ABSTRACT NUMBER: ESOC2026A850 ASPECTS AS A SURROGATE MARKER OF CORE-PERFUSION MISMATCH IN LATE-WINDOW LARGE VESSEL OCCLUSION STROKE

Abstract Background and aims In acute stroke due to large vessel occlusion (AIS-LVO), identifying a core-perfusion mismatch guides reperfusion therapy in extended time windows, but perfusion imaging is not universally available. We assessed whether the ASPECTS score is associated with core-perfusion mismatch and could serve as a surrogate imaging marker. Methods We retrospectively analyzed a large multicenter international cohort of consecutive AIS-LVO (ICA or M1) imaged 4.5-24 hours from last time seen well, with baseline CT- or MR-perfusion imaging. Core-perfusion mismatch was defined as a volume ≥ 15 mL and ratio ≥ 1.8 using automated software. Receiver operating characteristic (ROC) analyses were performed to identify optimal ASPECTS cutoffs for mismatch classification. Results Overall, 531 patients were included (CT, n = 182; MRI, n = 349). Core-perfusion mismatch was present in 86% of the CT cohort and 62% of the MRI cohort. In the MRI cohort, ASPECTS predicted mismatch with an AUC-ROC of 0.843 (95%CI, 0.797-0.890), with an optimal cutoff of ≥6, yielding a sensitivity, specificity, positive and negative predictive value (PPV and NPV) of 0.87, 0.72, 0.84, and 0.77, respectively. In the CT cohort, the AUC-ROC was 0.801 (95%CI, 0.698-0.904), with an optimal cutoff of ≥7, yielding a sensitivity, specificity, PPV and NPV of 0.74, 0.81, 0.96, and 0.34, respectively. Conclusions The high PPV of ASPECTS≥6 on MRI and ≥ 7 on CT supports their potential use as pragmatic inclusion criteria for patient selection in future late-window AIS-LVO trials when perfusion imaging is unavailable, as surrogate markers of salvageable tissue. Conversely, low ASPECTS values –especially on CT– do not exclude core-perfusion mismatch. Conflict of interest Nothing to disclose.