12-lead electrocardiogram-based artificial intelligence model accurately predicts near-term atrial fibrillation in patients with embolic stroke of undetermined source

Embolic stroke of undetermined source (ESUS) composes about 17% of ischemic strokes, and identifying undiagnosed atrial fibrillation (AF) is crucial for determining treatment strategies.1 Although guidelines recommended implantable loop recorders (ILR),2 cost and invasiveness often limit their use. This study investigates the efficacy of an artificial intelligence (AI) model predicting AF from sinus rhythm (SR) electrocardiograms (ECG) in ESUS patients and its impact on ILR implantation decisions.