Virtual In-Silico Modeling Guided Catheter Ablation Predicts Effective Linear Ablation Lesion Set for Longstanding Persistent Atrial Fibrillation: Multicenter Prospective Randomized Study

Abstract

<b>Objective:</b> Radiofrequency catheter ablation for persistent atrial fibrillation (PeAF) still has a substantial recurrence rate. This study aims to investigate whether an AF ablation lesion set chosen using <i>in-silico</i> ablation (V-ABL) is clinically feasible and more effective than an empirically chosen ablation lesion set (Em-ABL) in patients with PeAF. <b>Methods:</b> We prospectively included 108 patients with antiarrhythmic drug-resistant PeAF (77.8% men, age 60.8 ± 9.9 years), and randomly assigned them to the V-ABL (<i>n</i> = 53) and Em-ABL (<i>n</i> = 55) groups. Five different <i>in-silico</i> ablation lesion sets [1 pulmonary vein isolation (PVI), 3 linear ablations, and 1 electrogram-guided ablation] were compared using heart-CT integrated AF modeling. We evaluated the feasibility, safety, and efficacy of V-ABL compared with that of Em-ABL. <b>Results:</b> The pre-procedural computing time for five different ablation strategies was 166 ± 11 min. In the Em-ABL group, the earliest terminating blinded <i>in-silico</i> lesion set matched with the Em-ABL lesion set in 21.8%. V-ABL was not inferior to Em-ABL in terms of procedure time (<i>p</i> = 0.403), ablation time (<i>p</i> = 0.510), and major complication rate (<i>p</i> = 0.900). During 12.6 ± 3.8 months of follow-up, the clinical recurrence rate was 14.0% in the V-ABL group and 18.9% in the Em-ABL group (<i>p</i> = 0.538). In Em-ABL group, clinical recurrence rate was significantly lower after PVI+posterior box+anterior linear ablation, which showed the most frequent termination during <i>in-silico</i> ablation (log-rank <i>p</i> = 0.027). <b>Conclusions:</b> V-ABL was feasible in clinical practice, not inferior to Em-ABL, and predicts the most effective ablation lesion set in patients who underwent PeAF ablation.