Noncontrast Computed Tomography versus Computed Tomography Angiography Source Images for Predicting Final Infarct Size in Anterior Circulation Acute Ischemic Stroke: a Prospective Cohort Study

ARTICLE IN PRESS

Noncontrast Computed Tomography versus Computed Tomography Angiography Source Images for Predicting Final Infarct Size in Anterior Circulation Acute Ischemic Stroke: a Prospective Cohort Study Amritendu Mukherjee, DM,* Prakash Muthusami, MD, PDCC,† Aneesh Mohimen, MD,* Srinivasan K, MD, PDCC,* Babunath B, DAMIT,* Sylaja PN, DM,‡ and Chandrasekharan Kesavadas, MD*

Background: There has been a recent debate regarding the superiority of computed tomography angiography source images (CTASIs) over noncontrast computed tomography (NCCT) to predict the final infarct size in acute ischemic stroke (AIS). We hypothesized that the parenchymal abnormality on CTASI in faster scanners would overestimate ischemic core. Methods: This prospective study assessed the correlation of Alberta Stroke Program Early CT Score (ASPECTS) on NCCT, CTASI, and computed tomography perfusion (CTP) with final infarct size in patients within 8 hours of AIS. Follow-up with NCCT or diffusion-weighted magnetic resonance imaging (MRI) was performed at 24 hours. Correlations of NCCT and CTASI with final infarct size and with CTP parameters were assessed. Subgroup analysis was performed in patients who underwent intravenous thrombolysis or mechanical thrombectomy. Inter-rater reliability was tested using Spearman’s rank correlation. A P value less than .05 was considered statistically significant. Results: A total of 105 patients were included in the final analysis. NCCT had a stronger correlation with the final infarct size than did CTASI (Spearman’s ρ = .85 versus .78, P = .13). We found an overestimation of the final infarct size by CTASI in 47.6% of the cases, whereas NCCT underestimated infarct size in 60% of the patients. NCCT correlated most strongly with CBV (ρ = .93), whereas CTASI correlated most strongly with CBF (ρ = .87). Subgroup analysis showed less correlation of CTASI with final infarct size in the group that received thrombolysis versus the group that did not (ρ = .70 versus .88, P = .01). Conclusion: In a 256-slice scanner, the CTASI parenchymal abnormality includes ischemic penumbra and thus overestimates final infarct size—this could result in inappropriate exclusion of patients from thrombolysis or thrombectomy. Key Words: 256-Slice computed tomography—ASPECTS—noncontrast CT—CT angiography source image—final infarct size. © 2016 National Stroke Association. Published by Elsevier Inc. All rights reserved.

From the *Department of Imaging Sciences and Interventional Radiology, Sree Chitra Tirunal Institute for Medical Sciences and Technology, Trivandrum, India; †Department of Diagnostic Imaging, The Hospital for Sick Children, Toronto, Ontario, Canada; and ‡Comprehensive Stroke Care Program, Department of Neurology, Sree Chitra Tirunal Institute for Medical Sciences and Technology, Trivandrum, India. Received June 16, 2016; revision received August 6, 2016; accepted September 15, 2016. Address correspondence to Amritendu Mukherjee, DM, Department of Imaging Sciences and Interventional Radiology, Sree Chitra Tirunal Institute for Medical Sciences and Technology, Trivandrum, India. E-mail: [email protected]. 1052-3057/$ - see front matter © 2016 National Stroke Association. Published by Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jstrokecerebrovasdis.2016.09.026

Journal of Stroke and Cerebrovascular Diseases, Vol. ■■, No. ■■ (■■), 2016: pp ■■–■■

1

ARTICLE IN PRESS A. MUKHERJEE ET AL.

2

Introduction Imaging plays a key role in the management of acute ischemic stroke (AIS). Computed tomography (CT) is presently the imaging modality of choice for the evaluation of patients with AIS.1-4 The Alberta Stroke Program Early CT Score (ASPECTS) is a simple and reproducible grading system developed to assess early ischemic changes in noncontrast computed tomography (NCCT),1,5,6 and is currently used worldwide in the decision algorithm for thrombolysis and mechanical thrombectomy in anterior circulation AIS.7 Several studies in the last decade have shown that hypoattenuation on computed tomography angiography source images (CTASIs) correlates better with the final infarct size than does the abnormality on NCCT.8-14 The implication thereof was that CTASI reflects the ischemic core more faithfully than does NCCT. The thin slices acquired and whole-brain coverage in computed tomography angiography (CTA) are obvious advantages over NCCT and computed tomography perfusion (CTP). In fact, the American Heart Association recommendations include CTASI as a correlate of cerebral blood volume (CBV) and consider CTASI to be as sensitive as diffusion-weighted imaging (DWI), outside the brain stem and the posterior fossa, for AIS.1 In the light of this large body of evidence, a recent report15 that CTASIs were in fact cerebral blood flow (CBF) weighted rather than CBV weighted raised several questions as to the exact nature of this phenomenon.16-18 Theoretically, a CTA protocol optimized for newer scanners would result in the generation of CTASI before a contrast steady state has been established in brain parenchyma, thereby significantly overestimating infarct size.15,19 For stroke neurologists and neuroradiologists involved in the care of and decision making for revascularization in patients with AIS, it is imperative to understand what the implications of the aforementioned factors are, with faster scanners becoming the norm in radiology departments worldwide. We undertook a prospective study to determine the ability of CTASI and NCCT in a 256-slice CT to estimate the final infarct size, and to assess the correlation of NCCT and CTASI with CTP.

Materials and Methods Study Protocol and Patients This prospective study was approved by the Institution Review Board. All consecutive patients presenting within 8 hours’ onset of AIS between April 2014 and May 2015 were included in the study. Written informed consent was obtained from the patient or the accompanying person, when patient consent was not possible or unreliable. Clinical management of patients and treatment algorithms were not affected by participation in the study. A total of 208 patients were enrolled. A National Institute of Health Stroke Scale (NIHSS) score cutoff of 3 was used

to exclude minor strokes (n = 10). Patients were excluded if there was evidence of intracerebral hemorrhage on the NCCT (n = 38), if there were contraindications for administering a CT contrast agent (n = 3), if consent was not provided or obtained (n = 13), or if any of the CT images (NCCT, CTA, or CTP) were of poor quality due to patient motion or technical factors (n = 21). In addition, patients with vertebrobasilar strokes (n = 11) were excluded, as were patients who did not undergo followup (FU) imaging at 24-48 hours due to various reasons (n = 7). Baseline demographic data, including age; sex; side of involvement; risk factors including diabetes, hypertension, smoking and coronary artery disease; NIHSS score; ASPECTS; time from symptom onset to CT; tissue plasminogen activator treatment, either intravenous (IV) or intra-arterial; site of vessel occlusion; and mechanical thrombectomy, were recorded prospectively in our stroke database.

Imaging Our institutional acute stroke CT protocol on a 256slice CT scanner (Brilliance iCT; Philips Healthcare, Best, The Netherlands) included NCCT and CTA, with CTP additionally performed on a per-case consensus clinical decision between the stroke neurologist and the neuroradiologist. Given the clinical relevance to determine the correlate of a CTA study performed for maximal arteriographic information, we did not alter acquisition parameters that would otherwise achieve a contrast steady state at the time of CTA. Plain CT of the head was performed with the following parameters: 120 kV, 350 mAs, with a 5-mm section thickness parallel to the inferior orbitomeatal line, coverage from skull base to vertex, and an approximate radiation dose of 2.2 mSv. The CTA parameters were 120 kV, 400 mAs, .9-mm slice thickness with 50% overlap, pitch less than 1, gantry rotation time of .5 second, and coverage from the aortic arch to the skull vertex, and an approximate radiation dose of 5.8 mSv. Nonionic contrast (.7 mL/kg, iohexol, Omnipaque [GE Healthcare Co. Ltd., Shanghai, China]; 300 mg iodine/mL) was injected through an 18G cannula at 5 mL/s, followed by 40 mL saline chaser, also at 5 mL/s. Bolus tracking was used, with the tracker in the proximal descending thoracic aorta, a trigger threshold of 150 HU and a post-threshold delay of 5 seconds. The CTP study parameters were 80 kVp, 100120 mAs, and 1.5 seconds per rotation. Fifty milliliters of iodinated contrast agent was injected 5 seconds before the start of the first phase at a rate of 5 mL/s, followed by a 20 mL saline chaser at the same rate. A total of 40 cycles were acquired, for a total of 60 seconds, with 6 cycles of delayed imaging, with a cycle time of 30 seconds each, resulting in a total CTP time of 240 seconds and an approximate radiation dose of 2.6 mSv. A single operator (B.B., CT technologist with 8 years’ experience) used

ARTICLE IN PRESS CTASI OVERESTIMATES INFARCT CORE

Philips IntelliSpace Portal v4.0 (Philips Healthcare) to calculate CTP parametric maps (CBF, CBV, mean transit time [MTT], and time to peak [TTP]). Arterial input and venous output functions were obtained from the ipsilateral anterior cerebral artery and the midportion of the superior sagittal sinus, respectively. FU with NCCT or diffusion-weighted magnetic resonance imaging was performed at 24 hours to assess the final infarct size. NCCT of the head was performed with the same parameters as above. DWI was performed in a 1.5-T scanner (Somatom; Siemens, Erlangen, Germany) with a multishot echo-planar spin-echo sequence with 4800-millisecond repetition time, 126-millisecond echo time, 23-cm field of view, 178 × 178 matrix, 5-mm section thickness, 1.5-mm gap, b values of 0 and 1000, and 20 directions.

Image Analysis All images were interpreted separately by 2 neuroradiologists (A.M. and A.M.) who were blinded to clinical information other than the side of involvement. Images from the NCCT, CTASI, CTP, and FU imaging of each patient were interpreted at an interval of at least 7 days in between reading sessions, with randomization of images for interpretation performed offline by a CT technologist (B.B.) who was part of the study group. Before the start of the study, both raters underwent a joint training session on 15 previous cases of AIS to harmonize image interpretation. All images were interpreted with the same window settings for both NCCT and CTASI (window level, 40 HU; window width, 40 HU). The same types of monitors and lighting conditions were used for the interpretation of all the images. Observers interpreted not more than 20 sequences per day to maintain a good level of concentration. ASPECTS was documented for NCCT, CTASI, and CTP parametric maps (CBV, CBF, MTT, and TTP), as well as for the FU NCCT/DWI. Initial NCCT was assessed for the presence of a hyperdense vessel. CTA images were analyzed for the presence, site, and extent of an intraluminal thrombus. Our stroke protocol at the start of the study did not include delayed CTA imaging for collateral assessment. The FU images were also assessed for hemorrhagic transformation and graded (HI1/HI2/PH1/ PH2) for the same.20,21

Statistical Analysis Demographic, clinical, and imaging results were expressed as mean ± 2 standard deviations for continuous variables and as percentages for categorical variables. The two raters’ scores for each modality (NCCT, CTASI, and CTP parametric maps) were correlated using Spearman’s rank correlation. A mean ASPECTS score of the 2 readers was used for statistical analysis. Spearman’s rank correlation with Fisher’s z-transformation was performed between NCCT ASPECTS and FU ASPECTS, and between CTASI ASPECTS and FU ASPECTS. In the patients

3

who underwent perfusion, CTP parameters were correlated against NCCT ASPECTS and CTASI ASPECTS using Spearman’s rank correlation. For subgroup analysis, we divided our cohort into 2 groups: group 1, which received thrombolysis, and group 2, which did not receive thrombolysis. Spearman’s rank correlation analyses of NCCT ASPECTS with FU ASPECTS and CTASI ASPECTS with FU ASPECTS were done separately in these 2 groups. Correlation analyses were done separately for these subgroups according to time to CT scan after symptom onset. The SPSS statistical package (SPSS v 22.0; SPSS, Chicago, IL) was used for all statistical analyses. A P value less than .05 was considered to indicate a significant difference.

Results A total of 105 patients were included in our final analysis. All patients received NCCT and CTA at the time of admission, and 25 patients had CTP in addition. There were 53 left hemispheric strokes and 52 right-sided strokes. Most patients (44.2%) had a cardioembolic source for the stroke, whereas in 27.3% of the patients, the etiology was undetermined. Demographic and clinical data are shown in Table 1. Median NIHSS score at presentation was 11 (range 3-26) (Fig 1). Median modified Rankin Scale (mRS) scores were 4 at presentation, 3 at discharge, and 2 at 90 days. The mRS score at 90 days was available in 99 patients (6 patients were lost to FU and there were 7 deaths [median mRS score = 6], 5 during hospital admission and 2 during FU). The time from symptom onset to CT was less than 180 minutes in more than 60% of our patients. Imaging data are represented in Table 2. Sixty-five (62%) of 105 patients received some form of treatment for recanalization (50 patients [77%] underwent IV thrombolysis, 2 patients [3%] received intra-arterial thrombolysis, 8 patients [12%] received bridging IV thrombolysis followed by mechanical thrombectomy, and 5 patients [8%] received mechanical thrombectomy alone). Of the 50 patients who underwent IV thrombolysis, 5 patients received IV thrombolysis within 1.5 hours of symptom onset, 22 patients between 1.5 and 3.0 hours, and 23 patients between 3.0 and 4.5 hours. Mechanical thrombectomy was performed with stent retrievers in all cases (Solitaire FR [eV3, Plymouth, MN] in 10 cases, Trevo Retriever [Stryker Neurovascular {Fremont, CA 94538, United States}] followed by Solitaire FR in 1 case and Revive stentriever [Codman Neurovascular {Codman & Shurtleff, Inc. 325 Paramount Drive Raynham, MA 02767}] in 2 cases). Of these 13 cases, thrombolysis in cerebral infarction (TICI) 3 recanalization was obtained in 5 cases, TICI 2b in 4 cases, TICI 2a in 1 case, TICI 1 in 1 case, and recanalization could not be achieved in 2 cases. FU imaging was NCCT in 90 patients, whereas DWI was performed in 15 cases. Spearman’s correlation for ASPECTS from both raters for NCCT, CTASI, CBV, CBF, MTT, and TTP showed

ARTICLE IN PRESS A. MUKHERJEE ET AL.

4

Table 1. Demographics and clinical details of the study population N = 105

Demographic Age, mean (SD, range) Female, n (%) Right hemispheric stroke, n (%) Stroke etiology (TOAST classification), n (%)

Associated risk factors, n (%)

NIHSS score (mean, median)

58 (14.6, 19-84) years 38 (36.2) 53 (50.5) Large-artery atherosclerosis (embolic/thrombosis) Small-vessel occlusion (lacunar) Cardioembolic (high/medium risk) Other determined Indeterminate Hypertension Diabetes Coronary artery disease Smoking Presentation 24 h Discharge 90 days

22 (20.9) 10 (10) 44 (41.9) 1* (.95) 27 (25.7) 61 (58.1) 49 (46.7) 31 (29.5) 24 (22.8) 12.5, 11 9.1, 9 6.8, 6 4.3, 2

Abbreviations: CTA, computed tomography angiography; NIHSS, National Institutes of Health Stroke Scale; SD, standard deviation; TOAST, Trial of Org 10172 in Acute Stroke Treatment. *Diagnosed as internal carotid dissection by CTA and carotid Doppler studies.

excellent correlation for all studies (Spearman’s ρ = .90, .94, .99, .99, and .99 for NCCT, CTASI, CBV, CBF, and MTT, respectively). Spearman’s rank correlation analysis showed that NCCT ASPECTS had a stronger correlation with FU ASPECTS (ρ = .85, P < .001) than did CTASI ASPECTS (ρ = .78, P < .001), although Fisher’s z-transformation showed nonsignificance of this difference of correlations (P = .13). The reason for this difference of correlation with the final

Figure 1.

infarct size, we found, was overestimation of infarct size by CTASI in 49 (46.7%) of 105 patients (by a mean of 3.8 ± 3.45 ASPECTS points in these 49 patients (Table 3), whereas NCCT underestimated the final infarct size in 63 (60%) of 105 patients (by a mean of 3.67 ± 2.86 ASPECTS points in these 63 patients). This trend is evident on Bland– Altman analysis (Fig 2), wherein both NCCT and CTASI showed a strong agreement with FU ASPECTS; however, with a tendency for NCCT to underestimate infarct size

NIHSS scores at presentation and follow-up in our cohort. Abbreviation: National Institutes of Health Stroke Scale.

ARTICLE IN PRESS CTASI OVERESTIMATES INFARCT CORE

5

Table 2. Imaging characteristics of the study population Imaging parameter

N = 105

Time from stroke onset to CT scan (min) Dense MCA sign, n (%) MCA dot sign, n (%) Location on CTA, n (%)

162.8 ± 78.3 32 (30.5) 19 (18.1) Not seen Proximal ICA Terminal ICA M1—MCA Proximal M2—MCA

36 (34.2) 23 (21.9) 22 (20.9) 47 (44.7) 21 (20)

the final infarct size in groups a and b are shown in Table 5. A total of 23 patients had hemorrhagic transformation of the infarct on FU imaging. Of these 23 patients, 6 had HI1 (hemorrhagic infarct) hemorrhagic transformation, 11 had HI2, 1 had PH1 (parenchymal hemorrhage), and 5 had PH2. Of these 23 patients, 17 had received thrombolysis. Two patients had symptomatic HT (≥4 points increase in the NIHSS score). The rate of symptomatic HT after revascularization in our cohort was 3.3% (2 in 60 patients).

Discussion

Abbreviations: CT, computed tomography; CTA, computed tomography angiography; ICA, internal carotid artery; MCA, middle cerebral artery.

(mean NCCT and FU ASPECTS difference = +2.12), and for CTASI to overestimate infarct size (mean CTASI and FU ASPECTS difference = −1.12). In the 25 patients who underwent CTP, we found that maximal correlations of CTP parameters were as follows (Table 4): CBV correlated most strongly with NCCT (ρ = .93, P < .001), whereas CBF correlated most strongly with CTASI (ρ = .88, P < .001). Fisher’s z-transformation showed nonsignificance for the difference between these correlation values, although for CBV there was a tendency toward significance (P = .06). Subgroup analysis between NCCT and CTASI with FU imaging for patients who did (group 1) and did not (group 2) receive thrombolysis (5 patients who received only mechanical thrombectomy were excluded from this analysis) showed that there was less correlation of CTASI with the final infarct size in group 1 (n = 57) than in group 2 (n = 43), with ρ = .70 and .88, P = .01. Among patients where overestimation was seen by CTASI (n = 49), the mean overestimation in group 1 (n = 26) was higher than that in group 2 (n = 23) (overestimation by 2.8 ± 2.0 versus 1.8 ± 1.2 ASPECTS points, P < .05). Of the 25 patients who underwent CTP, 7 received thrombolysis (group a), 17 did not receive thrombolysis (group b), and 1 received only mechanical thrombectomy and was excluded from this subanalysis. The correlates of CTP parameters with

We have shown that in a 256-section CT scanner, both NCCT and CTASI correlate strongly with final infarct size in anterior circulation AIS. There is a tendency for NCCT to underestimate the final infarct size, whereas CTASI tends to overestimate the final infarct size. In addition, we have shown that on these state-of-the-art scanners, CTASI correlates strongest with CBF, whereas NCCT correlates strongest with CBV, in agreement with the findings of an earlier study in a 64-section CT scanner.15 With the current global trend in radiology departments of adopting faster scanners (64-, 128-, 256-, and 320-slice scanners), our findings assume critical clinical significance, inasmuch as not accounting for this underestimation and overestimation of irreversibly infarcted brain by NCCT and CTASI, respectively, would result in a number of otherwise eligible candidates being inappropriately included or excluded from revascularization. This would have major implications on daily clinical practice as well as in acute stroke trials designed to offer revascularization on the basis of NCCT or CTASI as irreversible infarction.11 For a few years now, it has been held that CTASI serves as a better surrogate for regions of brain parenchyma with severely reduced CBV than does NCCT.8,10-13,22 Several stroke centers are questioning or ignoring CTP data completely, preferring to avoid the extra time, radiation, and postprocessing involved. The thin sections acquired by CTA, complete brain coverage, and the relatively constant appearance of hypoattenuation irrespective of water content make CTASI a popular tool for the estimation of early ischemic change. Studies that reported the CTASI–final

Table 3. Measure of overestimation/underestimation of final infarct size by NCCT and CTASI NCCT versus FU ASPECTS

Overestimated infarct size No difference Underestimated infarct size

N = 8 (1.26 ± .49 ASPECTS points) N = 34 N = 63 (3.67 ± 2.86 ASPECTS points)

CTASI versus FU ASPECTS

Overestimated infarct size No difference Underestimated infarct size

N = 49 (3.80 ± 3.45 ASPECTS points) N = 29 N = 27 (2.61 ± 2.06 ASPECTS points)

Abbreviations: ASPECTS, Alberta Stroke Program Early CT Score; CTASI, computed tomography angiography source image; FU, followup; NCCT, noncontrast computed tomography.

ARTICLE IN PRESS 6

A. MUKHERJEE ET AL.

Figure 2. Bland–Altman analysis for agreement of (A) NCCT ASPECTS and (B) CTASI ASPECTS with FU imaging. Note the positive bias of NCCT (mean of differences = 2.12) indicating a tendency to underestimate infarct size, whereas CTASI shows a negative bias (mean of differences = −1.12), indicating a tendency to overestimate infarct size. Both modalities, however, show a strong agreement with FU ASPECTS within the limits of 95% confidence intervals. Abbreviations: ASPECTS, Alberta Stroke Program Early CT Score; CTASI, computed tomography angiography source image; FU, follow-up; NCCT, noncontrast computed tomography; SD, standard deviation.

infarct correlation were, however, carried out in older CT scanners whose acquisition speed for an angiographic study of the head and neck would allow for the achievement of a steady state of contrast in the voxels of interest. Our finding in a 256-section scanner that there was in fact less correlation between CTASI and the final infarct size, due to a higher degree of overestimation, in patients who underwent thrombolysis (group 1) than in those who did not (group 2) supports our hypothesis that the parenchymal abnormality on CTASI in new scanners includes the ischemic penumbra. It has been aptly pointed out that

the CTA protocol adopted for imaging plays a pivotal role in the extent of hypoattenuation on CTASI, in that a slower injection protocol would permit CTASI to be volume weighted.17,23 However, in a real-world setting, the role of a CTA in AIS remains the delineation of arterial anatomy—optimization to provide images uncontaminated with extra-arterial information requires fast table movements, high pitch, and faster injection rates, factors that make CTASI flow dependent. There is limited literature on the CTP correlates of ASPECTS on NCCT and CTASI. Two recent studies, one

ARTICLE IN PRESS CTASI OVERESTIMATES INFARCT CORE

7

Table 4. Spearman’s correlation of CTP parameters with NCCT and CTASI CTP parameter (N = 25) CBF CBV MTT TTP

NCCT

CTASI

P value

ρ = .80, P < .001 ρ = .93, P < .001 ρ = .80, P < .001 ρ = .80, P < .001

ρ = .88, P < .001 ρ = .80, P < .001 ρ = .85, P < .001 ρ = .85, P < .001

.36 .06 .60 .60

Abbreviations: CBF, cerebral blood flow; CBV, cerebral blood volume; CTASI, computed tomography angiography source image; CTP, computed tomography perfusion; MTT, mean transit time; NCCT, noncontrast computed tomography; TTP, time to peak. P values are determined by Fisher’s z-transformation.

Table 5. Spearman’s correlation of CTP parameters with final infarct size

CTP parameter

Thrombolysis (group a, n = 7)

No thrombolysis (group b, n = 17)

CBF CBV MTT TTP

ρ = .37, P = NS ρ = .56, P = NS ρ = .37, P = NS ρ = .37, P = NS

ρ = .78, P < .001 ρ = .94, P < .001 ρ = .76, P < .001 ρ = .76, P < .001

Abbreviations: CBF, cerebral blood flow; CBV, cerebral blood volume; CTP, computed tomography perfusion; MTT, mean transit time; TTP, time to peak.

with 28 patients8 and another with 56 patients,3 both on 16-slice CT scanners, found that CTASI correlates better with unsalvageable tissue than does NCCT. Another recent retrospective analysis of 64 AIS patients imaged on a 64-section scanner15 found a strong correlation of CTASI with CBF (r = .89, P < .0001), whereas NCCT correlated with CBV (r = .79, P < .0001). In our study too, we found similar CTASI–CBF and stronger NCCT–CBV correlations, although the differences did not reach significance due to the small number (25) in this subgroup. With the evolution of faster scanning technologies, we believe that hypoattenuated areas on CTASI include the ischemic penumbra, and also perhaps benign oligemia. It is evidently important that stroke centers around the world should not interpret images produced by new-generation CT scanners according to data in the literature based on older machines. The prospective nature of our study resulted in homogeneity of imaging, data collection, and interpretation methods, although some decisions, such as performing a CTP, were still done on a per-case basis, so as to minimize radiation and maintain ethical clinical standards, resulting in a smaller subgroup with CTP. In addition, it can be argued that parenchymal volumes would provide more precise comparisons of extent of involvement than

ASPECTS. Our reason to use ASPECTS was, however, deliberate—first, we wished to assess the images as they would be in a real-world setting, and we did not believe that the increased tediousness of quantifying infarct volumes would provide added clinical benefit. Second, although it would have been helpful to compare our data with FU angiography for vessel recanalization, this was not performed due to reasons of practicality. Last, we did not include collateral scores, which have been shown by recent trials to be important in determining the outcome of revascularization and prognosis.24-26 Our results question the popular notion that CTASI correlates with final infarct size more robustly than NCCT in the setting of AIS. We recommend the exercise of caution in replacing CTASI for either final infarct size or ischemic penumbra in a modern multislice CT scanner, although we acknowledge that it correlates stronger with the latter. In this regard, it will be interesting to study the correlates of NCCT and CTASI with CTP parametric maps in modern CT scanners in larger homogeneous patient groups.

Conclusion Contrary to current guidelines,1,2 which state that CTASIs faithfully represent irreversibly infarcted brain tissue in AIS, our study shows that this relationship cannot be considered relevant for faster CT scanners. The overestimation of the final infarct size by CTASI on state-of-the-art CT scanners and the correlation of CTASI with CBF suggest representation of the ischemic penumbra with or without benign oligemia. Given the growing popularity of faster scanners, it is important for the stroke community to bear this fact in mind while interpreting CT images for AIS. Acknowledgments: We would like to thank Dr. Sankara Sarma (Professor, Achutha Menon Center for Health Science Studies, SCTIMST, Trivandrum) and Mr. Jaykumar P. (Junior Laboratory Assistant, Department of Biochemistry, Trivandrum Medical College) for their assistance with the statistical analysis.

References 1. Jauch EC, Saver JL, Adams HP, et al. on behalf of the American Heart Association Stroke Council. Guidelines for the early management of patients with acute ischemic stroke: a guideline for healthcare professionals from the American Heart Association/American Stroke Association. Stroke 2013;44:870-947. 2. Powers WJ, Derdeyn CP, Biller J, et al. 2015 American Heart Association/American Stroke Association focused update of the 2013 guidelines for the early management of patients with acute ischemic stroke regarding endovascular treatment: a guideline for healthcare professionals from the American Heart Association/ American Stroke Association. Stroke 2015;46:3020-3035.

ARTICLE IN PRESS 8 3. Parsons MW, Pepper EM, Chan V, et al. Perfusion computed tomography: prediction of final infarct extent and stroke outcome. Ann Neurol 2005;58:672-679. 4. Wang X-C, Gao P-Y, Xue J, et al. Identification of infarct core and penumbra in acute stroke using CT perfusion source images. AJNR Am J Neuroradiol 2010;31:34-39. 5. Barber PA, Demchuk AM, Zhang J, et al. 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;355:16701674. 6. 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 Am J Neuroradiol 2001;22:1534-1542. 7. Higashida RT, Furlan AJ. Trial design and reporting standards for intra-arterial cerebral thrombolysis for acute ischemic stroke. Stroke 2003;34:e109-e137. 8. Lin K, Rapalino O, Law M, et al. Accuracy of the Alberta Stroke Program Early CT Score during the first 3 hours of middle cerebral artery stroke: comparison of noncontrast CT, CT angiography source images, and CT perfusion. AJNR Am J Neuroradiol 2008;29:931-936. 9. Camargo ECS, Furie KL, Singhal AB, et al. Acute brain infarct: detection and delineation with CT angiographic source images versus nonenhanced CT scans. Radiology 2007;244:541-548. 10. Coutts SB, Lev MH, Eliasziw M, et al. ASPECTS on CTA source images versus unenhanced CT: added value in predicting final infarct extent and clinical outcome. Stroke 2004;35:2472-2476. 11. Schramm P, Schellinger PD, Fiebach JB, et al. Comparison of CT and CT angiography source images with diffusionweighted imaging in patients with acute stroke within 6 hours after onset. Stroke 2002;33:2426-2432. 12. Ezzeddine MA, Lev MH, McDonald CT, et al. CT angiography with whole brain perfused blood volume imaging: added clinical value in the assessment of acute stroke. Stroke 2002;33:959-966. 13. Schramm P, Schellinger PD, Klotz E, et al. Comparison of perfusion computed tomography and computed tomography angiography source images with perfusionweighted imaging and diffusion-weighted imaging in patients with acute stroke of less than 6 hours’ duration. Stroke 2004;35:1652-1658.

A. MUKHERJEE ET AL. 14. Lev MH, Segal AZ, Farkas J, et al. Utility of perfusionweighted CT imaging in acute middle cerebral artery stroke treated with intra-arterial thrombolysis: prediction of final infarct volume and clinical outcome. Stroke 2001;32:2021-2028. 15. Sharma M, Fox AJ, Symons S, et al. CT angiographic source images: flow- or volume-weighted? AJNR Am J Neuroradiol 2011;32:359-364. 16. Watanabe M, Qureshi AI. Are CT angiography source images accurate for evaluating infarct volume? J Neuroimaging 2013;23:163-164. 17. Pulli B, Yoo AJ. CT angiography source images with modern multisection CT scanners: delay time from contrast injection to imaging determines correlation with infarct core. AJNR Am J Neuroradiol 2012;33:E61. 18. Kloska SP. CT angiographic source images with modern multisection CT scanners: appropriate injection protocol is crucial. AJNR Am J Neuroradiol 2011;32:E93. 19. Konstas AA, Goldmakher GV, Lee T-Y, et al. Theoretic basis and technical implementations of CT perfusion in acute ischemic stroke, part 2: technical implementations. AJNR Am J Neuroradiol 2009;30:885-892. 20. Trouillas P, von Kummer R. Classification and pathogenesis of cerebral hemorrhages after thrombolysis in ischemic stroke. Stroke 2006;37:556-561. 21. Zhang J, Yang Y, Sun H, et al. Hemorrhagic transformation after cerebral infarction: current concepts and challenges. Ann Transl Med 2014;2:81. 22. Bhatia R, Bal SS, Shobha N, et al. for the Calgary CTA Group. CT angiographic source images predict outcome and final infarct volume better than noncontrast CT in proximal vascular occlusions. Stroke 2011;42:1575-1580. 23. Pulli B, Schaefer PW, Hakimelahi R, et al. Acute ischemic stroke: infarct core estimation on CT angiography source images depends on CT angiography protocol. Radiology 2012;262:593-604. 24. Liebeskind DS. Collateral lessons from recent acute ischemic stroke trials. Neurol Res 2014;36:397-402. 25. Rusanen H, Saarinen JT, Sillanpää N. Collateral circulation predicts the size of the infarct core and the proportion of salvageable penumbra in hyperacute ischemic stroke patients treated with intravenous thrombolysis. Cerebrovasc Dis 2015;40:182-190. 26. Bang OY, Saver JL, Kim SJ, et al. Collateral flow predicts response to endovascular therapy for acute ischemic stroke. Stroke 2011;42:693-699.