ID: 4344760 PREDICTION OF UNDIAGNOSED ATRIAL FIBRILLATION BY ARTIFICIAL INTELLIGENCE USING 12-LEAD SINUS RHYTHM ELECTROCARDIOGRAMS IN PATIENTS WITH EMBOLIC STROKE OF UNDETERMINED SOURCE

In patients with embolic stroke of undetermined source (ESUS), undiagnosed atrial fibrillation (AF) is a possible cause of stroke, and the use of an implantable cardiac monitor (ICM) is recommended for its detection. However, the invasiveness and cost limitations of ICM make its use in all ESUS patients impractical. Therefore, there is a need to develop less invasive and cost-effective screening methods that can complement or replace ICM. Recently, AI techniques to estimate AF risk from sinus rhythm (SR) electrocardiogram (ECG) have attracted attention, but their usefulness in ESUS patients has not been fully tested.