Construction and Evaluation of an Intelligent Decision System for Reperfusion Therapy in Acute Ischemic Stroke

This multicenter, cluster-randomized controlled trial will evaluate the effectiveness and safety of the LingBao System, an AI-enabled clinical decision support platform for reperfusion therapy in acute ischemic stroke (AIS). Twenty certified stroke centers will be randomized 1:1 to LingBao-assisted care or standard care. Consecutive patients aged 18 years or older who present within 24 hours of symptom onset or last-known-well and are evaluated for intravenous thrombolysis and/or endovascular therapy will be prospectively enrolled. The study initially plans to enroll approximately 3,000 patients from about 20 certified stroke centers, with approximately 150 patients per center. Because this is a cluster-randomized trial, a pre-specified blinded sample size re-estimation will be performed after approximately 40% of participants have completed the 90-day mRS assessment. The re-estimation will be based only on pooled, blinded information, including cluster size, cluster-size variability, intracluster correlation, follow-up completeness, missingness of the primary outcome, and the overall distribution of the 90-day mRS. No between-group treatment effect will be examined, and any sample size adjustment will only allow an increase in enrollment. At intervention sites, clinicians may use the LingBao System during their routine workflows. The platform integrates routinely available clinical and imaging data, automatically estimates onset-to-treatment windows, screens contraindications, and provides evidence-based, guideline-concordant recommendations for reperfusion therapy; all treatment decisions remain at physician discretion. The primary endpoint is the 90-day modified Rankin Scale (mRS) score analyzed by ordinal shift. Secondary endpoints include workflow metrics (door-to-needle time and door-to-puncture time), reperfusion treatment rates, early neurological improvement, symptomatic intracranial hemorrhage, and mortality. The findings will provide real-world eviden