Non-Contrast CT versus CT perfusion for predicting post-thrombectomy outcomes in anterior circulation large vessel occlusion stroke
摘要
While CT perfusion (CTP) is the current standard for endovascular thrombectomy (EVT) selection in extended time windows with acute ischemic stroke with large vessel occlusion (LVO-AIS), the added prognostic value of CTP over non-contrast CT (NCCT)-based models remains debated, particularly when combined with clinical variables.
MethodsIn this dual-center retrospective cohort study, we analyzed 216 consecutive anterior circulation LVO-AIS patients undergoing EVT between May 2021 and March 2024. Univariate and multivariate analysis were performed on both NCCT-derived and CTP-derived imaging parameters alongside clinical variables, with 90-day functional independence (modified Rankin Scale [mRS] 0–2) as primary endpoint. The ROC curves of the two models were computed using 5-fold cross-validation, and differences in the area under the curve (AUC) were evaluated using DeLong’s test.
ResultsFunctional independence (mRS 0–2) was achieved in 42.6% (92/216) of the patients. The NCCT-clinical model incorporated ASPECTS, baseline NIHSS and diabetes status, and the CTP-clinical model incorporated ischemic core volume, baseline NIHSS and diabetes status. The NCCT-clinical model demonstrated AUC of 0.819 and the CTP-clinical model 0.834. The difference of AUC showed no statistical significance between NCCT-clinical model and the CTP-clinical model (ΔAUC = 0.015, P = 0.964). The NCCT-clinic model demonstrated comparable performance to CTP-clinic model, including accuracy (74.0% vs. 75.8%), sensitivity (80.5% vs. 82.1%) and specificity (65.7% vs. 67.7%).
ConclusionThe NCCT-clinical model achieved comparable ability to predict prognosis compared to the CTP-clinical model in extended time windows with LVO-AIS.