Abstract WMP16: Eloquence-Weighted Imaging Improves Clinical Outcomes Prediction in Endovascular Stroke Therapy

Abstract

Background: Larger infarct size at presentation as determined by ASPECTS is associated with reduced likelihood of good outcome in acute ischemic stroke. However, infarct volume alone explains only a modest fraction (∼30%) of the variation in stroke outcome. Incorporating the relative eloquence of each ASPECTS region may improve the predictive power. Methods: In the combined database of the SWIFT and STAR trials, we identified patients treated with the Solitaire stent retriever. Using the 24hr CT scan, a multivariate linear regression was used to determine the relative contribution of each ASPECTS region, separately in each hemisphere, to freedom from disability (mRS 0-2) at 90 days. The coefficients from the regression were used to create an Eloquence-weighted ASPECTS score (EL-ASPECTS), which was compared against the original in predicting outcome based on the presentation CT scan. Results: Among 254 patients treated with ET, average age was 68, 64% were female, and NIHSS was mean 16 (SD +/- 5). Mean ASPECTS at presentation was 8.2 and 6.4 at 24 hrs. The most commonly involved ASPECTS regions were the lentiform nuclei (70%), insula (55%), and caudate (52%). In multivariate analysis, for the right hemisphere on 24hr CT, preservation of M1 (OR 1.6) and M4 (OR 1.2) regions were most strongly predictive of good outcome. For the left hemisphere on 24 hr CT, preservation of M3 (OR 2.6), and M5 (OR 2.5) and involvement of M2 (OR -1.9) were most predictive. Eloquence weights were assigned to all 20 R/L ASPECTS regions to create EL-ASPECTS. EL-ASPECTS, compared with original ASPECTS, demonstrated improved discrimination for independent functional outcome for right hemisphere (C-statistic 0.78 vs. 0.69), left hemisphere (0.78 vs. 0.72), and all stroke patients (0.76 vs. 0.70). On presentation CTs, multivariate analyses including age and presentation NIHSS demonstrated EL-ASPECTS but not original ASPECTS was predictive of good clinical outcome (OR 1.65, p<.01). Higher C-statistic values were seen with EL-ASPECTS in analysis of presentation CT scans. Conclusions: Incorporation of regional weighting into ASPECTS improves the ability to predict who will achieve independent functional outcomes with endovascular therapy in acute ischemic stroke.