Abstract DP242: New Non-contrast CT Tools to Predict Clinical Outcomes After Endovascular Therapy for Patients with Large Infarct Core

Background: Recent trials have demonstrated that endovascular therapy (EVT) improves outcome for patients with large core infarcts. These studies have demonstrated that current imaging methods like the Alberta Stroke Program Early CT Score (ASPECTS) and CT Perfusion may not accurately identify which patients benefit from emergent reperfusion. Here, we evaluate a novel, low-cost biomarker called Net Water Uptake (NWU), derived from non-contrast CT scans using an automated algorithm, and a novel scoring system from non-contrast CT called the Visual Hypodensity Score (VHS) to identify patient’s who may benefit from large core EVT. Methods: We included anterior circulation LVO AIS patients with large core infarcts (ASPECTS 0 to 5) from a prospectively collected registry who underwent EVT. Blinded raters scored ASPECTS and VHS from non-contrast CT, NWU was computed with an automated algorithm. Interrater reliability was assessed using kappa statistics. Multivariable logistic regression adjusted for age, NIHSS, IV thrombolysis, time from last-known-well to groin puncture, occlusion location, and TICI. The primary outcome was to compare performance of blinded rater ASPECTS, VHS, and automated NWU for predicting 90-day functional outcome (mRS 0-3 vs.4–6) after EVT. Results: Among 114 patients with large core infarct, median age was 65 [IQR 53-75], median NIHSS was 19 [IQR 16-22] (see Table 1). Interrater reliability was fair overall for ASPECTS (κ=0.31) and VHS (κ=0.24). In adjusted models, ASPECTS (OR 0.84, p=0.16) and VHS (OR 1.11, p=0.17) were not significantly associated with outcome, whereas NWU in the M3 region was significant (OR 1.30, p=0.02). Full model results are in Table 2 and interrater reliability in Table 3. Conclusion: In large core infarct EVT patients, ASPECTS and VHS showed only fair interrater reliability and were not significant predictors of 90-day mRS after adjusting for important clinical factors including age, NIHSS, and TICI. Automated NWU was statistically significant only in M3 for predicting poor 90-day mRS. More tailored and reliable imaging biomarkers may be needed to guide EVT selection in this population.