Clinical application of virtual antiarrhythmic drug test using digital twins in patients who recurred atrial fibrillation after catheter ablation

Abstract

Abstract Funding Acknowledgements Type of funding sources: None. Background We previously reported a proof-of-concept study that digital twins can test antiarrhythmic drug (AAD) effects on atrial fibrillation (AF) after integrating the patients’ cardiac CT imaging and electroanatomical map (EAM). In this study, we evaluated the reproducibility of the virtual AAD test (V-AAD) in the clinical setting in patients who recurred after AF catheter ablation (AFCA) and took AADs. Methods This single-center retrospective study included 232 patients (71.6% male, 60.5±10.2 years of age, 40.9% paroxysmal AF) who were prescribed AADs after AFCA due to real or high chance of recurrence. Clinical AADs (C-AAD) were chosen at the discretion of the primary care physician blinded to V-AAD test. All those included patients underwent V-AAD test using digital twins reflecting the patients’ atrial anatomy, histology, and electrophysiology by integrating cardiac CT and EAM. The ion currents for each AAD (amiodarone, sotalol, dronedarone, flecainide, and propafenone; 2 doses for each drug) were defined based on previous publications. We defined the most efficient V-AAD as the V-AAD that terminated virtual AF the fastest. We evaluated the association between AF recurrence rate while using C-AAD and the most efficient V-AAD, which showed excellent termination/ defragmentation of AF in the digital twins. Results In the clinical setting, the most effective V-AAD was administered in 72 patients (31.0%), and not effective V-AADs in 61 patients (26.3%). During the median 782 (IQR 582-1047) days follow-up, AF recurrence was observed in 132 patients (56.9%) (under the class IC 44.4% and class III 57.4%). The recurrence rate within a year after using the effective V-AAD was 40.9%, and that of using the ineffective V-AAD was 54.1%, respectively (log-rank p=0.046). The recurrence rate within a year in patients with ≥2 effective V-AADs and those with 0 or 1 were 42.4% and 59.3%, respectively (log-rank p=0.056). Conclusions The digital twins-guided V-AAD test was feasible for evaluating the efficacy of multiple AADs in patients with AF who recurred after AFCA. We need a prospective randomization study to evaluate the prediction power of the V-AAD test.