Neuroradiology (2011) 53:79–88 DOI 10.1007/s00234-010-0702-9

DIAGNOSTIC NEURORADIOLOGY

Clinical outcome and imaging follow-up in acute stroke patients with normal perfusion CT and normal CT angiography Bernd Eckert & Tobias Küsel & Andreas Leppien & Peter Michels & Axel Müller-Jensen & Jens Fiehler

Received: 3 December 2009 / Accepted: 5 April 2010 / Published online: 27 April 2010 # Springer-Verlag 2010

Abstract Introduction Acute stroke multimodal CT imaging (MMCT: non-enhanced CT, CT angiography, and CT perfusion (CTP)) may show normal results despite persistent clinical stroke. We prospectively evaluated the sensitivity/specificity of MMCT infarct detection and the clinical outcome in patients with normal MMCT findings. Methods From April 2007 to April 2008, all patients with acute hemispheric stroke within 6 h of symptom onset who underwent complete MMCT and MRI follow-up imaging were included. MMCT analysis included occlusion type, early infarct hypodensities (EIH), mean transit time (MTT), and cerebral blood volume (CBV) maps according to Alberta Stroke Program Early CT Score (ASPECTS). Clinical assessment included symptom onset to CT scan (≤3 h/>3 h), the National Institute of Health Stroke Scale score (admission/discharge), and the modified Rankin scale (mRS) 90 days after stroke onset. Results One hundred seven were included (mean age, 68.4 years; ≤3 h, n = 84; >3 h, n = 23; intravenous thrombolysis (IVT), n=51; ≤3 h, n=40; >3 h, n=11). In patients with normal MMCT on admission (n=54), followB. Eckert (*) : A. Leppien Department of Neuroradiology, Asklepios Klinik Altona, Abt. für Neuroradiologie, Paul-Ehrlich-Str. 1, 22763 Hamburg, Germany e-mail: [email protected] T. Küsel : P. Michels : A. Müller-Jensen Department of Neurology, Asklepios Klinik Altona, Hamburg, Germany J. Fiehler Department of Neuroradiology, University Hospital Hamburg-Eppendorf, Hamburg, Germany

up MRT detected brain infarctions in 23 patients (lacunar strokes, n=16; infratentorial strokes, n=4; territorial infarction, n=3). Sensitivity/specificity/positive predictive value/ negative predictive value of any infarct detection was 69.5%/99.8%/99.9%/57.2% and of a any territorial infarct detection was 93.9%/99.9%/99.9%/93.6%, respectively. In univariate regression analysis (time to CT scan, ≤3 h/>3 h; IVT: yes/no; ASPECTS EIH/CBV/MTT, 10/<10), only the evidence of normal CTP (ASPECTS MTT=10) had a statistically significant impact (p=0.02) on a good outcome (mRS 0.1). Conclusion MMCT sensitivity in acute lacunar or infratentorial stroke was poor. But, we found a high specifity and a fairly good sensitivity in territorial infarct detection. In acute stroke patients with normal MMCT findings on admission, a good clinical prognosis can be expected. Keywords Perfusion CT . Acute stroke . ASPECTS . Thrombolysis . CT angiography

Introduction The first and most important aim in acute stroke imaging is to rule out intracranial bleeding and other pathologies than ischemic stroke. Compared with conventional nonenhanced CT (NECT), recent multimodal CT (MMCT) imaging, including CT angiography (CTA) and CT perfusion (CTP), has improved diagnostic sensitivity and accuracy of ischemic lesions in acute stroke patients [1, 2], but the clinical impact especially of CTP still has to be proven clearly. In a large series with MRI confirmation, even a valid penumbral tissue imaging combining relative mean transit time (MTT) and absolute cerebral blood volume (CBV) could be demonstrated [3]. A recent study

80

has demonstrated that the most accurate assessment of the site of occlusion, infarct core, salvageable brain tissue, and collateral circulation in acute stroke patients is afforded by a combination of CTP and CTA [4]. However, depending on the type of the multidetectorrow CT technology, CTP only covers limited areas of the brain. The commonly used 16- to 64-slice CT scanners can detect 2- to 4-cm brain tissue. Infratentorial, lacunar, or peripheral territorial strokes outside the CTP slices may be missed and result in a negative CTP despite persistent clinical symptoms. Diagnosis and therapeutic approach in those patients remain a clinical problem. In chronic or subacute infarction, CTP alone may also demonstrate falsenegative results. High accuracy in all aspects of acute stoke diagnosis demands the combined analysis of all NECT, CTA, and CTP. The purpose of the present study was to evaluate the specifity and sensitivity of MMCT imaging and to analyze the clinical outcome of acute stroke patients with normal MMCT findings. The study was designed as a prospective trial during a 1-year period with a standardized neuroradiologic MMCT imaging and interpretation protocol, using an established protocol of intravenous thrombolysis (IVT), thus reflecting real-life clinical conditions in a highvolume neurovascular center.

Methods Inclusion criteria From April 2007 to April 2008, all acute stroke patients were included in this study if the following conditions were met: 1. The patient presented with an acute, persistent neurological deficit suggesting hemispheric stroke within 6 h of symptom onset. 2. A NECT scan excluded intracerebral hemorrhage. 3. Patient underwent complete multimodal CT imaging with native CT, CTA, and PCT. 4. Follow-up MRI was performed within 2 to 5 days to confirm the extent of final infarction. In case of clinical contraindication to MR imaging, the follow-up scan was carried out with a CT scan. CT imaging The CT examination was performed with a 40-rowmultidetector CT scan, Brilliance 40 (Philips, Eindhoven, The Netherlands). All patients with suspected acute stroke and without notable renal insufficiency or contrast agent allergy underwent the following imaging protocol: NECT

Neuroradiology (2011) 53:79–88

and PCT at four cross-sectional positions and CTA of the cervical and intracranial vessels. Imeron 400 (Bracco Imaging, Konstanz, Germany) was used as contrast agent in CT angiography and in perfusion CT. Overall imaging time took 15–20 min. Imaging protocol Native CT scan: sequential imaging, 120 kV, 500 mA; slice reconstruction: cerebrum—5 mm, skull base 2.5 mm, window setting; cerebrum—35/65 HU, skull base, 35/ 80 HU. Perfusion CT, 80 kV, 200 mA. The 40-ml bolus of Imeron 400 was applied into an antecubital vein, using a power injector with an injection rate of 5 ml/s. CT scanning was started 7 s after the injection of the contrast material. Rotation time, 1.5 s for 60 s; reconstructed slice thickness, 10 mm. Brilliance 40 CT technology permitted the assessment of four adjacent 10-mm slices, allowing the assessment of 4-cm scan length. The PCT slices were done in the same inclination angle as the NECT scan and were positioned above the sinus frontalis. PCT maps covered brain sections from the basal ganglia up to the centrum semiovale. CTP post-processing: PCT data were analyzed by using a prototype-software (Philips Medical System, Best, The Netherlands). The software relies on the central volume principal [5] and applies a closed form (non-iterative) deconvolution to calculate the MTT. This operation requires a reference arterial input function, selected by the software in a region of interest that the user draws around the anterior cerebral artery (ACA). The CBV map is calculated from the areas under the time enhancement curves. A simple equation combining CBV and MTT values allows the calculation of cerebral blood flow (CBF=CBV/MTT) [5]. Finally, the software provides four different maps for each 10-mm section: MTT, CBF, CBV, and TTP (time to peak). CT angiography, 120 kV, 200 mA. The 50-ml bolus of Imeron 400 was applied into an antecubital vein, using a power injector with an injection rate of 5 ml/s. The scan started automatically 5 s after the bolus had reached the defined slice in the aortic arch (bolus-tracking program). The CT angiography covered the extracranial vessels from the aortic arch up to the cranial crest. Reconstruction thickness of the transversal planes was 1 mm; analysis included coronary and sagittal reconstructions and threedimensional MP range reconstructions of the intracranial vessels. Clinical assessment Clinical assessment included the National Institute of Health Stroke Scale (NIHSS) on admission and at discharge.

Neuroradiology (2011) 53:79–88

Patients with severe comorbidities with contraindications against intravenous thrombolysis were not included in the study. Blood pressure values and serum glucose levels at admission were not evaluated systematically. Clinical follow-up was measured with the modified Rankin scale (mRS) 90 days after stroke onset by telephone interview. The study was approved by the ethical committee of the “Hamburger Ärztekammer”. CT data interpretation Two senior neuroradiologists (B. E. and A. L.) prospectively analyzed all multimodal acute CT scans on admission, including the Alberta Stroke Program Early CT Score (ASPECTS) scoring. At night and during the weekend, MMCT findings were evaluated by telephone (B.E. and A.L.) with neuroradiologically trained radiologic staff on duty. In unclear cases, the neuroradiologist came into the hospital for final diagnostic decision. All other cases were confirmed the next morning by the expert neuroradiologist before any follow-up imaging had been performed. Due to the limited PCT area, volumetric or whole brain assessment could not be performed. For this reason, semiquantitative assessment using the ASPECTS [6] was applied, because PCT levels covered both representative brain areas at the basal ganglia and the centrum semiovale. Multimodal CT diagnostic included the evaluation of early infarct hypodensities (EIH), MTT, and CBV maps according to the ASPECTS [7]. The assessment lead to a score ranging from 10 (no lesion) to 0 (lesions include all 10 areas: caudate nucleus, lenticulate nucleus, internal capsule, insular cortex, and M1 to M6 cortical regions). The arterial territories of ACA and the posterior cerebral arteries (PCA) were included in the diagnostic process, especially to define an extensive early infarct, but were not measured with the ASPECT score. Patients with an occlusion of the basilar artery in the CTA were not included. Main diagnostical CT criteria included EIH on the NECT scan, the MTT lesion to identify the penumbral tissue, severe CBV reduction (<2 ml/mg brain tissue), and the type of vessel occlusion. If the EIH ASPECTS was 7 or less, the patient was considered to have an elevated risk of a stroke of more than one third of MCA territory. However, the decision to withdraw IVT due to extensive early infarct signs was done by the neuroradiologist in the individual case. Penumbral tissue in case of late (>3 h) treatment onset was defined as a mismatch between the ASPECTS of MTT lesion and the extent of CBV/EIH lesion by the expert neuroradiologist. In case of different CBV and EIH findings, EIH was defined as the infarct core. The Philips software, demonstrating automatically calculated summary maps of penumbra and infarct core, were not used for the diagnostic decision.

81

The type of arterial vessel occlusion was defined according to the CTA and CTP findings and categorized as follows: Type 1: Occlusion of the intracranial carotid T bifurcation Type 2: Tandem ICA and proximal MCA occlusion Type 3: Proximal MCA occlusion(M1 to proximal M2 (trifurcation) segment) Type 4: Peripheral MCA occlusion, detected on CTP Type 5: No CTA occlusion, no CTP lesion Intravenous thrombolysis, treatment protocol Treatment decisions on thrombolytic therapy were based on the clinical status, assessed with NIHSS, time window considerations, and the findings of multimodal CT imaging. Within 3 h of symptom onset, patients presenting with a persistent neurological deficit of NIHSS≥4 and without contraindication against thrombolytic therapy were treated with intravenous rt-PA, according to the national and European guideline of thrombolytic stroke treatment. All acute stroke patients being potential candidates for thrombolytic treatment within 6 h of symptom onset were prospectively included during a 1-year period. According to the study of Barber et al. [8], 32% of patients who were excluded from IVT due to mild or significantly improving neurologic symptoms on admission were either dependent at discharge or dead during hospitalization. Persistent vessel occlusion with initially sufficient but unstable collateralization is suspected as one major pathology in those cases. For this reason, our acute stroke diagnosis and treatment protocol also included patients with improving symptoms or mild clinical deficits NIHSS<4. IVT was considered as a treatment option in case of persistent vessel occlusion. In the extended time window from 3 to 6 h after symptom onset, thrombolytic treatment was considered, if the multimodal CT imaging revealed EIH<1/3 of the MCA territory and if penumbral tissue was clearly identified with perfusion imaging (MTT lesion > EIH lesion). All diagnostic and therapeutic decisions were done with one neuroradiologist (B.E. or A.L.) and one experienced stroke neurologist (B. M. or A. M.-J.). At night and during weekends, treatment decisions were done by telephone with the radiologic and neurologic staff on duty. Bleeding complications Symptomatic intracranial hemorrhage (SICH) was defined as evidence of intracranial bleeding, combined with a clinical deterioration of NIHSS≥4. Parenchymal hemorrhage (PH, blood clot not exceeding 30% of the infarct area with mild

82

space occupying effect, without clinical deterioration [9]) was also assessed. Follow-up imaging Follow-up imaging was performed with MRI within 2– 5 days in order to identify the extent of final infarction. MR imaging was provided with a 1.5-T MRI (Intera, Philips, Eindhoven, The Netherlands) and contained the following sequences: transversal diffusion-weighted imaging (DWI: TR 3,726 ms, TE 81 ms), transversal T2-weighted imaging (TR 4,249 ms, TE 100 ms), transversal flair (TI 2,000 ms, TR 6,000 ms, TE 100 ms), sagittal T1-weighted imaging (TR 472 ms, TE 81 ms), and a time of flight MR angiography (TR 17 ms, TE 7 ms) with multiplanar reconstructions. In case of clinical contraindications, follow-up imaging was done with a NECT scan as described above. Recanalization success, which is restricted to a subcohort of patients with detected CTA occlusion, was not systemically evaluated. All CTP-negative patients, however, were examined by MRI. The extent of the final infarction was also evaluated with the ASPECT score by one of two expert neuroradiologist not blinded to the initial data. Additional detected infarcts within or beyond the CTP (ASPECTS) slices were defined as (1) supratentorial lacunar infarction, (2) brainstem infarction, and (3) territorial infarction. Clinical outcome analysis Statistical analysis of clinical outcome was performed after dichotomizing the treatment variables as follows: time window (symptom onset to CCT), ≤3 h/>3 h; IVT: yes/no; ASPECTS EIH on admission, 10/<10; ASPECTS CBV on admission, 10/<10; ASPECTS MTT on admission, 10/<10. Neurologic outcome 90 days after stroke was dichotomized in good (mRS 0.1) and poor outcome (mRS 2–6). Statistical analysis Differences in detection rate of cerebral infarcts with the multimodal CT imaging were compared with Fischer’s exact test. The neurological outcome analysis related to different treatment factors was calculated with univariate regression analysis using SPSS 13 (Chicago, IL, USA).

Results The present study prospectively included 107 patients (54 female, 53 male, mean age 68.4 years, range 26–87 years).

Neuroradiology (2011) 53:79–88

Multimodal CT imaging was performed within 3 h of symptom onset in 84 patients and beyond 3 h of symptom onset in 23 patients. MRS after 90 days revealed favorable outcome (mRS 0.1) in 54 patients and unfavorable outcome (mRS 2–5) in 50 patients; three patients died. Table 1 summarizes the main clinical and neuroradiological data. On admission, a CTP lesion was detected in 53 patients. CTA disclosed 34 patients with a proximal vessel occlusion. CTP detected 19 additional patients with a perfusion deficit due to peripheral MCA occlusion. In 54 patients, multimodal CT imaging did not detect any lesion (CTP negative). The ASPECT score of EIH, MTT, and CBV on admission and the final infarct related to the occlusion type is documented in Fig. 1. NIHSS on admission and discharge related to the occlusion type is documented in Fig. 2. IVT was performed in 51 patients (time window ≤3 h: n= 40, >3 h: n=11). In the extended time window (>3 h), 12 patients were not treated with IVT (EIH >1/3: n=1, no penumbral tissue: n=11). The underlying vessel occlusion in the IVT treated patients were as follows: occlusion of the intracranial carotid T bifurcation: n=5; tandem ICA and MCA occlusion: n=4; proximal MCA occlusion: n=18; peripheral MCA occlusion: n=14; no occlusion: n=10. IVT protocol violations (n=4) One patient with severe clinical symptoms (NIHSS 12) was treated beyond the 3-h window despite negative CTP (decision by neurologist). In clinical and MRI follow-up, postictal Todd’s palsy was identified as the underlying pathology (90-day mRS=2 as before the ictus). One patient with proximal MCA occlusion (NIHSS=22) was treated with IVT beyond 3 h, despite EIH>1/3 of the MCA territory, which was not correctly identified by the radiologist on duty but corrected by the neuroradiologist who came into the hospital. IVT was begun after first radiologic assessment but immediately stopped after neuroradiologic evaluation. MRI follow-up revealed a PH without clinical deterioration (90-day mRS = 2). Two patients with a NIHSS<4 were treated with IVT, one due to a severe dysphasia and the other due to a proximal MCA occlusion with evidence of penumbral tissue. Bleeding complications SICH occurred in one patient with a carotid T occlusion, treated with IVT within 3 h after symptom onset (90-day mRS=5). PH without clinical deterioration occurred in seven patients (proximal MCA occlusion (n=6); peripheral MCA occlusion (n=1)). IVT was performed in six of these seven patients.

83

Mean ASPECTS-Score related to occlusion type 10 9

Mean ASPECTS

8 7 6 5 4 3 2 1 0 Carotid-T

ICA+MCA EIH

MCA prox. CBV

MCA distal No occlusion

MTT

End

EIH: Early Infarct Hypodensities; CBV: Cerebral Blood Volume ; MTT: Mean Transient Time; ASPECT: Alberta Stroke Program Early CT Score

Fig. 1 Mean ASPECTS Score related to occlusion type. EIH early infarct hypodensities, CBV cerebral blood volume, MTT mean transit time, ASPECT Alberta Stroke Program Early CT Score

Infarct detection Follow-up MRI revealed infarctions in all CTP positive patients and detected brain infarctions in 23 out of 54 CTP negative patients (lacunar strokes inside the perfusion map: n=16, see Fig. 3; infratentorial infarction: n=4, territorial infarction beyond the initial perfusion maps: n=3 (MCA:

Mean NIHSS related to occlusion type 20 18 16 14

Mean NIHSS

NIHSS National Institute of Health Stroke Scale, ASPECTS Alberta Stroke Program Early CT Score, IVT intravenous thrombolysis, EIH early infarct hypodensities, CBV cerebral blood volume, MTT mean transit time, mRS modified Rankin scale

2.1 5.2 4.0 3.3 2.3 1.0 8.1 3.5 5.5 6.1 8.4 9.7 7.2 2.7 1.7 3.1 7.4 10.0 8.8 4.7 8.3 6.5 9.3 10.0 9.2 5.8 8.3 8.3 9.5 10.0 51 5 4 18 14 10 Total Carotid T ICA + MCA Proximal MCA Distal MCA No occlusion

107 6 6 22 19 54

68.4 68.2 65.7 70.8 69.9 67.1

84 6 6 17 12 43

23 0 0 5 7 11

8.3 19 15.7 13.6 7.5 4.4

56 1 2 4 5 44

ASPECTS mean CBV IVT Mean age Samples Occlusion type

Table 1 Clinical and multimodal CT findings.

≤3 h

>3 h

NIHSS on admission

No IVT

ASPECTS mean EIH

ASPECTS mean MTT

ASPECTS mean End

mRS 90 days mean

Neuroradiology (2011) 53:79–88

12 10 8 6 4 2 0 Carotid-T

ICA+MCA

MCA prox.

Admission

MCA distal No occlusion

Discharge

NIHSS: National Institute of Health Stroke Scale

Fig. 2 Mean NIHSS related to occlusion type. NIHSS National Institute of Health Stroke Scale

84

n=2, ACA; n=1)). In 31 patients, follow-up imaging revealed no signs of any infarction. Postictal Todd paresis due to an epileptic seizure as reason of the acute stroke symptoms was diagnosed by clinical findings and electroencephalogram in six patients. Compared with the neurologic status before the qualifying stroke event, mRS after 90 days remained stable in all six patients. TIA was diagnosed in the remaining 25 patients. Statistical analysis of CTP detection of any infarct, using Fischer’s exact test, revealed a positive predictive value (PPV) of 99.9%, a negative predictive value (NPV) of 57.2%, a sensitivity of 69.5%, and a specificity of 99.8%. Statistical analysis of CTP detection of a territorial infarct, using the Fischer’s exact test, revealed a PPV of 99.9%, a NPV of 93.6%, a sensitivity of 93.9%, and a specificity of 99.9%. Stroke with negative CTP and NIHSS≥4 on admission Mean NIHSS score was 4.4 in CTP negative patients versus 12 in patients with positive CTP findings (Table 2). Out of the 54 patients with negative CTP imaging on admission, 30 patients presented with a stable neurological deficit of NIHSS≥4 (mean NIHSS=7). Follow-up MRI revealed lacunar infarcts in 12 patients and brainstem infarcts in four patients. Territorial infarction beyond the

Fig. 3 A 60-year-old man suffering right-sided hemiparesis (NIHSS=6). CT imaging was performed 2.5 h after symptom onset. a–d Acute MMCT findings showing a residual lacunar infarct in the right internal capsule but no evidence of left hemispheric ischemia. a Native CT scan, b CTP source image in the identical level with c CBV map and d MTT map. e–f MRI 2 days after symptom onset demonstrating an acute lacunar infarction in the left capsula interna. e Diffusion-weighted imaging. f FLAIR sequence

Neuroradiology (2011) 53:79–88

initial perfusion maps was found in one patient. In six patients with negative follow-up MRI, postictal Todd’s palsy could be diagnosed. TIA with negative follow-up MRI was diagnosed in the remaining 11 patients. In 10 patients, IVT was performed (<3 h: n=9, >3 h: n=1, see “IVT protocol violations (n=4)”). Neurologic outcome was favorable (mRS 0–1) in 18 patients (60%) and unfavorable (mRS 2–6) in 12 patients (40%), including two patients with a mRS of 2 before the qualifying stroke. With IVT (n=10), neurologic outcome was favorable in five and unfavorable in another five patients. Without IVT (n=20), neurologic outcome was favorable in 13 (65%) and unfavorable in seven (35%) patients. In patients with negative perfusion CT and NIHSS<4 on admission (n=24, no IVT), all but one patient had a good neurologic outcome. This patient was discharged without neurological symptoms and expired 2 months later for cardiac reasons. In patients with positive perfusion CT (ASPECTS MTT<10, n=53) clinical outcome was good in 12 (23%) patients and poor in 41 (77%) patients, respectively. With NIHSS≥4 (n=49, mean NIHSS 13), good outcome was seen in eight (16%) patients and poor outcome in 41 (84%) patients. With IVT (n=39) good outcome was achieved in eight patients (21%) and poor outcome in 31 patients (79%), respectively. Without IVT (n=10), poor outcome was seen all patients. In the four patients with a

Neuroradiology (2011) 53:79–88 Table 2 CTP/NIHSS on admission and neurologic outcome.

ASPECTS Alberta Stroke Program Early CT Score, CTP CT perfusion (normal: MTT=10, pathologic: MTT<10), NIHSS National Institute Health Stroke Scale, IVT intravenous thrombolysis, mRS modified Rankin scale

85 Treatment factor

Samples

Good outcome mRS 0.1

Poor outcome mRS 2–6

Normal CTP total NIHSS≥4 IVT No IVT NIHSS<4 IVT No IVT Pathologic CTP total NIHSS≥4 IVT No IVT NIHSS<4 IVT No IVT

54 30 10 20 24 0 24 53 49 39 10 4 2 2

41 18 5 13 23 0 23 12 8 8 0 3 1 2

13 12 5 7 1 0 1 41 41 31 10 1 1 0

Univariate logistic regression analysis was performed to identify factors affecting neurological outcome. Significance level was predefined as p<0.05 (Table 3).

Related to good outcome (mRS 0.1) versus poor clinical outcome (mRS 2–6), univariate regression analysis revealed no significant differences for the treatment factors: time window (p=0.82), IVT (p=0.35), ASPECTS EIH on admission (p= 0.58) and ASPECTS CBV on admission (p=0.71). Only the results of ASPECTS MTT on admission were found to be statistically significant (p=0.02). Acute stroke patients with negative perfusion CT on admission (ASPECTS MTT=10)

Table 3 Treatment factors and neurologic outcome.

Total

NIHSS<4, clinical outcome was good in three cases (IVT: n=2, see “IVT protocol violations (n=4)”). Clinical outcome analysis

Outcome factors

n

Neurologic outcome Good, mRS 0.1

mRS modified Rankin scale, IVT intravenous thrombolysis, ASPECTS Alberta Stroke Program Early CT Score, EIH early infarct hypodensities, CBV cerebral blood volume, MTT mean transient time a

Statistically significant

Poor, mRS 2–6

n

%

n

%

53

49.5

54

50.5

Total, p value Time window, p=0.82 ≤3 h

107 84

42

50

42

50

>3 h IVT, p=0.35 Yes No EIH ASPECTS, p=0.58 10 <10 EIH ASPECTS, p=0.58 10 <10 MTT ASPECTS, p=0.02a

23

11

47.8

12

52.2

51 56

15 38

29.4 67.8

36 18

70.6 32.2

74 33

47 6

63.5 18.2

27 27

36.5 81.8

70 37

46 7

65.7 18.9

24 30

34.3 81.1

54 53

41 12

75.9 22.6

13 41

24.1 77.4

10 <10

86

were found to have a significantly better clinical outcome than patients with a detected hypoperfusion (ASPECTS MTT<10).

Discussion The present study is the first study to systematically address MRI follow-up and clinical outcome in acute stroke patients with normal MMCT findings, including NECT, CTA, and a normal perfusion CT, covering a 4-cm brain tissue. A normal MMCT on admission was found to be a frequent finding in 54 out of 107 included patients. We found a high accuracy to detect territorial infarction but only a 70% sensitivity to detect any infarction. The most frequent reason of false-negative MMCT was found to be lacunar infarcts within the CTP perfusion maps. The present prospective trial also identified a normal MMCT being a significant predictor of a good clinical outcome (mRS 0.1) at 3 months. The present study included acute stroke patients within 6 h of symptom onset, inclusive patients with improving symptoms, or mild clinical deficits NIHSS<4. With these inclusion criteria, a normal MMCT was found in half of the enrolled patients. Follow-up MRI detected brain infarcts in 23 of the 54 patients with normal MMCT, the majority being lacunar infarcts (n=16) within the CTP perfusion maps. Infratentorial infarcts were seen in four cases, and territorial infarcts beyond the perfusion maps were detected in another three cases. Compared with imaging follow-up, a normal MMCT revealed a 100% specifity and a 70% sensitivity of any infarct detection. After excluding infratentorial and lacunar strokes, however, the sensitivity to detect embolic infarcts increased to 94%. Out of the 31 patients with negative follow-up MRI, Todd’s palsy was diagnosed in six patients [10], and TIA was diagnosed in the remaining 25 patients. The data are in accordance with a recent single center study [11] comparing CTP and DWI in acute stroke patients. Any stroke sensitivity with CTP, covering 2 cm brain tissue, was found to be only 50%, but after excluding small vessel stroke, 71 out 77 DWI positive cortical strokes were detected in CTP, increasing embolic stroke sensitivity to 92%. Of 265 patients without DWI lesion, none had shown CTP abnormalities. In another MRI-confirmed study, CTP covering a 24-mm brain tissue revealed an infarct sensitivity of 64.6% [1]. Excluding small vessel infarcts, territorial infarct sensitivity was found to be 85%. Using a toggling table technique with a 64-slice scanner, the z-axis coverage can be increased up to 80 mm [12]. But, even with a 320detector row CT [13] enabling whole brain coverage, the detection of acute lacunar or pontine small vessel infarcts within the scanned CTP volume may remain a diagnostic

Neuroradiology (2011) 53:79–88

problem. In the present study, all missed lacunar infarcts were located within the CTP perfusion maps. In acute stroke patients with normal MMCT findings, immediate subsequent MRI may detect additional lesions that warrant specific treatment, but acute small vessel stroke may even be missed with MRI. In a prospective, blind comparison of NECT versus MRI in an emergency assessment of patients with suspected acute stroke, MRI had a sensitivity of 83% and NECT of 26% relative to the final clinical diagnosis of acute stroke [14]. False-negative MRI were associated with brain stem location, time from symptom onset less than 3 h, and mild clinical symptoms (NIHSS score <4). In another MRI study in 401 patients with suspected stroke, 26% of the enrolled patient had an initial negative DWI [15]. Of the 26 patients with a final clinical stroke diagnosis, 19 patients had either MRI follow-up evidence or clinical diagnosis of brain stem or lacunar stroke. The present study also evaluated the clinical outcome at 3 months. With our MMCT-based IVT protocol up to 6 h after symptom onset, SICH occurred in one patient with a carotid T occlusion (2.2%). In the univariate treatment factor analysis of our non-randomized study population, time from symptom onset (≤3 versus >3 h) and IVT were not found to have a significant impact on clinical outcome. The analysis of the MMCT findings were dichotomized in normal (ASPECT=10) or pathologic (ASPECT<10) results of the three brain tissue parameters EIH, CBV, and MTT to evaluate the predictive value of a normal result. Patients with normal EIH or CBV on admission had a better clinical outcome than patients with a detected EIH or CBV lesion. But, these differences were not found to be statistically significant. Only the finding of a normal CTP on admission (MTT-ASPECT=10) was found to have a statistically significant impact on good clinical outcome (mRS 0.1). A NIHSS≥4 is considered a cut-off value to indicate IVT in acute stroke. In treatment protocols based on NECT only, the exclusion of ICH suffices to indicate IVT within 3 h of symptom onset. A normal CTP on admission is not considered a contraindication against IVT, but the clinical benefit has not yet been proven. In the present study, 23 out of 24 patients with normal CTP and a NIHSS score <4 showed a good clinical outcome at 3 months. Out of the patients with a normal CTP and a NIHSS >4 (n=30), 60% still had a good clinical outcome compared with only 23% of patients with pathologic CTP findings on admission. IVT was performed in 10 patients with a normal CTP without any bleeding complications. A good clinical outcome was observed in 50% of the IVT patients, compared with 65% of the 20 patients without IVT treatment. Further imagingbased trials with a randomized design will be necessary to answer the important question whether IVT may be indicated in acute small vessel stroke, which is the

Neuroradiology (2011) 53:79–88

underlying pathology in most of the cases with normal MMCT findings. In another study analyzing the outcome of acute stroke patients with normal CTA results on admission, nearly 1/3 of patients who met the clinical criteria for IVT within 3 h of symptom onset were negative for arterial occlusion. Compared with patients with a CTA occlusion, patients without CTA occlusion had a lower NIHSS on admission (median 15 versus 11), a higher rate of favorable outcome (32% versus 45%), and a lower mortality rate (37% versus 15%). Control CT revealed lacunar infarcts in 18% and peripheral cortical stroke in 44% of the patients [16]. Limitations The major methodic limitation of the present study is the limited brain coverage of CTP, which will remain a substantial problem of acute stroke CT imaging in the next years. The inclusion criteria were not based on a randomized protocol but on our standard IVT protocol, which may have biased the enrollment towards patients with a low NIHSS on admission. This single-center study included 107 patients within 1 year, but the number of patients is still too small for general conclusions. Recanalization success is a decisive outcome factor in patients with a CTA occlusion, but cannot be measured in patients with normal MMCT findings, and thus, recanalization was not systematically evaluated. The ASPECTS method is the only method for semiquantitative evaluation of EIH and CTP results [17], but still comprises major restrictions also in the data interpretation of the present study. The degree of “substantial” involvement of an ASPECTS station has not yet been defined. A hemispheric pure ACA or PCA infarct with pathologic EIH and CTP findings may result in a normal ASPECTS scoring, which only refers to MCA territories. Clinical assessment included NIHSS on admission, mRS at 3 months, and time from symptom onset to CT scan. Blood pressure values and serum glucose levels on admission or other vascular comorbidities were not systematically evaluated.

Conclusion MMCT imaging in acute stroke, including NECT, CTA, and CTP, covering the brain sections from the basal ganglia to the centrum semiovale has a poor sensitivity in lacunar or infratentorial stroke detection but an acceptable high accuracy to detect a territorial hypoperfusion. Our results suggest that in acute stroke patients with a normal MMCT on admission, a good clinical prognosis can be predicted with high probability. The benefit of IVT in those cases

87

with underlying small vessel disease must be clarified in further imaged-based randomized trials.

Conflict of interest statement of interest.

We declare that we have no conflict

References 1. Lin K, Do KG, Ong P, Shapiro M, Babb JS, Siller KA, Pramanik BK (2009) Perfusion CT improves diagnostic accuracy for hyperacute ischemic stroke in the 3-hour window: study of 100 patients with diffusion MRI confirmation. Cerebrovasc Dis 28:72–79 2. Lovblad KO, Baird AE (2009) Computed tomography in acute ischemic stroke. Neuroradiology 52:175–187 3. Wintermark M, Flanders AE, Velthuis B, Meuli R, van Leeuwen M, Goldsher D, Pineda C, Serena J, van der Schaaf I, Waaijer A, Anderson J, Nesbit G, Gabriely I, Medina V, Quiles A, Pohlman S, Quist M, Schnyder P, Bogousslavsky J, Dillon WP, Pedraza S (2006) Perfusion-CT assessment of infarct core and penumbra: receiver operating characteristic curve analysis in 130 patients suspected of acute hemispheric stroke. Stroke 37:979–985 4. Tan JC, Dillon WP, Liu S, Adler F, Smith WS, Wintermark M (2007) Systematic comparison of perfusion-CT and CTangiography in acute stroke patients. Ann Neurol 61:533–543 5. Wintermark M, Fischbein NJ, Smith WS, Ko NU, Quist M, Dillon WP (2005) Accuracy of dynamic perfusion CT with deconvolution in detecting acute hemispheric stroke. AJNR Am J Neuroradiol 26:104–112 6. Barber PA, Demchuk AM, Zhang J, Buchan AM (2000) Validity and reliability of a quantitative computed tomography score in predicting outcome of hyperacute stroke before thrombolytic therapy. ASPECTS study group. Alberta Stroke Programme Early CT Score published erratum appears in Lancet 2000 Jun 17;355 (9221):2170. Lancet 355:1670–1674 7. Lee JY, Kim SH, Lee MS, Park SH, Lee SS (2008) Prediction of clinical outcome with baseline and 24-hour perfusion CT in acute middle cerebral artery territory ischemic stroke treated with intravenous recanalization therapy. Neuroradiology 50:391–396 8. Barber PA, Zhang J, Demchuk AM, Hill MD, Buchan AM (2001) Why are stroke patients excluded from TPA therapy? An analysis of patient eligibility. Neurology 56:1015–1020 9. Berger C, Fiorelli M, Steiner T, Schabitz WR, Bozzao L, Bluhmki E, Hacke W, von Kummer R (2001) Hemorrhagic transformation of ischemic brain tissue: asymptomatic or symptomatic? Stroke 32:1330–1335 10. Mathews MS, Smith WS, Wintermark M, Dillon WP, Binder DK (2008) Local cortical hypoperfusion imaged with CT perfusion during postictal Todd's paresis. Neuroradiology 50:397–401 11. Rai AT, Carpenter JS, Peykanu JA, Popovich T, Hobbs GR, Riggs JE (2008) The role of CT perfusion imaging in acute stroke diagnosis: a large single-center experience. J Emerg Med 35:287– 292 12. Youn SW, Kim JH, Weon YC, Kim SH, Han MK, Bae HJ (2008) Perfusion CT of the brain using 40-mm-wide detector and toggling table technique for initial imaging of acute stroke. AJR Am J Roentgenol 191:W120–W126 13. Salomon EJ, Barfett J, Willems PW, Geibprasert S, Bacigaluppi S, Krings T (2009) Dynamic CT angiography and CT perfusion employing a 320-detector row CT: protocol and current clinical applications. Klin Neuroradiol 19:187–196

88 14. Chalela JA, Kidwell CS, Nentwich LM, Luby M, Butman JA, Demchuk AM, Hill MD, Patronas N, Latour L, Warach S (2007) Magnetic resonance imaging and computed tomography in emergency assessment of patients with suspected acute stroke: a prospective comparison. Lancet 369:293–298 15. Sylaja PN, Coutts SB, Krol A, Hill MD, Demchuk AM (2008) When to expect negative diffusion-weighted images in stroke and transient ischemic attack. Stroke 39:1898–1900

Neuroradiology (2011) 53:79–88 16. Mikulik R, Goldemund D, Reif M, Aulicky P, Krupa P (2009) Outcome of patients with negative CT angiography results for arterial occlusion treated with intravenous thrombolysis. Stroke 40:868–872 17. Lin K, Rapalino O, Lee B, Do KG, Ong P, Shapiro M, Babb JS, Siller KA, Pramanik BK (2009) Correlation of volumetric mismatch and mismatch of Alberta Stroke Program Early CT Scores on CT perfusion maps. Neuroradiology 51:17–23