Abstract Number: Esoc2026ys190 Heterogeneous Treatment Effects of Intravenous Thrombolysis in Patients With Acute Ischemic Stroke

Abstract Background and aims Thrombolysis has been a standard treatment for patients with acute ischemic stroke (AIS). The four intravenous thrombolytic agents used in ischemic stroke—alteplase, tenecteplase, reteplase, and prourokinase—have been tested by several randomized controlled trials, while the optimal agent for intravenous thrombolysis in patients with AIS remains unclear. The project aimed to develope an individualized treatment selection strategy based on heterogeneous treatment effects evaluated by a machine learning algorithm. Methods We develop a machine learning model in a derivation cohort splited from individual patients data pooled from four trials (TRACE-2 trial n = 1,430; ORIGINAL trial n = 1,489; RAISE trial n = 1,412; PROST-2 trial n = 1,552) to predict individual treatment effects of four agents from intravenous thrombolysis on modified Rankin Scale (mRS) score of 0-1 at 90 days. The optimal treatment strategy will be recommanded by the model and will be evaluated in simulation on the validation cohort. Results A framework for deriving and validating individualised treatment effects and approaches to drive precision treatment decisions for intravenous thrombolysis in patients with AIS has been outlined. Conflict of interest Aoming Jin: nothing to disclose