ORIGINAL CONTRIBUTION

(1)

Perfusion Computed Tomography in Transient Ischemic Attack

Shyam Prabhakaran, MD, MS; Sohal K. Patel, MD; Jordan Samuels, MD;

Bethany McClenathan, MD; Yousef Mohammad, MD; Vivien H. Lee, MD

Background:Diffusion- and perfusion-weighted imaging after transient ischemic attack (TIA) has been well studied, while less data exist on perfusion computed tomographic (PCT) imaging.

Objectives:To examine the frequency of PCT abnormalities in patients with anterior circulation TIA and to identify factors associated with the presence of PCT abnormality.

Design:Retrospective study.

Setting:Academic hospital.

Patients:Sixty-five consecutive patients admitted to Rush University Medical Center, Chicago, Illinois, between June 1, 2007, and November 30, 2009, for anterior circulation motor or aphasic TIA in whom PCT was performed.

Main Outcome Measures:Using an automated software algorithm, perfusion abnormality was defined as brain tissue associated with a mean transit time greater than 145%

of that of the contralateral hemisphere and cerebral blood

volume greater than 2.0 mL/100 g. Demographic, risk factor, clinical, radiographic, and in-hospital outcome data were reviewed.

Results:Of 65 patients with anterior circulation TIA who underwent PCT (median age, 62.4 years; 49.2% male), 22 (33.8%) had focal perfusion abnormalities. The presence of motor symptoms (95.5% vs 67.4%, P=.01), multiple (⬎1) episodes (18.2% vs 2.3%, P=.04), ipsilateral arterial stenosis greater than 50% or occlusion (77.3% vs 11.6%, P⬍.001), large-artery atherosclerosis subtype (59.1% vs 9.3%, P⬍.001), and subsequent in-hospital events (22.7%

vs 0%, P=.001) were more frequent in those with perfu- sion abnormality.

Conclusions:On acutely performed PCT, one-third of patients with hemispheric TIA have perfusion abnormalities. Perfusion abnormality may mark patients at greater risk for subsequent early deterioration. This requires further study.

Arch Neurol. 2011;68(1):85-89

I

NSIGHTS FROM DIFFUSION-WEIGHTED

imaging (DWI) studies have shaped our understanding of the pathogenesis and natural history and have led to a new definition of transient ischemic attack (TIA).¹Al- though DWI identifies cytotoxic damage and defines infarction (ie, stroke) with rare exceptions,²penumbra without core in- farct (perfusion abnormality with no areas or only reversible areas of restricted diffusion) should be the signature imaging cor- relate of transient brain ischemia in the acute setting. Thus, perfusion imaging abnormalities might serve as an objective marker of TIA.³

Several magnetic resonance perfusion (MRP) studies^4-6have detected hypoper- fused (ischemic) regions, in addition to smaller areas of restricted diffusion, after

TIA. Perfusion computed tomography (PCT) has been less rigorously assessed in TIA⁷despite being more widely available in the emergency department setting than MRP. We, therefore, assessed the prevalence of perfusion abnormalities as detected by PCT in consecutive patients with TIA admitted to Rush University Medical Center. The objectives of this study were to describe the frequency of PCT abnormality in anterior circulation TIA and to identify factors associated with the presence of PCT abnormality.

METHODS

In June 2007, we instituted a clinical protocol using PCT in patients with suspected acute ischemic stroke or TIA presenting within 24 hours of symptom onset. All patient data were Author Affiliations:

Department of Neurological Sciences, Rush University Medical Center, Chicago, Illinois.

(2)

prospectively entered into a stroke registry database. A TIA was defined as any syndrome of focal neurologic dysfunction as- cribable to a vascular territory and lasting less than 24 hours.

Diagnosis of TIA was definite if an appropriate acute ischemic lesion was seen on brain imaging (DWI or PCT) and probable if alternative causes were excluded and there was agreement by 2 stroke neurologists (S.P. and V.H.L.).

We retrospectively identified all probable and definite TIA cases diagnosed between June 1, 2007, and November 30, 2009.

Study inclusion criteria were (1) PCT within 24 hours of symptom onset; (2) transient language, speech, or motor symptoms lasting less than 24 hours; and (3) anterior circulation localization. The last criterion was determined by a stroke neu- rologist (S.P.) based on clinical symptoms and signs on hospital admission and confirmed, if possible, by neuroimaging results (ie, diffusion or perfusion abnormality). Cases were also excluded if symptoms were sensory only or if the perfusion study was of poor quality. Last, contraindications to iodinated contrast included previous documented contrast allergy, receipt of contrast within the previous 24 hours, a baseline creatinine level greater than 1.5 mg/dL (to convert to micromoles per liter, multiply by 88.4), and pregnancy.

Baseline data included demographics (age, sex, and race/

ethnicity), clinical features (symptoms, number and duration of episodes, initial National Institutes of Health Stroke Scale score, hospital admission blood pressure, and calculated ABCD²[age (A), blood pressure (B), clinical features (C), and symptom duration and diabetes (D)] score⁸), medical history (vascular risk factors, including hypertension, diabetes mellitus, dyslipid- emia, atrial fibrillation, current smoking, and coronary artery disease), radiographic features (computed tomography [CT] and magnetic resonance imaging [MRI] results, including PCT abnormalities, intracranial or extracranial large-artery stenosis greater than 50% or occlusion, and DWI abnormalities), and in- hospital events (recurrent TIA or ischemic stroke). Recurrent TIA required transient focal language or motor symptoms as defined for entry TIA. Ischemic stroke was defined as persistent symptoms lasting longer than 24 hours associated with new in- farcts on repeated DWI. Based on the clinical and diagnostic tests, etiology was classified as large-artery atherosclerosis, small- vessel disease, cardioembolism, or undetermined/other mechanism. The study was approved by the Rush University Medical Center institutional review board.

The PCT studies were conducted using a 64-section scan- ner (Philips Medical Systems, Cleveland, Ohio). A bolus of 40 mL of intravenous contrast material was infused at a rate of 4 to 5 mL/s. Perfusion maps were generated using Philips Bril- liance software (Philips Medical Systems), producing 4 sections separated by 10 mm. Manual postprocessing was performed only to define arterial and venous inputs for the automated PCT variable calculations. Using an automated algorithm, areas of perfusion abnormality were defined by a mean transit time greater than 145% of that of the contralateral hemisphere and cerebral blood volume greater than 2.0 mL/100 g (Figure). We chose this threshold because of previous validation as optimally representing tissue at risk for infarction in acute stroke.³The automated algorithm using the previously mentioned variable thresholds yielded brain tissue maps in 4 sections, outlining ischemic tissue in green.

Using univariable analyses, we aimed to evaluate factors associated with perfusion abnormality. The proportions, means, or medians of relevant variables among those with and without PCT abnormality were compared using the Fisher exact,␹², t, and Mann- Whitney tests, as appropriate. A multivariable logistic regression model was developed to explore potential independent pre- dictors of PCT abnormality. P⬍.05 was considered statistically significant. All statistical analyses were performed using a software program (SPSS, version 14.0; SPSS Inc, Chicago, Illinois).

RESULTS

We identified 122 patients during the study diagnosed as having probable or definite TIA and evaluated within 24 hours of symptom onset. Of these 122 patients, 65 (53.3%) met the inclusion criteria for the study (Table 1).

The most frequent reason for exclusion was posterior circulation TIA (15.6%). Degraded or nondiagnostic PCT accounted for 9.8% of exclusions, with the most com- mon cited reason being improper timing of contrast- enhanced bolus due to congestive heart failure (n = 7).

Median time from symptom onset to PCT was 758 minutes (interquartile range, 403-1325 minutes). No patient experienced a study-related complication during hospitalization.

The mean age of the cohort was 62.4 years; 49.2% were male and 49.2% were white. Symptoms had resolved completely at the time of presentation in 58.5% of patients.

Perfusion abnormality was detected in 22 patients (33.8%).

On univariable analysis, patients with focal hypoperfusion were more likely than those without focal hypoperfusion to present with motor deficits (95.5% vs 67.4%, P = .01) and to report multiple (⬎1) episodes since onset (18.2% vs 2.3%, P = .04). Patients with perfusion abnor- mality were also more likely to have ipsilateral arterial stenosis greater than 50% or occlusion (77.3% vs 11.6%, P⬍.001), large-artery atherosclerosis subtype (59.1% vs 9.3%, P⬍.001), and a higher occurrence of subsequent in-hospital events (22.7% vs 0%, P = .001). Other fac- tors, including demographics, medical history, time to perfusion imaging, and clinical features, such as blood pressure and ABCD²score, were not associated with the presence of PCT abnormality (Table 2). In an explor- atory multivariable logistic regression model, the presence of an ipsilateral occlusive lesion was associated with PCT abnormality (adjusted odds ratio, 25.8; P⬍.001) in- dependent of motor symptoms (P = .07), multiple epi- sodes (P = .60), and large-artery atherosclerosis subtype (P = .88).

All 5 patients who experienced in-hospital events (4 recurrent TIAs and 1 ischemic stroke) had perfusion abnormality, and in 2 patients, in-hospital recurrence occurred in the absence of large-vessel occlusive disease and DWI abnormality. Recurrent events were referable to the areas of perfusion abnormality and were similar to the initial presentation. Bed rest and volume expansion with saline bo- luses were instituted in each patient with recurrent symptoms; induced pharmacologic hypertension was also instituted for 24 hours in the 1 patient with ischemic stroke.

Repeated PCT (a median of 3 days after initial PCT) in these 5 patients showed delayed but complete resolution of the perfusion abnormality.

In patients with TIA who also underwent subsequent MRI ([50], 76.9%; median time to MRI, 33 hours 20 minutes), DWI abnormalities were noted in 72.2% of those with and 25.0% of those without PCT abnormal- ity (P = .002). In this subset, 13 patients (26.0%) were PCT⫹/DWI⫹, 8 (16.0%) were PCT−/DWI⫹, 5 (10.0%) were PCT⫹/DWI−, and 24 (48.0%) were PCT−/DWI−.

The combined yield of a PCT or DWI abnormality was 52.0%.

(3)

COMMENT

In the largest study to describe the frequency of perfusion abnormalities in patients with hemispheric TIA using CT, we observed that 1 in 3 had evidence of focal perfusion abnormality. Furthermore, perfusion abnormality was observed in 28.9% of patients in whom symptoms had resolved completely. These results are consistent with those of previous studies^4-6of perfusion abnormality after TIA using MRP, suggesting that the 2 modalities may perform similarly in the detection of acute anterior circulation ischemia. In the present patients in whom both PCT and DWI were performed, the positive yield of PCT alone was 36.0% compared with 42% for DWI alone, with a combined yield of 52.0%. Given its widespread availability in the emergency setting, PCT may be the preferred imaging modality in patients with TIA.

Patients with demonstrable perfusion abnormality were more likely to have ipsilateral arterial stenosis or occlu-

sion, motor symptoms, multiple presenting TIA episodes, and a large-artery atherosclerosis mechanism. Other studies^4-7,9have observed similar associations. Although a hemodynamic process should be assessed in every patient with TIA, these cumulative findings suggest that perfusion abnormality can be seen even in the absence of large-artery hemodynamic compromise. Distal branch occlusive lesions at the arteriolar or capillary level (micro- circulation) not visible on noninvasive vascular imaging may be operant in these cases.¹⁰

Studies^11-14have shown that the presence of DWI abnormality after classic TIA is a predictor of short- and long-term stroke risk. From these data, the risk of stroke after DWI-positive TIA is as high as 30% in the first 90 days. However, had these been instead classified as mi- nor ischemic strokes by virtue of irreversible tissue damage on neuroimaging, the risk of stroke after DWI-negative TIA (true TIA by the new definition) would have been exceedingly low. Identifying tissue at risk for infarction or the ischemic penumbra after true TIA should better

2 1 2

1 2 1

A B C

D E F

Artery

Vein

Figure. Computed tomographic angiography and perfusion computed tomography (7 hours after onset) and magnetic resonance imaging and magnetic resonance angiography (8 hours after onset) results in a patient with transient aphasia and right hemiparesis (complete resolution by 5.5 hours). A-C, Mean transit time and cerebral blood volume maps (arrows show areas of mean transit time delay with preserved volume) used to generate the outline of the ischemic penumbra (green area) in the superior division of the middle cerebral artery (MCA) territory. D, Computed tomographic angiogram shows no evidence of proximal occlusion in the left MCA (arrow). E, Diffusion-weighted image shows a small left insular area of restricted diffusion (arrow). F, Magnetic resonance angiogram shows no evidence of large-vessel occlusion or stenosis in the left MCA (arrow).

(4)

stratify stroke risk radiographically and may better predict risk of clinical deterioration. All 5 patients with recurrent cerebrovascular events had PCT abnormalities, and in 2 of these 5 patients, no large-vessel or DWI lesion was found. Those with rapid radiographic resolution of the focal ischemia may harbor lower risk of subsequent stroke, whereas those with persistent tissue at risk (irrespective of symptom resolution at the time of imaging or the presence of large-artery occlusive disease) may mark those with limited cerebrovascular re- serve and an elevated risk of clinical deterioration.

Perfusion CT has several advantages over MRP.

First, the techniques are based on helical and spiral CT scanners that are widely available in hospitals, including emergency departments. The median time from symptom onset to PCT was considerably shorter (approximately 12 hours compared with nearly 24 hours) than in some MRP studies. Furthermore, PCT is a rapid test, requiring only approximately 40 seconds for data acquisition; can be performed in sequence with other techniques, such as CT angiography, without an over- dose of contrast material; and has the potential for quantitative and qualitative interpretation. In addition, it can be performed in patients with MR contraindications, such as claustrophobia and metal implants. The major disadvantage of PCT is the limited amount of brain that can be studied.¹⁵Calculation of brain ischemic volume remains limited in cases in which the ischemic lesion extends beyond the covered area. In recent years, multisection scanners of 128 sections and higher have made these limitations less problematic. Other dis- advantages, relative to MRI, are the risks of x-ray radiation and the use of iodinated contrast agent.¹⁵Notably, no contrast-related complications occurred in this study, a result attributable to the strict exclusion criteria (15 patients had prespecified contraindications).¹⁶ However, because serial kidney function testing and long-term follow-up for adverse events were unavailable and given the recent Food and Drug Administration

Table 1. Study Exclusions and Inclusions

Initial Diagnosis of Probable/Definite TIA Within 24 h on Symptom Onset

Patients, No.

(N = 122) Excluded

Posterior circulation localization 19

Technical errors or poor quality 12

Serum creatinine⬎1.5 mg/dL 11

Pure sensory symptoms 9

Contrast within 24 h for other reason, ie, pulmonary embolism

3

Contrast allergy 1

Patient medically unstable for transport 1

Patient refused 1

Subtotal 57

Included

Definite TIA 30

Probable TIA 35

Subtotal 65

Abbreviation: TIA, transient ischemic attack.

SI conversion factor: To convert creatinine to micromoles per liter, multiply by 88.4.

Table 2. Univariable Comparison of Baseline Characteristics in 65 Patients With and Without PCT Abnormality

Characteristic

PCT Abnormality

Present (n = 22)

PCT Abnormality

Absent (n = 43)

P Value Age, mean (SD), y 61.0 (14.8) 63.1 (15.2) .59

Ageⱖ60 y, No. (%) 12 (54.5) 29 (67.4) .31

Male sex, No. (%) 9 (40.9) 23 (53.5) .34

Race/ethnicity, No. (%) .50

White 10 (45.5) 22 (51.2)

Black 10 (45.5) 13 (23.3)

Hispanic 2 (9.1) 6 (14.0)

Other 0 2 (4.7)

Medical history, No. (%)

Current smoking 9 (40.9) 11 (25.6) .21

Hypertension 19 (86.4) 31 (72.1) .20

Diabetes mellitus 3 (13.6) 10 (23.8)^a .52 Coronary artery disease 4 (18.2) 9 (20.9) ⬎.99 Atrial fibrillation 3 (13.6) 5 (11.6) ⬎.99 Hypercholesterolemia 12 (54.5) 30 (69.8) .23

Previous TIA 6 (27.3) 6 (14.0) .19

Previous ischemic stroke 6 (27.3) 8 (18.6) .42 Clinical features

Motor symptoms, No. (%) 21 (95.5) 29 (67.4) .01 Speech or language

symptoms, No. (%)

11 (50.0) 29 (67.4) .17

Sensory symptoms, No. (%) 10 (45.5) 14 (32.6) .31 Duration of last TIA, median

(IQR), min

22.5 (13.3-125.0)

60.0 (15.0-180.0)

.23

TIA duration, No. (%) .24

⬍10 min 5 (22.7) 4 (9.3)

10-59 min 9 (40.9) 16 (37.2)

⬎59 min 8 (36.4) 22 (52.5)

⬎1 episode within the previous 48 h, No. (%)

4 (18.2) 1 (2.3) .04

Symptoms resolved, No.

(%)

11 (50.0) 27 (62.8) .32

Initial NIHSS score, median (IQR)

0.0 (0.0-1.0) 0.5 (0.0-1.3) .54

Initial BP, mean (SD), mm Hg

Systolic 148.8 (32.1) 151.1 (32.4) .79

Diastolic 82.2 (17.9) 81.6 (21.3) .91

⬎140/90, No. (%) 16 (72.7) 25 (58.1) .25 ABCD²score, median (IQR) 4.5 (3.0-5.0) 5.0 (4.0-5.0) .33 ABCD²score of⬎3, No. (%) 16 (72.7) 35 (81.4) .53

TIA etiology, No. (%) ⬍.001

Large-artery atherosclerosis

13 (59.1) 4 (9.3)

Lacunar 0 13 (30.2)

Cardioembolic 2 (9.1) 5 (11.6)

Unknown/other 7 (31.8) 21 (48.8)

Radiographic features Onset-to-PCT time⬍12 h,

No. (%)

9 (40.9) 23 (53.5) .34

Onset-to-PCT time, median (IQR), min

809 (477-1364)

709 (382-1298)

.77

DWI abnormality, No. (%)^a 13 (72.2) 8 (25.0) .002 Ipsilateral large-artery

stenosis⬎50%, No. (%) 17 (77.3) 5 (11.6) ⬍.001 Outcomes, No. (%)

In-hospital deterioration 5 (22.7) 0 .001

Abbreviations: ABCD², age (A), blood pressure (B), clinical features (C), and symptom duration and diabetes (D); BP, blood pressure; DWI, diffusion-weighted imaging; IQR, interquartile range; NIHSS, National Institutes of Health Stroke Scale; onset-to-PCT time, time from symptom onset to percutaneous computed tomography; TIA, transient ischemic attack.

aFifty patients underwent DWI (18 with PCT abnormality and 32 without PCT abnormality).

(5)

warning regarding PCT and excess radiation risks to patients (http://www.fda.gov/medicaldevices/safety /alertsandnotices/ucm185898.htm), the safety of PCT in TIA requires further study.

This study has several limitations besides the small sample size and the absence of long-term follow-up. Be- cause only anterior circulation TIAs were studied, the util- ity of PCT in posterior circulation TIAs is unknown. Given the challenges of posterior fossa imaging with CT, MRP may be better suited for this purpose. Although automated processing of PCT data has been shown to have less variability in parameter measurements and improved interobserver agreement,¹⁷the selected parameter thresholds, around which there is considerable debate, can have a significant effect on the reported prevalence of perfusion abnormality. In addition, because serial perfusion imaging was not performed systematically in each patient, we cannot be certain that hypoperfusion seen with chronic occlusive disease did not lead to an overestimate of the prevalence.⁹Alternatively, given the limited brain coverage (4 sections compared with newer-generation scanners capable of 16 and 32 sections), the present reported detection rates may actually underestimate the true prevalence of perfusion abnormality after acute TIA. Nondiag- nostic results were noted in nearly 10% of all patients, most of which were attributable to inaccurate bolus timing due to congestive heart failure. Finally, the treating physi- cians were not blinded to the PCT results; this may have affected management decisions. These factors should be considered in future studies.

Overall, this study demonstrates that PCT detects perfusion abnormalities in carefully selected patients with anterior circulation motor or aphasic TIA, similar to that observed in MRP studies. Given the association with early deterioration and its widespread availability in the emergency department, PCT also has potential in guiding early management and therapeutic interven- tions. Further study that includes adjustment for other factors in larger cohorts is necessary to confirm these findings.

Accepted for Publication: May 3, 2010.

Correspondence: Shyam Prabhakaran, MD, MS, Depart- ment of Neurological Sciences, Rush University Medi- cal Center, 1725 W Harrison St, Ste 1121, Chicago, IL 60612 ([email protected]).

Author Contributions: All authors had full access to all the data in the study and take responsibility for the in- tegrity of the data and the accuracy of the data analysis.

Study concept and design: Prabhakaran and McClenathan.

Acquisition of data: Prabhakaran, Samuels, and McClenathan. Analysis and interpretation of data: Prab- hakaran, Patel, Samuels, Mohammad, and Lee. Drafting

of the manuscript: Prabhakaran and Patel. Critical revi- sion of the manuscript for important intellectual content:

Prabhakaran, Samuels, McClenathan, Mohammad, and Lee. Statistical analysis: Prabhakaran. Administrative, tech- nical, and material support: McClenathan and Lee. Study supervision: Prabhakaran and Mohammad.

Financial Disclosure: None reported.

REFERENCES

1. Easton JD, Saver JL, Albers GW, et al; American Heart Association; American Stroke Association Stroke Council; Council on Cardiovascular Surgery and Anesthesia; Coun- cil on Cardiovascular Radiology and Intervention; Council on Cardiovascular Nursing;

Interdisciplinary Council on Peripheral Vascular Disease. Definition and evaluation of transient ischemic attack: a scientific statement for healthcare professionals from the American Heart Association/American Stroke Association Stroke Council; Coun- cil on Cardiovascular Surgery and Anesthesia; Council on Cardiovascular Radiol- ogy and Intervention; Council on Cardiovascular Nursing; and the Interdisciplinary Council on Peripheral Vascular Disease. The American Academy of Neurology af- firms the value of this statement as an educational tool for neurologists. Stroke.

2009;40(6):2276-2293.

2. Inatomi Y, Kimura K, Yonehara T, Fujioka S, Uchino M. Hyperacute diffusion- weighted imaging abnormalities in transient ischemic attack patients signify ir- reversible ischemic infarction. Cerebrovasc Dis. 2005;19(6):362-368.

3. Wintermark M, Flanders AE, Velthuis B, et al. Perfusion-CT assessment of in- farct core and penumbra: receiver operating characteristic curve analysis in 130 patients suspected of acute hemispheric stroke. Stroke. 2006;37(4):979-985.

4. Krol AL, Coutts SB, Simon JE, Hill MD, Sohn CH, Demchuk AM; VISION Study Group. Perfusion MRI abnormalities in speech or motor transient ischemic attack patients. Stroke. 2005;36(11):2487-2489.

5. Mlynash M, Olivot JM, Tong DC, et al. Yield of combined perfusion and diffusion MR imaging in hemispheric TIA. Neurology. 2009;72(13):1127-1133.

6. Restrepo L, Jacobs MA, Barker PB, Wityk RJ. Assessment of transient ischemic attack with diffusion- and perfusion-weighted imaging. AJNR Am J Neuroradiol.

2004;25(10):1645-1652.

7. Lu J, Li KC, Hua Y. Primary study on imaging in transient ischemic attacks. Chin Med J (Engl). 2005;118(21):1812-1816.

8. Johnston SC, Rothwell PM, Nguyen-Huynh MN, et al. Validation and refinement of scores to predict very early stroke risk after transient ischaemic attack. Lancet.

2007;369(9558):283-292.

9. Chaves CJ, Staroselskaya I, Linfante I, Llinas R, Caplan LR, Warach S. Patterns of perfusion-weighted imaging in patients with carotid artery occlusive disease.

Arch Neurol. 2003;60(2):237-242.

10. del Zoppo GJ, Mabuchi T. Cerebral microvessel responses to focal ischemia.

J Cereb Blood Flow Metab. 2003;23(8):879-894.

11. Ay H, Koroshetz WJ, Benner T, et al. Transient ischemic attack with infarction: a unique syndrome? Ann Neurol. 2005;57(5):679-686.

12. Coutts SB, Simon JE, Eliasziw M, et al. Triaging transient ischemic attack and mi- nor stroke patients using acute magnetic resonance imaging. Ann Neurol. 2005;

57(6):848-854.

13. Cucchiara BL, Messe SR, Taylor RA, et al. Is the ABCD score useful for risk stratifi- cation of patients with acute transient ischemic attack? Stroke. 2006;37(7):1710- 1714.

14. Purroy F, Montaner J, Rovira A, Delgado P, Quintana M, Alvarez-Sabı´n J. Higher risk of further vascular events among transient ischemic attack patients with diffusion- weighted imaging acute ischemic lesions. Stroke. 2004;35(10):2313-2319.

15. Wintermark M, Sesay M, Barbier E, et al. Comparative overview of brain perfu- sion imaging techniques. Stroke. 2005;36(9):e83-e99 [published online ahead of print August 11, 2005]. doi:10.1161/01.STR.0000177884.72657.8b.

16. Smith WS, Roberts HC, Chuang NA, et al. Safety and feasibility of a CT protocol for acute stroke: combined CT, CT angiography, and CT perfusion imaging in 53 consecutive patients. AJNR Am J Neuroradiol. 2003;24(4):688-690.

17. Soares BP, Dankbaar JW, Bredno J, et al. Automated versus manual postprocessing of perfusion-CT data in patients with acute cerebral ischemia: influ- ence on interobserver variability. Neuroradiology. 2009;51(7):445-451.