Abstract 464: Non‐contrast CT ASPECTS versus CT Perfusion for Patients with Large Infarct Core Undergoing EVT

Introduction The Alberta Stroke Program Early CT Score (ASPECTS) from non‐contrast CT (NCCT) and CT perfusion (CTP) estimates have been used in recent trials to triage patients with large vessel occlusion (LVO) for endovascular therapy (EVT). Recent EVT trials demonstrated that patients with large infarct core (ASPECTS 0 to 5), can still benefit from EVT. The reliability and upper limit of CTP core volume estimates in this population remains unclear. We aimed to assess how ASPECTS and CTP core correlate with one another in a real‐world cohort of patients undergoing large‐core EVT, and which modality performs better for 90‐day outcome prediction. Methods From a multicentre, prospectively collected registry, consecutive patients were identified with pre‐treatment large infarct core, defined by NCCT ASPECTS, who underwent EVT. Included patients underwent both NCCT and CTP prior to treatment and had 90‐day modified Rankin Scale (mRS) recorded. Baseline demographic and clinical characteristics were compared for those with good versus poor primary outcome (mRS 0‐3 versus 4‐6). ASPECTS‐CTP correlation was evaluated with Spearman's coefficient. Three multivariable logistic regression models (ASPECTS‐based, CTP‐based, combined ASPECTS+CTP) evaluated predictors of poor outcome, with discrimination assessed by area under the receiver operating characteristic curve (AUROC) and DeLong test, and validated by 300 bootstrap resamples. Results Among 64 patients who underwent large‐core EVT, the median age was 64 years, 56.2% were female, and median NIHSS was 18 IQR 15‐22. Poor outcome patients (mRS 4‐6) were older and had larger CTP cores, while other baseline factors did not differ (see Table 1). CTP core volume was frequently underestimated but overall showed a significant negative correlation with NCCT ASPECTS (ρ = ‐0.459, p < 0.001). On multivariable regression, age and CTP core volume independently predicted poor 90‐day outcome, whereas ASPECTS, NIHSS, and IV thrombolysis did not (see Table 2). Final model performance was good, with AUROCs above 0.8, and after 300 bootstrap resamples, optimism‐corrected AUROCs remained high (0.787‐0.840), supporting good discriminative ability despite the small cohort. Conclusion In a multicentre registry cohort of patients who underwent EVT with large presenting infarct core, defined by ASPECTS 0‐5, the ASPECT score showed a significant negative correlation with CT perfusion‐derived core infarct volume. Multivariable regression revealed no significant difference in outcome discrimination between ASPECTS, CTP, or their combination when predicting 90‐day mRS adjusted for confounders. Age and CTP core volume consistently emerged as independent predictors of poor outcome. image