By: Peter D. Schellinger, M.D., PH.D.

Noncontrast CT imaging is the most widely used diagnostic tool for stroke imaging,1 mainly due to its almost 100% sensitivity for intracerebral hemorrhage (ICH), the most important differential diagnosis in ischemic stroke.2 The advent of new MR techniques such as perfusion- and diffusion-weighted imaging, however, has revolutionized diagnostic imaging of stroke.1,3

DWI may delineate infarcted brain tissue within minutes, and PWI defines the area of cerebral hypoperfusion. The absolute volume difference or ratio of PWI and DWI reveals the ischemic tissue at risk of irreversible infarction.4,5 MRI further allows a definitive diagnosis of ICH within the first hours of stroke on susceptibility-weighted T2*-weighted sequences, due to a profound signal loss caused by paramagnetic effects of deoxyhemoglobin, even when it is present only in traces.6,7 Detection of old as well as new microbleeds and early hemorrhagic infarction by T2*WI may allow exclusion from thrombolysis of patients who have an excessive bleeding risk.8

Several studies have reported early findings of stroke MRI within the first six to 12 hours, demonstrating the feasibility and practicality of this method in the setting of acute stroke and thrombolytic therapy.9-11 In essence, the presence of a vessel occlusion on MR angiography is associated with a PWI/DWI mismatch, the MRI setting that defines the ideal candidate for thrombolysis. In addition, early recanalization achieved by thrombolysis results in significantly smaller infarcts and a significantly better clinical outcome.

With regard to the controversy regarding the utility of stroke MRI because of its low availability,12 we believe that the constellation of the DWI, PWI, MRA, T2WI and PWI/DWI mismatch allows early identification of patients in whom outcome and final infarct size-ultimately, the patient's fate-have not yet been determined. With an increasing distribution and 24/7 availability of stroke MRI, the identification of patients who are suitable for thrombolytic therapy and of those who are not will lead to increased benefit and reduced complications. Furthermore, the rather strictly defined therapeutic window may be qualified and individualized according to the findings in each patient. Finally, stroke MRI may be useful in identifying patients with extensive infarctions, with or without a PWI/DWI mismatch, which might be salvaged by more aggressive means.

One serious drawback of stroke MRI remains its overall low availability, even though three to five years have passed since the first reports of its implementation in the clinical routine. Modern CT scanners of the third, fourth, or even fifth generation (volume scanners), however, are less expensive and are available even in smaller community hospitals, where they are used primarily for extracranial scanning. Acute stroke is not treated only at specialized academic medical centers; the majority of patients first present in local general hospitals that have no MR facilities.13 Several meetings have included discussions on how to increase and enhance the diagnostic reach of CT-based stroke imaging to obtain better information about the physiological parameters (ICH or ischemic stroke, size of infarct core and penumbra, vessel status) identified on stroke MRI.

NONCONTRAST CT

A standard CT protocol includes axial scanning with a slice thickness of 3 to 5 mm for the posterior fossa and 8 to 10 mm for suprasellar structures. The differentiation of ischemic stroke and ICH by clinical means alone is impossible. Although early deterioration of vigilance, vomiting, anterior circulation as opposed to posterior circulation syndrome, or hypertensive crisis may hint of ICH, these signs and symptoms can also be seen in ischemic stroke. An effort was made in the early 1990s to define the early signs of ischemic infarction on CT scans, based on CT images collected in thrombolysis trials.

On the other hand, the erroneous belief that CT is insensitive to changes within the first hours after ischemic stroke is still widely held. Even newer textbooks and recent papers in highly regarded non-neuroradiological journals claim that CT is negative with regard to ischemic tissue changes during the first 12 hours following symptom onset.14,15

Intravenous thrombolytic therapy with rtPA requires documentation of an acute-onset focal neurological deficit with a baseline National Institutes of Health Stroke Score (NIHSS), adhering to the general blood pressure guidelines, accounting for general recombinant tissue plasminogen activator contraindications, and exclusion of ICH by CT. These are the criteria by which rtPA was proved effective within three hours in the NINDS trial in 1995 and approved by the FDA in 1996.16 Noncontrast CT-based diagnostic stroke workup is the absolute minimum requirement.

The neuroradiological community has generally accepted in the last few years that CT can demonstrate early infarct signs within the first six hours after stroke. With the availability of high-quality CT scanners, more investigators have reported positive CT findings in early ischemic stroke. The sensitivity of these findings varies widely, ranging from 12% to 92%, depending on the infarct signs, the exact time window of the investigated population, and the authors. The most common sign of an early infarct is a gray-matter hypodensity, which develops in the early stages of an infarction and can be subtle and thus difficult to detect.

The development of extracellular edema, due to a masking of cortical or deep nuclear structures, can be recognized even during the first few hours following stroke onset. This phenomenon is best demonstrated in cases of acute occlusion of the main stem of the middle cerebral artery (MCA), where the parenchymal hypodensity shows up first in terminal artery supply areas such as the basal ganglia.2,17,18

Depending on the origin of the recurrent artery of Heubner, this hypodensity also involves the head of the caudate nucleus. In cases with insufficient leptomeningeal collaterals, the primary hypodensity involves cortical structures such as the insula (the "loss of the insular ribbon sign")19 or other territories of the MCA. There is a correlation between the size of early hypodensities and the risk of a secondary hemorrhage and clinical outcome.2,20

Patients with signs of profound ischemia with a strong hypodensity should not be given rtPA even within the three-hour time window because of an excessive risk of ICH. The sensitivity of early CT findings in the six-hour time window for stroke diagnosis, however, was estimated to be around 60% (with an expert panel) and 45% to 50% (at specialized stroke centers), as opposed to DWI (90% with an expert panel and 80% at specialized centers),21 and probably worse in the three-hour time window only. A post hoc analysis of the NINDS trial CT data (n = 616) yielded a 31% sensitivity for early ischemic changes, a mild correlation with acute NIHSS, but no effect on clinical outcome or on secondary ICH rate.22

The ASPECTS score is a new instrument for the improvement of CT rating.23,24 It divides the MCA territory into 10 regions of interest as seen on two standardized axial CT slices (basal ganglia and lateral ventricles). The entire MCA territory is allotted 10 points, one for each area, and a single point is subtracted for each of the defined regions if ischemic lesions are seen there. Interrater variability statistics showed that the interobserver reliability of ASPECTS was higher than that of the 1/3 MCA rule, although other preliminary data contradict these findings (Werner Hacke, personal communication). The baseline ASPECTS value correlated inversely with the severity of stroke on the NIHSS (r = -0.56, p < 0.001) and predicted functional outcome and symptomatic ICH (p < 0.001; p = 0.012). A sharp increase in dependence and death occurs with an ASPECTS score of 7 or less.

While the ASPECTS score may be superior to the 1/3 MCA rule, it is more a refinement of that rule than a completely new development. Furthermore, it has not been validated for patients outside the three-hour time window.

Another simple method for improving the diagnostic accuracy of noncontrast CT is the use of nonstandard variable window width and level review settings.25 Sensitivity and specificity for stroke detection were 57% and 100%, respectively, with standard viewing parameters, but sensitivity increased to 71% without loss of specificity (p = 0.03) with narrow window and variable settings.

The sensitivity of noncontrast CT for ICH is high. Although they are clearly defined, the diagnostic impact of early infarct signs is debatable because these signs are often subtle and require a high level of experience for detection and interpretation. A large area of manifest hypodensity exceeding a third of the MCA territory, however, should be regarded as a contraindication for thrombolytic therapy.

CTA/CTA SOURCE IMAGES

Thanks to spiral CT technology, CT angiography has become widely available for evaluation of the circle of Willis. This technique can provide accurate information on stenoses or occlusions in the basal arteries of the brain in acute ischemic stroke patients.26,27 Direct comparison of CTA and Doppler ultrasound suggests that CTA compares favorably with ultrasound and can reliably detect intracranial stenosis, emboli, and aneurysms of a moderate or larger size.28

The method is noninvasive, safe, and not as operator-dependent as Doppler ultrasound. In a large series of stroke patients, none had immediate adverse reactions after administration of intravenous nonionic iodinated contrast material.27 Newer generations of CT scanners allow for lower contrast doses. While older studies reported an increase of infarct size after administration of ionic contrast material, an experimental study clearly showed that bolus injection of nonionic contrast material did not affect infarct volume or worsen the symptoms of cerebral ischemia,29 although this should be more firmly established.

CTA is superior to Doppler ultrasound in the assessment of basilar artery patency in patients with the syndrome of acute basilar artery ischemia, particularly in patients with distal basilar artery occlusion.30

In addition to assessing a major vessel occlusion, CTA has the potential to provide information about the quality of the collateral circulation. Contrast enhancement occurs in arterial branches beyond the occlusion in patients with good leptomeningeal collaterals. This degree of enhancement can be taken as an estimate of the collateral blood flow.26,31 CTA may have limitations, however, such as being less reliable in showing branch occlusions of the MCA or other smaller vessels distal to the circle of Willis.27 But many colleagues suggest that it may be superior to fast MRA techniques (personal communications).

On the other hand, CTA can easily be performed directly after noncontrast CT. In a standard CTA protocol, 65 mL of a nonionic contrast medium is injected into a cubital vein at a rate of 5 mL/sec, using an injection pump. After a delay of 15 to 20 seconds, spiral scanning is performed with a slice thickness of 2 mm, an index of 1.5 mm, and a spiral pitch of 1.25 at 130 kV and 125 mA. CTA source images (CTA-SI) and 3D reconstructions of the data sets can be used for diagnosis.

In addition to the vessel status, CTA-SI can provide indirect information about the collateral circulation and may improve the contrast of perfused and malperfused brain areas, increasing the sensitivity for early ischemic changes not seen on noncontrast scans.26,31 Analysis of CTA-SI must be clearly differentiated from perfusion CT, in which, analogous to PWI, a contrast bolus tracking method is applied and hemodynamic parameters may be assessed.32 CTA-SI analysis is a stronger predictor of clinical outcome than the initial NIHSS and may predict final infarct volume and clinical outcome.

Patients with recanalization do not experience infarct growth, whereas those without complete recanalization do.33 Schramm et al therefore investigated whether CTA-SI allows detection of ischemic brain lesions in patients with acute ischemic stroke, whether its sensitivity is comparable to that of DWI, whether the hypoperfused brain area seen on CTA-SI correlates with the final infarct, and whether the qualitatively assessed collateral status reflects the risk of infarct growth.34

Schramm's group analyzed clinical and imaging findings of 20 consecutive stroke patients (seven women, 13 men; mean age 61 plus/minus 10 years) imaged with both modalities within six hours (45 minutes to five hours and 45 minutes) after stroke onset. CT was performed within the first 45 minutes to four hours and 45 minutes after stroke onset (mean 2.83 plus/minus 1.33 hours), followed by MR (range from one hour and 15 minutes to five hours and 30 minutes after symptom onset; mean 3.38 plus/minus 1.37 hours). The time interval between CT and MR ranged from 15 minutes to one hour (0.55 plus/minus 0.25 hour).

Of the 20 patients, 16 had a vessel occlusion seen on both CTA and MRA. All but one patient had an abnormal initial DWI scan. Five of these patients showed good collaterals and 11 showed poor collaterals surrounding the lesion site. All vessel occlusions detected on CTA were seen on MRA at the same location. In every patient, the status of the collateral vessels surrounding the lesion was determined on the CTA-SI; seven patients showed good intravascular enhancement of the perilesional vessels and were classified as "good," and 13 showed poor enhancement around the lesion site and were classified as "poor."

CTA-SI lesion volumes did not differ significantly from DWI lesion volumes (p = 0.601) at baseline either in patients with poor collaterals (p = 0.807) or in those with good collaterals (p = 0.6). In patients with poor collateral vessel status, initial CTA-SI lesion volumes differed significantly from T2WI lesion volumes on day five (p = 0.0058), whereas in patients with good collaterals, no significant difference was found (p = 0.176). Day-one DWI lesion volumes differed significantly from T2WI lesion volumes on day five in patients with poor collaterals (p = 0.0035), whereas the difference did not reach statistical significance in patients with good collaterals (p = 0.176), signaling profound lesion growth in patients with poor collateral status.

In all patients, the lesion volume measured on CTA-SI significantly correlated with the initial lesion volumes on DWI (p < 0.0001, r = 0.922) and with the outcome lesion volume on day-five T2WI (p = 0.013, r = 0.736). Furthermore, patients with good collaterals uniformly had a significantly better clinical outcome on day 90, as measured by four neurological and outcome scales (NIHSS, SSS, BI, and mRS, all plesser than or equal to 0.012).

A poor collateral status also predicted a significantly worse clinical outcome (p = 0.025 and p = 0.001) for two dichotomized clinical outcomes (mRS 0 to 1 and mRS 0 to 2). Day-five T2WI lesion volumes differed between the subgroups and was significantly larger in patients with poor collaterals (p = 0.024).

These results lead decisively to the conclusion that in the management of acute stroke, when MRI is not available, the decision to initiate thrombolytic treatment should be based on the clinical findings and CT scanning, including CTA.

As the aim of thrombolytic therapy is to obliterate the thromboembolus, the ability of CTA to detect intracranial vessel occlusion suggests that it is a useful screening tool for identifying patients in whom intravenous or intra-arterial thrombolysis is appropriate.35 Since the therapeutic time window for thrombolytic therapy is only three hours or up to six hours in selected patients, the need for an improved, CT-based diagnostic tool is evident. A normal noncontrast-enhanced CT scan in acute stroke does not imply insensitivity of the method; in fact, it represents the favorable situation in which ischemic edema has not yet developed and the chance to avoid irreversible damage is still good.

Unenhanced CT does not show the arterial occlusion itself; it does not even show the extent of disturbed cerebral perfusion. One might ask if standard postcontrast CT should not suffice to visualize the extent of ischemic tissue. In comparison to postcontrast CT, however, CTA has the advantage of showing the leptomeningeal collaterals surrounding the lesion of reduced parenchymal enhancement, which can be seen with both methods. In patients with a poor collateral vessel status, CTA-SI may provide information similar to that of the PWI-DWI mismatch concept, analogous to the findings of Jansen et al4 and Schellinger et al.

The volume of the affected brain area that has inadequate blood supply can be estimated by the difference between the CTA-SI lesion volumes and the brain area supplied by the occluded artery, taking into account the qualitative assessment of the collateral status. Patients with poor collaterals seem to represent those who may have a PWI-DWI mismatch in stroke MRI, and patients with good collaterals represent those without tissue at risk (i.e., small stroke, lacunar stroke, or tissue at risk already completely infarcted).

The combination of CT and CTA may also be more cost-effective than stroke MRI.

DYNAMIC PERFUSION CT

Whereas CTA and CTA-SI provide angiographic information solely about the circle of Willis and collateral vessels, perfusion CT (PCT) delivers information about the cerebral blood flow (CBF). The necessary perfusion software is not yet available on many CT scanners serving emergency units, however. Dynamic CT during the first pass of a bolus of iodinated contrast agent by unidirectional x-ray tube rotation results in images that can be used to generate functional maps of cerebral blood volume (CBV), CBF, or time-to-peak enhancement (TTP). These functional maps of PCT enable the stroke physician to assess cerebrovascular parameters and their changes in acute stroke patients.36,37

Areas with reduced CBF can be shown immediately after vessel occlusion with high contrast to areas with normal perfusion. In a study conducted by Koenig et al, perfusion CT was performed within six hours of symptom onset in 32 patients with acute stroke symptoms. Comparison with SPECT performed immediately after perfusion CT and with follow-up CT showed good correspondence in 81% of the patients.

In another study, the same authors compared the findings on different functional maps of cerebral PCT in stroke patients with early ischemic changes on conventional CT. They retrospectively evaluated the baseline CT scans of 45 acute stroke patients with respect to early CT findings. They compared the extent of cerebral ischemia for each patient as shown on the maps of CBF, CBV, and TTP and assigned the severity of ischemia to one of three levels, based on the findings of the CBF image. Conventional CT was performed within two hours from symptom onset in 75% of these patients. Of 45 patients, 29 showed early signs of ischemia on conventional CT, while PCT revealed cerebral ischemia in all patients. Approximately the same rate of incidence of severe ischemia was found in patients with early CT changes (55.2%) and those with normal findings (43.8%).

If the perfusion impairment was judged mild or moderate, the extent of the hypoperfused area was significantly larger on the CBF and TTP images than on the CBV map. This was significantly different in patients with severe hypoperfusion, in whom a complete correspondence of the affected area between the three functional maps was found.

The authors concluded that the use of conventional CT for assessment of stroke in the hyperacute phase is limited, and PCT yields excellent information regarding the severity and extent of ischemia.

In addition, the use of various perfusion maps may help differentiate the core of infarction from the ischemic penumbra zone. Besides the delineation of hypoperfused brain tissue, the characterization of ischemia with respect to severity is of major clinical relevance because the degree of hypoperfusion is the most critical factor in determining whether an ischemic lesion turns into an infarct or returns to normal brain tissue.

In yet another study, Koenig et al attempted to determine whether measurements of the relative CBF, relative CBV, and relative TTP can be used to differentiate areas undergoing infarction from reversible ischemic tissue.39 PCT was used to calculate rCBF, rCBV, and rTTP values from areas of ischemic cortical and subcortical gray matter in 34 patients with acute hemispheric ischemic stroke less than six hours after onset. Results were obtained separately from areas of infarction and noninfarction, according to the findings on follow-up imaging studies. The efficiency of each parameter to predict tissue outcome was tested.

A significant difference was found between infarct and peri-infarct tissue for both rCBF and rCBV but not for rTTP. Threshold values of 0.48 and 0.60 for rCBF and rCBV, respectively, were found to discriminate best between areas of infarction and noninfarction, with the efficiency of the rCBV slightly superior to that of rCBF. The prediction of tissue outcome could not be increased by using a combination of various perfusion parameters. Thus, the assessment of cerebral ischemia by means of perfusion parameters derived from perfusion CT provides valuable information to predict tissue outcome.

Rother et al investigated 22 patients with CTP within 143 plus/minus 96 minutes of stroke onset and with CT within the first six hours of symptom onset and before the start of treatment in a consecutive clinical series.40 The area of the perfusion deficit from TTP maps, hemispheric lesion area from follow-up CT, final infarct volume, and stroke recovery (NIHSS) were assessed. Eighteen patients had perfusion deficits in the MCA territory and corresponding hypoattenuation in follow-up CT. Three patients with normal PCT findings showed lacunar infarctions or normal findings on follow-up CT. In one patient, PCT did not reveal a territorial deficit above the imaging slice. The overall sensitivity and specificity of PCT for detection of perfusion deficits in patients with proven territorial infarction (n = 18) on follow-up CT were 95% and 100%, respectively.

New and currently available advances in CT technology, including dynamic scanning with multisection data acquisition (multidetector CT), may further increase the value of this technique and provide information about the 3D extent of cerebral ischemia. Electron-beam CT is also capable of performing multislice dynamic CT with calculation of absolute CBV and CBF. In a small series of 11 patients with acute stroke symptoms, EBCT demonstrated reduced CBV and CBF in the four patients with proven infarcts on follow-up CT. High image noise limited the demarcation of ischemic tissue, however.41

In some patients in the studies conducted by Koenig et al,32,38,39 the ischemia was located outside the scanning level of perfusion CT and was therefore missed. This demonstrates the disadvantages of the method, which is that only one brain section (a maximum of two 1-cm slices with MDCT) can be evaluated with each bolus injection. In some of Koenig's patients, however, ischemia was located outside the scanning level of perfusion CT and was therefore missed. It is possible to obtain a second bolus injection to perform another perfusion CT, but it is impossible to evaluate the total brain.

Another disadvantage is that PCT, like PWI, but unlike PET and SPECT, renders only semiquantitative information about CBF.

INFORMED CONSENT

The following recommendations are not based on prospective randomized data and do not meet the criteria of an officially approved therapy. But in light of a substantial and still growing body of evidence in favor of the procedures, these recommendations may be seen as an expert opinion and provide the rationale for an individual therapeutic approach in an institutional protocol. The fact must be stressed when informed consent is obtained that an individual therapy is being offered that is based on advanced knowledge but that does not meet the criteria of drug approval institutions and therefore may be associated with a higher risk of hazardous if not fatal side effects.

Conversely, patients should be informed that the drawback of a later onset of therapy may be outweighed by a sophisticated diagnostic imaging procedure that tells the physician whether or not to treat them. Patients and their relatives should be informed not only about the hazards of thrombolytic therapy within or outside the three-hour time window but also about its potential benefits and the risk of not being treated.

ADVANCED STROKE CT ALGORITHM

Based on the information derived from CT/CTA/CTA-SI/PCT, patients can be assigned to one of four distinct categories:

- patients with infarctions of less than 33% of the MCA territory according to CT/CTA-SI;

- patients with large infarctions exceeding 50% of the MCA territory according to DWI;

- patients with MCA occlusions distal to the lenticulostriate branches and with the presence of a CT/CTA/CTA-SI/PCT mismatch; and

- patients with distal internal carotid artery or proximal MCA occlusions and the presence of a CT/CTA/CTA-SI/PCT mismatch.

Thrombolytic therapy should be an urgent consideration within the three-to-six-hour time window in infarctions of less than 33% of the MCA territory according to CT/CTA-SI. In the second category, the indication for thrombolysis is, in our opinion, moderate at best, and is based not on evidence but on questionable pathophysiological reasoning and thus may be a matter of individual debate. Intravenous and/or intra-arterial thrombolysis is indicated and essential in patients with MCA occlusions distal to the lenticulostriate branches and with presence of a CT/CTA/CTA-SI/PCT mismatch. Recanalization should be achieved by all means in patients with distal internal carotid artery or proximal MCA occlusions and the presence of a CT/CTA/CTA-SI/PCT mismatch.

We recommend against thrombolysis in patients with large infarctions exceeding 50% of the MCA territory according to DWI. The indication for intravenous thrombolysis is a matter of debate, but it is not performed in our center in the following patients:

- those without CT/CTA/CTA-SI/PCT mismatch but with vessel occlusion;

- those with CT/CTA/CTA-SI/PCT mismatch but without (identifiable) vessel occlusion (e.g., distal MCA branch);

- those without proof of vessel occlusion and without mismatch (e.g. lacunar stroke); and

- those with CT/CTA/CTA-SI/PCT mismatch, vessel occlusion, and CTA-SI lesion volume between 33% and 50% of the MCA territory.

References

1. Schellinger PD, Fiebach JB, Hacke W. Imaging-based decision making in thrombolytic therapy for ischemic stroke: present status. Stroke 2003;34(2):575-583.

2. Von Kummer R, Bozzao L, Manelfe C. Early CT diagnosis of hemispheric brain infarction. Berlin: Springer Verlag, 1995:1-95.

3. Hacke W, Warach S. Diffusion-weighted MRI as an evolving standard of care in acute stroke. Neurology 2000;54(8):1548-1549.

4. Jansen O, Schellinger PD, Fiebach JB, et al. Early recanalization in acute ischemic stroke saves tissue at risk defined by MRI. Lancet 1999;353:2036-2037.

5. Schellinger PD, Fiebach JB, Jansen O, et al. Stroke magnetic resonance imaging within 6 hours after onset of hyperacute cerebral ischemia. Ann Neurol 2001;49(4):460-469.

6. Fiebach JB, Schellinger PD, Gass A, et al. Stroke MRI is accurate in hyperacute intracerebral hemorrhage. A multicenter study on the validity of stroke imaging. Lancet 2003: submitted for publication.

7. Schellinger PD, Jansen O, Fiebach JB, et al. A standardized MRI stroke protocol: comparison with CT in hyperacute intracerebral hemorrhage. Stroke 1999;30:765-768.

8. Kidwell CS, Saver JL, Carneado J, et al. Predictors of hemorrhagic transformation in patients receiving intra-arterial thrombolysis. Stroke 2002;33(3):717-724.

9. Kulkens S, Schwark C, Schellinger PD, et al. Thrombolysis in ischemic stroke 3 to 6 hours after symptom onset using an MR-based algorithm. International Stroke Conference, Phoenix, AZ, Feb. 13-15, 2003.

10. Rother J, Schellinger PD, Gass A, et al. Effect of intravenous thrombolysis on MRI parameters and functional outcome in acute stroke < 6h. Stroke 2002;33(10):2438-2445.

11. Schellinger PD, Jansen O, Fiebach JB, et al. Monitoring intravenous recombinant tissue plasminogen activator thrombolysis for acute ischemic stroke with diffusion and perfusion MRI. Stroke 2000;31(6):1318-1328.

12. Zivin JA, Holloway RG. Weighing the evidence on DWI: caveat emptor. Neurology 2000;54(8):1552.

13. Handschu R, Garling A, Heuschmann PU, et al. Acute stroke management in the local general hospital. Stroke 2001;32:866-870.

14. Gilman S. Imaging the brain. First of two parts. NEJM 1998;338(12):812-820.

15. Gilman S. Imaging the brain. Second of two parts. NEJM 1998;338(13):889-896.

16. The National Institute of Neurological Disorders and Stroke rt-PA Stroke Study Group. Tissue plasminogen activator for acute ischemic stroke. NEJM 1995;333(24):1581-1587.

17. Bozzao L, Bastianello S, Fantozzi LM, et al. Correlation of angiographic and sequential CT findings in patients with evolving cerebral infarction. AJNR 1989;10(6):1215-1222.

18. Von Kummer R, Nolte PN, Schnittger H, et al. Detectability of cerebral hemisphere ischaemic infarcts by CT within 6 h of stroke. Neuroradiology 1996;38(1):31-33.

19. Truwit CL, Barkovich AJ, Gean-Marton A, et al. Loss of the insular ribbon: another early CT sign of acute middle cerebral artery infarction. Radiology 1990;176(3):801-806.

20. Bozzao L, Angeloni U, Bastianello S, et al. Early angiographic and CT findings in patients with hemorrhagic infarction in the distribution of the middle cerebral artery. AJNR 1991;12(6):1115-1121.

21. Fiebach JB, Schellinger PD, Jansen O, et al. CT and diffusion-weighted MR-imaging in randomized order: DWI results in higher accuracy and lower interrater variability in the diagnosis of hyperacute ischemic stroke. Stroke 2002;33(9):2206-2210.

22. Patel SC, Levine SR, Tilley BC, et al. Lack of clinical significance of early ischemic changes on computed tomography in acute stroke. JAMA 2001;286(22):2830-2838.

23. Barber PA, Demchuk AM, Zhang J, Buchan AM. 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. Lancet 2000;355(9216):1670-1674.

24. Pexman JH, Barber PA, Hill MD, et al. Use of the Alberta Stroke Program Early CT Score (ASPECTS) for assessing CT scans in patients with acute stroke. AJNR 2001;22(8):1534-1542.

25. Lev MH, Farkas J, Gemmete JJ, et al. Acute stroke: improved nonenhanced CT detection-benefits of soft-copy interpretation by using variable window width and center level settings. Radiology 1999;213(1):150-155.

26. Knauth M, Kummer RV, Jansen O, et al. Potential of CT angiography in acute ischemic stroke. AJNR 1997;18:1001-1010.

27. Shrier DA, Tanaka H, Numaguchi Y, et al. CT angiography in the evaluation of acute stroke. AJNR 1997;18(6):1011-1020.

28. Lubezky N, Fajer S, Barmeir E, Karmeli R. Duplex scanning and CT angiography in the diagnosis of carotid artery occlusion: a prospective study. Eur J Vasc Endovasc Surg 1998;16:133-136.

29. Doerfler A, Engelhorn T, von Kummer R, et al. Are iodinated contrast agents detrimental in acute cerebral ischemia? An experimental study in rats. Radiology 1998;206(1):211-217.

30. Brandt T, Knauth M, Wildermuth S, et al. CT angiography and Doppler sonography for emergency assessment in acute basilar artery ischemia. Stroke 1999;30(3):606-612.

31. Wildermuth S, Knauth M, Brandt T, et al. Role of CT angiography in patient selection for thrombolytic therapy in acute hemispheric stroke. Stroke 1998;29(5):935-938.

32. Koenig M, Klotz E, Luka B, et al. Perfusion CT of the brain: diagnostic approach for early detection of ischemic stroke. Radiology 1998;209(1):85-93.

33. Lev MH, Segal AZ, Farkas J, et al. Utility of perfusion-weighted CT imaging in acute middle cerebral artery stroke treated with intra-arterial thrombolysis: prediction of final infarct volume and clinical outcome. Stroke 2001;32(9):2021-2028.

34. Schramm P, Schellinger PD, Fiebach JB, et al. Comparison of CT and CT angiography source images with diffusion-weighted imaging in patients with acute stroke within 6 hours after onset. Stroke 2002;33(10):2426-2432.

35. Moonis M, Fisher M. Imaging of acute stroke. Cerebrovasc Dis 2001;11:143-150.

36. Hamberg LM, Hunter GJ, Halpern EF, et al. Quantitative high-resolution measurement of cerebrovascular physiology with slip-ring CT. AJNR 1996;17(4):639-650.

37. Hamberg LM, Hunter GJ, Kierstead D, et al. Measurement of cerebral blood volume with subtraction three-dimensional functional CT. AJNR 1996;17(10):1861-1869.

38. Koenig M, Banach-Planchamp R, Kraus M, et al. CT perfusion imaging in acute ischemic cerebral infarct: comparison of cerebral perfusion maps and conventional CT findings. Rofo Fortschr Geb Rontgenstr Neuen Bildgeb Verfahr 2000;172:219-226.

39. Koenig M, Kraus M, Theek C, et al. Quantitative assessment of the ischemic brain by means of perfusion-related parameters derived from perfusion CT. Stroke 2001;32(2):431-437.

40. Rother J, Jonetz-Mentzel L, Fiala A, et al. Hemodynamic assessment of acute stroke using dynamic single-slice computed tomographic perfusion imaging. Arch Neurol 2000;57(8):1161-1166.

41. Bruning R, Penzkofer H, Schopf U, et al. Calculation of absolute cerebral blood volume and cerebral blood flow by means of electron-beam computed tomography (EBT) in acute ischemia. Radiologe 1998;38(12):1054-1059.

DR. SCHELLINGER is an assistant professor of neurology at the University of Heidelberg in Germany and a visiting fellow at the National Institute of Neurological Disorders and Stroke in Bethesda, MD.