation were included in the study (41 males, mean age 56 years; 33 with pulmonary ﬁbrosis, 29 with pulmonary vascular diseases and 20 patients with a variety of other lung diseases). The population was divided into two groups: 1) normal or mild PH patients with mPAPb35 mmHg (n=45); and 2) severe PH patients with mPAP 35 mmHg (n=37). Dimensions of the main pulmonary arteries, segmental, subsegmental and sub-subsegmental pulmonary arteries were recorded at predetermined levels. Measurements were correlated with mean pulmonary artery pressure (mPAP) in the two groups. Results: In contrast to a strong correlation between main pulmonary artery diameter and mPAP in patients with severe PH (r=0.43, p 0.001), no such correlation existed in the normal/mild PH group (r=0.25, p=0.09). However, signiﬁcant correlations were demonstrated between mean segmental, subsegmental and sub-subsegmental artery size and mPAP in normal/mild PH patients (r=0.43, p=0.02; r=0.47, p 0.005, r=0.44, p 0.01, respectively). No correlations between mPAP and these smaller pulmonary arteries were found in severe PH. Conclusion: The diameter of the main PA on MDCT cannot be considered a reliable indicator of mPAP in patients with mild PH. However, the segmental, subsegmental and sub-subsegmental arteries may reﬂect raised mPAP better in this group.

B-815

11:06

The effects of diffuse pulmonary ﬁbrosis on the accuracy of CT signs of pulmonary hypertension A. Devaraj, A.U. Wells, M. Meister, J. Wort, T. Corte, D.M. Hansell; London/UK ([email protected]) Purpose: To evaluate the accuracy of pulmonary artery signs on multi-detector CT (MDCT) of pulmonary hypertension in patients with ﬁbrotic lung disease. Methods and Materials: Sixty-two patients (30 males, mean age 56.4 years) who had undergone both thoracic MDCT and right heart catheterisation were evaluated. The population consisted of two groups: the ﬁrst group comprised 33 patients with ﬁbrotic lung disease (FLD), the second, a group of 29 patients with a variety of pulmonary vascular diseases (PVD) (including 16 with idiopathic pulmonary arterial hypertension). Dimensions of the main pulmonary arteries, segmental, subsegmental and sub-subsegmental arteries were measured at predetermined levels. Correlations were made between absolute measurements and mean pulmonary artery pressure (mPAP) in the two groups. Results: Compared to good correlations between main pulmonary artery diameter and mPAP in patients with PVD (r=0.51, p 0.005), there was no such correlation

A

B

C

D

E

F

G

329

Tuesday

Purpose: To compare the effect on image quality and radiation dose of individual adjustment of low kilovoltage for multidetector-row CT (MDCT) angiograms of the chest acquired with single- and dual-source CT technology. Methods and Materials: This study was based on the comparative analysis of two-paired populations of 44 patients of less than 80 kg body weight, evaluated with dual-source (Deﬁnition, Siemens; Group 1) and single-source (Sensation 64, Siemens; Group 2) CT with an individual adaptation of the kilovoltage according to the patient’s weight. Both groups were otherwise scanned with similar parameters (collimation: 64x0.6 mm; rotation time: 0.33 s; 4D automatic modulation of milliamperage), except for the pitch value of 1.5 in Group 1 and 1.2 in Group 2. Results: The mean DLP value was statistically signiﬁcantly lower in Group 1 compared to Group 2 (116.16p53.98 vs 160.5p77.8 mGy cm; p=0.0028) with similar subjective image quality and no signiﬁcant difference in the objective image quality assessed by measuring the noise level within the trachea (16.3p3.9 vs 16.4p4.7 HU; p=0.93) and descending aorta (27.4p6.3 vs 26.5p6.8 HU; p=0.58).The levels of diagnostic reference (i.e., DLP values of 75% of CT scans) in the studied population were 135 mGy cm in Group 1 and 204 mGy cm in Group 2. Conclusion: Dual-source CT allows more extensive use of low kilovoltages than single-source CT without impairment in image quality, thus reducing the overall radiation dose delivered during standard CT angiograms of the chest.

B-813

Scientific Sessions in patients with FLD (r=0.23, p=0.18). In FLD mean segmental artery size provided the best individual correlate with mPAP (r=0.61, p 0.001); this correlation was less strong in the PVD group (r=0.45, p=0.03). No relationship between the subsegmental and sub-subsegmental arteries and mPAP existed in either group. Conclusion: The widely used sign of main pulmonary artery dilatation is not a reliable marker of pulmonary hypertension in patients with FLD. The mean segmental artery size may provide a better indication of mPAP in patients with FLD.

B-816

11:15

Pulmonary hypertension: Correlation of CINE-MRI and phase-contrast MRI with six-minute-walk-test and parameters from right heart catheterization C. Fink1, H. Leuchte2, L. Bitto2, J. Behr2, M. Reiser2, S. Schoenberg1; 1Mannheim/DE, 2 Munich/DE (christian.ﬁ[email protected]) Purpose: To assess the correlation of CINE MRI and phase-contrast (PC) MRI with six-minute-walk-test (6MWT) and parameters from right heart catheterization (RHC) in pulmonary hypertension (PAH). Methods and Materials: 38 patients with PAH (idiopathic PAH, n=15, associated PAH, n=3; chronic thromboembolic hypertension (CTEPH), n=20) were examined with CINE-MRI and PC-MRI at 1.5 T. The following quantitative hemodynamic parameters were assessed: right and left ventricular ejection fraction (RVEF, LVEF), peak velocity, average velocity, and average ﬂow in the main pulmonary artery and aorta. These parameters were correlated to 6MWT and parameters from RHC (mean pulmonary arterial pressure (mPAP), pulmonary vascular resistance (PVR), right atrial pressure (RAP), and cardiac output (CO)). Results: MRI data did not correlate with 6MWT. We observed a moderate positive correlation of the LVEF with PVR (all: r=0.47, PAH: r=0.39, CTEPH: r=0. 51, p 0.05). The average blood ﬂow in the pulmonary artery showed a good inverse correlation with mPAP (all: r=-0.75, PAH: r=-0.76, CTEPH: r=-0. 66, p 0.05) and PVR (all: r=-0.83, PAH: r=-0.81, CTEPH: r=-0.75, p 0.05). There was also a moderate inverse correlation with RAP (all: r=-0.50, PAH: r=-0.39, CTEPH: r=-0.56, p 0.05), and a good correlation with CO (all: r=0.69, PAH: r=0.70, CTEPH: r=0.66, p 0.05). Moreover, we observed a moderate correlation of the average blood ﬂow in the aorta with mPAP, PVR, RAP and CO. Conclusion: Average blood ﬂow derived from PC-MRI showed the best correlation with conventional RHC measurements. No meaningful correlation can be found between CINE-MRI and RHC. No functional MRI measurement correlates with the 6MWT.

B-817

11:24

Quantitatively assessed 3D dynamic perfusion MRI of the lung for disease severity evaluation in patients with collagen vascular diseases Y. Ohno, H. Koyama, S. Matsumoto, D. Takenaka, Y. Onishi, M. Nogami, K. Sugimura; Kobe/JP ([email protected]) Purpose: To evaluate the capability of quantitatively assessed 3D dynamic perfusion MRI for assessment of disease severity and progression to pulmonary hypertension (PAH) in connective vascular disease (CVD) patients. Methods and Materials: Eighteen gender- and age-matched CVD patients without and with PAH and nine healthy volunteers underwent 3D dynamic perfusion MRI (TR 2.7 ms/TE 0.6 ms/Flip angle 40°, 100-240 mm slab thickness, 10-12 partitions), Doppler echocardiography, and pulmonary function test. All CVD patients with PAH also underwent right heart catheterization. Disease severity of CVD patients was assessed in terms of diffusing capacity for carbon monoxide (%DLco) and estimated systolic pulmonary arterial pressure (sPAP), and progression of PAH in terms of pulmonary arterial pressure (mPAP) and pulmonary vascular resistance (PVR). Quantitatively assessed pulmonary perfusion parameter maps were generated, and mean pulmonary blood ﬂow (mPBF), pulmonary blood volume (mPBV) and mean transit time (mMTT) were determined as averages of all ROI measurements. To determine correlations between perfusion parameters and disease severity in CVD patients, mPBF, mPBV and mMTT were statistically compared with sPAP and %DLCO. To determine capability of progression to PAH for CVD patients with PAH, mPBF, mPBV and mMTT were statistically correlated with mPAP and PVR. Results: All pulmonary perfusion parameters correlated signiﬁcantly with sPAP and %DLCO (p 0.05). mPBF and mPBV correlated signiﬁcantly with mPAP and PVR (p 0.05). Conclusion: Quantitatively assessed 3D dynamic perfusion MRI can be used for assessment of disease severity and progression of PAH in CVD patients.

B-818

11:33

Treatment monitoring in patients with pulmonary arterial hypertension (PAH) using quantitative 3D MR pulmonary perfusion S. Ley, E. Gruenig, F. Risse, N. Ehlken, D. Mereles, H.-U. Kauczor; Heidelberg/DE ([email protected]) Purpose: Pulmonary arterial hypertension (PAH) is mainly a small vessel disease with reduced perfusion and exercise capability. To monitor the exercise improvement under therapy, exercise tests like the 6-min walk distance (6MWD) are used. A new possibility to directly and quantitatively assess the effect of therapy on the small vessels is 3D-MR-perfusion. Therefore, the goal of our study was to compare the assessment of treatment response in patients with PAH using MR-perfusion with the 6MWD. Methods and Materials: Twenty patients with PAH were included in this study of whom 10 received no therapy (control), and 10 received therapy. Three weeks after the initial examination, patients were re-examined using 6MWD and 3D-MRperfusion (1.5 T, FLASH3D, TREAT, TE/TR 0.8/1.9 ms; FA 40º; 3.5x1.9x4 mm3, each TA 1.5 s, inspiratory breath-hold; 0.1 mmol/kg Gd-DTPA). Evaluation of 3D-perfusion was done with an in-house written software using a singular-value-decomposition technique. Quantitative whole lung evaluation comprised pulmonary blood volume (PBV [ml/100 ml]) and mean transit time (MTT [s]). Results: The 6MWD was unchanged in the control-group (difference 17 m, p=0.3, sign-test) while it improved signiﬁcantly in the therapy-group (91 m, p=0.004). The PBV remained stable in the control-group (0.2 ml/100 ml, p=0.3) and changed signiﬁcantly in the therapy-group (2.1 ml/100 ml, p=0.02). The control-group showed a signiﬁcantly prolonged MTT (1.1 s, p=0.02) indicative for disease progression. The MTT remained stable in the therapy-group (0.4 s, p=0.3). Conclusion: While in the PAH control-group pulmonary perfusion was stable or even deteriorated, it improved in the therapy-group. Therefore, therapeutic effects in PAH patients seem to be related to the small pulmonary arterial vasculature.

B-819

11:42

Pulmonary parenchyma perfusion imaging with SSFSE-FAIR, compared with 3D DCE MRI L. Fan1, S.Y. Liu1, F. Sun2, X.S. Xiao1; 1Shanghai/CN, 2Beijing/CN ([email protected]) Purpose: To evaluate pulmonary parenchyma perfusion with FAIR and compare it with 3D DCE-MRI in healthy volunteers and in patients with pulmonary embolism or cancer. Methods and Materials: The optimized inversion time TI of FAIR was determined in twelve healthy volunteers. The other 10 healthy volunteers and 30 patients with pulmonary embolism or lung cancer underwent SSFSE-FAIR imaging followed by 3D DCE-MRI on a 1.5 T MR system. For all subjects, the homogeneity of FAIR and DCE-MRI perfusion was assessed. In case of perfusion abnormities, the contrast between normal lung and perfusion defects was quantiﬁed by calculating a normalized signal intensity ratio (SI normal-SI abnormal / SI normal). Paired Student’s t test was used for the statistical analysis. Results: High quality lung perfusion images were acquired with both FAIR with 1000 ms TI and DCE-MR. In volunteers, the signal intensity of FAIR perfusion images was homogeneous. Wedge-shaped perfusion defects were visualized in 10 pulmonary embolism and 12 central lung cancer cases inﬁltrating the pulmonary artery. There was no signiﬁcant statistical difference of signal intensity ratio between FAIR and DCE-MRI (p 0.05). In eight peripheral lung cancers, all lesions showed low perfusion against the higher perfused lung parenchyma in both FAIR and DCE-MRI. Conclusion: Pulmonary parenchyma perfusion imaging with FAIR and DCE-MRI was feasible and consistent, which has high diagnosis accuracy for the detection of perfusion abnormalities caused by pulmonary embolism or lung cancer. FAIR as a noninvasive technique could obtain the same diagnostic information when compared with DCE-MRI.

B-820

11:51

Characterization of regional alterations in pulmonary perfusion via MDCT in nonsmokers and smokers S.K. Alford, E.J.R. van Beek, M. Hudson, H. Baumhauer, G. McLennan, E.A. Hoffman; Iowa City, IA/US ([email protected]) Purpose: To determine normative values of perfusion parameters obtained from multidetector-row computed tomography (MDCT) imaging and to assess mean transit time (MTT) and pulmonary blood ﬂow (PBF) as early markers for pathological changes associated with inﬂammatory lung diseases. Methods and Materials: ECG-gated dynamic axial MDCT (Sensation 64, Siemens)

330

A

B

C

D

E

F

G

Scientific Sessions with a central bolus injection of Iohexol were acquired in 34 subjects (13 nonsmokers, 12 normal smokers, and nine smokers with CT-only ﬁndings of emphysema (SM-CTemph)). Quantitative image analysis determined MTT and PBF normalized to regional water content. Mean and coefﬁcient of variation (CV) were calculated globally and for left, right, non-dependent (ND) and dependent (D) regions. Results: Perfusion parameters were determined for nonsmokers (MTT: 4.66p1.05 sec, CV 0.45; PBF: 9.16p1.74 ml/min/g, CV: 0.40), normal smokers (MTT: 4.54p0.73 sec, CV: 0.46; PBF: 9.16p2.70 ml/min/g, CV: 0.38) and SM-CTemph (MTT: 4.28p0.74 sec, CV: 0.58; PBF: 7.45p1.55 ml/min/g, CV: 0.45). MTT and PBF measurements for left and right lung demonstrate no signiﬁcant differences. When comparing ND to D regions, the CV for MTT signiﬁcantly differed (nonsmokers: $CV: +0.13, p 0.003; normal smokers: $CV: +0.17, p 0.002; SM-CTemph: $CV: +0.19, p 0.008) indicating that, while CV is increased globally in SM-CTemph (p 0.004), there is a greater increase in CV in ND regions. Conclusion: MDCT-based perfusion measurements demonstrate signiﬁcant differences in heterogeneity of perfusion parameters in subjects with imaging-only based evidence of lung pathology and show potential for detection of early inﬂammatory changes in the lung.

10:30 - 12:00

Room R2

Cardiac

SS 1803b Valvular and cardiac function studies Moderators: N.L. Kelekis; Athens/GR G.K. Schneider; Homburg a.d. Saar/DE

B-821

10:30

Comparison of non-invasive MRI approaches to invasive hemodynamic measurements in the assessment of valvular aortic stenosis M. Weininger, C.O. Ritter, B. Lengenfelder, F. Sagmeister, W. Voelker, H. Koestler, D. Hahn, M.J. Beer; Wuerzburg/DE ([email protected]) Purpose: To investigate a non-invasive assessment of the aortic oriﬁce area (AVA) in patients with valvular aortic stenosis using functional MRI (AVAf) and MRI-planimetry (AVAp) in comparison to invasive hemodynamic measurements (AVAg). Methods and Materials: 21 patients (12 male, 9 female, mean age 65p12) with aortic stenosis (AHA classiﬁcation; mild: n=1; moderate: n=6; severe: n=14) were examined using a 1.5 T MRI scanner (Magnetom Symphony Quantum, Siemens, Germany) and a standardized scanning protocol (SSFP-cine-MRI, ﬂow quantiﬁcation and PSIR-late-enhancement imaging). Strictly based on clinical indications all patients underwent conventional cardiac catheterization prior to MRI. During this invasive procedure AVA was determined according to the Gorlin formula (AVAg). AVAf was calculated using the continuity equation based on measurements of LVOT diameter, ﬂow in LVOT, trans-stenotic ﬂow and diameter of the ascending aorta. Results: Mean values were as follows: AVAg 0.95p0.36 cm²; AVAf 0.92p0.46 cm²; AVAp 1.16p0.28 cm². Comparing AVAg and AVAf a signiﬁcant correlation was found (r=0.73; p 0.05). Functional MRI could reliably detect all patients with oriﬁce areas 1.0 cm², in whom valve replacement therapy was indicated (14/14, p 0.0004). In contrast correlation between AVAg and AVAp was poor (r=0.40; p 0.05). Furthermore, AVAp could not reliably differentiate severe aortic stenosis (AVA 1.0 cm²; 6/14, p=0.11). Conclusion: In patients with aortic stenosis high correlation was found between functional MRI and invasive measurements. Our data indicate that functional MRI is superior to MRI-planimetry for quantiﬁcation of aortic stenosis. MRI-planimetry might be impaired due to its limitation to clearly delineate the margins of the deformed cusps and the complex 3-D structure of the valve apparatus.

B-822

10:39

Purpose: To assess the diagnostic value of computed tomography (CT) for detection of valvular lesions in patients with infective endocarditis (IE) in comparison with transesophageal echocardiography (TEE) and intraoperative ﬁndings (OP). Methods and Materials: 26 patients with infective endocarditis (51 valves: 45

B-823

10:48

Preoperative evaluation of patients undergoing valvular replacement for aortic stenosis using 64-slice CT angiography I. Carbone, P. Sedati, H. Grazhdani, F. Ciolina, N. Galea, D. Cannata, C. Catalano, R. Passariello; Rome/IT ([email protected]) Purpose: Coronary artery disease (CAD) and aortic valve stenosis (AVS) share many risk factors. CAD should be ruled out in candidates for AVS surgery. We retrospectively assessed the role of ECG-gated 64-slice CT angiography (64-SCTA) in patients with AVS before surgery. Methods and Materials: Twenty-ﬁve patients with AVS underwent ECG-gated 64-SCTA (collimation=0.6 mm, 800 mAs and 120 Kvp, 110 mL Iomeprol-400+30 mL saline) of thoracic aorta, including the heart, before surgical valve replacement. Images were evaluated by two readers in consensus and compared with surgical ﬁndings in terms of the degree of AV calciﬁcation, valvular morphology, diameter of the AV annulus and sino-tubular junction, and valvular area planimetry. Quantitative evaluation of cuspid opening was also performed. Finally, the presence of CAD, thoracic aortic aneurysm and left ventricle hypertrophy were assessed. Results: Visualization of the AV without motion artifacts was possible in 22/25 patients. Valvular morphology was correctly assessed in 23/25 cases. The degree of AV calciﬁcation and the AV annulus and sinotubular junction diameters were correctly determined by 64-SCTA in all cases. The planimetric area of AV was assessed in 22/25 cases. CAD and AVS were correlated in 11/25 patients. Ascending aortic aneurysms requiring surgical replacement were detected in ﬁve cases. Signiﬁcant left ventricle hypertrophy was found in 13 patients. Conclusion: Preoperative evaluation of patients undergoing surgical replacement for AVS with 64-SCTA is feasible. The presence of CAD can be ruled out noninvasively giving to the surgeon important information regarding the status of the AV and thoracic aorta in a single exam.

B-824

10:57

Free-breathing multislice MRI for determination of cardiac functional parameters M.J. Beer, H. Stamm, A. Weng, C.O. Ritter, D. Hahn, H. Köstler; Würzburg/DE ([email protected]) Purpose: Realtime magnetic resonance imaging (MRI) techniques with parallel imaging are an alternative to assess cardiac function and morphology. Using a new free breathing technique we evaluated the accuracy of the method with regard to global left ventricular function. Methods and Materials: Twenty-two individuals (13 patients, nine volunteers) were examined on a 1.5-Tesla scanner. Realtime imaging was performed by a non-breath-hold TSENSE technique with a temporal resolution of 49 ms. Standard Cine MRI was a single-slice SSFP-technique. Ejection fraction (EF), end diastolic volume (EDV), end systolic volume (ESV) and left ventricular (LV) mass were evaluated for all data sets. Differences between the methods (mean, 95% conﬁdence interval, CI) were calculated. Results: In parallel imaging no artifacts occurred. All patients with an increased global mass, EDV, ESV and a decreased EF were detected with the multislice approach and data acquisition time decreased to less than 1/10 compared to

A

B

C

D

E

F

G

331

Tuesday

Diagnostic performance of multislice computed tomography in infective endocarditis: Comparison with transesophageal echocardiography and intraoperative ﬁndings G. Feuchtner1, P. Stolzmann2, W. Dichtl1, T. Schertler2, J. Bonatti1, H. Scheffel2, S. Müller1, A. Plass2, T. Bartel1, H. Alkadhi2; 1Innsbruck/AT, 2Zurich/CH ([email protected])

native/6 prosthetic) were examined with 64-slice CT or dual-source CT (SiemensTM): 2x32x0.6 mm; 0.33 s;120 kV/600-900 mAs or 350 mAs; eff.sl. 1 mm/inc. 0.7/ B25f and retrospective ECG-gating: 10% steps during RR-interval (including 4Dvalvular-analysis); 80-100 ml iodixanol (Visipaque320TM); 5-6 ml/s; bolus-tracking (CT-coronary-angiography protocol); and with TEE before surgery. Results: The diagnostic accuracy of CT for detection of vegetations was: sensitivity 96% (23/24), speciﬁcity 96% (26/27), PPV 96% (23/24), NPV 96% (26/27) (pervalve-based-analysis) and sens. 88% (36/41), spec. 93% (27/29), PPV 95% (36/38), NPV 83% (27/32) (per-vegetation-based-analysis) vs OP; but lower (p 0.001) than the accuracy of TEE. 37/42 (88%) vegetations in 23/24 (96%) patients were correctly identiﬁed with CT vs OP; four missed lesions were small ( 4 mm). Vegetation size measurements by CT correlated (r=0.97; p 0.001) with TEE (7.7p5 mm). Mobility of vegetations was correctly diagnosed in 21/22 (96%) patients with CT vs TEE. All four leaﬂet perforations (b2 mm) were missed by CT. Diagnostic accuracy of CT (9/9; 100%) for the detection of perivalvular abscess/pseudoaneurysm was signiﬁcantly (p 0.001) higher than that of TEE (8/9; 89%). CT provided more detailed morphological information regarding perivalvular involvement (myocardial/ pericardial/anulus/coronary sinus) than TEE. Conclusion: CT shows excellent results in detecting valvular vegetations in infective endocarditis but was inferior than TEE due to missed small lesions. CT was superior than TEE for evaluation of peri-valvular abcess/pseudoaneurysm, which could provide useful pre-operative information.

Scientific Sessions single-slice cine. The results with the realtime TSENSE and cine sequence, respectively, were for EF [%] 58.5p8.2 vs 58.6p7.5 with a difference of -0.1 (mean) [-5.24, 5.06] (95% conﬁdence interval), for ESV [ml] 64.3p32.4 vs 64.1p37.5 (difference 0.18 [-14.78, 15.14]), for EDV [ml] 150.5p45.4 vs 149.5p52.0 (difference 0.92 [-19.22, 21.07]), and for LV mass [g] 123.2p37.4 vs 120.0p43.7 (difference 3.16 [-18.37, 24.70]). Conclusion: Accurate evaluation of global function is possible using accelerated free-breathing realtime MR; especially, measurement of EF is sufﬁcient for all clinical questions.

B-825

11:06

Accuracy of quantitative measurement of mitral valve regurgitation using 64-section multi-detector row CT: Comparison with real-time threedimensional echocardiography Y.-K. Guo, Z.-G. Yang, Y. Li, Y. Wu, E.-S. Ma, X.-C. Zhang; Chengdu/CN ([email protected]) Purpose: The aim of this study was to prospectively determine whether ECG-gated 64-section multi-detector row CT (64-MDCT) can quantitatively assess mitral valve regurgitation volume (RV) and fraction (RF) and to compare them with real-time three-dimensional echocardiography (RT3DE) results. Methods and Materials: The study population consisted of 21 consecutive patients (8 men, 13 women; mean age 28.6 years p10.3; range 19-41 years) with isolated mitral valve regurgitation who underwent RV and fraction measurement by 64-MDCT and RT3DE. The results were compared with those of MRI as the standard of reference. The difference between left and right ventricular stroke volumes on 64-MDCT was calculated as RV. Regurgitation parameters of all patients by RT3DE were measured ofﬂine by using the average rotation method. Linear regression analysis and a paired Student's t test were used to compare the RV and RF calculated by 64-MDCT and RT3DE. Agreement analysis between 64-MDCT and RT3DE was performed with Bland and Altman analysis. Results: RV at 64-MDCT correlated well with that of MRI (r=0.95, mean difference 4.3 ml), and mean value of RV had no statistically signiﬁcant difference between the two methods (p 0.05). RT3DE underestimated the RV compared to MRI (t=6.18, p 0.05), but the correlation coefﬁcient was acceptable (r=0.86, mean differential 8.9 ml). RF measured by 64-MDCT had better agreement with that measured by MRI (r=0.91, mean difference -0.7%) than that obtained by RT3DE (r=0.84, mean difference -3.6%). Conclusion: 64-section MDCT can accurately assess the regurgitation volume and fraction in patients with isolated mitral valve regurgitation.

B-826

11:15

Assessment of aortic stenosis after aortic valve replacement: Comparative evaluation of dual-source CT and echocardiography T. Saam, M. Oberhoffer, C. Rist, N. Minaifar, F. Vogt, B. Reichart, C.R. Becker, M.F. Reiser, K. Nikolaou; Munich/DE ([email protected]) Purpose: To prospectively evaluate whether planimetric measurements of aortic valve area (AVA) with dual-source computed tomography (DSCT) correlate with measurements obtained by echocardiography in a group of patients after aortic valve replacement. Methods and Materials: 20 patients underwent DSCT of the heart (Somatom Deﬁnition, Siemens Medical Solutions, Forchheim, Germany) without heart rate control. All patients had undergone aortic valve replacement (homografts, mean time after sugery: 8p3 years). Maximum AVA in systole planimetrically measured with CT was compared with calculated echocardiographic AVA values determined with the continuity equation, using transvalvular pressure gradients. To assess intraand inter-reader reproducibility, the DCST data was re-analyzed by two readers four weeks after the initial review. Results: All DSCT datasets were of diagnostic image quality concerning valve depiction. Mean AVA as measured by DSCT was 2.8p0.9 mm² and 1.9p0.5 mm² by echocardiography. The planimetric evaluation of the CT data as compared to results of echocardiography showed a signiﬁcant correlation of the results (Pearson´s R=0.77, p 0.001). Intra- and inter-reader reproducibility were good with intra-class correlation coefﬁcients of 0.86 and 0.81, respectively, p 0.001. A high-grade valve stenosis (opening area 1.0 cm²) could be correctly excluded by DSCT in all patients. Conclusion: First experience indicates that DSCT is able to assess aortic valve opening area with high image quality and good intra- and inter-reader reproducibility, in subjects after aortic valve replacement. In terms of diagnostic accuracy, signiﬁcant aortic valve stenosis can be correctly excluded by DSCT.

332

A

B

C

D

E

F

G

B-827

11:24

Tricuspid annular plane systolic excursion measurement using MRI: A fast and easy tool to estimate right ventricular systolic function A.A. Azarine1, A. Ferreira1, X. Lyon2, A. Redheuil1, P. Garrigoux1, E. Mousseaux1; 1 Paris/FR, 2Lausanne/CH ([email protected]) Purpose: To test the accuracy of the tricuspid annular plane systolic excursion (TAPSE) measure as a simple and fast way to estimate the RV function, using MRI. Methods and Materials: 100 patients with congestive heart failure (CHF, n=75) or right ventricular arrhythmogenic dysplasia (ARVD, n=25) diagnosed by ARVD Task Force criteria underwent 1.5 Tesla cardiac MRI. MRI protocol included the quantiﬁcation of RV ejection fraction using contiguous short-axis slices and the measurement of TAPSE in a 4-chamber view. Results: TAPSE and RV ejection fraction were lower in patients with ARVD compared to CHF subjects, 13p5 vs 16p5 mm (p=0.010) and 35p13 vs 44p11% (p 0.001), respectively. TAPSE correlated signiﬁcantly with RV ejection fraction in the entire cohort (R=0.68, p 0.001) and in the ARVD (R=0.77, p 0.001) and CHF subgroups (R=0.60, p 0.001). There was no signiﬁcant difference between the correlation coefﬁcients of the two subgroups (z statistic=1.34, p=NS). A TAPSE cut-off value of 14 mm identiﬁed patients with RV ejection fraction 35% with a sensitivity, speciﬁcity and negative predictive value of 79, 80 and 92%, respectively. Conclusion: TAPSE correlates signiﬁcantly with RV ejection fraction measured by MRI and may be used in selected patients as a simple quantitative surrogate of RV systolic function.

B-828

11:33

Computing the left atrium volume from three orthogonal dimensions measured on T1-weighted BB images is not an accurate method for measuring left atrial size M. Burgmans, I. Hof, N. Prakken, P. Loh, B. Velthuis; Utrecht/NL ([email protected]) Purpose: CE-MRA volume rendering represents a validated standard for left atrium volumetry (LAV), but is time-consuming. We examined whether the volume of the LA could be calculated from measurements of three orthogonal dimensions of the left atrium. Methods and Materials: We studied 166 cardiac MRI-scans, which were performed before and after pulmonary vein antrum isolation (PVAI) in 79 patients with drug refractory atrial ﬁbrillation. On a 1.5 T scanner T1 BB images in the axial and coronal plane and a 3D gadolinium enhanced coronal T1-FFE MR angiography were obtained. The longitudinal (L), transverse (T) and anteroposterior (AP) diameters of the LA were measured on the T1BB images and used to compute the LA volume using the ovoid volume formula (P/6*[AP*L*T]). This volume was compared to CE-MRA volume rendering. Results: There was a moderately strong linear correlation (r=0.6) between the two methods. Bland-Altman analysis showed poor agreement between the two methods. Adjusting for the bias of the Vol T1BB measurement (which could be caused by the ovoid formula not being fully applicable to the left atrium) still resulted in variance of Vol T1BB approximately 35% above and below the Vol CE-MRA values. Conclusion: Computing the volume of the LA from three orthogonal dimensions measured on T1-weighted BB images can give an estimate of actual left atrial size, but is not accurate enough.

B-829

11:42

Is it possible to predict the degree of pulmonary regurgitation after tetralogy of Fallot repair by the pulsation of the central pulmonary arteries? L. Uliel1, N. Merchant2, E. Di Segni1, D. Chang2, O. Goitein1, N.S. Paul2, E. Konen1; 1Ramat-Gan/IL, 2Toronto, ON/CA ([email protected]) Purpose: Pulmonary regurgitation (PR) is a problem of important prognostic signiﬁcance in patients after tetralogy of Fallot (TOF) repair. Cardiac MRI (CMRI) is the modality of choice for the quantiﬁcation of PR after TOF repair. It has been shown with CMRI that regurgitant ﬂow volume weakly correlates with cross sectional area of central pulmonary arteries (PAs). The aim of this study is to evaluate whether PAs pulsation (percent systolic-diastolic cross sectional area change) correlates with calculated PR at CMRI. Such a parameter might be of clinical value when performing ECG-gated cardiac CT in patients after TOF repair. Methods and Materials: Phase-contrast sequence was performed in 69 patients after TOF repair. Fast-cine/phase-contrast technique was used to quantify the regurgitation of the main, right and left PAs and the cross sectional areas of these vessels during systole and diastole. Pearson correlation was used for statistical analysis.

Scientific Sessions Results: The correlation between the pulsation and the amount of regurgitation in the PAs was weak. The highest values were R=0.4239 and P 0.001 for the right pulmonary artery (RPA). In all seven patients with RPA area during systole twice or more than that in diastole, the regurgitation was at least 34%. Conclusion: Degree of PAs pulsatility does not allow accurate prediction of PR in most patients after correction of TOF. Thus, the potential use of cardiac CT in the evaluation of PR is limited.

B-830

11:51

Integration of cardiac computed tomography into pulmonary vein isolation in patients with paroxysmal atrial ﬁbrillation T.F. Weber1, H. Klemm2, A. Koops2, H.-D. Nagel2, G. Adam2, P.G. Begemann2; 1 Heidelberg/DE, 2Hamburg/DE ([email protected]) Purpose: Detailed anatomic information of the left atrium is necessary for performing radiofrequency ablation of atrial ﬁbrillation. In this study the impact of a preinterventionally acquired cardiac computertomography (CT) on pulmonary vein isolation (PVI) was assessed. Methods and Materials: Examinations of 54 patients with paroxysmal atrial ﬁbrillation undergoing PVI were analysed. In 27 patients a supplementary cardiac CT was obtained before PVI (CT-group, 12 women, 15 men, aged 59.7p9.9): 16x1.5 mm collimation, 0.2 pitch, 120 kV tube voltage, 400 effective mAs. Fluoroscopy time, effective dose and quantity of radiofrequency (RF) pulses of the following PVI were compared to 27 patients undergoing stand-alone PVI (11 women, 16 men, aged 62.0p9.9). Mann-Whitney tests served for statistical comparison. Results: CT datasets were successfully integrated into the ablation procedure of each patient. The mean quantity of RF pulses was signiﬁcantly lower in the CTgroup (22p8.0 vs 27p11.8, p=0.030), and a signiﬁcant reduction of ﬂuoroscopy time was found (41.8p12.0 vs 51.2p16.0 min, p=0.005). Effective doses of the catheter ablation differed in an equivalent dimension but not signiﬁcantly (14.9p10.0 vs 20.0p16.0 mSv, p=0.203). The mean additive effective dose of the cardiac CT was 8.5p0.3 mSv. Conclusion: CT-guided ablation requires less ﬂuoroscopy time than stand-alone PVI. Due to the multifaceted dependency of ﬂuoroscopy doses a consistent reduction of these was not observed. As a supplementary CT constitutes an additive dose, a CT atriography providing sufﬁcient image quality while reducing X-ray exposure has to be designed. The diminishment of RF pulses implies an increase in effectiveness and safety of the catheter ablation.

10:30 - 12:00

Room R3

Interventional Radiology

SS 1809 Vascular interventions (2) Moderators: P. Bosnjakovic; Nis/RS I.P. Vulev; Bratislava/SK

B-831

10:30

Carotid stent angioplasty during ﬁlter protection: Analysis of microscopical ﬁndings and clinical follow-up P.E. Huppert, D. Claus, G. Wiest; Darmstadt/DE ([email protected])

B-832

10:39

Transfemoral endovascular treatment of proximal common carotid artery lesions: A single centre experience on 153 lesions T. Paukovits, J. Haász, A. Molnar, Z. Szeberin, B. Nemes, D. Varga, K. Hüttl, V. Berczi; Budapest/HU ([email protected]) Purpose: To assess primary success and safety of percutan transluminal angioplasty (PTA) and/or stenting of the origin/proximal common carotid artery lesions (pCCA). Methods and Materials: Between 1994 and 2006, treatments on 147 patients (153 stenoses) (71 female; 121 left) with signiﬁcant diameter stenosis ( 70% in symptomatic, n=46; 85% in asymptomatic, n=101 patients) of pCCA were retrospectively reviewed. All but one procedure were done through transfemoral approach. Stents were implanted in 108 (70.5%) cases. Embolic protection devices were used in 16 cases. Follow-up included neurological examination, carotid duplex scan and ofﬁce/telephone interview. Results: Primary technical success was 98.7% (151/153). There were no lethal complications during the procedures. Periprocedural neurological ( 48 hours) complications were 3/153 (2.0%) major strokes, 4/153 (2.6%) TIAs (including 1 contralateral TIA), 1/153 (0.7%) dizziness. Eight (5.2%) puncture site bleedings, 1/153 (0.7%) bradycardia and 1/153 (0.7%) asthma cardiale were encountered. Follow-up examination (104/147 patients, 70.7%) revealed one contralateral TIA 19 months later and one minor stroke 52 months later in 2 asymptomatic patients. The 30 days death/all-stroke rate is 2.8% (3/109 procedures). Average follow-up time was 24.9 months. Cumulative primary patency was 95.7p2.1% at 1-year, 80.2p7.1% at 4-years, and 68.7p12.2% at 84 months. Cumulative secondary patency rate was 100% at 1 year, 87.7p7.0% at 4 years and 76.2p11% at 84 months. Log-rank test showed no statistical difference (p=0.881) in primary cumulative patency between PTA alone (n=30) or PTA/stent (n=79). Conclusion: Transfemoral PTA with or without stenting of the pCCA is a successful treatment with an acceptable safety level.

B-833

10:48

Tailored carotid stenting in 200 consecutive procedures: Is patient based endovascular treatment the method of choice for high surgical risk patients with carotid stenosis? C. Oláh, I. Lázár, S. Barta, R. Lovász, A. Valikovics; Miskolc/HU ([email protected]) Purpose: To report our single centre experiences with carotid stent procedures tailored to the anatomy and pathophysiology of every single patient. Methods and Materials: Between February 2002 and June 2007, 200 carotid stent procedures were performed. 83 patients (41.5%) were symptomatic and 117 asymptomatic (58.5%) for the ipsilateral carotid stenosis. The average percent of stenosis was 81.9% with symptomatic patients and 82.4% in asymptomatics. 42 patients (21%) had contralateral carotid occlusion and 58 patients (29%) had signiﬁcant contralateral stenosis. Altogether 82% of the patients were not candidates for surgery. Eight types of carotid stent and ﬁve different ﬁlter types were used. Embolic protection was introduced in 43.4% according to the decision of the interventionist. Results: All procedures were technically successful. One patient (0.5%) died in 48 hours in myocardial infarction. One major stroke (0.5%) and two minor strokes (1%) occurred in 30 days. Overall stroke and death rate was 2%. One clinically silent stent occlusion and one restenosis occurred during the follow-up. Conclusion: Carotid stenting is safe and efﬁcient procedure in all patients with signiﬁcant carotid stenosis. Our results suggest that high surgical risk patients have to be treated with carotid stenting. The need for embolic protection can be reliably judged by the interventionist.

B-834

10:57

Open versus closed cell design in carotid artery stenting I.Q. Grunwald, K. Karp, M. Politi, C. Krick, A. Haaß, W. Reith; Homburg a.d. Saar/DE ([email protected]) Purpose: To compare open versus closed cell designs in carotid artery stents concerning DWI lesions and clinical results. Methods and Materials: We retrospectively reviewed a nonrandomized cohort of 194 patients (143 men, 51 women) who underwent carotid artery stenting in our department AND received DWI before and after intervention. Open, semi-open and closed cell designs were used. Patient characteristics such as symptomatic and asymptomatic stenosis, degree of stenosis and ﬂuoroscopy time were taken into account along with clinical results (TIA, PRIND, Stroke).

A

B

C

D

E

F

G

333

Tuesday

Purpose: To evaluate the efﬁcacy of ﬁlter protection during carotid stent angioplasty (CSA) using an open cell design stent system by analysis of microscopical ﬁlter ﬁndings in correlation with morphology of stenoses and clinical results. Methods and Materials: 125 patients with atherosclerotic de novo carotid artery stenoses were treated by CSA using 7-9 mm/20-40 mm (diameter/length) nitinol open cell stent systems (Precise, Cordis) and 4-6 mm (diameter) ﬁlter protection devices (Angioguard, Cordis). 78 of 125 patients (62%) had symptomatic stenoses. Embolic material within extracted ﬁlters was analysed by microscopy. Carotid stenoses were analysed by Doppler ultrasonography and angiography. Neurological follow-up ranged from three months to 2.8 years. Results: 101 of 125 ﬁlters (81%) contained embolic material. Size of particles ranged from 50 to 1,000 μm. 35% of particles were larger than 300 μm. Embolic volume was signiﬁcantly larger in excentric compared to concentric stenoses but there was no correlation to the grade of stenoses and to the presence of calciﬁcations. Rate of ipsilateral stroke (TIA) was 3.2% (4%) with two strokes during the procedure. All patients suffering from procedure-related side effects also showed embolic material within the ﬁlter; no patient with an empty ﬁlter had any side effect.

Conclusion: Embolism occurs during CSA in the majority of cases if open cell design stent systems are used. Filter protection is effective in most but not in all of these patients.

Scientific Sessions Results: We found a signiﬁcant correlation between the free cell area and the number of new lesions detected on DWI after stent placement (p=0.027). There was also a positive correlation between the free cell area of the stents used and the total surface area of new lesions that occurred (p=0.024). Less lesions were found when using an open cell design. Conclusion: This is the only study so far comparing free cell area with DWI lesions. At the moment there is no foundation for the assumption that a closed cell design is superior to an open cell design. There are a number of factors such as the stent tip and stent ﬂexibility that may inﬂuence the outcome of such studies.

B-835

11:06

Unprotected CAS: Is it still an option? I.Q. Grunwald, K. Karp, C. Krick, P. Papanagiotou, A. Haaß, M. Müller, H. Schieber, W. Reith; Homburg a.d. Saar/DE ([email protected]) Purpose: To determine the incidence of new ischemic lesions found on diffusionweighted MR imaging (DWI) in nonselected patients after unprotected carotid artery stent placement. Methods and Materials: We retrospectively reviewed a nonrandomized cohort of 317 patients (240 men, 77 women) presenting with carotid occlusive disease who underwent unprotected carotid artery stent placement between 2003 and 2006. Mean degree of stenosis was 86.94p9.72 Â. In 197 patients (62.1%) DWI was obtained before and 24 hours after stent placement. New lesions were evaluated according to size and location. Results: In 59/197 patients (29.9%), new ischemic lesions were found on DWI, in the vessel dependent area. In 23/197 patients (11.7%) new ischemic lesions were found in the vessel independent area. Combined stroke/death rate was 3.63%. Conclusion: In unprotected CAS new ischemic lesions will occur in up to 34%. A randomized study should now show if these lesions occur less frequently in protected carotid artery stent placement.

B-836

11:15

Endovascular treatment of internal carotid artery stenosis: Immediate and eight years follow-up M. Stefanini, R. Gandini, S. Fabiano, M. Di Primio, E. Gaspari, G. Simonetti; Rome/IT ([email protected]) Purpose: Carotid artery stenting is emerging as a reasonable alternative to surgical endarterectomy not only in high-risk patients. However, few data are available on the long-term clinical efﬁcacy of carotid artery stenting, using protection device, and on the incidence of restenosis. Methods and Materials: During the last eight years, we have treated, with the use of cerebral protection devices, 853 consecutive patients (946 lesions) with stenosis of the internal carotid artery. The mean age of the patients was 69p14 years with 120 (14.06%) 80 years old. Only 443 patients were symptomatic (51.93%). All the patients were submitted prior the procedure to a US Doppler, Angio-MR and/or Angio VCT examination, and a neurological examination was performed routinely before and after the procedure. Results: Techincal success was achieved in 849 lesions (99.53%) and the cerebral protection was successfully applied in 879 procedures (92.9%). The 30-day incidence of stroke and death was 0.95% (nine patients). Complications consisted of death (two patients, 0.21%), major stroke (four, 0.42%) and minor stroke-TIA in 17 cases (1.79%). The follow-up consisted of clinical, neurological and US Doppler examination. The US Doppler control showed critical restenosis ( 80%) in seven cases. Conclusion: In our experience carotid artery stenting seems to be a safe and efﬁcient treatment of internal carotid artery stenosis with results that could be compared with surgical alternative. At the 8-year follow-up carotid artery stenting appeared effective in stroke prevention with low incidence of restenosis.

B-837

11:24

The Cook Celect™ ﬁlter: The UK and global experience so far C.D. Watts1, R. Uberoi1, S. Lyon2, R.W. Günther3, G. Riojas4, G.M.L.G. Avendano4, M.E.B. Lipp4, G. Plant5, J. Patel6, M.A. De Gregorio7, M.J. Lee8; 1Oxford/UK, 2Melbourne/AU, 3Aachen/DE, 4Mexico City/MX, 5 Basingstoke/UK, 6Leeds/UK, 7Zaragoza/ES, 8Dublin/IE ([email protected]) Purpose: To assess a new optional ﬁlter - the Cook Celect™ ﬁlter - for retrievability, complications, and safety. We will present our data on behalf of the ten global sites involved in this trial. Methods and Materials: This is a prospective, non-randomised, non-comparative, global study of the Cook Celect™ ﬁlter. Ethical approval has been obtained.

334

A

B

C

D

E

F

G

Patients with accepted indications for IVC ﬁlter placement, whether temporary or permanent, consented to enter the trial which started in 2005. Venacavagrams are performed at the time of insertion and at removal to calculate ﬁlter tilt and assess complications. This is supplemented by an AXR. The level of difﬁculty in removing the ﬁlter is recorded. The trial is divided into subgroups A or B dependent on whether the ﬁlter is inserted temporarily or permanently. Each group undergoes regular follow-up for ﬁlter position by AXR, and evidence of complication by clinical examination and IVC Duplex. Two post ﬁlter removal examinations are also performed. Results: The global ﬁgures for those ﬁlters (43) placed temporarily demonstrate a 95% overall successful removal rate with 100% at 26 weeks, 93% of ﬁlters with less than 10° of tilt, and 81% of ﬁlters being removed with no or minimal difﬁculty. There have been no major ﬁlter associated complications. We have placed eight ﬁlters locally with successful retrievals of the ﬁve ﬁlters intended for removal. Conclusion: At this time the Celect™ ﬁlter, a modiﬁcation of the Cook Tulip™ ﬁlter, shows promising longer term retrievability than other optional ﬁlters currently available.

B-838

11:33

Symptomatic lower and upper extremity deep venous thrombosis management: Catheter directed thrombolysis with or without pharmacomechanical thrombectomy U.C. Turba, A. Sim, M.A. Meuse, B. Arslan, J.F. Angle, U. Bozlar, A.H. Matsumoto, K.D. Hagspiel, M.D. Dake; Charlottesville, VA/US ([email protected]) Purpose: Current standard of care for extreme deep venous thrombosis (DVT) is anticoagulation. However, when DVT becomes symptomatic such as extremity swelling or even threatened limb ischemia due to venous congestion, more invasive treatment methods may be needed. Our purpose is to report instutional experience in this patient population. Methods and Materials: Following institutional approval, we identiﬁed 57 patients in whom standard anticoagulation therapy failed for DVT and further more invasive treatments were needed due to clinical symptoms. Patient demographics including indications for the procedures, complications, successes and failures were evaluated. Results: Fifty-seven patients (pts); 18 pts upper extremity and 39 pts lower extremity sypmtomatic DVT included in our study. Upper extremity: [18 pts, 22 procedures]. 14/22 received tissue plasminogen activator (t-PA), 8/22 had urokinase. 8/22 had pharmaco-mechanical thrombectomy (PMT) with catheter directed thrombolysis (CDT), 14 had CDT without PMT. 18/22 procedures had percutaneous transluminal angioplasty (PTA). Lower extremity: (39 pts, 39 procedures). 37/39 received t-PA, 2/39 had urokinase, 31/39 had mechanical thrombectomy, 22/39 received CDT with PMT with power pulse spray, 22/39 received CDT with PTA and subsequently 11/39 stent deployment. Upper extremity: Technical success, complete thrombus removal was 12/22 (55%), partial thrombus removal was 10/22 (45%). Lower extremity: Technical success, complete thrombus removal was 25/39 (64%), partial thrombus removal was 14/39 (36%). Conclusion: Patient selection is critical for catheter directed thrombolysis. Acute thrombus responds better to thrombolysis with PMT. Pharmaco-mechanical thrombectomy seems to shorten the procedure time in combination with catheter directed thrombolysis versus catheter directed thrombolysis alone.

B-839

11:42

Outpatient treatment of arterial inﬂow stenoses of dysfunctional hemodialysis access ﬁstulas by retrograde venous access puncture and catheterization T.F. Le Maire, L.E.M. Duijm, K.J.G. Receveur, A.V. Tielbeek; Eindhoven/NL ([email protected]) Purpose: To determine the feasibility of outpatient endovascular treatment of inﬂow stenoses in arteriovenous ﬁstulas (AVFs) through retrograde venous access catheterization. Methods and Materials: We consecutively included 22 dysfunctional AVFs with arterial inﬂow stenoses suspected at non-invasive radiologic imaging. Following retrograde venous access puncture, an interventional radiologist aimed to cross the arteriovenous anastomosis and advance a catheter into the aortic arch. After depiction of the complete vascular access tree, angioplasty and/or stent placement was attempted of stenoses showing 50% luminal diameter reduction at DSA. Results: In one AVF, a catheter could not be advanced into the aortic arch. DSA depicted 28 inﬂow stenoses in the remaining 21 AVFs. Clinical improvement was obtained in 18 out of 19 patients with a technically successful intervention ( 30% residual stenosis after angioplasty or stent placement). Access ﬂow of twelve patients improved from 431p150 ml/min to 818p233 ml/min; four patients with

Scientific Sessions steal symptoms became symptom free. One non-maturing ﬁstula was salvaged by angioplasty; access cannulation problems were solved in another patient. In one case, steal symptoms persisted after brachial artery stent placement. Two patients, in whom stent placement was not thought desirable, showed a 30% residual stenosis after angioplasty. No complications were observed at DSA and endovascular intervention. At 1-year follow-up, arterial restenosis was documented in one patient. Conclusion: Outpatient retrograde venous access puncture and catheterization, as an alternative to a potentially more hazardous brachial artery or more invasive femoral artery approach, should be considered for endovascular treatment of inﬂow stenoses.

B-840

11:51

Transvenous renal biopsy of kidney transplants via transfemoral approach A. Schmid, H.-L. Voit, C. Hugo, U. Baum, W.A. Bautz, M. Uder; Erlangen/DE ([email protected]) Purpose: Histological evaluation of renal allograft is of great importance for adequate treatment decisions in patients with transplant failure. However, especially in the early transplant period, various conditions like bleeding disorders, ascites or hematomas in the iliac fossa may increase risk of standard transcutaneous biopsy. Transjugular renal biopsy of native kidneys is an established procedure in patients with elevated risk for bleeding complications or in obese patients. The transjugular route is unsuitable, however, in patients with kidney transplants. The aim of our study was to evaluate the feasibility of renal biopsy of kidney transplants via the femoral vein as an alternative transvenous approach. Methods and Materials: We performed seven transvenous kidney biopsies of renal transplants in six patients (age 17-68 years) with the transjugular Renal Access and Biopsy Set (RABS) by Cook®. The ipsilateral femoral vein was used as vascular approach. Three to six passes to the upper pole of the transplant were made in each biopsy session. Results: In all cases access to the transplant vein via the transfemoral route was possible. Histological diagnosis was established in all patients. In one patient biopsy had to be repeated due to insufﬁcient tissue in the ﬁrst sample. Biopsies contained ﬁve glomeruli on average. Renal venogram indicated extracapsular perforation in one patient, no signiﬁcant extravasation was noted on follow-up in any case. Conclusion: Transfemoral biopsy of renal allografts is a useful alternative approach in patients at elevated risk for complications after transcutaneous biopsy.

Tuesday

A

B

C

D

E

F

G

335

Scientific Sessions

336

A

B

C

D

E

F

G

Scientific Sessions

Recommend Documents