Automated Detection of Macro-Reentrant Atrial Tachycardia Circuits Using LAT-Derived Graph Networks

Background: Accurate identification of macro-reentrant atrial tachycardia (AT) circuits is critical for successful ablation but remains challenging with conventional mapping techniques. The aim of this study was to automatically detect macro-reentrant AT loops from high-density local activation time (LAT) maps. Methods: We developed an algorithm for automated detection of macro-reentrant AT circuits using LAT-derived directed graphs. Compared to previous graph-based approaches, the algorithm is designed to identify the fastest-conducting reentrant pathways and cluster them by rotational orientation (clockwise vs. counterclockwise) to distinguish single- from dual-loop circuits. The algorithm was applied retrospectively to 60 macro-reentrant scar-related AT cases mapped with CARTO or Ensite from two institutions. The results were compared with blinded expert electrophysiologist annotations of loop location and single- vs. dual-loop classification. Results: The 60 cases included 16 right atrial and 44 left atrial ATs from 51 patients. Expert review identified 57% single-loop and 43% dual-loop circuits. Compared with expert annotation, the algorithm correctly identified anatomical loop locations with 88% accuracy and correctly distinguished single- vs. dual-loop ATs in 93% of cases. Conclusion: Our LAT graph-based algorithm automatically identified single- and dual-loop macro-reentrant AT circuits. Localizing these pathways may provide insight into circuit mechanisms and help guide ablation. CLINICAL PERSPECTIVE: We developed and validated a graph-based algorithm that automatically identifies macro-reentrant AT circuits from LAT maps with high agreement compared to expert electrophysiologist annotation.The algorithm localizes reentrant loop pathways and distinguishes single- from dual-loop circuits.Localizing these pathways may provide insight into circuit mechanisms and help guide ablation.