Abstract TP238: Comparison of Baseline ASPECTS, Ischemic Lesion Net Water Uptake, and CBF<30% for the Prediction of Failed Recanalization During Mechanical Thrombectomy

Introduction: Failed recanalization (Frecan; mTICI 0–2a) after endovascular thrombectomy (EVT) occurs in ≈20% of patients with acute ischaemic stroke due to large-vessel occlusion (AIS-LVO). We compared three admission imaging biomarkers-net water uptake (NWU), Alberta Stroke Program Early CT Score (ASPECTS) and CT-perfusion–derived relative ischemic core volume (relative cerebral blood flow < 30%, rCBF < 30%)-for their ability to predict Frecan. Methods: Consecutive anterior-circulation AIS-LVO cases treated within 24 h were retrospectively collected from two comprehensive stroke centers (January 2019–December 2024). NWU and ASPECTS were measured on non-contrast CT; rCBF < 30% volumes were obtained from CT perfusion. Univariable and multivariable logistic regression identified independent predictors of Frecan. Discriminative performance was evaluated with receiver-operatingcharacteristic (ROC) analysis; Youden’s index defined optimal cut-offs. Results: Among 688 patients (median age 75 years; 51.7% women), Frecan occurred in 133 (19.3%). Median NWU was higher in the Frecan group (8.08% vs 5.48%; p<0.001) as was median rCBF < 30% volume (13 mL vs 9 mL; p < 0.001); ASPECTS was lower (8 vs 9; p=0.023). On multivariable analysis, only NWU remained independently associated with Frecan (adjusted OR 1.16 per 1% increase; 95%CI 1.09–1.24; p<0.001). ROC curves showed superior predictive accuracy for NWU (AUC 0.659) compared with rCBF < 30% (AUC 0.542) and ASPECTS (AUC 0.557). An NWU threshold of 5.70% yielded 53.3% sensitivity and 74.4% specificity. Conclusions: NWU outperforms ASPECTS and rCBF < 30% for predicting failed recanalization after EVT. Quantitative oedema assessment may improve pre-procedural risk stratification in AISLVO.