Research Value of utilizing both ASPECTS and CT angiography collateral score for outcome prediction in acute ischemic stroke Dongbeom Song1 Kijeong Lee1 Eun Hye Kim1 Young Dae Kim1 Jinkwon Kim1 Tae-Jin Song1, Hye Sun Lee2, Hyo Suk Nam1, and Ji Hoe Heo1* Background Alberta Stroke Program Early CT Score (ASPECTS) represents the extent of irreversibly damaged tissue; while CT angiography collateral score (CTA-CS) denotes the degree of collaterals. Aims We investigated whether there is cumulative value in using both ASPECTS and CTA-CS for outcome prediction and attempted to determine the specific subgroup of patients who could benefit from successful reperfusion using these scores. Methods This is a retrospective observational study of stroke patients treated with intra-arterial reperfusion therapy for unilateral arterial occlusion in the anterior circulation. A favorable outcome was defined as modified Rankin Scale ≤ 2 at three-months. Receiver operating characteristic comparison analysis was performed to decide whether outcome predictability increases when ASPECTS and CTA-CS are used together. Classification and regression tree (CART) analysis was done to identify the variables that best predict outcome and define the specific subgroup of patients who could benefit from successful reperfusion. Results A total of 91 consecutive patients were included. Outcome predictability of ASPECTS with CTA-CS was better than that of ASPECTS (P = 0·088) or that of CTA-CS (P = 0·049). CART analysis revealed that ASPECTS > 5 was the primary determinant of favorable outcome, followed by CTA-CS > 1. Among 19 patients with ASPECTS ≤ 5, none had a favorable outcome. Successful reperfusion was associated significantly with favorable outcome in the 51 patients with ASPECTS > 5 and CTA-CS > 1, but not in the 21 patients with ASPECTS > 5 and CTA-CS ≤ 1. Conclusions Outcome predictability improves when using ASPECTS and CTA-CS together. Key words: acute stroke therapy, ASPECTS, collateral, CT angiography, thrombolysis, thrombectomy
better outcome compared to the old generation mechanical device (3,4), and the recently published report of the Multicenter Randomized Clinical Trial of Endovascular Treatment for Acute Ischemic Stroke in the Netherlands (MR CLEAN) has demonstrated improved clinical outcome of IART compared to the best medical care (5). Although reperfusion is the primary target of IART and a powerful predictor of clinical outcome, successful reperfusion does not always translate into improved outcome (6). Therefore, selection of patients who could be beneficial with reperfusion may be important to improve IART efficacy (7). While many imaging modalities have been tested as possible candidate selection tools, CT with CT angiography (CTA) has the advantages of faster scan time, straightforward interpretation, and wide availability (8). With CT, the Alberta Stroke Program Early CT Score (ASPECTS) quantifies the extent of irreversibly damaged tissue, in addition to the exclusion of haemorrhage (9). With CTA, collateral score (CS) can be measured, in addition to providing anatomical information of the arterial system and the location of the occlusion site. As ASPECTS and CTA-CS were both known to predict clinical outcome (10–14), they might be used together to define patients who could be beneficial with reperfusion. However, it was uncertain whether there would be cumulative value when utilizing the combination of these scores. Therefore, we investigated whether outcome predictability increases when using ASPECTS and CTA-CS together. In addition, we attempted to determine the specific subgroup of patients who could benefit from successful reperfusion utilizing these two scores together.
Introduction
Methods
As reperfusion rates have risen with recent advances in mechanical devices, intra-arterial reperfusion therapy (IART) is increasingly implemented (1,2). Stent retrievers have shown Correspondence: Ji Hoe Heo*, Department of Neurology Yonsei University College of Medicine 50-1 Yonsei-ro Seodaemoon-ku 120-752 Seoul Republic of Korea. E-mail: [email protected] 1 Department of Neurology, Yonsei University College of Medicine, Seoul, Korea 2 Department of Biostatistics, Yonsei University College of Medicine, Seoul, Korea Received: 26 September 2014; Accepted: 27 January 2015 Conflict of interest: None declared. Funding: This work was supported by a grant from the Korea Healthcare Technology Research and Development Project, Ministry for Health and Welfare, Republic of Korea (HI10C2020, HI08C2149). DOI: 10.1111/ijs.12505 © 2015 World Stroke Organization
Study subjects This is a retrospective observational study. Candidates for this study were consecutive patients with ischemic stroke of the unilateral anterior circulation and received IART between January 2009 and May 2012 in the Stroke Center at Severance Hospital of Yonsei University Health System, a 2000-bed tertiary hospital in Seoul, Korea. Patients who had CT with CTA performed before IART were included in the study. Patients with unavailable mRS at three-months were excluded. The Institutional Review Board of Yonsei University Health System approved this study and granted a waiver of consent because of the retrospective nature of the study. Reperfusion therapy and image study protocol The reperfusion therapy and imaging protocol of our center has been published previously (15). Briefly, patients presenting within 3 h of symptom onset were considered for IV tissue plasminogen activator [tPA (Actilyse, Boehringer-Ingelheim, Ingelheim, Germany)] after noncontrast CT (LightSpeed Plus, GE Healthcare, Milwaukee, WI, USA or SOMATOM Sensation 64, Siemens Vol ••, •• 2015, ••–••
1
Research Healthcare, Erlangen, Germany). If patients showed an unsatisfactory clinical response [defined as a less than 50% improvement as measured by the National Institutes of Health Stroke Scale (NIHSS)] at the end of IV t-PA infusion, IART was considered after follow-up CT with CTA. Those patients who presented between 3 and 6 h of symptom onset were also considered for IART after CT with CTA. Imaging analysis CT with CTA, which were taken just before IART, was used for imaging analysis. ASPECTS was measured with noncontrast CT using recently suggested methodology which does not account for isolated cortical swelling as an early ischemic change (16). CTA-CS was measured with reconstructed maximum intensity projection CTA images as follows: 0 = absent collaterals in >50% of middle cerebral artery second branch (M2) territory; 1 = diminished collaterals in >50% of M2 territory; 2 = diminished collaterals in 5 and CTA-CS>1, successful reperfusion was the important determinant of favorable outcome.
1-Specificity Fig. 2 Receiver operating characteristic curve of Alberta Stroke Program Early CT Score (ASPECTS), CT angiography collateral score (CTA-CS) and ASPECTS with CTA-CS to predict favorable outcome.
(62%) of 72 patients with ASPECTS>5 had favorable outcome. For the patients with ASPECTS>5, the most important determinant of favorable outcome was CTA-CS>1. Among 21 patients with ASPECTS>5 and CTA-CS≤1, 7 (33%) patients had favorable outcome, but among the 51 patients with ASPECTS>5 and CTACS>1, 38 (75%) patients showed favorable outcome. For the © 2015 World Stroke Organization
Effect of successful reperfusion on favorable outcome according to the baseline ASPECTS and CTA-CS Following the result of CART analysis, we categorized the study subjects into three subgroups according to their baseline ASPECTS and CTA-CS (patients with ASPECTS≤5, those with ASPECTS>5 and CTA-CS≤1, and those with ASPECTS>5 and CTA-CS>1) to investigate whether successful reperfusion was associated with favorable outcome in each subgroup. For 19 patients with ASPECTS≤5, we could not perform further analysis as no one had a favorable outcome in this subgroup. For 21 patients with ASPECTS>5 and CTA-CS≤1, successful reperfusion was not associated with favorable outcome, both in the univariate and multivariate analyses, adjusting for hypertension, atrial fibrillation, initial NIHSS, initial occlusion site and initial IART modality. For 51 patients with ASPECTS>5 and CTACS>1, successful reperfusion was significantly associated with favorable outcome both in the univariate and multivariate analyses (Table 1). Inter-rater agreement Inter-rater agreements for total ASPECTS (linear weighted κ, 0·654; 95% CI, 0·587–0·721), ‘ASPECTS>5’ (κ, 0·762; 95% CI, Vol ••, •• 2015, ••–••
3
Research
D. Song et al.
49% ASPECTS≤5
n = 91
ASPECTS>5
0%
62% CTA-CS≤1
n = 19
n = 72
CTA-CS>1
33% 75% Initial ICA occlusion
n = 21
Initial MCA occlusion
No successful reperfusion
n = 51
Successful reperfusion
17% 56% n = 12
n=9
50% n = 16
86% ASPECTS≤7 n = 35
60% n = 10
ASPECTS>7
96% n = 25
Fig. 3 Classification and regression tree analysis (CART) model demonstrating the best variables that predict clinical outcome. The white area of each pie chart represents the percentage of patients in that subgroup who had favorable clinical outcome. Table 1 Association between successful reperfusion and favorable outcome according to baseline ASPECTS and CTA-CS Group
OR (95% CI)
P
*Adjusted OR (95% CI)
P
ASPECTS5 and CTA-CS≤1 ASPECTS>5 and CTA-CS>1
NA 1·389 (0·194–9·967) 6·000 (1·536–23·438)
1·000 0·013
NA 0·508 (0·026–10·016) 8·212 (1·163–57·994)
0·656 0·035
*Adjusted for hypertension, atrial fibrillation, initial National Institutes of Health Stroke Scale, initial occlusion site and initial intra-arterial reperfusion therapy modality. ASPECTS, Alberta Stroke Program Early CT Score; CTA-CS, CT angiography collateral score; OR, odds ratio; CI, confidence interval; NA, not applicable.
0·607–0·916), ‘ASPECTS>7’(κ, 0·607; 95% CI, 0·447–0·767), total CTA-CS (linear weighted κ, 0·674; 95% CI, 0·576–0·772) and ‘CTA-CS>1’(κ, 0·697; 95% CI, 0·544–0·849) were all good. Interrater agreement for ‘ASPECTS>5 and CTA-CS>1’ (κ, 0·711; 95% CI, 0·565–0·856) was also good with 87·5% observed agreement.
Discussion This study showed that outcome predictability was improved by using ASPECTS and CTA-CS together compared to using one scoring system only. ASPECTS and CTA-CS represent different pathophysiological perspectives following the occlusion of the cerebral artery, in that ASPECTS provides extent of irreversibly injured brain area and CTA-CS provides degree of collateral blood flow. This construct accounts for the only moderate correlation between ASPECTS and CTA-CS and the better outcome predictability using both. We attempted to determine the specific subgroup of patients who might benefit from successful reperfusion using CART analysis. CART analysis is useful in identifying the variables that best predict outcome, and in classifying the study subjects into the mutually exclusive subgroups sharing common characteristics (21). CART model of our study revealed that ASPECTS>5 was the
4
Vol ••, •• 2015, ••–••
primary determinant of favorable outcome, followed by CTACS>1. Therefore, we could categorize study subjects into subgroups by these cut-points of ASPECTS and CTA-CS. The finding that none of the 19 patients with extensive infarct core (ASPECTS≤5) had favorable outcome, regardless of reperfusion achievement, is in line with the previous studies, which reported that large infarct core, assessed with diffusion-weighted magnetic resonance image, was associated with unfavorable outcome in spite of successful reperfusion (23,24). Although the cut-point differed by 1 point, the study analyzing the data from the Penumbra Pivotal Stroke Trial also demonstrated that none of the 28 patients with ASPECT≤4 had favorable outcome (11). However, the recently published article analyzing the Interventional Management of Stroke (IMS) III data demonstrated that 19 (20%) of 92 patients with ASPECTS≤4 had favorable outcome (12). It might be explained by the different sample size and the different baseline characteristics of the low ASPECTS subgroup in each study. Although optimal cut-point might be inconclusive considering the small sample size of our study, exclusion of patients with large infarct core assessed with very low ASPECTS might be reasonable approach to decrease futile reperfusion. For the 51 patients who do not have extensive infarct core (ASPECTS>5) and have good collateral (CTA-CS>1), successful © 2015 World Stroke Organization
Research
D. Song et al. reperfusion was the most important determinant for favorable outcome in the CART analysis. We could confirm that successful reperfusion was associated with favorable outcome for this subgroup of the patients in the univariate and multivariate analyses. These patients may have sufficient potentially reversible brain area because they do not have extensive infarct core, and good collaterals could maintain the viability of those areas until successful reperfusion is accomplished. Therefore, assertive effort should be made to achieve reperfusion in this subgroup of patients. For the 21 patients with potentially reversible brain area (ASPECTS>5) but poor collaterals (CTA-CS≤1), not successful reperfusion but initial occlusion site was the most important determinant of favorable outcome. Successful reperfusion was not significantly associated with favorable outcome also in the univariate and multivariate analysis for these patients. Considering the significant time interval from imaging scan to reperfusion, poor collateral may be insufficient to prevent conversion of ischemic penumbra to infarct core before final reperfusion is achieved. However, this result of subgroup analysis should be interpreted with caution considering this subgroup has only 21 patients. Moreover, from the pathophysiologic perspective that collaterals decide the rate of penumbra loss (25), we infer that successful reperfusion could be beneficial if it is achieved within a short period of time. In this context, immediacy of treatment may be particularly crucial in this subgroup. Although our CART analysis model did not classify a more specified subgroup of patients who would benefit from successful reperfusion— especially with regard to onset to reperfusion time—analysis of bigger data from multiple hospitals, may reveal whether rapid reperfusion within a certain time limit is beneficial even in the patients with poor collaterals. Inter-rater agreement is one of the major hindering factors when utilizing imaging scores (26). Inter-rater agreements of ASPECTS in our study were similar to those in the previous studies, which reported κ for ‘ASPECTS>7’ as 0·53 – 0·89 (9,26). Of note, inter-rater agreement of ‘ASPECTS>5’ was better than that of ‘ASPECTS>7’, the conventional cut-point of ASPECTS in our study. It might be attributed to the fact that subtle early ischemic change on noncontrast CT represented by high ASPECTS is relatively difficult to detect compared to more extensive and obvious early ischemic change represented by lower ASPECTS. Even though two imaging score systems are used, ‘ASPECTS>5 and CTA-CS>1’ has shown good inter-rater agreement with κ of 0·711 and the 87·5% observed rate of agreement. Our study had several limitations. First, our study was retrospective analysis of IART registry. Although all the consecutive patients from prospective IART registry who met the inclusion criteria of this study were enrolled and the decision to carry IART was determined by the written protocol, our study was not free from the patient and treatment selection biases. Second, the sample size was only modest and difference between the outcome prediction value of CTA-CS with ASPECTS and that of ASPECTS alone was marginally significant. In addition, the sample sizes of three subgroups defined by ASPECTS and CTA-CA cut-point were relatively small to perform subgroup analyses and that there © 2015 World Stroke Organization
were risk of committing type I or II error. Therefore, the results of our study, particularly regarding the cut-point of ASPECTS and CTA-CS to define subgroup of patients who could be beneficial with reperfusion should be interpreted with caution. Third, this was a retrospective analysis at a single center. The findings of our study could only be interpreted as a hypothesis generation and warrant validation with independent larger sample size data from multiple centers. Furthermore, our findings which were based on patients with IART are not applicable to patients who receive intravenous thrombolysis. Nevertheless, our findings showed that simultaneous consideration of two different scoring systems (ASPECTS and CTA-CS), representing different perspectives in the pathophysiology of acute cerebral infarct, improves the prediction of outcomes after IART.
References 1 Ellis JA, Youngerman BE, Higashida RT, Altschul D, Meyers PM. Endovascular treatment strategies for acute ischemic stroke. Int J Stroke 2011; 6:511–22. 2 Bivard A, Lin L, Parsonsb MW. Review of stroke thrombolytics. J Stroke 2013; 15:90–8. 3 Nogueira RG, Lutsep HL, Gupta R et al. Trevo versus Merci retrievers for thrombectomy revascularisation of large vessel occlusions in acute ischaemic stroke (TREVO 2): a randomised trial. Lancet 2012; 380:1231–40. 4 Saver JL, Jahan R, Levy EI et al. Solitaire flow restoration device versus the Merci Retriever in patients with acute ischaemic stroke (SWIFT): a randomised, parallel-group, non-inferiority trial. Lancet 2012; 380:1241–9. 5 Berkhemer OA, Fransen PS, Beumer D et al. A randomized trial of intraarterial treatment for acute ischemic stroke. N Engl J Med 2015; 372:11–20. 6 Meyers PM, Schumacher HC, Connolly ES Jr, Heyer EJ, Gray WA, Higashida RT. Current status of endovascular stroke treatment. Circulation 2011; 123:2591–601. 7 Balucani C, Grotta JC. Selecting stroke patients for intra-arterial therapy. Neurology 2012; 78:755–61. 8 Goyal M, Almekhlafi MA, Fan L et al. Evaluation of interval times from onset to reperfusion in patients undergoing endovascular therapy in the interventional management of stroke III trial. Circulation 2014; 130:265–72. 9 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; 355:1670–4. 10 Hill MD, Rowley HA, Adler F et al. Selection of acute ischemic stroke patients for intra-arterial thrombolysis with pro-urokinase by using ASPECTS. Stroke 2003; 34:1925–31. 11 Goyal M, Menon BK, Coutts SB et al. Effect of baseline CT scan appearance and time to recanalization on clinical outcomes in endovascular thrombectomy of acute ischemic strokes. Stroke 2011; 42:93–7. 12 Hill MD, Demchuk AM, Goyal M et al. Alberta Stroke Program early computed tomography score to select patients for endovascular treatment: interventional Management of Stroke (IMS)-III Trial. Stroke 2014; 45:444–9. 13 Maas MB, Lev MH, Ay H et al. Collateral vessels on CT angiography predict outcome in acute ischemic stroke. Stroke 2009; 40:3001–5. 14 Lima FO, Furie KL, Silva GS et al. The pattern of leptomeningeal collaterals on CT angiography is a strong predictor of long-term functional outcome in stroke patients with large vessel intracranial occlusion. Stroke 2010; 41:2316–22. Vol ••, •• 2015, ••–••
5
Research 15 Lee KY, Kim DI, Kim SH et al. Sequential combination of intravenous recombinant tissue plasminogen activator and intra-arterial urokinase in acute ischemic stroke. AJNR Am J Neuroradiol 2004; 25:1470–5. 16 Puetz V, Dzialowski I, Hill MD, Demchuk AM. The Alberta Stroke Program Early CT Score in clinical practice: what have we learned? Int J Stroke 2009; 4:354–64. 17 Souza LC, Yoo AJ, Chaudhry ZA et al. Malignant CTA collateral profile is highly specific for large admission DWI infarct core and poor outcome in acute stroke. AJNR Am J Neuroradiol 2012; 33:1331–6. 18 Jayaraman MV, Grossberg JA, Meisel KM, Shaikhouni A, Silver B. The clinical and radiographic importance of distinguishing partial from near-complete reperfusion following intra-arterial stroke therapy. AJNR Am J Neuroradiol 2013; 34:135–9. 19 DeLong ER, DeLong DM, Clarke-Pearson DL. Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach. Biometrics 1988; 44:837–45. 20 Breiman L, Friedman J, Stone CJ, Olshen RA. Classification and Regression Trees. Belmont, CA, Wadsworth, 1984. 21 Lemon SC, Roy J, Clark MA, Friedmann PD, Rakowski W. Classification and regression tree analysis in public health: methodological review and comparison with logistic regression. Ann Behav Med 2003; 26:172–81. 22 Song D, Kim BM, Kim DJ et al. Comparison of stent retriever and intra-arterial fibrinolysis in patients with acute ischaemic stroke. Eur J Neurol 2014; 21:779–84. 23 Yoo AJ, Verduzco LA, Schaefer PW, Hirsch JA, Rabinov JD, Gonzalez RG. MRI-based selection for intra-arterial stroke therapy: value of
6
Vol ••, •• 2015, ••–••
D. Song et al. pretreatment diffusion-weighted imaging lesion volume in selecting patients with acute stroke who will benefit from early recanalization. Stroke 2009; 40:2046–54. 24 Albers GW, Thijs VN, Wechsler L et al. Magnetic resonance imaging profiles predict clinical response to early reperfusion: the diffusion and perfusion imaging evaluation for understanding stroke evolution (DEFUSE) study. Ann Neurol 2006; 60:508–17. 25 Jung S, Gilgen M, Slotboom J et al. Factors that determine penumbral tissue loss in acute ischaemic stroke. Brain 2013; 136:3554–60. 26 Gupta AC, Schaefer PW, Chaudhry ZA et al. Interobserver reliability of baseline noncontrast CT Alberta Stroke Program Early CT Score for intra-arterial stroke treatment selection. AJNR Am J Neuroradiol 2012; 33:1046–9.
Supporting information Additional Supporting Information may be found in the online version of this article at the publisher’s web-site: Table S1. Baseline characteristics and outcome of patients with CT angiography and those without. Table S2. Baseline characteristics and the factors associated with favorable outcome.
© 2015 World Stroke Organization
better outcome compared to the old generation mechanical device (3,4), and the recently published report of the Multicenter Randomized Clinical Trial of Endovascular Treatment for Acute Ischemic Stroke in the Netherlands (MR CLEAN) has demonstrated improved clinical outcome of IART compared to the best medical care (5). Although reperfusion is the primary target of IART and a powerful predictor of clinical outcome, successful reperfusion does not always translate into improved outcome (6). Therefore, selection of patients who could be beneficial with reperfusion may be important to improve IART efficacy (7). While many imaging modalities have been tested as possible candidate selection tools, CT with CT angiography (CTA) has the advantages of faster scan time, straightforward interpretation, and wide availability (8). With CT, the Alberta Stroke Program Early CT Score (ASPECTS) quantifies the extent of irreversibly damaged tissue, in addition to the exclusion of haemorrhage (9). With CTA, collateral score (CS) can be measured, in addition to providing anatomical information of the arterial system and the location of the occlusion site. As ASPECTS and CTA-CS were both known to predict clinical outcome (10–14), they might be used together to define patients who could be beneficial with reperfusion. However, it was uncertain whether there would be cumulative value when utilizing the combination of these scores. Therefore, we investigated whether outcome predictability increases when using ASPECTS and CTA-CS together. In addition, we attempted to determine the specific subgroup of patients who could benefit from successful reperfusion utilizing these two scores together.
Introduction
Methods
As reperfusion rates have risen with recent advances in mechanical devices, intra-arterial reperfusion therapy (IART) is increasingly implemented (1,2). Stent retrievers have shown Correspondence: Ji Hoe Heo*, Department of Neurology Yonsei University College of Medicine 50-1 Yonsei-ro Seodaemoon-ku 120-752 Seoul Republic of Korea. E-mail: [email protected] 1 Department of Neurology, Yonsei University College of Medicine, Seoul, Korea 2 Department of Biostatistics, Yonsei University College of Medicine, Seoul, Korea Received: 26 September 2014; Accepted: 27 January 2015 Conflict of interest: None declared. Funding: This work was supported by a grant from the Korea Healthcare Technology Research and Development Project, Ministry for Health and Welfare, Republic of Korea (HI10C2020, HI08C2149). DOI: 10.1111/ijs.12505 © 2015 World Stroke Organization
Study subjects This is a retrospective observational study. Candidates for this study were consecutive patients with ischemic stroke of the unilateral anterior circulation and received IART between January 2009 and May 2012 in the Stroke Center at Severance Hospital of Yonsei University Health System, a 2000-bed tertiary hospital in Seoul, Korea. Patients who had CT with CTA performed before IART were included in the study. Patients with unavailable mRS at three-months were excluded. The Institutional Review Board of Yonsei University Health System approved this study and granted a waiver of consent because of the retrospective nature of the study. Reperfusion therapy and image study protocol The reperfusion therapy and imaging protocol of our center has been published previously (15). Briefly, patients presenting within 3 h of symptom onset were considered for IV tissue plasminogen activator [tPA (Actilyse, Boehringer-Ingelheim, Ingelheim, Germany)] after noncontrast CT (LightSpeed Plus, GE Healthcare, Milwaukee, WI, USA or SOMATOM Sensation 64, Siemens Vol ••, •• 2015, ••–••
1
Research Healthcare, Erlangen, Germany). If patients showed an unsatisfactory clinical response [defined as a less than 50% improvement as measured by the National Institutes of Health Stroke Scale (NIHSS)] at the end of IV t-PA infusion, IART was considered after follow-up CT with CTA. Those patients who presented between 3 and 6 h of symptom onset were also considered for IART after CT with CTA. Imaging analysis CT with CTA, which were taken just before IART, was used for imaging analysis. ASPECTS was measured with noncontrast CT using recently suggested methodology which does not account for isolated cortical swelling as an early ischemic change (16). CTA-CS was measured with reconstructed maximum intensity projection CTA images as follows: 0 = absent collaterals in >50% of middle cerebral artery second branch (M2) territory; 1 = diminished collaterals in >50% of M2 territory; 2 = diminished collaterals in 5 and CTA-CS>1, successful reperfusion was the important determinant of favorable outcome.
1-Specificity Fig. 2 Receiver operating characteristic curve of Alberta Stroke Program Early CT Score (ASPECTS), CT angiography collateral score (CTA-CS) and ASPECTS with CTA-CS to predict favorable outcome.
(62%) of 72 patients with ASPECTS>5 had favorable outcome. For the patients with ASPECTS>5, the most important determinant of favorable outcome was CTA-CS>1. Among 21 patients with ASPECTS>5 and CTA-CS≤1, 7 (33%) patients had favorable outcome, but among the 51 patients with ASPECTS>5 and CTACS>1, 38 (75%) patients showed favorable outcome. For the © 2015 World Stroke Organization
Effect of successful reperfusion on favorable outcome according to the baseline ASPECTS and CTA-CS Following the result of CART analysis, we categorized the study subjects into three subgroups according to their baseline ASPECTS and CTA-CS (patients with ASPECTS≤5, those with ASPECTS>5 and CTA-CS≤1, and those with ASPECTS>5 and CTA-CS>1) to investigate whether successful reperfusion was associated with favorable outcome in each subgroup. For 19 patients with ASPECTS≤5, we could not perform further analysis as no one had a favorable outcome in this subgroup. For 21 patients with ASPECTS>5 and CTA-CS≤1, successful reperfusion was not associated with favorable outcome, both in the univariate and multivariate analyses, adjusting for hypertension, atrial fibrillation, initial NIHSS, initial occlusion site and initial IART modality. For 51 patients with ASPECTS>5 and CTACS>1, successful reperfusion was significantly associated with favorable outcome both in the univariate and multivariate analyses (Table 1). Inter-rater agreement Inter-rater agreements for total ASPECTS (linear weighted κ, 0·654; 95% CI, 0·587–0·721), ‘ASPECTS>5’ (κ, 0·762; 95% CI, Vol ••, •• 2015, ••–••
3
Research
D. Song et al.
49% ASPECTS≤5
n = 91
ASPECTS>5
0%
62% CTA-CS≤1
n = 19
n = 72
CTA-CS>1
33% 75% Initial ICA occlusion
n = 21
Initial MCA occlusion
No successful reperfusion
n = 51
Successful reperfusion
17% 56% n = 12
n=9
50% n = 16
86% ASPECTS≤7 n = 35
60% n = 10
ASPECTS>7
96% n = 25
Fig. 3 Classification and regression tree analysis (CART) model demonstrating the best variables that predict clinical outcome. The white area of each pie chart represents the percentage of patients in that subgroup who had favorable clinical outcome. Table 1 Association between successful reperfusion and favorable outcome according to baseline ASPECTS and CTA-CS Group
OR (95% CI)
P
*Adjusted OR (95% CI)
P
ASPECTS5 and CTA-CS≤1 ASPECTS>5 and CTA-CS>1
NA 1·389 (0·194–9·967) 6·000 (1·536–23·438)
1·000 0·013
NA 0·508 (0·026–10·016) 8·212 (1·163–57·994)
0·656 0·035
*Adjusted for hypertension, atrial fibrillation, initial National Institutes of Health Stroke Scale, initial occlusion site and initial intra-arterial reperfusion therapy modality. ASPECTS, Alberta Stroke Program Early CT Score; CTA-CS, CT angiography collateral score; OR, odds ratio; CI, confidence interval; NA, not applicable.
0·607–0·916), ‘ASPECTS>7’(κ, 0·607; 95% CI, 0·447–0·767), total CTA-CS (linear weighted κ, 0·674; 95% CI, 0·576–0·772) and ‘CTA-CS>1’(κ, 0·697; 95% CI, 0·544–0·849) were all good. Interrater agreement for ‘ASPECTS>5 and CTA-CS>1’ (κ, 0·711; 95% CI, 0·565–0·856) was also good with 87·5% observed agreement.
Discussion This study showed that outcome predictability was improved by using ASPECTS and CTA-CS together compared to using one scoring system only. ASPECTS and CTA-CS represent different pathophysiological perspectives following the occlusion of the cerebral artery, in that ASPECTS provides extent of irreversibly injured brain area and CTA-CS provides degree of collateral blood flow. This construct accounts for the only moderate correlation between ASPECTS and CTA-CS and the better outcome predictability using both. We attempted to determine the specific subgroup of patients who might benefit from successful reperfusion using CART analysis. CART analysis is useful in identifying the variables that best predict outcome, and in classifying the study subjects into the mutually exclusive subgroups sharing common characteristics (21). CART model of our study revealed that ASPECTS>5 was the
4
Vol ••, •• 2015, ••–••
primary determinant of favorable outcome, followed by CTACS>1. Therefore, we could categorize study subjects into subgroups by these cut-points of ASPECTS and CTA-CS. The finding that none of the 19 patients with extensive infarct core (ASPECTS≤5) had favorable outcome, regardless of reperfusion achievement, is in line with the previous studies, which reported that large infarct core, assessed with diffusion-weighted magnetic resonance image, was associated with unfavorable outcome in spite of successful reperfusion (23,24). Although the cut-point differed by 1 point, the study analyzing the data from the Penumbra Pivotal Stroke Trial also demonstrated that none of the 28 patients with ASPECT≤4 had favorable outcome (11). However, the recently published article analyzing the Interventional Management of Stroke (IMS) III data demonstrated that 19 (20%) of 92 patients with ASPECTS≤4 had favorable outcome (12). It might be explained by the different sample size and the different baseline characteristics of the low ASPECTS subgroup in each study. Although optimal cut-point might be inconclusive considering the small sample size of our study, exclusion of patients with large infarct core assessed with very low ASPECTS might be reasonable approach to decrease futile reperfusion. For the 51 patients who do not have extensive infarct core (ASPECTS>5) and have good collateral (CTA-CS>1), successful © 2015 World Stroke Organization
Research
D. Song et al. reperfusion was the most important determinant for favorable outcome in the CART analysis. We could confirm that successful reperfusion was associated with favorable outcome for this subgroup of the patients in the univariate and multivariate analyses. These patients may have sufficient potentially reversible brain area because they do not have extensive infarct core, and good collaterals could maintain the viability of those areas until successful reperfusion is accomplished. Therefore, assertive effort should be made to achieve reperfusion in this subgroup of patients. For the 21 patients with potentially reversible brain area (ASPECTS>5) but poor collaterals (CTA-CS≤1), not successful reperfusion but initial occlusion site was the most important determinant of favorable outcome. Successful reperfusion was not significantly associated with favorable outcome also in the univariate and multivariate analysis for these patients. Considering the significant time interval from imaging scan to reperfusion, poor collateral may be insufficient to prevent conversion of ischemic penumbra to infarct core before final reperfusion is achieved. However, this result of subgroup analysis should be interpreted with caution considering this subgroup has only 21 patients. Moreover, from the pathophysiologic perspective that collaterals decide the rate of penumbra loss (25), we infer that successful reperfusion could be beneficial if it is achieved within a short period of time. In this context, immediacy of treatment may be particularly crucial in this subgroup. Although our CART analysis model did not classify a more specified subgroup of patients who would benefit from successful reperfusion— especially with regard to onset to reperfusion time—analysis of bigger data from multiple hospitals, may reveal whether rapid reperfusion within a certain time limit is beneficial even in the patients with poor collaterals. Inter-rater agreement is one of the major hindering factors when utilizing imaging scores (26). Inter-rater agreements of ASPECTS in our study were similar to those in the previous studies, which reported κ for ‘ASPECTS>7’ as 0·53 – 0·89 (9,26). Of note, inter-rater agreement of ‘ASPECTS>5’ was better than that of ‘ASPECTS>7’, the conventional cut-point of ASPECTS in our study. It might be attributed to the fact that subtle early ischemic change on noncontrast CT represented by high ASPECTS is relatively difficult to detect compared to more extensive and obvious early ischemic change represented by lower ASPECTS. Even though two imaging score systems are used, ‘ASPECTS>5 and CTA-CS>1’ has shown good inter-rater agreement with κ of 0·711 and the 87·5% observed rate of agreement. Our study had several limitations. First, our study was retrospective analysis of IART registry. Although all the consecutive patients from prospective IART registry who met the inclusion criteria of this study were enrolled and the decision to carry IART was determined by the written protocol, our study was not free from the patient and treatment selection biases. Second, the sample size was only modest and difference between the outcome prediction value of CTA-CS with ASPECTS and that of ASPECTS alone was marginally significant. In addition, the sample sizes of three subgroups defined by ASPECTS and CTA-CA cut-point were relatively small to perform subgroup analyses and that there © 2015 World Stroke Organization
were risk of committing type I or II error. Therefore, the results of our study, particularly regarding the cut-point of ASPECTS and CTA-CS to define subgroup of patients who could be beneficial with reperfusion should be interpreted with caution. Third, this was a retrospective analysis at a single center. The findings of our study could only be interpreted as a hypothesis generation and warrant validation with independent larger sample size data from multiple centers. Furthermore, our findings which were based on patients with IART are not applicable to patients who receive intravenous thrombolysis. Nevertheless, our findings showed that simultaneous consideration of two different scoring systems (ASPECTS and CTA-CS), representing different perspectives in the pathophysiology of acute cerebral infarct, improves the prediction of outcomes after IART.
References 1 Ellis JA, Youngerman BE, Higashida RT, Altschul D, Meyers PM. Endovascular treatment strategies for acute ischemic stroke. Int J Stroke 2011; 6:511–22. 2 Bivard A, Lin L, Parsonsb MW. Review of stroke thrombolytics. J Stroke 2013; 15:90–8. 3 Nogueira RG, Lutsep HL, Gupta R et al. Trevo versus Merci retrievers for thrombectomy revascularisation of large vessel occlusions in acute ischaemic stroke (TREVO 2): a randomised trial. Lancet 2012; 380:1231–40. 4 Saver JL, Jahan R, Levy EI et al. Solitaire flow restoration device versus the Merci Retriever in patients with acute ischaemic stroke (SWIFT): a randomised, parallel-group, non-inferiority trial. Lancet 2012; 380:1241–9. 5 Berkhemer OA, Fransen PS, Beumer D et al. A randomized trial of intraarterial treatment for acute ischemic stroke. N Engl J Med 2015; 372:11–20. 6 Meyers PM, Schumacher HC, Connolly ES Jr, Heyer EJ, Gray WA, Higashida RT. Current status of endovascular stroke treatment. Circulation 2011; 123:2591–601. 7 Balucani C, Grotta JC. Selecting stroke patients for intra-arterial therapy. Neurology 2012; 78:755–61. 8 Goyal M, Almekhlafi MA, Fan L et al. Evaluation of interval times from onset to reperfusion in patients undergoing endovascular therapy in the interventional management of stroke III trial. Circulation 2014; 130:265–72. 9 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; 355:1670–4. 10 Hill MD, Rowley HA, Adler F et al. Selection of acute ischemic stroke patients for intra-arterial thrombolysis with pro-urokinase by using ASPECTS. Stroke 2003; 34:1925–31. 11 Goyal M, Menon BK, Coutts SB et al. Effect of baseline CT scan appearance and time to recanalization on clinical outcomes in endovascular thrombectomy of acute ischemic strokes. Stroke 2011; 42:93–7. 12 Hill MD, Demchuk AM, Goyal M et al. Alberta Stroke Program early computed tomography score to select patients for endovascular treatment: interventional Management of Stroke (IMS)-III Trial. Stroke 2014; 45:444–9. 13 Maas MB, Lev MH, Ay H et al. Collateral vessels on CT angiography predict outcome in acute ischemic stroke. Stroke 2009; 40:3001–5. 14 Lima FO, Furie KL, Silva GS et al. The pattern of leptomeningeal collaterals on CT angiography is a strong predictor of long-term functional outcome in stroke patients with large vessel intracranial occlusion. Stroke 2010; 41:2316–22. Vol ••, •• 2015, ••–••
5
Research 15 Lee KY, Kim DI, Kim SH et al. Sequential combination of intravenous recombinant tissue plasminogen activator and intra-arterial urokinase in acute ischemic stroke. AJNR Am J Neuroradiol 2004; 25:1470–5. 16 Puetz V, Dzialowski I, Hill MD, Demchuk AM. The Alberta Stroke Program Early CT Score in clinical practice: what have we learned? Int J Stroke 2009; 4:354–64. 17 Souza LC, Yoo AJ, Chaudhry ZA et al. Malignant CTA collateral profile is highly specific for large admission DWI infarct core and poor outcome in acute stroke. AJNR Am J Neuroradiol 2012; 33:1331–6. 18 Jayaraman MV, Grossberg JA, Meisel KM, Shaikhouni A, Silver B. The clinical and radiographic importance of distinguishing partial from near-complete reperfusion following intra-arterial stroke therapy. AJNR Am J Neuroradiol 2013; 34:135–9. 19 DeLong ER, DeLong DM, Clarke-Pearson DL. Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach. Biometrics 1988; 44:837–45. 20 Breiman L, Friedman J, Stone CJ, Olshen RA. Classification and Regression Trees. Belmont, CA, Wadsworth, 1984. 21 Lemon SC, Roy J, Clark MA, Friedmann PD, Rakowski W. Classification and regression tree analysis in public health: methodological review and comparison with logistic regression. Ann Behav Med 2003; 26:172–81. 22 Song D, Kim BM, Kim DJ et al. Comparison of stent retriever and intra-arterial fibrinolysis in patients with acute ischaemic stroke. Eur J Neurol 2014; 21:779–84. 23 Yoo AJ, Verduzco LA, Schaefer PW, Hirsch JA, Rabinov JD, Gonzalez RG. MRI-based selection for intra-arterial stroke therapy: value of
6
Vol ••, •• 2015, ••–••
D. Song et al. pretreatment diffusion-weighted imaging lesion volume in selecting patients with acute stroke who will benefit from early recanalization. Stroke 2009; 40:2046–54. 24 Albers GW, Thijs VN, Wechsler L et al. Magnetic resonance imaging profiles predict clinical response to early reperfusion: the diffusion and perfusion imaging evaluation for understanding stroke evolution (DEFUSE) study. Ann Neurol 2006; 60:508–17. 25 Jung S, Gilgen M, Slotboom J et al. Factors that determine penumbral tissue loss in acute ischaemic stroke. Brain 2013; 136:3554–60. 26 Gupta AC, Schaefer PW, Chaudhry ZA et al. Interobserver reliability of baseline noncontrast CT Alberta Stroke Program Early CT Score for intra-arterial stroke treatment selection. AJNR Am J Neuroradiol 2012; 33:1046–9.
Supporting information Additional Supporting Information may be found in the online version of this article at the publisher’s web-site: Table S1. Baseline characteristics and outcome of patients with CT angiography and those without. Table S2. Baseline characteristics and the factors associated with favorable outcome.
© 2015 World Stroke Organization
