First clinical data on artificial intelligence‐guided catheter ablation in long‐standing persistent atrial fibrillation
Fabian Bahlke,Florian Englert,Miruna Popa,Felix Bourier,Tilko Reents,Carsten Lennerz,Hannah Kraft,Alex Tunsch Martinez,Marc Kottmaier,Jan Syväri,Madeleine Tydecks,Marta Telishevska,Sarah Lengauer,Gabriele Hessling,Isabel Deisenhofer,Nico Erhard
DOI: https://doi.org/10.1111/jce.16184
IF: 2.7
2024-01-12
Journal of Cardiovascular Electrophysiology
Abstract:Introduction Despite advanced ablation strategies and major technological improvements, treatment of persistent atrial fibrillation (AF) remains challenging and the underlying pathophysiology is not fully understood. This study analyzed the multiple procedure outcome and safety of catheter ablation of spatiotemporal dispersions (DISPERS) detected by artificial intelligence (AI)‐guided software in patients with long‐standing persistent AF. Methods and Results The Volta VX1 software was used for 50 consecutive patients undergoing catheter ablation for persistent AF. First, high‐density mapping (78% biatrial) with a multipolar mapping catheter was performed. In addition to pulmonary vein isolation (PVI), ablation of DISPERS was performed aiming at homogenizing, dissecting, isolating, or connecting DISPERS areas to nonconducting anatomical structures. Follow‐up contained regular visits at our outpatient clinic at 1, 3, 6, and 12 months including 7‐day Holter electrocardiograms. Patients were mainly suffering from long‐standing persistent AF (mean AF duration 50.30 ± 54.28 months). Following PVI, ablation of left atrial and right atrial DISPERS areas led to AF cycle length prolongation (mean of 162.0 ± 16.6 to 202.2 ± 21.6 ms after) and AF termination to atrial tachycardia (AT) or sinus rhythm (SR) in 12 patients (24%). No stroke or pericardial effusion occurred; major groin complications (pseudoaneurysm n = 1, atrioventricular fistula n = 1) were detected in two patients. After a blanking period of 6 weeks, recurrence of any atrial arrhythmia was documented in 26 patients (52%). The majority of patients presented with organized AT (n = 15) while AF was present in n = 9 patients and AT/AF was observed in n = 2 patients. Twenty‐two patients underwent reablation. During a mean follow‐up of 363.14 ± 187.42 days and after an average of 1.46 ± 0.68 procedures, 82% of patients remained in stable SR. Conclusion DISPERS‐guided ablation using machine learning software (the Volta VX1 software) in addition to PVI in long‐standing persistent AF ablation resulted in high long‐term success rates regarding AF and AT elimination. Most arrhythmia recurrences were reentrant AT. After a total of 1.46 ± 0.68 procedures, freedom from AF/AT was 82%. Despite prolonged procedure times complication rates were low. Randomized studies are necessary to evaluate long‐term efficacy of dispersion‐guided ablation using AI.
cardiac & cardiovascular systems