Abstract Number: Esoc2026a1712 Bayesian Ordinal Markov Transition Modeling of 2-Year Functional Outcomes After Collateral-Based Selection for Late-Window Endovascular Treatment in Ischaemic Stroke Patients

Abstract Background and aims Endovascular treatment (EVT) trials commonly report modified Rankin Scale (mRS) outcomes at discrete follow-up timepoints, providing limited insight into functional trajectories. We estimated longitudinal mRS trajectories up to 2 years using data from MR CLEAN-LATE. Methods MR CLEAN-LATE (NL58246.078.17; ISRCTN19922220) randomized 502 patients to EVT or standard care. The mRS was assessed at 90 days, 1 year and 2 years after randomization. Crude data were used to describe mRS distributions and survival stratified by 90-day mRS. Missing outcomes were handled using multiple imputation. A Bayesian Markov ordinal regression model was fitted to estimate transition probabilities between mRS states from 90 days to 1 year and from 1 to 2 years. Results The mRS was fully observed at 90 days and was available for 340 (68%) and 428 (85%) patients at 1 year and 2 years, respectively. Vital status was available for 499 (99%) patients at 2 years. Survival rates were highly associated with the 90-day mRS (Figure 1A). Figure 1B shows mRS distributions at 2-years by 90-day mRS score. Model-based transition probabilities per mRS score were highly similar between the EVT and control group and between the time periods 90 days to 1 year and 1 to 2 years. Steady-state probabilities were highest for favorable mRS scores (mRS 0-3: 22-35%). Mortality probabilities were highest for mRS 4 (23-29%) and mRS 5 (43-50%). Conclusions The estimated mRS transition probabilities up to 2 years post-stroke may be used for long-term prognostic modeling and health-economic evaluations. Conflict of interest W.H. van Zwam received consultation fees from Stryker, Nicolab and Cerenovus (paid to the institution), and is a DSMB member of the We-Trust study, Solonda study, and In Extremis study (all funding paid to the institution). A.A. Postma received an institutional grant from Siemens Healthineers and Bayer Healthcare. Figure 1 - belongs to Conclusions