$32

Oral Presentations

4. Fast spin-echo MR assessment of patients with poor outcome following cervical surgery.

1. Last progress in the imaging of spinal cord diseases

K-A. ThuomaJ, Wang Wu 1, R. Hedlund 2, W. Leszniewski 2, L. Vavruch 2. Departments of Diagnostic Radiology and Spinal Surgerye University Hospital LinkOping, Sweden.

D. Bal~riaux (HOpital Erasme, Brussels, Belgium) 2. CT-diskography with manometry and MRI as predictors of the outcome of lumbar percutaneous automated nucleotomy.

R. Dullerud, T. Amundsen, H. Lie, N. G. Juel, B. Magnaes Departments of Radiology, Section of Neuroradiology, Neurology, Physical Medicine and Rehabilitation and Neurosurgery, Ullevaal University Hospital and Chrismed, Oslo, Norway. Study design and objective: This prospective observational study was performed to assess whether CT-diskography (CTD),diskomanometry (DMM) or the MR signal intensity of the disks were reliable predictors of the outcome of nucelotomy. Material and Methods: 91 patients, 44 females and 47 males aged 18-68 years (mean 37.4) treated at 99 disk levels were included in the study. All had plain CT, MRI, CT-D and DMM performed prior to automated percutaneous nucleotomy with the nucleotome R system. Results: 69 (76 %) of the patients responded well to the treatment within 3 months and were classified as successes. Due to recurrencies the success rate at 1 year was reduced to 65 %. Except for low-significant better clinical results in patients with a pain response similar or identical to the presenting complaint, no association was demonstrated between diskographic or MRI parameters and clinical results. Conclusions: The results of this study does not justify the routine use of diskography, as a painful procedure with the potential risk of diskitis, in the work-up for nucleotomy.

OBJECTIVE: To review the pre- and postoperative cervical fast spin-echo (FSE) MR examinations and medical records in a large series of patients following anterior discectomy and fusion (ADF) with the Cloward technique and to identify postoperative complications. MATERIAL AND METHODS: The material comprised 146 consecutive patients which were operated on with ADF with Cloward technique for myelopathy, radiculopathy or myeloradiculopathy. Forty patients with un-satisfactory results were included in the study. Inclusion criteria: pre- and postoperative MRI and cervical radiography, lack of clinical improvement within 12 months after surgery, and postoperative CT. t.5 T equipment. RESULTS: Nineteen patients had initial improvement followed by deterio- ration of the preoperative symptoms, while the remaining 21 patients were not improved or worsened. Preoperative/ postoperative disc herniations (48/27) and bony stenosis (14/16) above, below or at the fused level were the most common findings. In 45 % surgery failed to decompress the spinal canal. In only four patients no cause of remaining myelopathy and/or radiculopathy was found. CONCLUSION: FSE MR can demonstrate a large variety of pathological findings in the great majority of patients with poor clinical outcome after ADF. However, CT shows bone changes better than FSE in axial projection. 5. MR findings before and after multiple segment decompression in cervical myelopathy: correlation with early clinical results

3. One dimensional Fourier analysis applied to M-mode sonograms of the spinal cord ill MMC.

R. Schumache/, K.Brgthl 2, D. RichterJ, 1Rb.-Abt. Univ.-Kinderklinik und 2Klinik f Neuroradiologie des Univ.-Klinikums Mainz, Langenbeckstr.1, D-55101 Mainz OBJECTIVE: Movements of the spinal cord are induced by arterial pulsation and breathing. These movements are reduced in patients with tethered cord syndrome, and in MMC patients with reduced cord motility tethered cord syndrome is more likely to develop. This study was done to evaluate whether motion patterns of spinal cord could objectively be analyzed using Fast Fourier Transformation (FFT). DESIGN AND METHODS: 153 patient examinations (age range 1-14 y) were done using a 5mhz linear transducer (B- and Mmode). The video signal was digitized and stored in a personal computer. FFTwas applied to the M-mode area of the image using an own software application (written in C). RESULTS: FFT shows characteristic patterns of spectra for groups of normal, reduced and missing cord motility. The FFA of normal cord motility showed a frequency peak at the frequency of the pacing pulse (heart beat) and extra 4-5 harmonic maxima up to 11 Hz. The group of reduced cord motility showed only 1 to 2 harmonic maxima. While an immobile cord showed 0 to 1 maximum. The maximum speed of transverse motion of the neural structures in the lumbar region is up to 25 mm/s. CONCLUSION: Motion patterns of the cord can be analyzed using FFT and characteristic frequency spectra can be shown. This allows to objectively and individuallyclassify the motion pattern in patients with MMC. This will hopefully be more fitting to the individual patient and produce a better prognosis with respect to his risk of developing a secondary tethered cord syndrome.

Sabitzer R, 1 Prayer D~3 HOrmann M, 2 Aichholz M f Sunder-Plasmann M, t Schindler E. 31Department of Neurosurgery, 2Department of Radiology, 3Department of Neuroradiology, University of Vienna~Austria OBJECTIVE: To investigate the usefulness of cervical titanium cage implants in patients with cervical myelopathy. DESIGN AND METHODS: 19 patients/40-83 yrs, mean 64 yrs, suffering from multisegmental spondylotic cervical myelopathy between 0.5 and 12 months, underwent MRI before and after spinal surgery. At 0.5% 1T and 1.5T superconducting units T1 weighted spin echo (SE)- and T2 weighted fast SE sequences were performed in sagittal and axial orientations. For surgical procedure an anterior approach was chosen; patients received a titanium cage replacing i or 2 vertebral bodies. RESULTS: Preoperative MR scans revealed spinal cord indentations at 1 to 3 levels in 19 cases, T2 weighted sequences showed intramedullary signal hyperintensity in 14 cases. In postsurgical MR studies of 13 patients cord indentations had disappeared partially or completely, while intramedullary signal changes remained widely unchanged in all but 4. Spasticity improved in 16 cases immediately after surgery while other clinical signs (weakness, radicnlopathy) did not change markedly within the first 16 postoperative weeks. CONCLUSION: Our findings suggest that patients with multisegmental spinal myelopathy may derive immediate profit from surgical decompression, even in presence of clinical- and MRIsigns of myelopathy.

$33 6. MRI of the 3 cases of spinal cord abscess

M. Braun*, J-C Lacour**, R. Anxionnat* J-C Huot ~ H. Yettou ~ J-C. Marchal ~ L. Picarcl*. *Departments of Diagnostic and Interventional Neuroradiology*, Neurology** and Neurosurgery ~ University Hospital Nancy, France Purpose: Spinal cord abscesses are rare (73 cases in the literature). We present 3 cases illustrating the various clinical features. Patients and methods: Three patients (one female and two male) were studied 24 hours to 4 day after the onset of the clinical symptoms. The entire spinal cord was explored with T1 and T2 sequences. Results: We present one case of acute form of spinal cord abscess (less than 7 days of evolution) and two subacute one (evolution within 6 weeks). The lesions involved the cervical cord in 2 cases and the cone in one. M R axial and sagittal planes show in all cases a central enhancing lesion within the gray matter surrounded with an important white matter edema extending along 8 to 10 myelomeres. In one patient (acute form), thrombosis of the superficial veins were noted adjacent to the absess. Two patients were operated on because of evidence of central necrosis, the third one had only a septic myelitis. Clinical outcome was very poor (two patients remained unchanged, the other died of septicemia). Conclusion: The pre-operative diagnosis of spinal cord abscess is difficult. M R I shows the location of the lesion, its extension and the eventual local complications (venous thrombosis). In the subacute clinical form, pre-operative differential diagnosis with glial tumor may be difficult.

7. Very rapid lumbar MR myelography with a HASTE sequence

K. Tsuchiya, Y Mizutani, T. Seki, J. Hachiya, K. Kobayashi. Department of Radiology, Kyorin University, Mitaka, Tokyo, Japan O B J E C T I V E : To assess the utility of M R myelography using a H A S T E (half-Fourier single shot turbo spin-echo) sequence in the diagnosis of lumbar spinal stenosis. M A T E R I A L S A N D M E T H O D S : H A S T E M R myelograms as well as routine M R images were reviewed in 17 patients with lumbar disk herniation(s) or canal stenosis. M R studies were performed on a 1.5-T Siemens Vision system. H A S T E myelographic images were obtained in anteroposterior, both oblique and lateral projections using parameters as follows: T R / T E ; 10.9/87 msec, FOV = 220-350 mm, imaging matrix = 242 x 256 and slice thickness = 20 mm. Fat suppression was always used. Scanning time for one projection was 2.02 sec. RESULTS: In all patients, we could obtain excellent myelogramlike images in which the spinal cord and nerve roots were discriminated from the cerebrospinal fluid. Indentation of the dural sac and/or nerve sleeve was well depicted. In four patients in whom conventional myelography was performed, its findings concurred with those of H A S T E M R myelography. C O N C L U S I O N : H A S T E M R myelography in conjunction with routine M R imaging could replace conventional lumbar myelography and postmyelographic CT.

oblique coronal image sets were reformatted with a maximum intensity projection (MIP) protocol. This image resembles a contrast myelogram and gives an excellent depiction of the intradural nerve-root and of the nerve-root sleeve. This sequence was added to our standard M R imaging protocol, consisting of a sagittal T1weighted sequence, a sagittal long-TR spin-density and T2 weighted sequence and an axial Tl-weighted sequence. In a retrospective study of 43 patients we evaluated the additional value of this MR radiculogram beside the standard examination. When there remained doubt with regard to nerve-root compression in the standard examination, M R radiculography was used to achieve a define verdict. RESULTS: in 44 % of the patients who were studied with the standard examination there was a need for further assessment with regard to nerve-root compression. In 68 % of these patients M R radioculography made a define verdict possible both in showing normal nerve-roots and compression. CONCLUSION: M R radiculography is a useful addition for the imaging of the lower lumbar spine in patients with sciatica and will further decrease the need for contrast myelography.

9. 3D myelography in the evaluation of cervical nerve root avulsion

R. Gasparotti, M. Pavia, E Pittiani, R. Brembilla. Dept. of Radiology University of Brescia PURPOSE: to assess the diagnostic accuracy of 3D myelography (3D MRM) as a non invasive tool for the evaluation of patients with traumatic injuries of the brachial plexus. D E S I G N A N D M E T H ODS: 20 patients with clinical and electromyographical evidence of traumatic brachial plexopathy were examined by 3D MRM, conventional cervical myelography and CT myelography between 3 weeks and nine months after trauma. 3D M R M was performed on a 1.5 Tunit with 3D CISS (Constructive Interference Steady State) sequences and a total scanning time of 8 minutes. 3D M R M findings were compared with conventional myelography and CT myelography findings. Surgical findings were available for comparison in 12 patients. RESULTS: 3D M R M allowed the identification of the whole spectra of traumatic lesions: pseudomeningoceles, radicular avulsions without disruption of the thecal sac and dural scars. 3D M R M proved to be superior to conventional myelography and comparable to CT myelography in the demonstration of traumatic pseudomeningoceles. 3D MRM brought to a slight overestimation of nerve root avulsion in comparison with either myelography or CT myelography and to underestimation of the presence and extension of the dural scars in comparison with conventional myelography. Imaging findings correlated well with surgical findings in the 12 operated patients. CONCLUSION: 3D M R M is able to detect noninvasively the vast majority of traumatic lesions that involve the proximal end of the brachial plexus in a single rapid examination. On the basis of our findings 3D M R M can be proposed as a screening examination for patients with traumatic injuries of the brachial plexus.

10. 3D-MR myelography (3D-MRM) in cases of severe spinal stenosis: Comparison with X-ray myelography

P A.M. Hofman and J.T. Wilmink. Department of Diagnostic Radiology, University Hospital Maastricht

K E W Eberhardt i, HP. Hollenbach 2, WJ Huk 1. tDiv. of Neuroradiology, Dept. of Neurosurgery, University of Erlangen-N~irnberg, Erlangen, Germany, 2Siemens Medical Engineering Group, Erlangen, Germany

OBJECTIVE: In order to increase the sensitivy of M R imaging with regard to nerve-root compression in patients with sciatica, we evaluated the additional value of M R radiculography. D E S I G N A N D M E T H O D S : A heavily T2 weighted FSE sequences was employed to produce a myelographic image of the dural sac (0.5 T system, T R / T E 6000/450, echo train 46 slice thickness 5 mm, with 2.5 mm overlap, field of view 15 cm, matrix 166X256, scan time 6:30 min.). The single slices of the left and right

Purpose: In the present study 3D-MR Myelography of the spine was compared with X-ray Myelography in cases of severe spinal canal stenoses and spondylolistheses. Material and Methods: In 50 patients with severe spinal stenoses (35 patients with a complete contrast block in X-ray Myelography) we compared in a prospective case study the sensitivity of the 3DM R M with X-ray Myelography. The results were compared with the intraoperative findings. The patients were examined with a 1.0

8. Optimizing the image of the intradural nerve root: the value of MR radiculography.

$34 T whole body MR-system (Siemens Magnetom Impact). For the examination of the lumbar spine a strongly T2*-weighted 3D-FISP sequence (TR = 73ms, TE = 21ms, c~= 7 deg) with fat saturation was applied in the sagittal orientation. For the thoracic and the cervical spine the same sequence (TR = 73ms, TE = 21ms, a = 5 deg) without fat saturation was used. The 3D-data sets were evaluated using a Maximum Intesity Projection (MIP) program. Results: The diagnostic sensitivity to detect nerve root compression syndroms in cases of severe canal stenoses and spondylolistheses is higher for 3D-MRM than for X-ray Myelography. Conclusions: The M R technique has advantages in patients with severe canal stenoses as well as spondylolistheses over X-ray Myelography. In the future the 3D-MRM could be the method of choice for these patients.

pseudocystic lesions in the temporal, frontal and subcortical region which were not present at the early stage, and which corresponded to the advanced neurological deterioration. M R performed in four children in the exacerbated stage of the disease showed the diffuse confluent white matter abnormalities with almost identical signal to that of the ventricle in the temporal and subcortical region. MRS in one child showed spectra without normal cerebral metabolites. On the basis of long-term follow up it might be assumed that the changes in the white matter as demonstrated by US, CT, M R and MRS in children examined for the unclassified megalencephalic leukodystrophies, are age related. Consequently, these changes might be considered the distinctive features in relation to Alexander's disease.

11. Development and genetics: Determining the body plan (In-

D. Prayer-Wimberger, K. Hittmai5 C WOber-BingOl, C. WOber, E. Schindler. Dept. of Neuroradiology, University of Wien, Austria

14. Virchow-Robin spaces in childhood migraine

vited lecture) Thomas P. Naidich (Miami, Florida, USA) 12. Central nervous system malformations. Intrauterine diagnosis with RMI.

De Tenyi A., Azambu]a N., - RadiologicaI Dep., Cuadro J., Alonso J., Bozzolo A., - Gyn Obst., Dep., Peha J. L. - Pediatric Dep., Facultad de Medicina, Montevideo - Uruguay. OBJECTIVE: The purpose of the present study is to evaluate the use and advantages of RMI in the diagnosis of malformations of the CNS in the fetus, in comparison with other imaging methods. M A T E R I A L S A N D M E T H O D S : RMI was performed in women during the last third of pregnancy. Previously, in all cases the diagnosis of a craniofacial or CNS malformation was made with ultrasound. A 1 Tesla Siemes Magnetom Impact superconductive magnet was employed, and at least three T1 weighted sequences in the axial, coronal and sagittai planes of the fetal head were performed. Some patients underwent also transvaginal ecography. In cases where interruption of pregnancy was decided, a CT study, using an Elscint Twin II scanner was performed. Images were obtained with the helicoidal mode in the axial plane, and were reformatted in the others. All imaging modalities were compared with each other and confronted with pathological findings. RESULTS: The preliminary results of a prospective study are presented with regard of the first 15 cases. CONCLUSIONS: RMI gives detailed and valuable information of CNS malformations in the fetus and it proved to be decisive in the management of pregnancy.

13. Neuroimaging distinctive features in unclassified megalencephalic leukodystrophies

N. Be~enski, V. Bo~njak, S. Cop, K. Or~olic, D. Mikulic. University Hospital Rebro, Dept. of Radiology, Zagreb, Croatia An attempt was made to demonstrate the possible distinction in different unclassified megalencephalic leukodystrophies by using US, CT, M R and MRS, since biochemical test in these entities are not specific. Seven children with progressive neurological disorders have been followed up clinically and by imaging form 2 to 12 years. The clinical course and the repeated US, CT and M R during the period of 8 and 12 years suggested infantile form of Alexander's disease in two children. The remaining five children have also been followed up fi'om 3 to 9 years. At the early stage of the disease (from 20 months to 3 years) when only mild neurological disturbances existed, CT showed hypodens lesions in the periventricular and subcortical white matter. In one child at the age of twenty months US showed granular hypoechoic appearance in the white matter which was consistent with the leukodystrophy. During the subsequent year repeated CT in this group of children showed along with the diffuse hypodense lesions in the white matter, the

OBJECTIVE: As a lack of pathological MRI-findings is common in childhood migraine, aim of this study was to evaluate accentuated visibility of the hemispherical perivascular CSF-space, the Virchow-Robin spaces (VR), as possible supportive diagnostic finding in childhood migraine. DESIGN AND METHODS: 58 children (3-14, mean 10.8 yrs.) suffering from different types of headache (duration 3 months to 5.2 yrs., mean 2.1 yrs), with normal clinical findings and normal E E G underwent MRI. Using a 1.5 tesla or 0.5 tesla superconducting system Tl-weighted (TR 700/TE 15 ms)T2 weighted (TR2200-2500/TE15.9 ms) and/or fast spin echo T2 weighted sequences were performed. VR-spaces were assessed to be accentuated, if appearing in clusters and/or featuring bright signal on spin density images. Age-matched controls consisted of 30 children without history of headache (3-14, mean 10.2 years). RESULTS: Clinically, patients were divided into a "migraine" group (31 patients: 19 with definite diagnosis, 12 with suspicious clinical diagnosis) and a "tension-headache" group including 27 children. Accentuated VR-spaces were diagnosed in 19/31 "migraine children" (61%), in 6/27 (22%) children with "tension headache", and in 8/30 (27 %) of the control group. CONCLUSION: In children accentuated VR-spaces, as visualized by MRI are significantly more common with migraine than with tension headache. Thus MRI may support the diagnosis of migraine even in advance of clinical confirmation. 15. Superior sagittal sinus size: a parameter for intracranial pressure changes in infants and children with hydrocephalus

I. Weitzner, JrJ, D. Leslie1, S. Rifkinson-Mann 2, Depts. of Radiology and Neurosurgerye, New York Medical College, Valhalla, New York, U S.A. OBJECTIVE: This study seeks to correlate changes in superior sagittal sinus (SSS) size with changes in intracranial pressure. Increased intracranial pressure causes decreased cerebral perfusion accounting for some of the clinical signs and symptoms. Decreased SSS size would be expected. Conversely, successful shunting should increase cerebral perfusion with a resultant increase in SSS size. D E S I G N A N D M E T H O D S : 21 consecutive patients who had undergone shunt procedures over a twelve month period were studied retrospectively. Their ages ranged from newborn to 14 years. 6 patients had multiple, 13 had single shunt procedures. Four variables were studied: 1-SSS size on axial CT scans were compared pre and post shunting. 2-Ventricular size changes were noted. 3Transcranial Doppler examinations were correlated when this was possible. 4-Clinical findings. RESULTS: An inverse relationship between SSS size and ventricular size existed in 18 of the 19 cases studied. (2 cases were eliminated). In 6 cases the ventricles were more responsive to

$35 pressure changes than the SSS. In 2 cases the SSS was more responsive than the ventricles. In 1 case there was discordance. 4 of 12 patients who had TCD examinations had positive correlation of the Resistivity Index (RI). As SSS size increased and ventricular size decreases, RI decreased. As SSS size decreased and ventricular size increased, the RI increased. In 8 cases where there were changes of ventricular and SSS size, the RI was normal. Finally, in every case where there were changes in SSS or ventricular size, there was positive correlation with clinical findings. CONCLUSIONS: 19 cases of obstructive hydrocephalus were studied pre and post shunting. Ventricular size and SSS size had a positive correlation in 18 cases. In 2 cases the SSS correlated better than ventricular size with the clinical status. Assessment of changes in SSS size on CT appears to be a useful parameter in some cases where decision-making may otherwise be difficult.

M R appearance of many intracranial tumors. Prediction of tumor malignancy is difficult on the bases of both imaging methods. Histopathologicaly our group consisted of 22 malignant ependymoma (E3), 18 E2, 3 E1 and 2 subependymoma. Five of the 29 infratentorial cases died due to postoperative complications, but no postoperative mortality occurred in the supratentorial cases. A t the time of this study nine patients died due to tumor progression, survival time was 3 to 84 months. CONCLUSION: CT and M R provide unreliable information in the assessment of the malignancy of intracranial ependynomas and in distinguishing it from other intracranial tumors. However, accurate localization of the tumor and the early detection of recurrences have prognostic value. Other factors as density, enhancement of the lesion or signal intensity seem to have no clinical value.

16. Dural Venous Thrombosis in Children T.A.G.M. Huisman z2, G.E Eich 1, E. Martin 2, U. E WilliI Departments of Diagnostic Radiology1 and Magnetic Resonance 2 University Children's Hospital Zurich, Switzerland

18. MRI follow-up after CNS prophylactic radiotherapy and chemotherapy for acute lymphoblastic leukemia in children C. Colosimo, G.M. Di Lella, M. Moschini, R. Riceardi*. Institute of Radiology and Department of Pediatric Oncology*. Catholic University. Rome, Italy

Introduction: Dural Venous Thrombosis (DVT) is a rare disorder in children. Clinical symptoms vary and depend on the extent of thrombosis. DVT is known to occur in mastoiditis, facial infections and tumors. Patients and Methods: From May 1994 to April 1995, we observed 10 children, 5 boys and girls (mean 4.6 years) with DVT. Four patients presented with mastoiditis, 3 with severe head-trauma, 2 with seizures and one with an intracranial tumor. All patients were examined by CT (Picker PQ-2000). Axial, 5 mm thick, contiguous slices were performed from the skull-base to the vertex. This was followed by contrast-enhanced 3 mm slices infratentorial and 5 mm slices supratentorial. Results: Contrast-enhancement together with high window/level setting demonstrated DVT in all patients. In 4 patients, mastoiditis lead to thrombosis of the adjacent dural sinus by continuity. In 3 patients epidural hematoma displaced the transverse sinus causing thrombosis by compression. In the 2 cases with seizures the pathogenesis of DVT is unclear. In one patient tumor compressed the superior sagittal sinus. Conclusions: Contrast-enhanced CT using thin sliced and high window/level setting is recommended for diagnosis of DVT. This technique is highly sensitive, easy to perform and readily available. It should be considered not only in infection and tumor but also in severe head-trauma. 17. Intracranial ependymomas in childhood: imaging manifestations and prognosis. J. Martos, K. Molndr, I. Dobronyi National Institute of Neurosurgery, Budapest, Hungary O B J E C T I V E : Our purpose in this study was to review the case histories of 45 children with intracranial ependymomas in order to establish more exact correlation between the imaging factors and the prognosis. D E S I G N A N D M E T H O D S : Between 1987 and 1994, 45 children (from 8 months to 13 years of age) were operated with intracranial ependymomas in our institute. Clinical data, including diagnostic approaches, operative findings, histology, postoperative treatment and follow up findings were available. Imaging appearance of the primary tumors and follow up CT scans were analyzed and compared with the clinical course. RESULTS: 29 of 45 patients had infratentorial, and 16 had supratentorial ependymomas. All children underwent CT investigation, and in 12 cases additional M R examinations were carried out. M R was most useful in operation planning, evaluating location and extent of the tumor, especially in the posterior fossa, but it did not help in differential diagnosis. Ependymomas can mimic the CT and

P U R P O S E : 1) to define the MRI modifications in the brain after CNS radiochemoprophylaxis in children with Acute Lymphoblastic Leukemia (ALL) and 2) to compare CT and M R I findings. PATIENTS and M E T H O D S : 20 off-therapy, disease-free, children treated for A L L were studied by MRI. In all A L L patients the treatment protocol included CNS prophylaxis, combining cranio-spinal irradiation and intrathecal metotrexate. Interval between CNS prophylaxis and MRI varied from 3 to 11 years. Every patient underwent unenhanced CT study yearly. MRI examination was obtained with 0.5 T superconductive unit, performing multiplanar T1-PD-T2 weighted as well T2*-weighted Gradient-Echo sequences. 6 cases were examined by multiple sequential MRI. RESULTS: CT showed intraparenchymal calcifications in 6 children while M R I demonstrated calcification only in 1 patient, exclusively as hyperintensities in the Tl-weighted SE images. Conversely MRI was able to display mild diffuse leukoencephalophathy in the subependymal white matter in 12/20 A L L patients. CSF ventricular and/or convexity spaces appeared dilated in 6/20 children. In children treated under the third year of age the thickness of paratrigonal white matter seemed to be slightly reduced. In one case CT and MRI gave evidence of development of a malignant glioma. CONCLUSIONS: M R I appears the method of choice to show white matter alterations following CNS prophylaxis in A L L children but it is completely ineffective in the diagnosis of iatrogenic calcifications.

19. MR of the spinal cord and dural sac in children with medulloblastoma and pinealoblastoma M. Braun* S. Gerber* Ch. Moret* S. Pouyoune* S. Bracard*, P. Chastagner**, J-C. Marchal ~ D. Sommelet** L. Picard*. *Departments of Diagnostic and Interventional Neuroradiology* Pediatric Neuro-oncology** and Neurosurgery ~, University Hospital Nancy, France Purpose: Evaluate the incidence of spinal cord and dural metastases as seen on MRI in children presenting with a medulloblastoma or pinealoblastoma. Patients and methods: In 24 children (aged 1 to 16) (18 medulloblastoma, 6 pinealoblastoma) the entire spinal cord and dural sac were systematically explored on post-contrast sagittal plane before and after each therapeutic step. The results were compared to the CSF analysis (cells and albumin). Results: During the evolution the 18 children with medulloblastoma, MRI revealed 9 cases of nodular (5 cases) or tumoral

$36 arachrtoiditis (4 cases), in the 6 children with pinealoblastoma one case of nodular metastases. We noted 2 cases of transitory false negative MRI and most important 9 cases where MR showed nodular lesions where no tumoral cells were found in the CSE High albumin levels ( < 1,5 g/l) is much better correlated to the presence of nodules and to a poor outcome (Ehrtich, 1978). Conclusion: MR of the spinal cord and dural sac is an important means of survey in children with neuroectodermal posterior fossa tumors allowing less frequent lumbar punctures.

20. Characteristics of spinal injuries in children

Gy. Katona Sote, Department of Diagnostic Radiology, Budapest, Hungary OBJECTIVE: To define the role of CT and MR investigations in the evaluation of spinal injuries in children. MATERIALS AND METHODS: 17 children (aged 2 weeks to 16 years) suffering from spinal injuries were studied in the late postinjury period: the earliest investigation occurred 2 weeks after trauma. 8 of them had cervical, 9 children had thoracolumbar injuries. 6 children were scanned (Somatom HiQ, Siemens) and 16 MR examinations were performed by using standard spin echo sequences (Magnetom SP 63, Siemens). RESULTS: On MR 8 children had cord abnormalities. MR revealed traumatic pseudomeningocele related to birth injury in one case. Kyphoscoliosis, subluxation and disk herniations were also found. 4 children had congenital malformation. CONCLUSIONS: Children may have spinal injuries which differ in appearance and in prognosis from those in adults. In some instance it can be difficult to decide whether the alteration is traumatic or whether the finding is a congenital malformation. MR provides a practical tool for diagnosis in pediatric trauma population too, Therefore I suggest that if needed MR should be performed as soon as possible.

21. Radiol0gic-pathologic correlations underlying the clinical presentation of epilepsy (Invited lecture)

Anne G. Osborn (Salt Lake City, Utah, USA) 22. Hippocampal sclerosis: comparison of visual evaluation, signaland planimetric measurements with :a high resolution T2-weighted turbo-spin-echo (TSE)-seqnence

B. Ostertun, H. Urbach, L. Solymosi. Dept. of Neuroradiology, University of Bonn, Germany Objective: Hippoc~mpal sclerosis (HS) is the most frequent cause for temporal lobe epilepsy. Its detection and exclusion of contralateral disease are essential for epilepsy surgery. Purpose of this study was to compare sensitivity of pure visual image evaluation with "objective" measurements of hippocampal signal and planes. Methods: Since 1993 hippocampi were examined with a coronal T2-weighted TSE scan (2 mm slicess) in presurgical work up and included in a long-term-study. So far we evaluated 23 patients with histopathological evidence of HS with unilateral seizure origin as follows: 19 preoperatively with purely visual judgement of hippocampal signal and size and 2) retrospectively (without knowledge of the resected side) with a) signal measurement and b) planimetric measurements in 5 representative slices. Results: 1) 95 % (22/23) of HS were correctly diagnosed and lateralized by pure visual evaluation. 2a) Normal relative signal was 738 + !08 (2SD) and for HS 882 + 182. 2b) Planimetry in the middle hippocampal portion showed 1,17 + 0,44 cm2 as normal (unaffected side). In HS 0,75 + 0,52 cm2 were measured. Signal judgement alone defined 15/23 HS (65%) as pathologic (p<0,05). Planimetric evaluation alone detected 13/23 HS (56,5 %). Both signal- and planimetric measurements lateralized correctly in 100 %.

Conclusions: Our preliminary analysis shows a positive predictive value for HS or 95 % for pure visual evaluation. Measurements of signal and planes correctly lateralize in 100 %. But without comparison of both sides, only relying on absolute signals or planes, the diagnosis can be established with p < 0,05 in only 65 or 56 %. The reason for this is a broad overlap of the 95 % intervals of normal and pathologic hippocampi. Therefore with these methods it seems impossible to rule out bilateral involvement in HS. Signal differences of both hippocampi proved to be the most reliable diagnostic feature of HS. Even with reduction of standard deviations of the measurements with higher numbers of patients it seems doubtful whether these time-consuming methods will proof superiority compared to image evaluation by an experienced neuroradiologist. 23. Advantages and limitations of ictal HMPAO SPECT examinations in the presurgical evaluation of patient with drug resistant partial seizures

K. Borbdly *^, P Ha!dsz** A. Balogh*** P. Barsi ~, J. Martos*, J. Vajda* S. Czirjdk*, P Bazs6 ^, G. Kolldr ^, V. Juhos*** A. Holl6** P Vdrady* 1. Nydry*. *National Institute of Neurosurgery, **National Institute of Neurology and Psychiatry, ***St. Istvdn Hospital, International Medical Center, Budapest, Hungary BACKGROUND: Because epileptogenic foci are associated with alterations in regional brain blood flow, ictal and baseline SPECT studies have come into wide use for presurgical evaluation, complemented by such other modalities as CT, MRI, and PET. Epilepsy surgery requires high accuracy in localization of loci. Even though MR! exquisitely demonstrates morphotogic changes, structural lesions may not always correlate with clinical, EEG and pathologic identification of foci. Functional imaging is a logical approach for detection of pathological cerebral function. However, the mapping of functional data from SPECT to specific anatomical structures is a difficult problem. PURPOSE: to assess the advantages and limitations of HMPAO SPECT giving the perfusion patterns in brain tissue within, neighboring and distant from the focus during baseline, spontaneous and provoked ictal studies to other modalities. METHOD: 32 patients were studied retrospectively, all with negative MRI findings. Baseline, spontaneous and/or provoked ictal SPECT studies were performed with a standard technique for each patient. Examinations were started 15 minutes after intravenous (iv) application of 740 MBq of Tc-99-m HMPAO. The provoked ictal studies were carried out with iv injection of 6-8 mg/kg of Metrazol. In 9 patients where FDG PETwas performed, the focal hypometabotism ,corresponded to ictal EEG findings, while in two cases demonstrated no changes. RESULTS: In 17 cases of ictal SPECT, we detected a demarcated supratentorial hyperperfusion demonstrating the epileptogenic area, with low inhomogenous cerebellar tracer uptake. In 18 cases of periictal SPECT, we detected both moderate inhomogenous cerebellar tracer uptake and moderate hyperperfusion in the epileptogenic region, with multifoca! supratentorial high tracer uptake. In 15 cases of fast iv injection of Metrazol, the patients demonstrated habitual seizures as verified clinically and by EEG; however, focal hypoperfusion was detected in the ictal SPECT images, which is the opposite of spontaneous ictal SPECT images. CONCLUSIONS: 1 / I n case of localizable foci, ictal SPECT reveals the epileptogenic area, 2/In lack of spontaneous seizure the use of provocating methods are advised, 3/ Because of the low anatomical resolution of SPECT, anatomical localization via exact matching of functional image data to morphological data is recommended. In these cases image fusion seems to be the most reliable solution.

$37

24. Correlation of clinical symptoms and neuroradioiogical findings in disorders of neuronal migration and gyrification

P Barsi (1,2), J. Ken~z (1,2), Gy. V6rallyay (1), Gy. Rdsonyi (3), P Haldsz (3), V. Farkas (4) (1) Dept. of Neuroradiology,.National Institute of Psychiatry and Neurology (2) CT & MR1 Laboratory, International Medical Center (3) Epilepsy Center, National Institute of Psychiatry and Neurology(4) II. Dept. of Pediatrics, Semmelweis Medical University Budapest, Hungary OBJECTIVE: To assess how clinical signs and symptoms, epileptologic data and imaging findings correlate in patients with disorders of the neuronal migration or gyrification, and what are the optimal imaging sequences on a 0.5 T M R I unit and on CT for this purpose. MATERIAL AND METHODS: We evaluated the MRI and/or CT examinations of 36 epileptic patients, in whom one or more types of migration-gyrificafion disorders were discovered, often combined with other developmental brain anomalies. The imaging material was rather diverse: besides conventional SE T1, PD and T2 sequences, heavily T2 weighted, IR and contrast-enhanced T1 sequences were also used in a number of cases; the conventional brain protocol was used in CT, sometimes complemented with thin slice series and contrast administration. Clinical history, neurological and epileptologic data were collected and correlated with imaging findings. RESULTS: CT could reveal most of the disorders, and thin slice series were very useful, but bony artifacts prevented the fine details being depicted near the skull base, primarily in the temporal lobes. Contrast administration rarely gave new hints. MRI was the tool for meticulous morphological analysis. The basic SE sequences showed most of the details, but IR and heavily T2 weighted sequences were more effective in delineating the border of gray and white matter, and revealing gliosis. Clinical data suggested that there was not always a close correlation between the extent of the disorder and the clinical and epileptologic state. In a lot of patients, epilepsy had a relatively late onset, and could be controlled medically. CONCLUSIONS: Attempting to find out the morphologic basis of epilepsy, one shall be aware of disorders of neuronal migration and gyrification as a possible cause, even in adult patients. CT is helpful, but M R I is the method of choice, The detailed demonstration of these disorders is important not only to rule out other pathologies, but also to provide all the possible morphological data for a possible surgical intervention in selected cases. 25. CT and MRI in 5 children with late-onset globoid cell leukodystrophy (Krabbe disease)

R.Barone(1), M.Beck(2), K.Briihl(3), PStoeter(3). University of Catania/Italy, Pediatric Clinic(l), University of Mainz/Germany, Pediatric Clinic(2) and Dep. Of Neurorad. (3)

the pyramidal tract. Calcifications and cerebellar lesions are less pronounced in the early-onset form.

26. Aspartylglucosaminuria ( A G U ) : MRI and histological find-

ings.

R. Raininko, T. Autti, M. Haltia, O. Salonen, S-L Vanhanen, M. Renlund, L. Sipila, H.J. Aronen, P. Santavuori. Depts. of Radiology, Child Neurology, Pathology and Paediatrics, University of Helsinki, Finland OBJECTIVE: To evaluate the general brain maturation and the effect of bone marrow transplantation (BMT) in A G U patients by means of MRI and to define the histopathological correlations. D E S I G N A N D M E T H O D S : Three patients with A G U and 10 age-matched controls were examined. One patient (1) was examined three times: before the BMT at the age of 2 years i month, and twice after BMT at the ages 4 years i month and 6 years 1 month. The second patient (2) was examined before and after the B M T a t the ages of 2 years 8 months and 4 years and 6 months. The third patient (3), who did not receive the BMT, was examined at the ages 3 years 6 months and 6 years 6 months. All living subjects were imaged at 1.0 T (SE 2500/22-90). Axial and coronal slices were used. After visual evaluation, the signal intensity was measured on T2-weighted images in numerous locations. The intensities were related to an internal reference (CSF). The post mortem material consisted of the brains of 4 adults with AGU. MRI was performed at 1.5 T (SE 3000/20-120). After MRI, histopathological specimens were taken from the gray and white matter. RESULTS: The first T2-weighted images of the children with A G U showed generally increased signal intensity in the white matter with peripheral foci of especially high signal intensity. Corpus callosum had lower signal intensity than the peripheral white matter. Deep gray matter gave impression of hypointensity compared to the peripheral white matter. Atrophy was not found. After BMT, only minor signal intensity changes were found. In contrast, in patient 3 (no BMT) the signal intensity of the peripheral white matter still was significantly higher ( < 2SD) at the age of 6,5 years than that of the controls and also higher than that of the deep grey matter. The adult patients showed no atrophy on the post mortem images. The signal intensity of the deep gray matter was equal to or lower than that of the white matter in 3/4 of cases. The white matter did not show any marked histopathological alterations The basic cortical cytoarchitecture was generally preserved but in all areas examined most nerve cell bodies were distended and rounded, and contained poorly defined vacuoles. The cytoplasm usually contained number of coarse pigment granules with the staining qualities of lipofuscin. CONCLUSIONS: A G U is basically a disease of the gray matter probably causing delayed brain maturation. BMT seems to normalize the brain maturation process.

Purpose: To show possible differences in imaging and course of early and late onset form of Krabbe disease.

Material and Method: In 5 girls with onset of Krabbe disease at the

27. MR findings in ceroid lipofuscinosis

age of 1,5-3 years, CT and / or MRI imaging was performed during stage 2 and 3 of the disease and compared to MRI findings of earlyonset type (5-month-old girl). Result: During its early stage, the late onset form showed discrete hypodensities (on T2 w) of the central white matter and pyramidal tracts as well as some signal reduction in the basal ganglia (deposition of Ca of Fe?). In advanced stages, progressive white matter lesions involved subcortical U-fibres and various fibre tracts of the basal ganglia and brain stem with accompanying grey matter atrophy. There were no disturbances of the blood-brain-barrier and no focal cerebellar lesions. White matter calcifications were present in 1 out of 4 CTs only. In the early-onset form, severe disturbances of myelinisation with cerebellar lesions were found in the early stage of the disease. Conclusion: In the late-onset form of Krabbe disease, lesions are seen first within the already myelinated central white matter and

M. Grboli, L. D'Incerti, E. Ciceri, L. Chiapparini, *N. Nardocci, M. Savoiardo. Dept. of Neuroradiology and *Pediatric Neurology. Istituto Nazionale Neurologico "C.Besta", Milano, Italy. OBJECTIVE: To review the M R studies of 8 cases of proved Ceroid Lipofuscinosis (CLF) to verify whether a pattern of distribution of MRI abnormalities could be identified, that could help the neuroradiologist in suggesting the diagnosis. PATIENTS A N D M E T H O D S : 18 M R examinations were performed in 8 patients with CLF from 1988 to 1994. The age of the patients ranged between 2 and 38 years. The patients were grouped as follows: 1 case of infantile form, 2 cases of late infantile form, 3 cases of juvenile and 2 cases of adult form. RESULTS: All the patients had cerebral and cerebellar atrophy. In all the cases thinning of the cerebral cortex was observed. Mild white matter hyperintensity in T2-w.i., particularly in the periven-

$38 tricular regions, was observed in all cases. In 2 cases, hypointensity in T2-w.i. was present in the thalami. CONCLUSIONS: Our findings are in agreement with those observed in the rare cases previously reported. The cerebral atrophy and the tNnning of the cortex are consistent with the pathological data of neuronal destruction, while the signal abnormalities of the white matter are compatible with wallerian degeneration of the fibers and gliosis. It should be noted that the white matter hyperintensity is lower than that commonly observed in leukodystrophies. The intracellular storage of lipofuscin, and perhaps of iron, can account for thalamic hypointensity. The MRI features observed in our cases are not specific, but, in the appropriate clinical situation, may suggest the diagnosis of CLF and, therefore, accelerate the diagnostic course.

28. Quantitative 1H Magnetic Resonance spectroscopy (MRS) of the brain in nonketotic hypergylcinemia (NKH) E. Martin, H. Rumper, G. Zeilinger, T. Huisman University Children's Hospital Zurich [E. M., H.R., T. H.] and Kantonspital Aarau [G.Z.], Switzerland Background: N K H G is an inborn error of amino acid metabolism in which large amounts of glycine (glyc) accumulate in the brain due to a molecular defect in the glycine cleavage system. High concentration (conc) of glyc are found in the CSE Aim: Can 1H-MRS detect differences in Glyc conc in various areas of the brain of neonates with N K H G in vivo? Subject and Intervention: 7-day old boy born at term after uneventful pregnancy and birth, progressive hypotonia and seizures during first week of life, artificial ventilation. At age 3 d. pp. elevated glyc levels in plasma 4176 mmol/L and urine 4966 mmol/mol creatinine lead to the diagnosis of NKHG. At day 7 pp. MRI was carried out and multivoxel quantitative 1H-MRS were obtained from volumes of 3.5ml in Frontal White Matter (FWM), ParietoOccipital WM (POWM), Basal Gangila (BG), Cerebellar WM (CBWM), and from i ml of CSF (lateral ventricle). Results: Glyc conc as measured by 1H-MRS were: FWM 4 mmol/ L, POWM 5 mmol/L, BG 5.5 mmol/L, CBWM 8 mmol/L and CSF < i mmol/L. Conclusions: In vivo 1H-MRS demonstrated that: 1) there are variations of glyc conc of up to 100 % in different areas of the brain; 2) the CBWM had the highest glyc conc and showed also an elevated lactate conc of 1 retool-L; and 3) there was barely glyc detectable in the CSF in which the simultaneous biochemically measured conc was 646 mmol/L. 29. Cerebral MR findings in patients with myotonic dystrophy Y Miaux, J.. Chiras, B. Eymart, N. Martin-Duverneuil, C. Cognard, C. Delaporte. Service de Neuroradiologie Charcot et Service de Neurologie, Unitd CNRS 1957, Groupe Hospitalier Pitid-Salp6tridre, 47 Boulevard de l'H@ital, Paris, France OBJECTIVE: To study MR cerebral abnormalities in patients with adult form of myotonic dystrophy (DM) and compare them with those in age-matched normal controls. DESIGN AND METHODS: Thirteen patients with adult form of DM (9 with paternal inheritance, 2 with maternal inheritance, 2 with unknown inheritance, age, 35.9 _+12.2 years) and thirteen agematched normal controls (age, 34.4 + 12.1 years) were studied with neuropsychological testing and MR imaging (1.5Tesla). Sagittal Tl-weighted and axial dual-echo long TR fast spin-echo images were obtained. All MR scans were analyzed by two neuroradiologists, three factors were measured (ventricular system, external CSF spaces, skull thickness) and focal white matter lesions were evaluated by mutual agreement. Statistical analysis was performed using the Student's test and chi-square test.

RESULTS: There was no significant differences in the measurements of ventricular system and external CSF spaces in patients with DM as compared with the age-matched controls. We found a significantly increased incidence of hyperintense areas on T2-weighted images in the white matter of patients with DM as compared with the age-matched controls. Five patients with DM (3 with paternal inheritance, 2 with unknown inheritance) showed hyperintense areas in the subcortical white matter of the temporal lobes. These findings were not associated with intellectual impairment, except in one case. This last patient with temporal lobe lesions on MRI had history of difficult birth and focal temporal epilepsy. There was marked thickening of the skull in three DM patients (frontal bone in 2 cases and whole skull in one case) with associated ossification of the falx in 2 cases. CONCLUSION: In this group of DM patients without intellectual impairment we found no significant atrophy and a significantly increased incidence of hyperintense subcortical areas in the temporal lobes.

30. The dopamin-receptor-SPECT with mI.IBZM in the evaluation of Parkinson's disease W. Reiche1, M. Grundmann 2, U. Piepgras 1. Institute for NeuroradiologS and Neurological Clinic 2, University Hospital, 66421 Homburg/Saar, Germany For effective drug therapy of Parkinson's disease it is necessary to distinguish between idiopathic and secondary genesis, la3I-IBZM is a radiolabelled Iodo-methyl-benzamid and binds specific to the cerebral Dopamin-receptor (D2) in the basal ganglia. The purpose of this study is to evaluate, if SPECT-examinations with 1231-IBZM of D2-receptors heIps to discriminate between idiopathic (IPD) and secondary Parkinson's disease (SPD). We investigated 21 patients (10 male, 11 female, medium age 55.5 + 11.9 years) - 9 with idiopathic (IPS), 8 with secondary Parkinson's disease (SPD), 1 with familiar chorea, 1 with drug induced parkinsonian syndrom and 2 controls - 90 minutes after i.v. application of 185 MBq 123I-IBZM with SPECT. After reconstruction with a butterworth filter (cutoff frequency 0.5) and attenuation correction (coef. 0.12 cm -1) we quantify the IBZM-basal gangliauptake as ratio to the frontal, D2-receptor free cortex (BG/FC). The patients with IPD as well as the controls revealea high specific IBZM-uptake in the basal ganglia compared to the adjacent frontal brain tissue (IPD: BG/FC=1.46+0.08; SPD: BG/ FC = 1.49 _+0.14) indicating an functional intact D2-receptor. A decreased IBZM-basal ganglia-uptake is found in cases with SPD ( B G / F C = l . 2 5 + 0 . 0 9 ) , in the case with chorea (BG/FC: right = 1.31, left = 1.29) and with drug induced parkinson (BG/FC: right = 1.18, left = 1.22). 4/9 patients with IPS present an interhemispheric side difference of 6-8 % with a higher basal ganglia uptake on that side, which is responsible for the individual neurological symptoms. We interprete this phenomena as D2-receptorstimulation, an answer to a high and longstanding neurotransmitter-deficit. This results demonstrate, that the IBZM-Dopamin-receptorSPECT permits a relative clear discrimination between idiopathic and secondary Parkinson's disease and gives important arguments for differential therapy.

31. Superficial siderosis of the CNS. Possibility of treatment of selected cases? M. Savoiardo, M. Grisoli, A. Erbetta, S. Giombini, D. Pelucchetti, *C. Agostinis, *V. Bonito. Istituto Nazionale Neurologico "C. Besta", Milano, *Ospedali Riuniti, Bergamo, Italy OBJECTIVE: To demonstrate that post-traumatic cases of CNS superficial siderosis (CNSSS) may benefit from repair of old dural lacerations.

$39 M E T H O D S : Fourteen patients with CNSSS were diagnosed by MRI. Three had tumors. Of the remaining 11 "idiopathic " cases, history of old trauma (that occurred 25-31 years earlier) was obtained in 4. Three had cervical root avulsion and one had a posterior fossa extradttrai cyst. On the hypothesis that badly repaired dura could be the source of repeated bleeding, two of these patients were operated. RESULTS: One patient with cervical root avulsion had repair of the dura (Savoiardo et al., A J N R 1993; 14: 1445, Bonito et al., J. Neurosurg. 1994; 80: 931). After initial worsening, he did not have any clinical improvement but the CSF became clear and colorless, without RBCs. The patient with posterior fossa extradural arachnoid cyst also had normal CSF postoperatively; in addition, he presented substantial improvement in ataxia and dysarthria, and a questionable improvement of hearing loss. CONCLUSIONS: CNSSS is a severely disabling, progressive disorder due to repeated microhemorrhages, most of the times from unknown source. In the post-traumatic cases with badly repaired dura, surgery may perhaps interrupt the progression of the disease; the improvement of our second case was unexpected and remains unexplained.

32. Present provision and expected development of neuroradiology in Europe (Invited lecture) K. Ericson (Stockholm, Sweden) 33. The economics of European neuroradiology - present and future (Invited lecture) I. Moseley (London, United Kingdom) 34. Training in neuroradiology: current state and future (Invited lecture) L Isherwood (Manchaster, United Kingdom) 35. Cost/Efficiency Study in Radiological Management of Lumbar Pathology? Hermanus N. MD*, David Ph. MD**, Abid A. MD* Bellemans M. MD* Baldriaux D MD**. * Departement of Radiology - Brugmann University Hospital (Brussels), **Departement of Neuroradiology - Erasmus University Hospital (Brussels) Objective In order to obtain a good cost/efficiency ratio of the lumbar pathology management, we propose a decision algorithm based on a statistical and technical evolution approach. Methodology We analyse the radiological reports of two kinds of University Hospitals, with regard to her v o l u m e , equipment and experience. Brugmann University Hospital : 600 beds, no neurosurgery department, 2 C.T., 1 M.R.I. since 1991. Sample of the study: 53 patients. Erasmus University Hospital : 900 beds, with neurosurgery department, 3 C.T., 2 M R I since 1985-1990. Sample of the study : 50 patients. We correlate the symptoms with the writed radiological reports of plain films, C.T. and M. R. I.. These informations are recorded on a PC data base. The computer allows for one or several clinical criteria to extract from the data base, a comparative analysis between the radiological technics related to the institution's characteristics, quality of clinical informations, radiological informations, the anatomical structures (discus, facettar joints . . . . ), the kind of practitioner, the ratio of out/in patients... Condnsiom The main point of this study is the essential role of clinical data in order to pursue the radiological examination until a satisfactory answer. C.T. is more accessible now (cost of installation, short time waiting for the patient). In the same period, we

have a quick M. R. I. evolution and a training period. For an useful approach, both medical and economical, two kinds of decision algorithm are proposed (one "real", one "ideal").

36. When should we enhance cranial CT scans? Ph. Demaerel, (3. Wilms, A.L. Baert Department of Radiology, University Hospitals K. U Leuven, Leuven, Belgium OBJECTIVE: To reassess the use of contrast agents for cranial CT in 1994. MATERIAL AND METHODS: Two thousand consecutive patients (1600 adults, 400 children) referred for cranial CT before and after contrast enhancement (CE) were prospectively examined. RESULTS: There were no abnormalities in 46 % of the cases (group I). In 14.5 %, there were lesions on the unenhanced scan but there was no contribution from the subsequent CE and a confident diagnosis was possible on the plain scan (group II). 17.5 % of the patients were referred for screening of metastases. The enhanced scan was normal in 77 % of them (group III). In 10.5 % of the cases the scan was requested following cranial surgery (group IV. In 4.5 % there was a suspicious lesion on the plain scan and no confident diagnosis could be given. There was enhancement in 53 % of the cases (group V). Some indications require the use of contrast, even if the unenhanced scan is normal (6 %, group VI). There was enhancement in 60 % of the cases. In 15 (0.7 %) patients, abnormalities were missed on the plain scan and visible on CE scans, but there was no selected indication for contrast administration (group VII). However, none of these findings required therapy. CONCLUSION: The availability of third and fourth generation CT scanners with their excellent spatial resolution should enable us to decrease the number contrast enhanced scans. We suggest to administer contrast if there is a suspicious lesion on the plain scan or in the appropriate clinical setting (group III-VI, 38.5 %). In all other instances there is no need for CE (61.5 %). If access to M R I is easy, a further reduction is even probable. At this moment we enhance 20 % of our patients. This protocol results in a major decrease in working costs and inevitably decreases the amount of side effects. 37. Glucose metabolic consequences of intracranial vascular malformations: pre- and postembolization PET FDG studies K. Borb~ly, A. Valavanis* GK. von Schulthess, M. Ladd, T. Berthold, C. Burger, A. Buck. Division of Nuclear Medicine and Department o f Medical Radiology, *Division of Neuroradiology, University Hospital, Zurich, Switzerland PURPOSE: In order to assess the metabolic consequences of cerebral malformations, nineteen patients with cerebral arteriovenous malformations (AVM)) and two with dural arteriovenous fistulae (DAVF) underwent Positron Emission Tomography (PET) exams with 18-F fluordeoxyglucose (FDG) before and/or after embolization. The data were correlated with morphological and clinical findings. Retrospectively, preembolization studies were examined to determine whether they may be used to predict the risk of so-called "threatened territories" neighboring and distant from the malformation, while postembolization data were analysed in order to determine the degree of glucose metabolic changes after embolization. METHOD: 16 preembolization studies and 11 postembolization P E T studies performed in 21 patients suffering from seizures (n = 12), bleeding (n = 4), headaches (n = 3), or dementia (n = 2) were analysed and compared to selective angiography, Magnetic Resonance Imaging (MRI), and clinical findings. Additionally, in five cases, a fusion between PET and MRI was performed. Six of the above patients had both pre- and postembolization PET studies, including a follow-up examination in one case of postembolization hemorrhage. PET to PET fusion was performed in

$40 three of these cases. Regions of interest were placed ipsilaterally and contralaterally in the brain tissue within, adjacent and remote fiom the vascular malformation and were normalized to the normal cerebellum, normal frontal region and whole brain. The diaschisis index was calculated by dividing the regional cerebral glucose metabolism in the diaschisis region by the metabolism in a normal cerebellar region. In patients with both pre- and postembolization PET studies, effects of embolization on glucose metabolism were evaluated by calculating the ratio of postembolization to preembolization data. RESULTS: All cerebral malformations were well demarcated in the PET images as hypometabolic areas, with various patterns and severities of hypometabolism adjacent and remote from the malformations. The adjacent/remote areas demonstrated no/or smaller changes in MRI. The PET data showed a good correlation with clinical findings. CONCLUSION: The preembolization PET FDG data reveal widespread metabolic consequences of cerebral malformations and suggest that they are associated with impairment of glucose metabolism and indirectly substantiate the traditional theories of AVM-related hemodynamic condition. Whether the postembolization metabolic depression in the surrounding areas and/or remote brain tissues is a transient or permanent result of embolization needs a longer follow-up. Our results suggest that functional changes in brain tissue as mapped by glucose metabolism are helpful in the preembolization/surgical evaluation of intracranial AVMs, particularly in medium and large AVMs.

DESIGN AND METHOD: Forty-five (45) patients (36 males, 9 females), mean age 66 years (range 35-87) with symptomatic or asymptomatie carotid stenosis underwent balloon angioplasty followed by elective stenting. Four patients underwent bilateral procedure. The study protocoi includes baseline and follow up independent neurologic assessment, complete brachiocephalic and cerebral angiography, MRI, TCD, duplex ultrasound and nuclear medicine studies. RESULTS: All procedures were angiographically and clinically successful. There were 4 neurological events. Of these, 3 patients recovered completely within 72 hours. The third patient with atrial fibrillation and mitral regurgitation four days after a successful procedure had a stroke from a cardiogenic embolus to the opposite middle cerebral artery. There were no post discharge complications. Follow up on angiography at 6 months is planned for all patients and is in progress. CONCLUSIONS: Based on this initial experience in an unselected patient group, this pilot study is intended to open the door for further PTA and stenting in atherosclerotic disease on the bifurcation of the carotid artery in a larger group of patients and to develop an alternative treatment for an endarterectomy. Based on this initial experience in an unselected patient group, angioplasty and elective carotid stenting appears feasible, effective and safe. 40. Aneurysm flow dynamics investigated in a transparent viscoelastic model

J. A. Horton, D. Vincent. Dept. of Radiology, Medical University of S. Carolina, Charleston, South Carolina, USA 38. Evolution of techniques for embolizing skull base meningiomas

William O. Bank, MD 1'3, Edwin R. Hudson, MD 1, Wayne J. Olan, MD 1, Enrique T. Gorbea, MD z, Maria A. Matsumoto, MD 2, Laligain N. Sekhar, MD s, Donald CC. Wright, MD ~ Interventionat and Therapeutic Neuroradiologj and Diagnostic Neuroradiology Sections 2, Department of Radiology, and Department of Neurological Surgery, The George Washington University Medical Centeg 901 23rd Street IV. W., Washington, DC 20037 Material & Methods: During the past 3 years we have performed endovascutar occlusion (embolization) in 102 patients with vascular skull base meningiomas, who required treatment of feeders arising from the cavernous internal carotid and ophthalmic arteries, in addition to the more commonly embolized branches of the external carotid system. Results: The rate of new skull base meningiomas to our center has remained relatively constant at 45 ( + / - 5) patients each year. Comparing the time periods of 5/93-4/94 with 5/94-4/95, our ability to catheterize meningohypophyseal branches has improved from 69 % to 81%. Our ability to achieve a microcatheter position from which embolization can be performed safely has improved from 58 % to 68 %. During the same time periods, our rate of ophthalmic artery catheterization has remained relatively constant at 95 %, while our ability to obtain safe position for embolization has improved from 7 % to 26 %. Conclusions: This evolution is due to technical refinements in the microcatheters, guidewires, and embolic agents, as well as our growing experience in how best to use them in this area. Microcatheter shape, use of "oversized" guidewires, and choice of embolic material will be discussed. 39. Angioplasty and stenting of atherosclerotic stenosis on the bifurcation of the carotid artery

Vitek JJ, Iyer SS, Yadav JS, Roubin GS University of Alabama at Birmingham Birmingham, Alabama OBJECTIVE: This study is intended to open the door for further PTA and stenting in atherosclerotic disease in the bifurcation of carotid artery and to develop an alternative treatment for endarterectomy.

OBJECTIVE: To examine flow patterns in aneurysms under pulsatile flow conditions in models that closely approxirnate real aneurysms. DESIGN AND METHOD: We used transparent viscoelastic aneurysm models made of silicone [Kim Knox, Golden Pacific Arts, Encinitas, California, USA] with flow generated by a pulsatile pump. Food coloring provided the flow stream marker. A camcorder recorded the results with eventual transfer to 1/2 inch tape using a frame counter for time measurement. RESULTS: We found two previously unknown properties of aneurysms. First, for aneurysms with similar sized necks, the angular flow velocity varies inversely with the size of the aneurysm. Consequently, shear stress is higher in small aneurysms than in large ones. The second finding was that in aneurysms with equal diameters but different neck widths, those with wide necks reach higher angular flow velocities. CONCLUSIONS: These results explain why small aneurysms can grow to large size, why they rupture, and why giant aneurysms have low bleed rates. Transparent viscoelastic models seem to mimic accurately the behavior of real intracranial aneurysms. The technique of using them to investigate highly nonlinear flow systems is elegant and opens a window to further investigations. 41. Intravascular ultrasound imaging of experimental aneurysms

Istvan Szikora MD, A jay K. Wakhloo MD PhD, Tamerla D. Chavis MD, Lee R. Guterman Phd MD, Leo N. Hopkins MD Neuroendovascular Laboratory, Department of Neurosurgery State University of New York at Buffalo School of Medicine and Biomedical Sciences OBJECTIVE: To evaluate the benefits of intravascular ultrasound imaging (IVUS) in demonstrating aneurysm morphology before and after endovascular treatment with Guglielmi detachable coils (GDC). DESIGN AND METHODS: Twenty four canine common carotid artery aneurysms were studied with both DSA and IVUS before treatment and ten aneurysms in 1 week and 4 or 8 weeks following endovascular treatment with a modified GDC system (Target Therapeutics, Fremont, CA). A Sonos Intravascular Imaging System (Hewlett-Packard, Andover, MA) was used, which employed a 4.8F transducer catheter, operating at 30 Mhz.

$41

RESULTS: Four aneurysms thrombosed spontaneously, five occluded completely and another five partially following en ~ dovascular treatment. In two cases, the parent artery thrombosed in one week after treatment. Measurements of the parent artery diameters proximai and adjacent to the aneurysm orifice correlated well on D S A and IVUS (less than 0.4 mm difference). The aneurysm ofirices were poorly visualized on D S A but clearly depicted by IVUS allowing for precise measurement of the diameter in one direction. Intrasaccular thrombi were clearly demonstrated by IVUS as echodens masses. Coils were well visualized as echodens, curvilinear structures within the aneurysm orifices, Neck remnants were demonstrated as imperfections of the coil surface within the orifice. The content of the aneurysm cavities displayed mixed echodensity following coil treatment in each case regardless of the angiographic result. In those cases in which the coils narrowed the crossection of the carotid artery by more than 50 % the parent artery thrombosed. C O N C L U S I O N : IVUS is superior to D S A for visualizing aneurysm orifice morphology and providing unique information about the configuration of coils within the orifice. This information might be important in planning and performing endovascular treatment with detachable coils. IVUS at 30 M H Z have a limited value in analyzing the content of a G D C treated aneurysm and depicting residual flow. 42. Effects of collagen in experimental swine aneurysms using independently detachable coils

Dawson RC, Krisht AF,, Shengelaia GG, Bonner GD. Emory University School of Medicine, Atlanta, Georgia Objective To assess the histological changes produced by collagen in experimental swine aneurysms using coils with controlled detachment features Design and Methods Sidewall carotid artery aneurysms were surgically created in 6 swine using venous pouches. Angiography confirmed patency in all aneurysms 2 weeks later. Two animals were placed in each of 3 groups. In group I the left carotid aneurysms were embolized with interlocking detachable collagen microcoils (IDCC), of platinum wire wound around a mandrel of bovine collagen. Approximately 25 % of coil surface was open to expose collagen to blood, Aneurysms were packed as tightly as possible, altough coil volume was less than 50 %. The right carotid aneurysms served as control. Group II animals had interlocking detachable microcoils (IDC) with no collagen, placed via microcatheter similar to the IDCC's. The left carotid aneurysms served as negative controls. Group IiI animals had IDCC's placed in the right carotid aneurysms and IDC's placed in the left carotid aneurysms. Thus 4 aneurysms were embotized with IDCC's, 4 aneurysms embolized with IDC's and 4 control aneurysms. Angiograms were obtained at 1, 4, 8 weeks post embolization. The animals were sacrificed 8 weeks post embolization, the carotid arteries harvested, placed in methacrylate blocks, sectioned on a special microtome and stained with a bone stain most suitable for this type preparation. Results (1) 3 control aneurysms remained patent, but progressively became smaller, 1 thrombosed. (2) CIIDMC's & IDMC's both produced total thrombosis; the CIIDMC's in a shorter period. (3) Grossly there was a harder, tougher consistency to the aneurysms filled with CIIDMC's. (4) There was no recanalization in either the CIIDMC's or the IDMC's. (5) Histologically the CIIDMC's showed a slightly more reactive pattern of remodeling. There was a minimal foreign body reaction. Conclusions (1) Bovine collagen induces a more active process, which may produce a stronger tissue complex.

(2) The model, due to its surgical creation may have defects related to the surgical trauma. We theorize a part in this process for platelet derived growth factor, present both in clot and secondary to surgical trauma. (3) This relationship may bear further investigation to improve resuits for related endovascular occlusive devices.

43. Coated stents: application and results under normal and pathological conditions in canine arteries

M. Schumacher, E Schellhammer, A.K. Wakhloo Section of Neuroradiology, University of Freiburg, 2nd Department of Surgery, University of Cologne, Germany Purpose: To investigate the use of different vein grafts as coating for stents in normal canine arteries, as well as experimental AVfistulae. Methods: 27 ballon-expandable tantalum, 2 self-expandable nitinol and 1 self-expandable steel stent were coated with either autologous (n = 16), denatured heterologous (n = 11) or denatured homologous vein grafts (n = 3). Using a transfemoral or transcarotid approach, stents were placed in the SeA (n = 7), the CCA (n = 5), the E C A (n = 1), the CFA (n = 2) and the I C A (n = 12), 4 experimental AV-fistulae of the common carotid artery were treated with autologus vein craft coated stents (tantal stents (n = 1), nitinol stent (n = 2), steel stent (n = 1). Results: All coated stents, placed with the use of a ballon catheter/introducer system (n = 6), showed acute (n = 4) or subacute vessel occlusion (n = 2). The outcome of stents delivered by a crochet technique (n = 24), cohered with the typ of vein graft chosen. 10 implanted autologous vein grafts showed 1 acute and 1 delayed vessel occlusion, others stayed patent without proof of lympho-plasmacellular reaction. 3 acute, 5 subacute and 3 delayed vessel occlusions were seen in denatured heterologous vein grafts (n = 11), which showed hyaloid transformation and lympho-plasmacellular formation. In 3 cases of denatured heterologous vein grafts 2 acute and 1 subacute occlusion were found. Although no signs of host-versus-graft reaction were displayed, inflammatory tissue response was seen. Occlusion of AV-fistula succeeded in 2 of 4 cases, non of them remained patent. Conclusion: Autologous vein grafts implanted in canine arteries have shown good acute and delayed results. However, preparation remains cumbersome, and surgical procedure is required for harvesting the grafts. Replacing this procedure by denatured vein grafts seems to be limited by cellular reaction. Furthermore, significant stenosis of the parent vessel is caused by the thickness of the vein graft coated stent. 44. Pressure measurements via different types of microcatheters

H. Henkes, K. U Wentz, B. Bellenberg, U Seiz, H.D. Papenfuj'3, D. Kiihne. Klinik fiir Radiologisehe Diagnostik und Neuroradiologie, Alfried Krupp Krankenhaus, Essen, Germany OBJECTIVE: An experimental setup, suitable to evaluate in vitro the reliability of intravascular pressure measurements via various microcatheters was designed. M A T E R I A L S A N D M E T H O D S : Flow conditions were simulated in a closed circular 6 mm tubing system, filled with water and driven by a pulsatile pump. Heights, widths and frequencies of the pressure pulses could freely be adjusted by a pressure sensor. The tips of different microcatheters (Tracker-10 and -18, Magic 1.5 and 1.8) were placed into the tubing system 5 em from that pressure control sensor. Pressure/time-curves were then measured with a second sensor system via the catheters. The pulse frequency was set to 60 per minute. The pressure peak amplitude was 75 mmHg and 150 mmHg. RESULTS: The pressure in the tubing system as given by the static pressure is dampened through the microcatheters. This results in a broadening of the pulse peaks and an attenuation of the peak amplitudes. This dampening effect correlates with the catheter lumina, i.e. the smaller the diameter the stronger the attenuation. The mean pressures, on the other hand, were measured reliably.

$42 CONCLUSION: Intravascular mean pressures and pressure changes can reliably be measured via the conventionally used microcatheters. This method, however, yields no valid data concerning pressure amplitude.

47. Unclippable aneurysms of the internal carotid artery, treated by balloon occlusion of the parent vessel: late clinical outcome

45. Laser-detachment of platinum microcoils: initial investigational experience

O B J E C T I V E : The late clinical outcome after balloon occlusion of the internal carotid artery (ICA) as treatment of unclippabte aneurysms arising from this vessel was evaluated. M A T E R I A L S A N D M E T H O D S : Between 4/1984 and 4/1995, 64 patients (47 female, 17 male; age range 16-81 years, mean _+SD 47 + 17 years) underwent endovascular balloon occlusion of one ICA as treatment of a cervical, cavernous or intradural aneurysm of this vessel. Follow-up data of 57 patients were obtained 1122 months (mean +_SD: 50 _+35 months) after the procedure; they included neurological examination and M R imaging. RESULTS: Compared to their state at admission, 27 patients were discharged in an unchanged condition, and 24 were improved. Improvements were primarily related to early regression of pain and cranial nerve palsies. A clinical follow-up examination including MR imaging was performed 1 to 122 months after ICA occlusion. No recurrence of the presenting symptoms was encountered and 29 patients were asymptomatic, Compared to their pre-therapeutic status, 24 out of 57 patients were improved but still symptomatic, 8 were basically unchanged and no patient was worse in terms of aneurysm-related symptoms. In 2 patients, new neurologic symptoms had emerged, related to their aneurysm or its treatment in 1 and unrelated in another 1 patient. M R follow-up studies revealed a remnant of the aneurysm sac with shrinkage and thrombosis in l l patients. In 37 patients the mass effect of the aneurysm was entirely resolved. Ischemic parenchymal lesions which were not present before the balloon occlusion were detected in 2 patient. CONCLUSION: Proximal balloon occlusion of the ICA for t h e treatment of unclippable aneurysms is effective. It eliminates the mass effect and the risk of aneurysm rupture. The procedure related mortality is below 2 %. In terms of clinical signs and symptoms, 1-122 months after ICA occlusion 51% of the patients were asymptomatic, 42 % were improved and only 14 % remained unchanged.

H. Henkes, H. Monstadt, K. U Wentz, D. Ki~hne. Klinilc fiAr Radiologische Diagnostik und Neuroradiologie, Alfried Krupp Krankenhaus, Essen, Germany OBJECTIVE: The aim of the investigation is the technical realization and approval of a new method for insertion and detachment of platinum microcoils and other vascular implants via standard microcatheters. M A T E R I A L A N D M E T H O D S : Flexible lightconducting fibres with 0,11 ram, 0,125 mm and 0,21 mm diameter were connected to platinum microcoils. It was examined whether an insertion wire, a lightconducting fiber and platinum microcoil can be advanced through a Tracker-18 ( T A R G E T ) microcatheter. The detachment of the attached coil from the lightconducting fiber was investigated. RESULTS: Platinum microcoils can be attached to available lightconducting fibers in a reliable and reproduceable manner, These attached coils together with the fiber can be advanced via a Tracker-18 microcatheter. Only extremely tortuous vessels may increase the friction to critical values. Nd-YAG and Ho-YAG laser sources were used. They allowed the instantaneous detachment of microcoils from 0.125 as well as from 0.21 mm fibres. In contrast to electrolytical detachment, no foreign substances are released to the blood stream, the detachment does not cause waiting time and does not activate thrombus formation. Due to specific technical features, injury to the vessel wall is avoided. C O N C L U S I O N : After further miniaturization and adaption of the lightconducting fibers to the required characteristics, laser detachment may have the potential to replace current methods such as electrolytical and mechanical coil detachment.

E. Berg-Dammer, H. Henkes, H. C Nasher, D. Kiihne.Klinik fi~r Radiologische Diagnostik und Neuroradiologie, Alfried Krupp Krankenhaus, Essen, Germany

48. Basilar top aneurysm: endovascular treatment vs. surgery 46. A new three-dimensional coil for endovascular treatment of intracranial aneurysms

J. A. Horton, D. Vincent Dept. of Radiology, Medical University of S. Carolina, Charleston, South Carolina, USA OBJECTIVE: Since aneurysms are usually nearly spherical structures, we designed coils with a spherical "memory" for placement within them. Our purpose was to test the stability with which they can be placed and remain within aneurysms. D E S I G N A N D M E T H O D : We used the viscoelastic models and pulsatile flow simulator described in the previous paper, and placed a Tracker-18 microcatheter [Target Therapeutics, San Jose, CA, USA] into the target aneurysm. We next introduced one or more coils into the aneurysm and examined changes in flow by injecting food coloring into the fluid stream. RESULTS: Spherical coils had the following properties: 1) They conform to the inside of aneurysms; 2) They are easily nestable, i. e., when one coil is already in place, the next coil fits smoothly and snugly within it; 3) When a critical number of coils are in place, flow in and out of the aneurysm is virtually stopped. CONCLUSION: Spherical coils are promising for the endovascular treatment of intracranial aneurysms. They can be nested within one another and even made in strings that deploy in a nestable fashion. A further iteration is placement of fibers on the final spheres to be deployed in a string so that the flowing bloodstream is not exposed to a thrombogenic (fibered) surface.

Groden, C., Raschdorff, C*, Freckmann, N.*, Herrmann, H.-D.*, Zeumer, H. Dept. of Neuroradiology, *Dept. of Neurosurgery, University Hamburg, Germany Objective: The purpose was to compare the effectiveness of surgical treatment (ST) and the endovascular treatment (ET) of the basilar top aneurysm (BATA). Design and Methods: Since 1980, 70 patients (p) with B A T A have been hospitalised. 7 P died before any treatment could be considered. 33 P were surgical and 30 endovascular candidates. Results : Surgery: 33 patients were considered for a surgical clip. 22 clippings of B A T A were performed. 21 of these were successful. 7 ST couldn't be done for various reasons. Complications: 5 patients had a rebleed before a stable condition for open surgery was achieved. 4 of them died as a consequence. 1 P died as a consequence of surgery. 3 P experienced disabling strokes. 8 P had cranial nerve lesions. Endovaseular: 30 patients were considered for ET (22 GDC-coils, 3 free coils and 3 balloons); 28 ET were performed. 26 of these were successful. 3 ET couldn't technically be performed. From the surgical point of view 10 P of this group were judged to be inoperable. Complications: No rebleed occured in advance of the procedure. 1 P died after balloon deflation, i fatal hemorrhage occured before a treatment with free coils could be completed. 1 P died after a successful ET due to respiratory problems, I P deceased 3 months

$43 after treatment for unknown reasons. 2 P had disabling strokes.1 P suffered cranial nerve lesion. Transient neurological symptoms occurred in 2 cases. Follow-up : No rebleed has occured during a period of up to 4 years yet. Conclusions: ET is an alternative to surgery. In contrast to surgery ET can be applied with minimal delay to hemorrhages also in those cases with a very high surgical risk or even inoperability.

49. A comparison of acute versus elective coil embolisation of intracranial aneurysms

J.. Byrne, A. Molyneux, S. Renowden, A. Hope. Department of Neuroradiology, Radcliffe Infirmary, Oxford, England OBJECTIVE: To compare the operative results and outcome at 6 weeks in patients treated following recent aneurysm rupture with those treated electively. METHOD: Two groups of 70 consecutive patients each were treated by endosaccular packing with Guglielmi detachable coils (Target Therapeutics). Patients in the acute group were treated within 3 weeks of confirmed SAH. They were recruited over 26 months; with mean age 53.12years and treated for 78 aneurysms in 76 procedures. The elective group were recruited over 28 months; with mean age 49.02years and treated for 76 aneurysms in 82 procedures performed at least 1 month after SAH. Patients in both groups were referred because of anticipated difficulty in surgically clipping. RESULTS: Procedural complications due to thromboemboli occurred with similar frequency in both groups but aneurysm rupture only occurred in the acute group (4 patients). Delayed ischaemic complications were also more common in this group. The morbidity and mortality was 16 % and 7 % in the acute and 4 % and 3 % respectively in the elective group. CONCLUSION: We conclude that despite the greater risk of complications in the acute period, coil embolisation is an acceptable alternative to surgical clipping. 50. CT, MRI, MRA and angiographical measurements before and after treatment of cerebral aneurysms with Guglielmi detachable coils (GDC).

G.Schroth, B.Frey, L.Remonda, C. Ozdoba, O.Heid, R. Baumgartner, N. Godoy Depts. of Neuroradiology, Neurology and Neurosurgery University of Bern, Switzerland. Purpuse: To compare the value of CT-angiography, M R I and M R A for planning and control of endovascular treatment of cerebral aneurysms, using electrolytical detachable coils. Design and Methods: 20 patients suffering from cerebral aneurysms were treated with GDC. Size and extend of the neck and fundus of the aneurysms as well as the diameters of the adjacent vessels were measured on the preoperative contrast enhanced CT, MRI, M R A and biplane angiography. Postoperatively the size of the G D C inside the aneurysms and the clinical, M R A and anglographical follow up were analyzed in comparision to the preoperative measurements. Results and Conclusions: CT-angiography using 3 dimensional display outlines exactly the free lumen of the aneurysm and its neck and the relationship to the adjacent yessels. By this way precise adaption of the optimal size of the first G D C is possible, which is important in early treatment of small aneurysms following acute subarachnoidal bleeding. Biplane angiography is most helpful during the endovascular procedure. Using ultrashort echo times, M R A seems to be the optimal method for non invasive follow up of aneurysms, treated with GDC.

5L Endovascular treatment of intracranial aneurysms with mechanically detachable MDS-N coils.

A. Bonafd* J. Berge** P. Holley*, Ph. Arrue*, C. Manelfe* G. Bourbotte*** *Toulouse, **Bordeaux, ***Montpellier P U R P O S E : Evaluation of the efficacy, reliability and safety of MDS-N coils (Balt, Montmorency, France) for occlusion of intracranial aneurysms. M E T H O D S A N D M A T E R I A L S : 33 patients, 20 female, 13 male were evaluated on the day of treatment using Hunt and Hess scale; there was 5 grade 0, 7 grade 1, 7 grade 2, 5 grade 3, 7 grade 4, 1 grade 5 cases. Aneurysms were classified according to their size in small ( < 12mm): 22 cases, large (< 12-25 <): 9 cases, giant ( < 25 mm): 2 cases. 24 were located in the anterior circulation, 9 in the posterior circulation. RESULTS: Technical failure occurred in 4 cases. Complete (13/29) or partial occlusion (16/29) were achieved in 29 cases, with a stable result obtained at a one month follow-up for 25 patients. Subsequently 4 patients underwent surgery. Coil migration in the parent artery (7 cases) resulted in severe deficit (2 cases) transient disability (2 cases) or went unnoticed (3 cases). Aneurysmal sac perforation occurred in one case. CONCLUSION: Prevention from rebleeding is effectively achieved with Tungsten microcoils aneurysmal occlusion. Further developments in mechanical detachment should be pursued to improve the reliability and the safety of the procedure.

52. MR neuro-angiography: An update (Invited lecture)

Z. Patay (Brussels, Belgium - Budapest, Hungary) 53. Blood flow evaluation in normal and pathologic carotid artery circulation with dynamic magnetic resonance flow imaging

K. W. Neff, M. Daffertshofer, S. Meairs, M. Hennerici, A. Schwartz. Dept. of Neurology, University of Heidelberg, Klinikum Mannheim, Germany Objective: The noninvasive measurement of blood flow velocity and volume flow in carotid artery circulation was performed with two different, independent MR techniques, dynamic M R inflow tracking and 2D phase-contrast cine M R imaging. Design and Methods: Carotid artery circulation of common (CCA) and internal (ICA) carotid artery was examined in 12 healthy volunteers and in 15 patients with carotid artery disease (ICA-stenosis at the level of the bifurcation, carotid artery dissection and distal ICA-stenosis or occlusion). Up to now 4 patients with ICAstenosis had flow evaluation before and after surgical intervention (TEA). Blood flow velocity and volume flow was measured with two different, independent M R techniques, dynamic M R inflow tracking, a modification of M R presaturation bolus tracking, and 2D phase-contrast cine MR imaging. Both techniques allow blood flow velocity and especially volume flow determination. Additional measurement of blood flow velocity was performed with Doppler ultrasonography. Results: Comparison of blood flow velocity for both M R techniques and Doppler ultrasonography in CCA and I C A demonstrated better correlation for dynamic M R inflow tracking (r < .93) than for phase-contrast velocity mapping (r < .82). Correlations for volume flow determination with dynamic M R inflow tracking and 2D phase-contrast MR imaging in CCA and I C A were r < .86. Clinical results for patients with carotid artery disease demonstrated preand poststenotic volume flow reduction in CCA and ICA. Volume flow reduction in patients with ICA-stenosis at the level of the bifurcation correlated well with the degree of stenosis. Both techniques documented a significant increase of flow after carotid artery surgery ( T E A of ICA) Conclusion: The results demonstrate that in vivo flow velocity and volume flow can be evaluated well with dynamic MR imaging. The presented studie quantified volume flow evaluation with two dif-

$44 ferent, independent MR techniques and allowed in vivo volume flow correlations.

56. MR angiographic screening for intracranial aneurysms: is it worthwhile?

N. J. Ferris. Department of Radiology, Alfred Hospital, Prahan, 3181, Melbourne, Australia 54. MR angiography of the carotid bifurcation: comparison of 2DTOF and 3D-TOF with variable flip angle (TONE)

G. Pittiglio, N. Anzalone, R. Chiesa ~ R. Castellano o, A. Grossi ~ G. Scotti. Depts. of Neuroradiology and Vascular Surgery ~ Scientific Institute HS Raffaele, University of Milan, Italy. PURPOSE: To compare 2D and 3D-TOF with variable flip angle (TONE) M R A with intra-arterious D S A in the evaluation of carotid bifurcation stenosis. M A T E R I A L A N D M E T H O D S : 40 consecutive patients, 73 carotid bifurcations, were studied up to now. Intra-arterious D S A and Doppler ultrasonography were performed in all the patients. Axial 2D-TOF (TR 30, TE 9, FA 35 ~ 54 siices, 3 mm thick, 1,5 mm overlap) with travel saturation and axial 3D-TOF (TR 33, TE 8, 64 partitions, 1 mm thick) with T O N E were performed with same FOV (200 mm) and matrix (252 x 256) using a transmitter neck coil. RESULTS: 40 carotids exhibited a stenosis greater than 70 %. 2DTOF misdiagnosed 13 carotid stenosis: 6 under, 7 over-estimation. 3D-TOF with T O N E misdiagnosed 7 carotid stenosis: 5 under- and 2 over-estimation (one occlusion vs. severe stenosis). CONCLUSION: 3D-TOF M R A with T O N E showed an overall better sensitivity than 2D-TOF M R A technique in the evaluation of carotid bifurcation stenosis. Nevertheless, considering stenosis greater than 70 %, sensitivity of 2D-TOF resulted to be slightly superior (97.5 % vs. 95 %) than 3D-TOE In conclusion only potential pitfall of 3D-TOF, even with TONE, seems to concern the diagnosis of very high grade stenosis with extremely slow flow that can be made more properly by 2D-TOE

55. High resolution magnetic resonance angiography in acute subaracnoid hemorrhage: Comparison with IA DSA and surgery

R. Gasparotti, M. Crispino, E Prandini, M. Bonetti, G. Galli*. Dept. of Radiology and* Neurosurgery, University of Brescia, Italy P U R P O S E : to assess the diagnostic accuracy of high resoiution Magnetic Resonance Angiography (HR M R A ) in patients with acute subarachnoid hemorrhage and to verify the clinical usefulness of targeted MIP projections reproducing the surgical anatomical view. D E S I G N A N D M E T H O D S : 53 patients with acute subarachnoid hemorrhage were examined with high resolution M R A within 2436 hours from bleeding and subsequently underwent intraarteriaI digital subtraction angiography (IA DSA). In adjunction to conventional axial, coronal and sagittal views, in each patient targeted MIP projections have been obtained with the aim of reproducing the surgical perspective of the pterional approach for the aneurysms of the Willis polygone. M R A and I A D S A findings were compared with operative findings. RESULTS: M R A detected 61 out of 63 aneurysms in 50 patients, with 2 false negatives. Fifty-one aneurysms were microsurgically clipped, while 4 occluded by endovascular treatment. Relative to IA DSA, M R A exhibited 96.8 % sensitivity and 100 % specificity. The smallest aneurysm detected by M R A was 2.5 mm in diameter. Compared with IA D S A M R A showed a tendence to underestimate the aneurysm size of the aneurysms with maximum diameter of 1 cm or more. MIP projections oriented along the surgical pterional perspective correlated well with the operative findings. CONCLUSIONS: this study demonstrated M R A feasibility and reliability also in emergency cases. Even though IA D S A is still recommended preoperatively, high resolution M R A can be considered a complementary examination which can be employed as a useful guidance to the surgical pterional approach, since it depicts vessels without superimposition and with degrees of rotation not usually obtainable with I A DSA.

O B J E C T I V E : To review our experience with the use of M R angiography (MRA) in screening for intracranial aneurysms. D E S I G N A N D M E T H O D S : The indications for all cerebral M R A studies performed at our institution from March 1993 to January 1995 were reviewed. All patients who had first degree relatives who were said to have had a history of subarachnoid haemorrhage, or proven intracranial aneurysm, were included. M R A was performed with a 3D time of flight technique in almost all cases. The original reports were reviewed retrospectively, and compared with findings at any subsequent angiography and surgery. R E S U L T S : Of 343 cerebral M R A studies performed in the study period, 72 met the criteria for inclusion in the study. 3 of the 72 studies ( 4.2 %) were positive for intracranial aneurysm ; one of these patients had 2 aneurysms~ All 4 aneurysms were confirmed at angiography, with no additional aneurysms found. Of the three positive cases, one patient has had her aneurysm successfully clipped. The patient with two aneurysms has had both clipped, sustaining moderate deficits on each occasion. The third patient has yet to undergo surgery. C O N C L U S I O N : The yield of M R A screening for aneurysms in our population appears low. Greater numbers will be required to allow accurate assessment of the cost-benefit ratio of this program (updated results will be presented).

57. MR-angiography versus CT-angiography in patients with giant aneurysms

K E W Eberhardt 1, HP Hollenbach 2, U B6ttcher 2, WJ HuU, P Weber2, J Regn 21Div. of Neuroradiology, Dept. of Neurosurgery, University of Erlangen-Narnberg, Erlangen, Germany, 2Siemens Meidcal Engineering Group, Erlangen, Germany Purpose: In a comparative case study 20 patients with giant aneurysms have been examined with Phase Contrast-MR-Angiography (PC-MRA) and CTA in conjunction with DSA. Material and Methods: The MR investigations were performed using a 1.5 T whole body MR-system (Siemens Magnetom Vision) and the CT examinations with a spiral CT scanner (Siemens Somatom Plus 4). We used for MR-Angiography 2D-PC- (TR = 79ms. T E = 7-11ms, a = 11 deg) and 3D-PC-(TR = 76ms, TE = 8-11ms, a = 15 deg) sequences. For both, the M R A and the CTA a maximum Intensity Projection Program was used. The flow velocity encoding varied between 5 and 100 cmAec. Results: Our results demonstrate that P C - M R A and CTA can be superior to D S A in the diagnosis of giant aneurysms. If the neck of an aneurysm cannot be visualized with DSA, PC can help to identify the branches and their relationship to the neck by analysis of the jet effect, but only CTA was able in these cases to visualize the neck of the aneurysms directly, using both, 3D- surface and 2D MIP techniques. Conclusions: PC-MR-Angiography and CTA offer very interesting new aspects in the diagnosis of cerebral vascular diseases. Both methods can be superior to D S A in patients with giant aneurysms and can help to better plan the therapeutic procedure. 58. 3D surface reconstruction MR angiography of intracranial aneurysms J-L Dumas, A Abdennebi, D Belin (1); R Cohen (2), J Moret (3), D

Goldlust. Services de Radiologie, Neurologie (1), Endocrinologie (2), H@ital Avicenne, Bobigny, France. Service de Neuroradiologie Interventionelle (3), Fondation A. de Rothschild, Paris, France. O B J E C T I V E : To determine the usefulness of three-dimensional phase-contrast (%D PC) M R A with surface reconstruction in

$45 characterizing intracranial saccular aneurysms greater than 10 mm in size. DESIGN AND METHODS: 6 patients referred for non invasive confirmation of unruptured aneurysm detected by conventional CT and/or MR imaging were studied by 3D PC (VENC of 20 cm/ sec, flow encoding 3 axes) in axial acquisition localized on the circle of Willis using a 1.5 T unit (Vectra, GE). We compared architectural informations obtained by 3D maximal intensity projection (MIP) and 3D surface reconstructions. Pretherapeutic conventional angiography was performed in 2 patients. RESULTS: In all cases, both reconstruction techniques equally confirmed the diagnosis, the number and location of aneurysms. The best analysis of size and shape of aneurysmal sac was obtained with 3D surface reconstruction. By using an endovascular approach this technique permitted to visualize the aneurysmal neck at the level of the lumen of the parent vessel. CONCLUSION: 3D PC with surface reconstruction is a very effective non invasive technique in aneurysms greater than 10 mm in size to predict therapeutic management, especially useful in cases of unruptured aneurysms.

59. 3D PC-angiography of cerebral AVM's combined with 3D rendering of cortical brain.

C. Fonda, *B. Bernardi, **R. Canapicchi, **M. Puglioli. Neuroradiology Misericordia e Dolce Hospital, Prato. *Neuroradiology Bellaria Hospital, Bologna. **Neuroradiology S. Chiara Hospital, Pisa. OBJECTIVE. To test the value of the PC-Angiography and of the combined 3D rendering of both vessels and brain surface in the evaluation of AVM's. MATERIAL AND METHODS. Fifteen patients with intracranial AVM's, evaluated by intra-arterial digital subtraction angiography (DSA), were also studied by PC-MRA. 2D-PC Angiographies were performed with different flow velocity encoding in order to choose the best compromise to visualize the arterial feedings, the venous dreinages, and the nidus of the AVM's. The complete representation of the intracranial circulation using 3DPC MRA, with appropriate flow velocity encoding, was followed by selective and/or subselective studies. These were obtained applying presaturation bands on undesirable vessels. The recent software applied on workstation allowed us to obtain high quality collapsed MRA with evidence of specific volumes and subvolumes. Simultaneos isocentric acquisition of stationary tissue images and PC Angiographic images were obtained. Stationary images that were partitions for 3D brain surface rendering were overlapped on the 3D angioarchitecture of the AVM. RESULTS AND CONCLUSIONS. The PC-Angiography with 3D rendering of both vessels and brain surface results in useful pictorial views of the anatomopathological topography of the AVM. The 3D PC selective MRA helps in the understanding of the different flow compartments of the intracranial AVM. DSA still remains the best m0dality to analyze the vascular structures of a AVM and is need to plan the surgical, radiosurgicat, and/or endovascular treatment. Nevertheless, the PC Angiography with 3D rendering of both vessels and br~in surface is a noninvasive modality helpful in demonstrating the complex anatomy of the cerebral AVM and surrounding brain surface.

60. Value of 3D high resolution imaging with CT angiography (CTA) and MR angiography (MRA) for neurosurgical operation planning

K.R.Atzor /1/, M.Mazdnek /2/, H. UKauczor /3/, J.Tintera /1/, J.Kalder /1/, K.Resch /3/, U Jendrysiak /4/, A. Pemeczky /2/, P. Stoeter /1/ /1/ Dept. of Neuroradiology, /2/ Dept. of Neurosurgery, / 3/Dept. of Radiology, /4/Institute of Med. Statistics and Documentation, University of Mainz, Germany PURPOSE: Spiral CT angiography (CTA) and high resolution MR angiography (MRA) without background suppression including postprocessing with 3D visualization was carried out in 20 patients with cerebral aneurysms and skull base tumors in order to compare the diagnostic sensitivity and value in planning minimal invasive neurosurgery. METHODS AND MATERIALS: CT technique: Spiral scan studies were performed on a Siemens Somatom Plus S scanner with contrast agent with a slice thickness of 1 mm. MR technique: High resolution imaging was performed by a gradient echo technique on a Philips ACS with a slice thickness of i mm, voxel size 0.4 x 0.6 x 0.8 (-1..2) mm. Reconstruction: Segmentation and 3Dreconstruction of vessels, tumor, nerves, bones and brain were done by thresholding and seeding with colour codation on an Allegro workstation (ISG) including animation of the 3D objects with an experienced neurosurgeon. RESULTS AND CONCLUSION: Optimizing CT and MR protocols for high resolution imaging and improving postprocessing procedures with the assistance of neurosurgical experience and monitoring during operation will provide additional information to minimize surgical trauma in removing aneurysms and tumors located mainly in surgically difficult regions such as the cranial base. Although a trained neurosurgeon will be able to understand 50-60 2D images from CT and MR by transforming them mentally into a 3D aspect of the situs, the comprimation of the important information into a few 3D images is extremely useful for a preoperative discussion of different approaches and for fast orientation during the operation procedure. Because this postprocessing is connected with loss of information the basic diagnosis should be made from the original slices. In addition the neuroradiologist must be able to differentiate between surgically important and less important information to work out useful and correct 3D objects and views. Segmentation of the surgically relevant information is a great responsibility, It also affords profound neurosurgical and interventional knowledge and experience to avoid errors and transmission of false information necessitating close cooperation and teamwork.

61. The veins of the posterior fossa. MR features and topographical anatomy.

M. Braun, Z Roland, S. Bracard, P. Bertaux, L. Picard. Department of Diagnostic and Interventional Neuroradiology, University Hospital Nancy, France OBJECTIVES: To determine the normal MR patterns and the principal landmarks of the veins of the posterior fossa relative to the arteries, to the brain and to the cerebellar structures, PATIENTS AND METHODS: 40 consecutive patients were evaluated in this prospective study. Axial and frontal planes after Gd intravenous injection (0,2 mmol/kg) were obtained with an MR scanner performing at 1.5 T. 3D fast gradient echo imaging sequences (Tlw) were used allowing 128 continuous slices (thick.: 1,2 mm). All patients presented with normal posterior fossa structures. RESULTS: After injection, the veins enhanced constantly and recognized on the basis of their topography and connections. The arteries were separately studied even in high signal with the 3D Gradient Echo sequence. In each cases, the presence or the absence of enhancement within the veins were noted. The following veins of the brain stem were studied on each side and were present in all cases: the perioccipital sinus, the vein of the lateral recess of

$46 the fourth ventricle, the superior petrosaI vein, the lateral mesencephalic vein, the cerebellar and vermian veins, the precentral vein; the following veins were unconstantly found: inferior petrosal vein (10 %), tonsillar vein (30 %), superficial veins of the pons (40 %), lateral medullopontine veins (13 %). CONCLUSIONS: The topography of the arteries and veins is important to know and to identify on native axial, sagittal or frontal MR planes. During the stereotactical procedures, the needle course is easily planed to avoid the main vessels. As pre-operative angiograms are less and less performed, the main connexions between a tumor and the surrounding veins or arteries must be evaluated on simple M R scans.

62. Sinus venous thrombosis: a new diagnostic approach with digital subtraction magnetic resonance venography

K.W. Neff, S. Meairs, J. K~thnen, M. Hennerici, A. Schwartz. Dept. of Neurology, University of Heidelberg, Klinikum Mannheim, Germany Objective: A new digital subtraction MR venography technique was investigated to obtain a diagnostic-quality MR venogram. Design and Methods: Digital subtraction MR venography was performed using a pre- and postgadolinium strongly Tl-weighted M P R A G E sequence (turbo gradient echo, TR/TE/flip/FoV/matrix/ Aq. - 10ms/4ms/ 10deg/250 mm/130"256/1). Slice-thickness was 1.25 mm, 128 slices, aquisition time 3min. This technique was applied to ten healthy volunteers (mean age 46 years, range 1868 years) and eigth patients (5 female, 3 male, mean age 43 years, range 17-66 years) with sinus venous thrombosis (SVT) in addition to routine MRI of the brain including gradient echo 'time of flight' (Flash-2D) MR venography. Patients with SVT had up to three follow-up investigations to demonstrate recanalization. M R venography volume data were obtained from digital subtraction of precontrast M P R A G E images from corresponding M P R A G E images acquired after administration of 20ml Gadolinium-DTPA. Raytraced volumetric rendering on MR venography volume data was performed using a 4 • 4 transformation matrix describing rotations and translations applied to the volume data. Results: In all normal controls and in patients with SVT digital subtraction MR venography was superior to conventional postprocessed (MIP 3D) 'time of flight' M R venography in depiction of sinus venous anatomy and in demonstration of internal and cortical venous structure. A pathological sinus signal with occlusion or venous flow reduction of the thrombosed vessels was depicted in all patients with SVT. The direct visualization of thrombosed material in one or more sinuses was possible for all patients with SVT in post-contrast M P R A G E images. Conclusion: Application of digital subtraction M R venography obtained from postprocessed pre- and post-contrast strongly T1weighted volume M P R A G E imaging improves M R I in demonstration of cerebral venous anatomy and depiction of SVT and results in a diagnostic quality M R venogram.

63. Radiology of early stroke (Invited lecture)

K. Sartor (Heidelberg, Germany) 64. Measurement of cerebral blood flow in focal ischemia: correlation of perfusion-weighted MRI with autoradiography

W. Reith:, S. Heiland, H. Schr6ck 2, M. Forstingt, A. Bergmann 1, W. Kuschinsky 2, K. Sarto/. 1Dept. of Neuroradiology, and 2 Dept. of Physiology, University of Heidelberg Medical School, D-69120 Heidelberg, Germany -- . Introduction: Perfusion imaging is a non-invasive method to assess regional hemodynamics. The present study had the aim to compare MRI perfusion imaging parameters with quantitative CBF measurements using the iodoantipyrine method.

Material and Methods: In 31 rats ischemic infarction was induced by encdovascular occlusion of the middle cerebral artery. Perfusion imaging was performed after 30 rain to 240 min occlusion time using a T2*-weighted FLASH-sequence acquiring 20 coronal images during administration of a Gd-DTPA bolus. The observed signal intensities were converted pixel-by-pixel into concentrationtime-curves. From these curves rrCBV, MTT, and rrCBF were calculated. Then 14C iodoantipyrine/100g body weight was injected intravenously to measure CBF quantitatively. Autoradiographie and M R data were compared in identical coronal sections of the same animals. Results: MR perfusion measurements showed that MTTin ischemic tissue was significantly increased. RrCBF in the ischemic hemisphere was decreased to 85 % of the non-ischemic hemispheric value after 30 rain occlusion time and decreased to 41% after 180 rain occlusion time. Residual blood flow in the ischemic hemisphere was 30 % less with autoradiography than the rrCBF measured with MR for short occlusion times ( < 60 min). For occlusion times of more than 120 min no significant differences between both methods occured. The regional extent of reduced perfusion as determined by autoradiography correlated with that determined by M R perfusion imaging within the limits of spatial resolution for long as well as for short occlusion times. Conclusion: The results show a high correlation of regional cerebral blood flow as determined by MR perfusion imaging and autoradiography during focal ischemia. It is concluded that M R perfusion imaging is a technique providing reliable information about perfusion. Therefore M R perfusion imaging appears to be well suited for therapy monitoring. 65. Reperfusion in focal cerebral ischemia: Comparison between autoradiography and MR perfusion imaging

M. Forstingd, H. SchrOckz, W. Reith 1, S. Heiland 1, A. Bergmann l, W. Kuschinsky 2, K. Sarto/ Department of Neuroradiology: and Physiology 2, University of Heidelberg, Germany Objective: Ischemic brain injury may be reduced if blood flow is restored after a brief period of ischemia. We wanted to evaluate the local changes of cerebral blood flow after varying times of transient occlusion of the middle cerebral artery (MCA) using the iodantipyrine method (gold standard for rCBF) and M R perfusion imaging, Design and Methods: In 44 rats ischemia was induced by endovascular M C A occlusion. Reperfusion after 30 to 240 minutes was achieved by withdrawing the occluder within the MR unit. MR perfusion imaging was done using the bolus tracking method and a T2*-weighted F L A S H sequence with a time resolution of 1.6 s/ image. MR signal intensities were converted to concentration-time curves and perfusion maps. For technical reasons autoradiography took place about 25 rain. after the last M R measurement. Results: M R perfusion imaging clearly demonstrated the ischemie area during occlusion in each animal. A fraction of less than 3 % of the hemisphere showed a reduction of R C B F to < 20 ml/min/100 g for an occlusion time of less than 120 min, whereas this area increased to 40 % after an occlusion time of 240 mill. There was an excellent correlation of both methods used, Conclusiom MR perfusion imaging correlates well with absolute CBF values obtained by autoradiography. In this model reperfusion after more than 2 hours of M C A occlusion results in an aggravation of ischemia. M R perfusion imaging is a reliable tool to monitor cerebral perfusion during ischemia and reperfusion and offers the opportunity for clinical application in therapeutic trials.

$47

66. Multislice Diffusion and Perfusion Weighted MRI for Monitoring Unfractionated Heparin Treatment in a Thromboembolic Stroke Model

S. Heiland 1, B. Elger2, M. Forstingfl, C. Gehrlein 1, W. Reith 1, J. Seega2, K. Sarto/1Dept. of Neuroradiology, University of Heidelberg Medical School, and 2Research and Development, Knoll A G, Ludwigshafen, Germany Purpose: In this study we used multislice diffusion and perfusion weighted echo planar imaging (EPI) to monitor the evolution of ischemia over time and to monitor the perfusion state in a thromboembolic stroke model in rats when the animals were treated either with heparin of saline. Material and Methods: In 20 rats embolic ischemia was induced by injecting blood clot suspension into the ICA. 90 minutes after injection of the embolus, animals were either treated with continuous intravenous infusion of saline (N = 10) or with a bolus of 2000 IU/kg unfractionated heparin (Liquemin N25000, HoffmannLa Roche) followed by infusion of 500 IU/kg heparin over 60 minutes (N = 10). MR data were obtained using a G E CSI II 2.0 T imaging spectrometer beginning 60 min after onset of ischemia up to 180 min after occlusion. Eight slice diffusion weighted EPI were acquired with 10 increasing b-values. From these data the apparent diffusion coefficient (ADC) was calculated. Bolus tracking was performed using 4 slice gradient recalled EPI. Images were performed acquired continuously 5 s before and up to 15 s after bolus injection of Gd-DTPA. From the perfusion-weighted images we calculated the cerebrovascular parameters. Results: Although there was no significant improvement in perfusion after treatment, 3 out of 10 treated animals showed a marked reduction of MTT and a significant increase in rrCBV and rrCBE A D C maps showed the evolution of ischemic volume over time. There was no significant difference in A D C of heparin and saline treated rats. Conclusion: Multislice diffusion- and perfusion mapping yield reliable parameters for early diagnosis and monitoring of focal cerebral ischemia. These imaging modalities and calculation of A D C maps and perfusion maps can be used to monitor different treatment strategies and may help to find optimal treatment regimens for stroke.

CONCLUSIONS: In some difference to the few reports about LIF in combined occlusions of the distal internal carotid and middle cerebral artery we observed a good recanalization and clinical outcome in 1/3 of patients. LIF is recommended in this condition, as an unfavorable natural course with large infarctions and severe brain edema is probable. 68. Effects of thrombolytic treatment on the hyperdense middle cerebral artery sign.

C Manelfe* E. yon Kummer** S. Bastianello*** L. Bozzao*** *Dept. of Neuroradiology, HOpital Purpan, Toulouse, France, **Neuroradiology, Heidelberg, Germany, ***Neuroradiologia, La Sapienza, Roma, Italy, All from ECASS Group. OBJECTIVE: The hyperdense middle cerebral artery sign (HMCAS) is considered to be specific for middle cerebral artery (MCA) trunk occlusion. We studied the effects of systemic thrombolytic treatment on HMCAS. D E S I G N A N D M E T H O D S : Three neuroradiologists blind to clinical information prospectively evaluated the CT scans of < 6 hours, day 1, and day 7 after hemispheric stroke of the intent-to treat group (620 patients) treated with intravenous recombinant tissue plasminogen activator (rt-PA, 1,1 mg/kg) (n = 313) or placebo (n = 307) in a double blind, randomized multicenter trial. RESULTS: Prior to therapy, 61/307 of patients of the placebo group (20 %) and 46/313 patients of the rt-PA group (15 0 showed the H M C A S on initial CT scan (p < 05). In the placebo group, the number of patients with H M C A S dropped to 16 ((48/307) at day 1 and to 8 % (24/307) at day 7. In the rt-PA treated group, the number of patients with the H M C A S dropped to 9 % (28/313) at day i (p < 05 vs. placebo) and 5 ((16/131) at day 7. CONCLUSIONS: After ischemic stroke the incidence of the H M C A S decreases considerably within one week. The decrease is significantly accelerated by systemic thrombolytic therapy with rtPA within the first 24 hours. This may reflect the effect of rt-PA on the thromboembolus of the M C A trunk. CT is able to show effects of thrombolytic treatment after ischemic stroke.

69. Early CT predictors for parenchymal hemorrhage in stroke patients treated with systemic rt-PA. 67. Local intraarterial fibrinolysis in combined occlusions of the distal internal carotid and middle cerebral artery

H. Urbach, B. Ostertun, L. Solymosi. Dept. of Neuroradiology, University of Bonn, Germany OBJECTIVE: Assessment of efficacy and safety of local intraarterial fibrinolysis (LIF) in combined occlusion of the distal internal carotid and middle cerebral artery. Aim of the LIF is the early recanalization of the occluded vessel. LIF must be considered as a supportion of the spontaneous recanalization, which usuallly does not happen fast enough to prevent stroke. The need for it is therefore especially given in occlusions with large thrombotic masses. METHODS: In 12 patients (7 m, 5 f; 26-75, mean 51 yrs) with combined occlusions of the distal internal carotid and middle cerebral artery LIF was started within 90-270 (mean 185) rain after onset of symptoms. An infusion of up to 106 I U Urokinase was given through a microcatheter for at most 2 hours. RESULTS: 4 patients had a good clinical outcome with minor neurological deficits (Barthel index < 85), 4 patients showed major neurological deficits after 3 months (Barthel index < 45), 4 patients died, in one of them an intraparenchymatous hemorrhage in a M C A infarction occured. A strong correlation of recanalization and clinical outcome was found in 10/12 patients. In one patient with excellent leptomeningeal collateral supply a good clinical outcome was reached despite moderate recanalization, in one patient with good recanalization an intraparechymatous hemorrhage occcured.

Bozzao L, Bastianello S, Colonnese C, Manelfe C, yon Kummer R, Fieschi C for the ECASS Group. Chair of Neuroradiology, Dept. Neurological Sciences, University Rome, Italy. Objective: Aim of the study is to evaluate the relationship between the frequency of parenchymal hemorrhage in placebo and rt-PA treated acute ischemic stroke patients and the presence of early CT hypodensity at entry scan performed within 6 hours from the onset of symptoms. Materials and Methods: 615 acute ischemic stroke patients (intent to treat) have been enrolled in a multicentric study (310 rt-PA treated and 305 placebo patients). Entry CT scans were evaluated by local neuroradiologists looking for the presence of early CT hypodensity exceeding the 33 % of the M C A vascular territory. According to the CT inclusion criteria of the study, patients with early CT hypodensity more than 33 %, should have been excluded. At the end the study CT scans were reviewed, blind to clinical and therapeutical data, by the CT panel reading of the ECASS study (BL,MC,vKR) in order to evaluate the number of intraparenchymal hemorrhage (PH) and the extent of early CT hypodensity at entry CT scan; a target population was so individualized. Results: 20 PH were observed in the intent to treat placebo group, while 61 PH were seen in the treated group. Early CT hypodensity exceeding 33 % was seen by the CT panel reading in 66 patients, 30 of the placebo and 36 of the rt-PA treated group. PH was seen in 2 of the placebo and in 11 of the rt-PA treated group. In the target population we have therefore 18 and 50 PH in the placebo and rtPA treated group respectively.

$48 Conclusions: Our data show that systemic rt-PA treatment increases the occurrence of PH in acute ischemic stroke. In the treated population (intent to treat) bleeding occurrence was higher in patients with early CT hypodensity at entry scan (30 %, 11 out of 36 vs. 18 %, 50/274). A careful evaluation of CT findings at entry CT scan seems to be very important in selecting patients for rt-PA trials.

70. CT monitoring of parenchymal hypodensity in ischemic stroke: effects of thrombolytic treatment

R. yon Kummer, S. Bastianello, L. Bozzao, C. Manelfe, for the ECASS Group. Dept. of Neuroradiology, University of Heidelberg, Germany OBJECTIVE: To study the effects of systemic thrombolytic treatment on the extent of hemispheric infarction. DESIGN AND METHODS: Three neuroradiotogists blind to clinical information prospectively evaluated the CT scans of < 6 hours, day 1, and day 7 after hemispheric stroke of the intent-totreat group (620 patients) treated with intravenous recombinant tissue plasminogen activator (rt-PA, 1,1 mg/kg) (n = 313) or placebo (n = 307) in a double blind, randomized multicenter trial. RESULTS: Prior to therapy, 168/307 of patients of the placebo group (55 %) and 167/313 patients of the rt-PA group (53.4 %) showed no parenchymal hypodensity on initial CT scan. In the placebo group, the number of patients without parenchymal hypodensity dropped to 18.6 % (57/307) at day 1 and to 16 % (49/307) at day 7. In the treated group, significantly more patients showed a normal CT scan at day 1 (26,5 %). At day 7, with 22 % (69/313) the incidence of normal CT scans tended to be higher (p < .1) in the rtPA group. CONCLUSIONS: Systemic thrombolytic therapy with rt-PA increased the proportion of patients without parenchymal hypodensity significantly by 7.9 % one day after ischemic stroke. This difference was slightly smaller at day 7. CT is able to show the efficacy of thrombolytic treatment after ischemic stroke.

71. Antiphospholipid antibodies primary syndrome (AAPS)= neuroradiologic findings in 11 symptomatic patients.

M Boukobza, JP Guichard, J Guigui, F Woimant* D Reizine, JJ MerlandDept of Neuroradiology and of *Neurology, Hopital Lariboisiere Paris, 75010 France. Purpose: The aim of this study is to describe the neuroradiologic findings in patient with AAPS. Material and Methods: 11 patients (22 to 47 years) were evaluated by CT, MRI (T1 and T2-weighted images, MR Angiography and 5 underwent cerebral angiography. Results: Abnormal findings on MRI occurred in 9/11 patients. Findings included infarcts (n = 7), multiple in one case, large in two cases, thalamic in one, of the brain stem in two and one hemorrhagic venous infarction in one case, in relation with lateral sinus thrombosis, visualized on MR. Multiple lacunar infarct, were detected in one case. Cortical atrophy was posent in 3 cases. Areas of white matter hyperintensity on T2-wi were found in two cases. Abnormal angiographic findings included arterial narrowing that simulated vasculitis (n = 1) and diffuse arterial narrowing (n = 1). Conclusion: AAPS should be suspected when no cause is apparent for either ischemic cerebrovascular disease or cerebral venous thrombosis in young adults.

72. Contrast enhanced MRI of acute and subacute stages of thromboembolic brain infarction, using Bis-Gd-MP and USPIO's as MR contrast agent

G. Schneider, C. Hayd, T. Hagen, D. Gohl, *W. Ebert, *D. Pfefferer, *W. Semmler, B. Kramann. Dep. of Diagnostic Radiology, University Hospital, 66421 Homburg/Saar, Germany, *IDF (Institut far Diagnostiforschung) an der FU Berlin, 14050 Berlin, Germany OBJECTIVE: To investigated the contrast behaviour of Bis-GdMesoporphyrine (Bis-Gd-MP) and USPIO's in contrast enhanced MRI of acute and subacute stages of exp. thromboembolic brain infarction in rabbits. METHODS: Exp. brain infarctions in rabbits were induced by transfemoral catheterization of the internal carotid artery and injection of embolization particles. The contrast behaviour of USPIO's (200 ~tmol/kg BW)was investigated 8 hours, 2 and 11 days after the onset of brain infarction, applying T2w SE-Sequences (TR/TE = 2500/90) before and up to 60 rain after intravenous contrast media administration (CMA). Bis-Gd-MP (50 ~mol/kg BW) was administered to animals 11 days after induction of brain infarction, using T l w SE-Sequences (TR/TE = 350/15) before and in 15 rain intervals up to 3 hours after CMA and a late postcontrast image after 16 h. RESULTS: In unaffected brain tissue USPIO's lead to a strong postcontrast decrease of SI (40 %) in T2w MR-images, whereas SI of infarcted brain tissue in acute stages (8 h resp. 2 days after brain infarction) nearly remained unchanged, resulting in a strong increase of the contrast ratio (CR) between unaffected brain tissue and infarcted areas (precontrast: 1.2-1.4, postcontrast: 2.1-3.2). 11 days after induction of brain infarction, also a slight postcontrast decrease of SI in infarcted areas (16 %) could be observed, but still a strong increase of the CR could be demonstrated. In T l w BisGd-MP enhanced MRI a specific accumulation of the contrast medium in necrotic tissue could be observed, demonstrating an increase of SI up to 56 %. This contrast enhancement resulted in an increase of the CR between unaffected and infarcted brain tissue from 1.1-1.2 precontrast to 1.6-1.8 16 h after CMA. The results of contrast enhanced MRI showed good correlation with histologic findings. CONCLUSIONS: USPIO's and Bis-Gd-MP seem to be promising new contrast agents in MRI of acute and subacute stages of brain infarction, allowing a much better assessment of the amount of infarcted brain tissue comPared with unenhanced T1 w or T2 w images. 73. Prevalence of vertebral artery dissection in Wallenberg syndrome: Neuroradiological analysis of 93 patients

T. Hosoya, M. Nagahata, K. Yamaguchi. Department of Radiology, Yamagata University School of Medicine, Yamagata, Japan

Objective: To assess the prevalence of vertebral artery dissection in Wallenberg syndrome (lateral medullary syndrome). Design & Methods: Vertebral artery dissection has been emphasized as a cause of Wallenberg syndrome (Hosoya et al., AJNR 1994; 15:1161-1165). This study assessed the prevalence of vertebral artery dissection based on a review of a large number of patients with Wallenberg syndrome. Records of 93 patients (70 men, 21 women, 2 patients not identified by gender: mean age at ictus, 58.1 years) with symptoms of Wallenberg syndrome who underwent MRI and/or vertebral angiography in 26 hospitals were reviewed retrospectively. Study items including the follows: (A) headache as clinical signs; (B) intramural hematoma on Tl-weighted images as an MR finding; (C) double lumen, intimal flap, pearl&string sign, and string sign as angiographic findings. These abnormal findings suggesting vertebral artery dissection were scored as 1 or 2 points. According to the total scores, each patient was classified as "definite dissection", "probable dissection", "suspected dissection" or "no dissection". Results: Twenty three of 93 patients were classified as "definite dissection", 23 as "probable dissection", 27 as "suspected

$49 dissection" and 20 as "no dissection". The frequency of definite or probable dissection in Wallenberg syndrome was estimated 50 % or more. Conclusions: Vertebral artery dissection seems to be the most frequent cause of Wallenberg syndrome. 74. Subcortical hemorrhage in dural sinus thrombosis

S.E.S. Ng*, R.I. Grossman** H.K. Boey* W. Y.. Cheong*** K. P Tan****. Depts. of Diagnostic Imaging, *Tan Tock Seng Hosp., Singapore, **Hosp. U. of Pennsylvania, USA, ***Advaced Imaging Cente~ Singapore, ****Singapore Gen. Hosp., Singapore OBJECTIVE: Dural sinus thrombosis (DST) is a potentially fatal disease with protean and non specific manifestations. Expeditious diagnosis is crucial, since prompt treatment can improve survival and lower morbidity. Studies on DST have focused largely on findings within the venous sinuses. However these findings often require contrast enhancement, specialized acquisitions and additional views for confirmation. This study sets out to establish the utility of subcortical hemorrhage (SCH) as a diagnostic sign for DST. D E S I G N S A N D METHODS: The images of thirty seven cases of DST diagnosed by MR and/or CT imaging over a four year period were evaluated retrospectively. CT scan was performed on all cases while twenty four had MRI. The records were also reviewed for any delay in the diagnosis, the reasons and subsequent steps taken to establish DST. RESULTS: The incidence of SCH was 46 % (seventeen eases). SCH was the solitary finding in 12 cases (71%). Of these 12 cases the diagnosis of DST was entertained in only 2 cases (16.7 %). This lead to a delay in diagnosis, and clinical deterioration, for 10 cases. Diagnosis of DST was clinched through subsequent MRI (5 cases) and angiography (5 cases). CONCLUSIONS: SCH is an important sign in the diagnosis of DST. It occurred as the solitary initial finding in 12 of our cases. The delay in in the diagnosis of DST for 10 cases could have been avoided had the possibility of DST, based solely on SCH, been considered.

75. Pre- and postoperative findings in cavernomas of the CNS diagnostic limitations and pitfalls

M. Mull, J. Reinhardt, H. Bertalanffy* A. Thron. Departments of Neuroradiology and Neurosurgery* Technical University (RWHT) Aachen, Germany Objective: Magnetic resonance imaging (MRI) is considered as the diagnostic modality of choice in cavernomas of the CNS. Even though their MRI pattern is regarded as pathognomonical MRI findings can widely variate and lead to misdiagnosis. Material and methods: MRI findings (1.5 T unit; TI-, PD-, T2-sequences) were reviewed in 65 patients (age range 0.4-79 years, mean age 41.7 years) between 1987 and 1994 suspicious of cavernoma or proved to be cavernoma following surgery. Two main patient groups were evaluated: 1 - operative group, n = 47 (72 %); 2 - conservative group, n = 18 (28 %). Follow-up MRI studies were done in 27 patients (42 %), preoperatively (n = 8) up to 72 months, postoperatively (n = 19) up to 146 months. Results: MR image pattern with characteristic intralesional aspect - defined as inhomogeneous, hyper-/isointense center - was missed in 40 To of cavernomas with acute hemorrhage. A perilesional ring due to hemosiderin deposition was evident in 90 % of all cavernomas. 42 cavernomas were identified histologically. In 4 of these patients preoperative diagnosis was tumor-bleeding (2), glioblastoma or meningioma. In 5 patients with preoperatively suspected cavernoma histology revealed oligodendroglioma, ependymoma, varix-like vascular malformation and AVM. Postoperative control MRI examinations revealed CSF-like parenchymal defects in 70 %, in 21% only hypointense hemosiderin stains could be found. 6 of 19 (31%) patients were suspicious of residual caverno-

ma based on persisting hyperintense lesions in T1-(5/6), PD-(5/6) and T2-(6/6) sequences. In this group 1 patient had a further intralesional hemorrhage after 33 months due to residual cavernomr, in 1 patient reoperation could not confirm suspected cavernoma. Conclusions: Cavernomas have an inhomogenous image pattern depending on flow, intra-/extralesional hemorrhage and calcifications. MRI diagnosis is especially beset with particular problems in cases with lesional hemorrhage and in the postoperative course. Careful follow-up studies (TI-, PD-, T2-sequences) are often necessary to detect CH and to differentiate postoperative changes from residual cavernoma. 76. General trends in the development of neuroradiology (Invited

lecture) 77. MR-guided intervention in neuroradiology and head and neck procedures (Invited lecture)

E A. Jolesz (Boston, Massachusetts, USA) 78. The precentral knob: a characteristic anatomical structure representing the motor hand area

T. Yousry, D. Schmidt, H. Alkadhi, U.D. Schmid* PA. Winkler*, A. Peraud*. Department of Radiology and Neurosurgery* Ludwig Maximilian University, Munich, Germany PURPOSE: It was shown by various methods (among others fMRI) that the motor hand area is located in the upper genu of the precentral gyrus. Its typical anatomical shape could represent a landmark that might help identifying the precentraI gyrus. M A T E R I A L A N D METHODS: 1) 8 patients and 10 volunteers were examined by fMRI (1,5 T Magnetom, spoiled FLASH). In all volunteers and in 5/8 patients the thus defined motor hand area was localized, assessed and described. 2) 20 volunteers and 10 patients were examined with MPRAGE. The previously described area was identified and measured in detail. 3) 5 anatomical specimens were examined by MRI using M P R A G E and then dissected. 4) The presence of this landmark was assessed on 50 routine MRI examinations. RESULTS: 1) The motor hand area was localized by fMRI on a characteristic knob-like structure in the upper genu of the precentral gyrus. This knob was present in all the volunteers and in 6/8 patients. Tumor distortion prevented the identification in 2 hemispheres. 2) The knob was omega shaped in 52 and epsilon shaped in 6 hemispheres. 3) This knob was present in all the anatomical specimens. 4) On 50 routine examinations this knob was always present and correctly predicted the exact localization of the precentral gyrus with a high interrater reproducibility. CONCLUSION: A knob in the precentral gyrus is the location of the motor hand area. This structure is so simple, characteristic and constant that it helps identifying the precentral gyrus directly in most of the examinations and therefore is of major help in daily work.

79. Functional magnetic resonance activation mapping in essential tremor

K.C. Seelos, S. E Bucher. Depts. of Radiology and Neurology, University of Munich, Germany OBJECTIVE: To detect cerebral activation patterns during postural tremor activity using functional magnetic resonance imaging (fMRI) in patients with essential tremor (ET), a monosymptomatic action tremor. D E S I G N A N D METHODS: Seven patients with ETwere studied at 1.5 T (Siemens Vision) using a RF-spoiled FLASH pulse sequence with first order motion compensation (TR/TE = 47/30,

$50 a = 10 ~ for fMRI. Serial images were continuously acquired at the same location with (1) patients at rest, (2) during maintained posture of the arm and (3) while having the wrist passively oscillated at tremor frequency (6-8 Hz). Activation maps were created by correlation analysis. RESULTS: ET was associated with bilateral cerebellar activation in all patients (7/7). There was also significant activation in the primary motor cortex (6/7), red nucleus (6/7) and globus pallidus (5/7) and contralateral thalamus (2/7). No activation was found in the olivary nuclei or the medulla. Passive wrist oscillation revealed only contralateral cerebellar activation. The number of activated pixels in the cerebellum was significantly lower (P < 0.001) in passive wrist oscillation (mean = 9 + 2) compared to involuntary tremor (mean = 15 _+3). Passive oscillation led to bilateral activation of the thalamus in 2 cases. CONCLUSION: The fMRI findings suggest a central origin of postural tremor activity. In contrast to a study with lSFDG-PETwe failed to demonstrate activity in the inferior olivary nucleus. The significantly larger activation of the cerebellum during involuntary tremor rather than during passive wrist oscillation suggests that (abnormal) proprioceptive input alone is not responsible for the cerebellar overactivity in ET.

80. Functional magnetic resonance imaging of motor, verbal and auditory tasks

C.Manelfe, I.Berry, E.Mueller, JM Franconi, K Boulanouar, JF Demonet, E Chollet, O.Rascol, M. Clanet. Neuroradiology, Neurology and INSERM U230, Toulouse (France), Siemens (France and MRA1 Erlangen, Germany) OBJECTIVE: to test the functional capabilities of patients with a protocol of stimulation performed with a recently implemented multislice echo-planar functional MRI device. D E S I G N A N D M E T H O D S : examinations were performed on 25 normal volunteers and patients at 1.5 tesla (Magnetom Vision, Siemens, Erlangen, Germany) with an EPI-FID sequence. The parameters used for acquisition are: flip angle 90 ~, effective echo time of 66 ms, matrix 66 x 64 (with a zero filling interpolation). 10 slices can be sampled in 1.2 s. Up to 80 measurements are acquired during the acquisition paradigm time. Time between measurement is 3 s. Subjects are requested to rest and activate 4 times during 30 s phases. Motor activations are finger oppositions. Verbal tasks are verb generation related to proposed nouns and verbal fluence on proposed topics. Audition experiment tests both primary and association function with the comparison of tone and word recognition tasks. The post-processing algorithm used is a pixel-by-pixel Z score calculation between stimulation and rest states. The 10 functional images archived are conventional gradient echo images with superimposition of an hypersignal area on the pixels with Z scores superior to a fixed threshold. RESULTS: F M R I proved itself easy to use with normal volunteers and with most patients. Patients with severe motor deficits showed decreased motor cortex activation consistent with their handicap. Patients moderately handicapped had larger activation areas involved than normal volunteers. This could be consistent with the recruitment of alternate pathways for functional systems. CONCLUSION: at this preliminary stage, F M R I implemented with multislice capabilities appears to have sufficient practicability of clinical use and seems promising for functional area identification and plasticity evaluation.

81. Auditory Cortex Activition in Deaf Subjects During Cochlear Electrical Stimulation: Evaluation by Functional MRI

Y Berthezene1, E. Truy e, A. Morgon 2, M.H. Giard3, E Turjman1, M. Hermier~, J. M. Franconi 4, J. C. Froment1 H@ital Neurologique et Neurochirurgica159, boulevard Pinel, 69003 Lyon FRANCE (1), Department dOto-Rhino-LaryngoIogie, HOpital Edouard Herriot Lyon FRANCE (2), INSERM U280 Lyon (3), Siemens (4) Objective: To detect activation in the auditory cortex during cochlear electrical stimulation in deaf patients. Material and Methods: The stimulating electrode was gently inserted under local anesthesia through the round window membrane in the scala tympani of the cochlea (n = 6). The stimuli were generated by H O R T M A N N ' s Cochlear Nerve Test (Hortmann, Hannover, Germany), which delivers electrically isolated constantcurrent square pulses ranging from 0 to 500 micro-Amperes and from 0.125 to 8 milliseconds. M R imaging was performed with superconductive unit operating at 1.5T using a standard circulary head coil. Images were acquired with a RF spoiled 2D F L A S H sequence (TR 91 msec; TE 60 msec; a 40~ Functional scans were acquired in an oblique axial plane running parallel to the Sylvian fissure with one slice just below it. Three to four consecutive series of 6-14 images each were obtained in minutes (without auditory stimulation followed by stimulation respectively) Pre and post stimulation images were compared with the Z statistics. Results: During electrical cochlear stimulation all patients described "auditory" sensations with activation of the superior temporal regions bilaterally. Regions beside the auditory cortices showed significant activation, but the regions of activation were less widespread than in the auditory cortex. Conclusion: These preliminary data indicate that fMRI can detect activation in the auditory cortex during cochlear electrical stimulation in deaf patients. 82. First localization of posterior eye field in a human using functional MRI

E. A. Cabanis, C. Desrosier, M. T. Iba-Zizen, A. Tourbah, R. Muri, C. Pierrot-Deseilligny Dept. of Neuroradiology, Ctre Hosp. Nat. Opht. XV-XX, Fdddration de Neurologie, La Pitid-Salp6triOre, Paris, France OBJECTIVE: In the primate the posterior eye field (PEF) was localized to the intraparietal sulcus. We used M R F in PEF localization in human. These result are shown for the first time in human. M A T E R I A L A N D M E T H O D S : A first group of 95 people was studied in 18 months (1993-1994) for visual field M R F evaluation and methodology, with Signa 1.5 Tesla (GEMS), and a light emitting diode as a source of stimulation. This second group of 9 volunteers was studied for oculomotor cortical field localization. The planes are parallel to the Ca-Cp line, passing by intraparietal sulcus and the calcarin scissure. Three procedures are used for acquisitions: 1 ~ saccades vs rest in 3 patients; 2 ~ fixation vs rest in 3 patients, 3 ~ saccades vs fixation in 3 patients. Data analysis was performed on Advantage Windows work stations with CD (HIA, V D G ) software. RESULTS: 1 ~ Saccades vs rest: significant signal was localized in the prefrontal field around the precentral sulcus; in the horizontal part of the intraparietal sulcus and in the primary visual field; 2 ~ Fixation vs rest has shown a significant signal in the primary visual field; 3 ~ Saccades vs fixation has shown a significant signal in the frontal eye field and the horizontal part of the intraparietal sulcus. CONCLUSION: M R F allows, for the first time in our knowledge in human, to localize the posterior eye field (PEF) to the horizontal part of the intraparietal sulcus.

$51 83. Unbiased measurement of neocortex development using MRI

CA.J. Broere*, LMP Ramos**, TW Polder**, Department of Neurology* University Hospital Nijmegen and Radiology, University Hospital Utrecht, The Netherlands Introduction Loss of neocortex volume assessment is important in neurodegenerative diseases, like Alzheimer's Disease. In this study a sensitive unbiased method to measure neocortex volume was designed, based on routine M R I examinations. Using this method results are obtained together with an estimation of the coefficient of error (CE) in the individual measurements. The volumetric results are independent of the geometry of the structure, which made it the only suitable one for analysing the geometrically very complex neocortex. Materials and Methods Axial and frontal MRI scans (Gyroscan, 1.5 Tesla) are obtained from normal persons (109 females, ages: 2 weeks - 68 years; 92 males, ages: 3 weeks - 67 years). Using the method of systhematic sampling combined with volumetry measurements, based on Cavalieri's Principle, an unbiased estimation of neocortex volume is obtained, with an accuracy of each measurement, expressed as the coefficient of error (CE) of less than 5 %. Results The method of measurement proves reliable within a CE of 5 %, without any observer bias or the M R I echo sequence chosen. The interobserver variability proved less than i %. The variability in the results are mainly due to biological variance. Statistical analysis showed the best data fitting curves to be hyperbolic.Analysis of the data obtained from persons over 18 years of age showed no departure from linearity of the data. Statistical analysis showed no significant difference between both sexes of slope or Y-intercept, but the slope in both cases was negative and different from zero (males: ~0.061 cc/month of age; females: -0.15 cc/month of age, with a significant difference from zero in females (r = 0.25, P = 0.035, Turkey-Kramer multiple compafisions test). Conclusion The reported method provides an accurate and fast estimation of neocortex volume and provides a possibility of accurate assessment of morphologic changes, such as neocortex atrophy, enlargement of the ventricular system, expressing the volume of these structures in actual volumetric units.

84. Three-dimensional (3D) neurosurgical planning and neuronavigation in eloquent brain areas

Knauth M 1, Wirtz R e, Tronnier 1/2, Albert F Ke, Forsting M 1, Kunze Se, Sartor K 1 Departments of NeuroradiologS and Neurosurgery2, University of Heidelberg Medical School, D-69120 Heidelberg, Germany Purpose: When operating in or near eloquent brain areas, tissue damage must be kept minimal. To achieve this goal two methods were used in combination: preoperative 3D planning and intraoperative neuronavigation. The accuracy of both methods was determined by comparison with the intraoperative situs. Material and Methods: 28 patients with a lesion in or near an eloquent brain area were examined with a contrast-enhanced 3DFTMRI sequence (Picker Vista, 1.0 T; slice thickness 1.0-1.6 mm). Based on this dataset 3D reconstructions were calculated (Allegro workstation, ISG). Reconstructions were focused on: (1) relation of tumor to external and bony landmarks; (2) cortical veins around lesion; and (3) relation of tumor to eloquent brain areas. The patients were operated using a neuronavigational device (Viewing Wand, ISG). The accuracy of preoperative 3D reconstructions and had the accuracy of registration for neuronavigation was assessed by measuring the deviation from the intraoperative situs at the beginning and during the operation.

Results: Preoperative planning allowed reliable visualization of the relationship of the tumor to external and bony landmarks, cortical veins and gyral pattern. An exception were four patients with complete sulcal effacement due to tumor or edema. The accuracy of the 3D reconstructions as assessed by the average deviation from the intraoperative situs at the beginning of the operation was less than 2.5 ram. However, the accuracy of neuronavigation decreased during operation to an average deviation of 4.3 mm due to brain shift, brain swelling, use of spatulas, and tumor removal. Conclusion: Important information for neurosurgical planning can be obtained from 3D-evaluation of high resolution MRI datasets. Most valuable for surgical planning is the simultaneous 3D display of the gyral pattern and the lesion. Knowing what the gyral pattern will look like after opening of the dura improves the orientation of the neurosurgeon, thus reducing the invasiveness of the operation. The main role of neuronavigation is to determine the exact site and direction of craniotomy and corticotomy. The value of neuronavigation decreases during the course of the operation due to "surgical"changes in anatomy. A possible solution to this problem would be to "update" the neuronavigation dataset by intraoperative MRI.

85. The 3D brain surface rendering in the MR evaluation of cortical dysplasias.

B. Bernardi, * C. Fonda, **R. Canapicchi, *** R. Guerrini. Neuroradiology Bellaria Hospital, Bologna. *Neuroradiology Misericordia e Dolce Hospital, Prato. **Neuroradiology S. Chiara Hospital, Pisa. ***INPE Universita' di Pisa, IRCC Stella Marls, Pisa. OBJECTIVE: To determine the role of the 3D surface rendering images in the evaluation of the broad spectrum of the cortical malformations. M A T E R I A L A N D M E T H O D S : Thirteen patients with all diffuse forms of cortical dysplasia, bilateral perisylvian syndroma, hemimegancephaly, other unilateral hemispheric dysplasias, and forms of focal dysplasia (Taylor's type, unilateral perisylvian) were evaluated by MR 3D brain surface rendering. Volume 3D gradient echo acquisitions with thin overcontigous partitions were obtained. We used a repetition time (TR) of 16 to 32, echo time (TE) of 7 to 8, and 20 ~ flip angle. In some cases the 3D brain surface reformations were combined with a PC Angiography obtained by isocentric acquisitions. The last cases were integrated with a IR-TSE acquisition using a T R 3400, TE 30, TI 300, and 3 mm thickness with gap 0. R E S U L T S A N D CONCLUSIONS: MR is a sensitive modality to display all types of cortical dyplasia. Volume 3D gradient echo acquisition show the irregular white matter-gray matter junctions and 3D reformations define the topographic location of dysplastic cortex with relation to the surface of the brain and its landmarks. 3D brain surface rendering is helpful to detect any disorder of the gyral pattern and can aid in the surgical planning of related epilepsy. The IR TSE sequences clearly demonstrate gray-white differentiation; this is key in the detection and accurate diagnosis of cortical anomalies. This sequence allows a curved reformatting of the cortical surface. Finally, 3D PC Angiography can be used to confirm that anomalous venous drainages are common in areas of dysplastic cortex.

86. MR! 3D tissue segmentation analysis of intracranial lesions

Carlos E Gonzalez, M.D., Simon Vinitski, Ph.D. Division of Neuroradiology, Division of Magnetic Resonance Imaging Thomas Jefferson University Hospital, Philadelphia, PA, USA P U R P O S E The aim of this work was to develop an accurate method for tissue classification and to apply it to characterize and measure brain lesions. M E T H O D S A N D M A T E R I A L S Tissue segmentation based on 2D and 3D feature map derived from high resolution M R images

$52 was performed in phantoms, normals, humans and patients particularly those with multiple sclerosis (MS) and brain tumors. Three inputs: proton density, T2- and, as a third, Tl-weighted MRI were utilized. Statistical and anisotrophic diffusion filters were applied to the data and k-Nearest Neighborhood segmentation algorithm was utilized. RESULTS The inclusion of T1 based images onto segmentation produced dramatic improvement in tissue identification. Our technique utilizing all three inputs provided better segmentation than that based on any combination of two inputs. The inclusion of T1 based images into segmentation produced dramatic improvement in MS tissue identification particularly between different areas within the same plaque. We presume these represent different stages of disease. Reproducible and accurate measurements of regional MS lesion load (temporal, parietal frontal, occipital) were obtained in each patient. The majority of lesions were identified in parietal occipital lobes. Results were consistent with the patients clinical evolution of the disease over a six months period. In benign tumors, we identified tumor volume prior to the injectionof gadolinium-DTPA. In malignant tumors, up to four abnormal tissue were identified: 1) solid tumor core, 2) cyst, 3) edema in white matter and 4) edema in gray matter. Subsequent neurosurgery confirmed the presence and location of the tissues. CONCLUSION The proposed technique of tissue segmentation based on 3D features map allows detailed tissue characterization and thus encourages further clinical investigation.

87. Imaging possibilities of the bordering territories of neuroradiology and head and neck radiology (Invited lecture)

J. Casselman (Brugge, Belgium) 88. Labyrinthine malformations: embryological correlations and flow-sensible imaging.

L. Man]re', E Riggio*, T. Angileri, S. Ferrara*, M. Tortorici* A. E. Cardinale. Department of Radiology "P Cignolini" - University of Palermo *ENT Department- University of Palermo Objective: To evaluate "flow-sensible" MRI capability vs High Resoluted CT in inner ear imaging of congenital malformations in the assessment of congenital and acquired sensorineural hearing loss (SNHL), and to correlate malformative condition of membranous labyrinth to the halted time of inner ear embriologycal development. Design and methods: During a 12-months period 182 patients complaining of SNHL and/or vertigo underwent HRCT and ~'flowsensible" MR imaging: we totally found 7 inner ears (1.8 %) presenting labyrinthine malformations (cochlea aplasia, bilateral semicircular canals aplasia in a Noonan's syndrome, large vestibular aqueduct syndrome, cystic lateral semicircular canal + vestibule dysplasia, Mondini malformation, single postero-superior semicircular canal in a Goldenhar's syndrome). MRI examination of the membranous labyrinth was performed using a middle-field unit and dedicated surface coil, with a volumetric GRASS sequence, sufficiently sensible to the slow flow of endolabyrinthine fluids using a very short TR, a large flip angle and a long TE. MIP reprocessing was also performed for 3D imaging of malformative membranous labyrinth. Results: GRASS imaging was always able to depict malformative conditions previously detected on HRCT, thanks to the higher spatial resolution and natural contrast between the high signal intensity of endolabyrinthinefluids and the hypointensity of the stationary tissue surrounding the membranous labyrinth. All the detected malformations perfectly fitted with an arrested inner ear development at a specific time among the IV and the VII week of fetal life. Conclusion: In conclusion, although HRCT is able to sharply depict the bony labyrinth, no information is detected concerning the small tissue structures forming the membranous labyrinth: how-

ever, membranous labyrinth imaging is fundamental if a cochlear implant is planned. MRI using a "flow-sensible" technique, as GRASS sequence, also thanks to MIP 3D reconstructions, revealed as a reliable and non-ionizing imaging modality in membranous labyrinth congenital malformations, alIowing to depict all the malformative condition previously seen on CT.

89. High resolution MRI of the endolymphatic duct and sac in cerebellopontine angle surgery

K.E.W. Eberhardt 1, R. Naraghi 1, H.P Hollenbach 2, W.J. Huk 1. 1Div. of Neuroradiology, Department of Neurosurgery, University Erlangen-Ni~rnberg, Erlangen, Germany, 2Siemens Medical Engineering Group, Erlangen, Germany Purpose: Preservation of hearing is an important issue in cerebellopontine angle surgery. To achieve this the precise imaging of the labyrinthine structures especially the endolymphatic duct and sac is necessary. In this study we present two MR methods which visualize these structures directly. Material and Methods: 20 healthy volunteers and 8 patients with acoustic neurinomas and useful hearing were examined with a 1.5 T whole body MR system (Siemens Magnetom Vision). A strongly T2*-weighted 3D-PSIF sequence (TR = 17ms, TE = 7ms, c~= 90 deg) as well as a CISS-scheme (TR = 14ms, TE = 6ms, c~= 80 deg) were applied in the transverse orientation. The acquired 3D-data sets were evaluated using a Maximum Intensity Projection (MIP) program. Results: In all cases it was possible with both sequences to visualize all labyrinthine components especially the endolymphatic duct and sac. Anatomical details like the two sinuses, the isthmus region and the endolymphatic sac could be identified. As a result relevant distances were determined and used for hearing preserving surgery and protection of the labyrinthine structures in patients with acoustic neurinomas. Conclusion: MRI is a useful method to visualize all structures of the membranous labyrinth in one projection. As a result of this diagnostic procedure further improvement of hearing preservation in surgery of cerebellopontine angle processes can be expected. 90. Evaluation of ocular melanoma: MR imaging compared to ultrasound associated with surgical correlation

Gupta, KL, Reuther III, WL. Department of Radiology, Tulane University Medical Center, New Orleans, Louisiana 70112 OBJECTIVE: Delineating the optimal MR imaging sequences and usefulness of ultrasound in evaluating ocular melanoma. DESIGN AND METHODS: Intraocular melanoma is the most common intraocular malignancy in adults accounting for 80 % of primary ocular malignancies. Magnetic resonance imaging was compared to ultrasound in the efficacy of evaluating, detecting characterizing 183 ocular melanoma cases. The findings, size of tumor, presence of retinal detachment and hemorrhage are analyzed. RESULTS: Heavily T-1 weighted sequences were considered more sensitive than wither T-2 or proton density images in revealing information regarding deeper lesions and providing superior contrast resolution and tissue characteristics. Ultrasound, however, is invaluable in that it is cost effective, easily provided on an outpatient basis and provides higher resolution in detecting smaller lesions. CONCLUSION: MRI and US are both excellent imaging modalities in the medical care and follow-up of the pre and post surgical ocular melanoma. Knowledge of each modalities efficacy and limitations facilitates optimal cost effective patient care.

$53

91. Magnetization transfer imaging; the effect of the prepulse parameters on the magnitude of the magnetization transfer effect.

93. Comparison of triple dose vs standard dose Gadolinium-DTPA for detection of MRI enhancing lesions in multiple sclerosis

P. A.M. Hofman and J.T. Wilmink. Department of Diagnostic Radiology, University Hospital Maastricht

A. Campi, M. Filippi* T. Yousry** C. Kandziora**. B. Colombo* R. Voltz*** V. Martinelli* S. Spuler *** G. Comi* G. Scotti. Depts of Neuroradiology and *Neurology, IRCCS H.S. Raffaele, University of Milan, Baly; Depts of **Radiology and ***Neurology, Klinikum Grosshadern, University of Munich, Germany.

OBJECTIVE: In order to arrive at an optimal image sequence for the evaluation of tissue magnetization transfer differences we designed an experiment to establish the effect of sequence parameters on the magnitude of the magnetization transfer contrast. D E S I G N A N D M E T H O D : MR data were acquired at 0.5 Tesla system. A gradient -echo sequence (TR 100 msec, T E 12 msec, flip angle 7 degree) was modified with a sinc shaped prepulse. The effect of the pulse duration, average amplitude and frequency offset on the magnetization transfer were determined. To measure the magnetization transfer effect data were acquired without the use of the prepulse. The experiments were performed with a phantom containing denatured ovalbumin and water, Althought the experiments were not performed on humans the specific absorbtion rate (SAR) was kept within the limits of the U.S. F D A guidelines. The magnetization transfer effect increases nonlineary with the pulse amplitude, with a maximum decrease in signal of 72 %. With a given average amplitude an increase in pulse duration decreases the magnetization transfer effect. The magnetization transfer effect is inversely proportional to the frequency insignificant with a frequency offset greater than 500 kHz. Plots of the magnetization transfer ratio versus the parameter demonstrate these effects. C O N C L U S I O N : The imaging of magnetization transfer effects with a sinc shaped off resonance prepulse is easily implemented and may be a useful tool in clinical radiology.

92. Magnetization transfer MRI with Gd-BOPTA/Dimeg and GdDTPA/Dimeg in a rat brain tumor model

PM. Castelli*, E Maggioni* P Marzola*, M. Daprd* C. Barbieri** FM. Cavagna*. *Milano Research Centre, Bracco S.p.A., Milan, Italy, **School of Medicine, University of Milan, Milan, Italy OBJECTIVE: The aim of this work was to compare the increase in the lesion to brain (L/B) ratio (Fisher et al., S M R M 1993:183), obtained with Magnetization Transfer (MT) in a rat brain tumor model after administration of Gd-BOPTA/Dimeg and Gd-DTPA/ Dimeg. DESIGN A N D METHODS: Cerebral tumors were induced by stereotaxic injection of 5 x 106 glioma C6 live cells suspended in 4 ml of growth medium into the right frontal lobe of male SD rats (Cavagna et al., SMR 1994: 263). Four animals were imaged using a S.M.I.S. animal system operating at 2T. Each rat received 0.1 mmol/kg i.v. of Gd-BOPTA/Dimeg followed by the same dose of Gd-DTPA/Dimeg after 3 hours. Tl-weighted Spin Echo (SE) (TR/TE = 350/10 (ms); FOV = 4 x 4 cm2; matrix (128) 2) and MTSE images (pulsed off-resonance saturation) were acquired prior to and up to 3 h after contrast Quantitative image analysis was performed using the program "Analyze | on a Sun Sparc 10 workstation (Sun Microsystems Corp., Mountain View, CA). RESULTS: In conventional SE imaging the L/B ratios were higher after Gd-BOPTA/Dimeg than after Gd-DTPA/Dimegat any time point (1.58 + 0.09 and 1.25 _+0.15, respectively, at 12 minutes after contrast), With MT, the L/B ratio increased further but much more after Gd-BOPTA/Dimeg, with a plateau at 2.62 + 0.25 between 10 and 20 minutes, while it peaked at about 1.98 + 0.20 ten minutes after Gd-DTPA/Dimeg. CONCLUSIONS: Gd-BOPTA]Dimeg proved to be superior to Gd-DTPA/Dimeg in enhancing the conspicuity of rat brain tumors in SE and MT-SE images. MT is more effective after Gd-BOPTA/ Dimeg than after Gd-DTPA/Dimeg. Plasma clearance in rats is much faster for Gd-BOPTA/Dimeg than for Gd-DTPA/Dimeg, while this is not the case in man (Lorusso et al., J M R I 1993; 3:155). The superiority of Gd-BOPTA/Dimeg in delineating brain tumors should therefore be borne out even stronger in this species.

OBJECTIVE: This study was performed to evaluate whether a triple dose (TD) of gadoliniumm-DTPA (Gd-DTPA) alone or in combination with delayed scanning increases the sensitivity of brain magnetic resonance imaging (MRI) for detecting enhancing lesions in patients with multiple sclerosis (MS). DESIGN AND METHODS: Tl-weighted brain MRI scans were obtained in 2 sessions for 22 patients with clinically definite MS. In the first session, one scan was obtained 5 to 7 minutes after the injection of 0.1 mmol/kg Gd-DTPA (standard dose, SD). In the second session, 6 to 24 hours later, one scan before and two scans after 5 to 7 minutes (for all the patients) and one hour (for 11 patients) after the injection of 0.3 mmol/kg Gd-DTPA (TD) were obtained. RESULTS: Eighty-three enhancing lesions were detected in 14 patients when the SD of Gd-DTPA was used. The numbers of enhancing lesion increased to 138 (average i n c r e a s e = 6 8 % ; p = 0.001) and the numbers of patients with such lesions to 18 (increase = 28 %) when the T D was used. In addition, the total area per patient occupied by such lesions was greater (p < 0.0001) and signal intensity higher (p = 0.0001) on the TD scans than the SD scans. There was an increase in the number of large enhancing lesion (p = 0.03) in the scans obtained one hour after the injection of the TD. CONCLUSIONS: These data indicate that in patients with MS a TD of Gd-DTPA can reveal many more enhancing lesion, which also appear larger. This suggests that the pathological nature of "active" lesions in MS is heterogeneous which might have major impact on planning clinical trials. 94. An in vivo transmetallation study of MRI contrast agents in human serum and urine

N.R. Puttagunta, Ph.D. z, W.A. Gibby, M.D. z'e, G.T Smith/, s Clar~. ~Magnetic Research, Inc., Provo, Utah, U S. A., 2Utah Valley Regional Medical Center, Provo, Utah, U S. A., 3Dept. of Chemistry, Brigham Young University, Provo, Utah, U. S. A. OBJECTIVE: The purpose of this study was to determine in vivo transmetallation of M R contrast agents from the change in Zn and/ or Cu levels in serum and urine. D E S I G N A N D M E T H O D S : A total of 31 patients in three groups received Gd D T P A (Berlex) -10, Gd D T P A B M A (Winthrop) 10, and Gd H P D O 3 A (Squibb) - 11, at 0.1 mM/kg. Urine was collected prior to, and for three hours post injection. Serum samples were obtained immediately prior to, and 15 minutes post injection. Zn and Cu analyses were performed using ICP in a blinded fashion. RESULTS: No statistical differences were observed in acute serum Cu or Zn levels between the agents. Gd DTPA B M A causes a 3.8 fold increase in total urinary Zn excretion compared to Gd D T P A and 26 fold increase compared to Gd H P D O 3 A . No significant difference between the three agents was observed in urine Cu levels. CONCLUSION: The increased urinary Zn excretion following Gd DTPA and Gd DTPA B M A injection could be due to transmetallation with Gd chelate, excess free chelate, or Ca chelate. These findings concur with studies showing residual Gd in rats (Wedeki n g e t al., M.R.I. 1992; 10:641) and our in vitro studies (Gibby et al., JMRI, 1994; 4(P):71), which shows Gd H P D O 3 A to be kinetically most stable.

$54

95. Subtraction spiral CT-angiography (SSCTA) in radiosurgery of intracranial AVMs

K. Heimberger, K. Kitz* H. GOrzer, E. Schindler. Dept. of Radiology, Div. of Neuroradiology, and Dept. of Neurosurgery*, University Hospital, Vienna, Austria OBJECTIVE: A subtraction algorithm was added to Spiral CTAngiography and results compared with I A D S A and M R A to examine its role in Gamma-Knife therapy of partially embolized intracranial AVMs. D E S I G N A N D M E T H O D S : 21 patients with partially embolized intracranial AVMs were examined with SSCTA, I A D S A and M R A before and after Gamma-Knife therapy. SSCTA was performed with the Siemens Somatom Plus S, using 3 mm slice thickness, a table speed of 3 mm/sec; 3 mm collimation, one gantry rotation per second, a pitch of one, a reconstruction increment of 1 mm, an X-ray delay of 16 to 20 seconds, iv contrast flow of 4 mm/ sec, a total contrast volume of 100 ml; data acquisition time was 3 x 30 seconds. Subtraction data of position identical non-contrast and contrast-enhanced slices were reconstructed in the 3D-surface rendering technique and with the maximum intensity projection. 3D-TOF was used for MRA. RESULTS: Exact and identical data were rendered for size and location of AVMs by I A D S A and SSCTA. Stereotactic data of M R A , however, showed line distortions of up to 2 mm. In patients with metal castings (coils, clips), M R A was unsuitable. SSCTA was tess sensitive in nidus compartments with very slow blood flow; but shrinkage of AVMs (six months to two years after radiosurgery) was clearly detected by IADSA, SSCTA and MRA. CONCLUSION: SSCTA is a simple and precise method for Gamma-Knife planning and can also be used for postoperative and postradiosurgery controls in outpatients with partially embolized AVMs.

96. Cerebral cavernomas and telangiectasias as secondary disease following acute lymphatic leukaemia (ALL) and cranial radiation

K. Brgthl(1), T. Humpl(2), P Gutjahr(3), J. Bohl (3), P. Stoeter(1) Dept. of neuroradiology (1), Pediatric Clinic (2), and Dept. of Neuropathology (3) University of Mainz, Germany Purpose: To point to the appearance of cerebral cavernomas and telangiectasias as a secondary, possibly treatment induced disease after ALL. Material/Method: Four children (3M, 1F) aged 4-10 years with A L L were treated by cranial radiation and chemotherapy. A t that time CT findings were nm-mal. Results: After 2-11 years, all children suffered an intracerebral haemorrhage caused by cavernomas (3 cases) and telangiectasias (1 case) of different localisation but always within the former irradiated area. The CT/MRI findings were verified by histology. Conclusions: Similar to secondary neoplasms, radiation is regarded as a possible cause of induced telangiectasias and cavernomas which were not seen after A L L treated solely by chemotherapy. The presence of a secondary vascular "malformation" should be discussed as a possible differential diagnosis in the case of an intracerebral bleeding as a late complication after ALL. 97. Inirasellar arachnoid cysts

P Flandroy*, A. Dubu&son**, A. Meurice* D. Martin**, A. Stevenaert**. University Hospital SART TILMAN, Lidge, Belgium, *Dept of Medical Imaging (Pr.R.EDondelinger), **Dept of Neurosurgery (Pr.A.Stevenaert) O B J E C T I V E : Intrasellar arachnoid cysts (IAC) are rare. The radiologicat data of 8 cases documented by surgical and pathological observations are reported. Possible physio-pathological mechanisms are discussed.

D E S I G N A N D M E T H O D S : Among 754 transphenoidal operative procedures performed in our institution from 1971 to 1994, 5 men and 3 women (age range : 20-62 years) showed an IAC. Six patients presented with symptoms mimicking those of a non-secreting pituitary adenoma. Two cysts were incidentally discovered. A C T scan was performed in 7 patients. Six patients underwent an M R examination using Tl-weighted sequences before and after gadolinium injection in all and T2-weighted sequences in 2. R E S U L T S : A hypodense intrasellar non-enhancing lesion was seen on all CT examinations with suprasellar extension in 4 patients. On MRI, the lesion appeared homogeneously more intense than the CSF on T1 in 5 out of 6 cases. All these lesions were ovoid and a huge suprasellar extension was seen in 5 cases. In all cases, the lesion moulded the medial aspect of the cavernous sinuses. C O N C L U S I O N : MRI may suggest the diagnosis of an I A C when showing a cystic, balloon-shaped, intrasellar non-enhanced lesion often with suprasellar extension, iso- or most often hyperintense in T1 to the CSE

98. Evolution of high-intensity basal ganglia lesions on Tl-weighted MR images in neurofibromatosis type 1.

H. Tareda 1,2,A.J. Barkovich ~'3,M. S. B. Edwards 3, S. E Ciricillo 3, H. Matsushima~ Y Hiramatsu 21Dept. of Radiology and 3Neurological Surgery, University of California, San Francisco, USA 21I. Dept. of Radiology, Toho University, Tokyo, Japan OBJECTIVE: To characterize the temporal evolution of the foci of T1 shortening in basal ganglia lesions of patients with neurofibromatosis type 1 (NF-1). DESIGN AND METHODS: We retrospectively reviewed MR images in 37 patients with NF-1 with special attention to signal changes in the basal ganglia. RESULTS: Regions of short T1 in the globus pallidus were observed in eight patients. In two of three patients in whom foci of T1 shortening were not identified on the initial imaging study, T1 shortening developed while T2 prolongation diminished following an initial increase. In the third patient, T1 shortening and T2 prolongation appeared simultaneously. Sequential scans in the other five patients, in whom areas of increased signal intensity in globus pallidus were present on both T1- and T2-weighted images on the initial MR, showed diminution of size of the region of T2 prolongation in one patient, mixed pattern of change of size of the region of T2 prolongation in one patient, and no change of the region of T2 prolongation in one patient; during the periods of these T2 changes, the areas of T1 shortening showed no significant interval change. CONCLUSION: The foci of prolonged T2 relaxation in the basal ganglia appear to evolve in a manner similar to the foci of T2 prolongation in the white matter of the posterior fossa. However, the corresponding foci of short T1 in the basal ganglia evolve with a different time course. The foci of short T1 develop at a later time than the T2 prolongation and progress but do not appear to regress. Possible causes of the T1 shortening are remyelination and calcification. 99. MRI detection of haemosiderin and gliosis in follow up examinations after brain trauma

Polonara G, Mauro A. M., Salvolini U *. Department of Neurology and Neuroradiology* University of Ancona, Italy PURPOSE: Identify the best sequences to study brain lesions after a trauma and search for particular lesions with prognostic value like haemosiderin and gliosis. MATERIALS AND METHODS: We evaluated 62 patients affected by brain injury. Each patient was submitted the same day to EEG, CT, M R I and clinical state examination. These examinations were repeated within the first two weeks, after 6-8 weeks, after 1214 weeks and after i year. MRI studies were performed on 1.0 T

$55 unit using a sagittal SE TlW, and axial and coronal TSE PD and T2W and an axial and/or coronal G E T2*W. The additional sequences used, when necessary, were coronal and/or axial SE T l W for best visualization of extracranial haemorrhage or an IR "dark fluid" (tFLAIR) that makes gliosis more conspicuous. RESULTS: We found extracerebral haematomas in 20 patients and intracerebral lesions in 49 patients. In 11 cases MRI was negative. Intraaxial lesions were diffuse axonal injuries (47,5 %), contusions (44,3 %), subcortical gray-matter injuries (5 %), brainstem injuries (3,2 % %). Haemosiderin was found in 14/62 patients using TSE T2W and 38/62 patients using G E T2*W. Only the biggest areas which were seen by G E T2*W were also visible on TSE images. In two patients haemosiderin was more evident with TSE where deposited close to the petrosal bone and frontal paranasal sinuses, because of artifacts related to the presence of air into these structures. CONCLUSIONS: G E T2* is more sensitive than conventional SE and TSE sequences for detection of haemosiderin. The presence of haemosiderin may determine the onset of post-traumatic epilepsia. 6 of the 62 patients had a late epileptic crisis and 1 developed epilepsia. All of them had multiple areas of haemosiderin, with or without gliosis. Gliosis was detected in 27 patients. Gliosis was better demonstrated in late follow up studies using t F L A I R sequences than TSE, PD or T2W.

M E T H O D A N D M A T E R I A L S : MRI of 19 patients were available for this investigation, because the part of the plexus was distinct with enhancement, and because those patients complaining of symptoms to be related to the eighth nerve, as tinnitus and/or hearing disturbance on the one side. Dizziness a n d / o r vertigo was not necessary for the case selection. Those cases who had any c-p angle tumor were not included in 19 cases. M R imaging was made in the plane axial to the brainstem long axis and in that coronal to the axis. Slice thickness was either 3 mm or 4 mm. RESULT: Nine of 19 cases were picked up, as they showing asymmetric extension of the plexus. In none of these 9 cases neurovascular compression of the cochlear nerve was observed. Seven cases had high pitch hearing loss on audiogram examination of the same side as that of the plexus approximation to the lemmoglial junction. tn three of these 7 cases the I-wave of A B R (Acoustic Brainstem Response) was elongated on the side. C O N C L U S I O N : There can be an entity in which B O C H D A LEK'S flower basket gives effect with its CSF secretion on the lemmoglial junction of the cochlear nerve, when the one side choroid plexus lying in very approximation to the junction. In these cases tinnitus and high pitch hearing disturbance occurs most probably.

100. CSF study with MRI in patients with craniosynostosis. G. Trasimeni, C.Di Biasi, M.Iannilli, A. Raco* G.F.. Gualdi and A.J. Raimondi* MR and CT Unit 1 ~ Medicine and *Pediatric Neurosurgery University of Rome "La Sapienza "

Posters

O B J E C T I V E : The purpose of this study is to evaluate the CSF state in patients with craniosynostosis with MRI. Subarachnoid and ventricular enlargement is reported in craniosynostosis as an expression of the asimmetric skull growth (compensatory spaces) Indications for decompressive surgery in craniosynostosis are for cosmetic results and to avoid intracranial hypertension on both short and long term basis. D E S I G N A N D M E T H O D S : We studied the CSF distribution by MRI in pre- and post-operative status in 10 consecutively Patients with craniosynostosis. We studied 4 Plagiocephaly, 2 Scafocephaly, 1 Apert, 1 Crouzon, 1 Trigonocephaly and 1 Macrocephaly. RESULTS: In all but one we found dilation of the subarachnoid spaces; 7 Patients had also ventriculomegaly. According to Raimondi's hydrocephalic score (Raimondi, CNS 1994; 10:2) we classified grade "0" one patient, grade "1" two patients and grade "2" seven patients. Eight out of the nine hydrocephalic patients had surgery and repeated MRI three months later. In post-op MRI examination we found subarachnoiod spaces size reduction in 6 patients, proving the usefulness of surgery in re-equilibrating CSF dynamism. C O N C L U S I O N : We think that CSF enlargement and ventriculomegaly in this patients is related to asymmetric skull growth and represents both a pathologic and pathogenetic condition. M R may help to decide upon surgery in some cases in which the skull is not too dysmorphic but the CSF dynamism is altered; and it may prove the efficacy of surgery in operated patients.

102. Acrylic vertebroplasty in spinal metastases (about 48 cases) J. Chiras, C Cognard, M. T. Sola-Martinez, A. Weill, M. Rose, E. Enkaoua, J.M. Simon. La Salp~tri~re Hospitak Neuroradiology Charcot, Radiotherapy, Orthopedic Surgery, Paris, France

101. Bochdalek's flower basket. Its pathogenetic significance. Y. Kuru, T. O'uchi, A. Tokumaru, Z. Horichi. Dept. of Radiology, Kameda General Hospital, Higashi-cho 929, Kamogawa City, 296Japan O B J E C T I V E : The extraventicular protrusion of the fourth ventricle choroid plexus (BOCHDALEKsches BlumenkSrbchen, flower basket) can be effecting with its CSF secretion on the lemmoglial junction of the cochlear nerve and may cause derangement of nerve conduction as well as some complaints relevant to the affection. Thin slicing M R I with Gd-DTPA enhancement can prove approximation of that plexus to the junction.

I. Spine and spinal cord

INDICATIONS. To establish the contribution of percutaneous vertebroplasty in management of patients with vertebral metastases. D E S I G N A N D M E T H O D Percutaneous vertebroplasty is a recently introduced technique which strengthened the vertebral body by the way of fullfilling it with acrylic cement. Thirty-seven patients presenting with 48 metastatic vertebrae were treated by this way, including cervical and dorso-lumbar spine. The vertebroplasty was performed as an alternative or after other treatments in 25 cases, in association with surgery in 4 cases and with radiotherapy in 8 cases. R E S U L T S In most cases, the release of pain was very important: complete release (27/33 cases) imcomplete (2 cases); temporary release was observed in 1 case dut to epidural involvement by the metastase. In 3 cases there was no significant release. Two patients died a few days after the plasty and therefore were not considered for pain evaluation: one of acute pneumocystis carinii infection misdiagnosed at the time of procedure, the other of pulmonary embolism related to decubitus and not to migration of cement. We observed 3 cases of transient radiculalgia and 2 cases of transient dysphagia after cervical vertebroplasty. We did not observed any cord compression. C O N C L U S I O N This new technique of treatment gives very satisfactory results in management of patients with spinal metastases and can be used in association with radiotherapy and or surgery.

103. Chiari I deformity: clinical and spinal MR evaluation followup A. Campi, S Gerevini, 1M Filippi, G Scotti. Departments of Neuroradiology and 1Neurology, Scientific Institute HS Raffaele, University of Milan, Italy OBJECTIVE: 1) To differentiate by spinal M R and clinical findings the different subgroups of patients classified under "Chiari I