Po-678-08 neural network model for automatic discrimination of atrial fibrillation episodes detected by an insertable cardiac monitor

Insertable Cardiac Monitors (ICMs) are commonly used to detect atrial fibrillation (AF). The “key episodes” feature introduced in Abbott’s Jot Dx™ ICM reduces electrogram (EGM) review burden with no delay to diagnosis. AF episodes with the fastest mean rate, longest duration, and 2nd longest duration are selected as Key episodes. An artificial intelligence (AI)/machine learning (ML) episode discriminator may further reduce false positive AF episodes, thereby increasing the percentage of clinically actionable true AF episodes.