Emergency Triage To Predict Collaterals In Acute Ischemic Stroke: Imaging Bests Clinical Factors In The Stratis Registry.

Background: Collateral grade is an established determinant of outcomes in acute ischemic stroke (AIS). The triage, workflow and therapeutic options for AIS may be tailored to collaterals and identifying key predictors of collateral status may therefore be crucial. We hypothesized that degree of collateral circulation prior to endovascular therapy in AIS may be predicted at the time of initial patient evaluation and triage. Methods: The STRATIS Registry showed that timelines, technical, and functional outcomes could be effectively attained in a large real-world cohort of endovascular therapy. Baseline clinical and imaging predictors of core lab adjudicated collateral grade (ASITN) by conventional angiography were determined, using multivariate modeling. Results: 586 STRATIS subjects (67.5 ± 15.2 years, 52.7% male) presenting with AIS at 147.4 ± 101.8 min from symptom onset (TFSO) and median NIHSS score 17.0 (range 8.0,30.0) were analyzed. Collateral grade was poor (ASITN 0-1) in 81, moderate (ASITN 2) in 297 and good (ASITN 3-4) in 208. Baseline stroke severity inversely correlated with collaterals (NIHSS per point, OR 0.946, 0.916-0.977, p=0.001), yet no clinical variables such as age, sex or co-morbidities were predictive of collateral status. Less severe ASPECTS at imaging triage (median 9, range 2-10) was associated with better collateral grade (ASITN 0-1, median 7 (2-10); ASITN 2, 8 (3-10); ASITN 3-4, 9 (5-10), p<0.001) and the strongest predictor of collaterals during triage (per point, OR 1.608, 1.399-1.849, p<0.001). Interestingly, the predictive nature of ASPECTS was not modified by TFSO (p=NS). Specific ASPECTS regions (all cortical M1-M6, but no subcortical) affected by early ischemia were also predictive of collateral grade. In particular, insular ASPECTS changes at imaging triage was the strongest predictor of collateral grade (ASITN 0-1, 53/66 (80.3%); ASITN 2, 139/258 (53.9%); ASITN 3-4, 41/194 (21.1%), p<0.001). Conclusions: Imaging, using only ASPECTS, during triage strongly predicts collateral grade, irrespective of time from symptom onset. Clinical variables, however, may not be used to accurately predict collaterals in the real-world practice of endovascular therapy.