Digital twin-guided virtual drug test predicts post-ablation recurrence via rotor suppression in atrial fibrillation
T Hwang, O Kwon, K Im, J Huh, D Ahn, H Park, M Kim, D Kim, J Park, H Yu, T Kim, J Uhm, B Joung, H PakAbstract
Background
Accurate prediction of the most effective antiarrhythmic drug (AAD) following atrial fibrillation (AF) ablation offers substantial clinical value. Nevertheless, in vivo pharmacologic testing and rotor mapping present significant technical and ethical challenges.
Purpose
This proof-of-concept study applied patient-specific digital twin simulations to model drug-induced changes in stable phase singularities (PS, rotors) and evaluated their association with post-ablation recurrence, thereby establishing a framework for personalized rhythm management after AF ablation.
Methods
Patient-specific digital twin models were reconstructed using electroanatomical maps acquired during AF ablation. Virtual AF was induced under baseline and four simulated AAD conditions (low and high concentrations of amiodarone and dronedarone). PS were computationally identified from digital twin simulations using phase mapping, and stable PS were defined as those persisting for ≥2 s within a 10 mm-diameter spatial area. Rotor areas were quantified as the ratio of stable PS area to the total left atrial (LA) surface area excluding pulmonary vein regions. Rotor suppression was defined as a ≥1% reduction in this ratio under AAD conditions compared with baseline. Patients were classified as responders or nonresponders, and post-ablation AF or atrial tachycardia (AT) recurrence was compared between the two groups.
Results
Fifty patients (mean age 62.0 ± 10.0 years; 14.0% women; 64.0% with persistent AF) who were treated with amiodarone or dronedarone during the 3-month blanking period were included. Baseline stable PS were identified in 48 patients (96%), with a mean PS-to-LA surface area ratio of 2.4 ± 2.3%. Rotor suppression was observed in 30 patients (60%), with a mean decrease in the ratio of 3.1 ± 1.6%, whereas nonresponders (n = 20, 40%) exhibited minimal change (0.3 ± 0.3%). Responders had significantly smaller LA diameter (42.2 ± 5.8 mm vs. 46.2 ± 5.8 mm; p = 0.022) and higher mean LA voltage (1.3 ± 0.5 mV vs. 0.7 ± 0.4 mV; p < 0.001) than nonresponders. During a 1-year follow-up, patients with rotor suppression demonstrated a lower AF/AT recurrence rate (log-rank p = 0.013). Rotor suppression remained an independent predictor of recurrence (adjusted HR 2.75, 95% CI 1.18–6.42; p = 0.019) and predicted 1-year AF/AT recurrence (AUC 0.762, 95% CI 0.625–0.898).
Conclusion
Digital twin–guided analysis enabled quantitative assessment of rotor suppression and revealed its significant association with post-ablation AF/AT recurrence. Rotor suppression during virtual AAD testing emerged as a predictive marker of clinical outcome, indicating that digital twin simulations could provide a mechanistic and personalized approach to post-ablation rhythm management.Flowchart of the Study