Arrhythmias & electrophysiology
Machine Learning Integration of Ablation Quality Metrics for Predicting Atrial Fibrillation Recurrence
- multicenter, retrospective cohort study
- development and external validation of a machine learning model
- 5-fold cross-validation for internal validation
+2 more
- development and external validation of a machine learning model
- 5-fold cross-validation for internal validation
Authors' conclusion
Integrating ablation quality metrics yielded an validated model that showed strong discrimination for 1-year AF recurrence, outperforming clinical features alone. This approach highlights the prognostic value of intraprocedural quality assessment and may support personalized patient management.
Study context
- Population
- Patients aged 18 to 80 years undergoing their first atrial fibrillation (AF) ablation in 14 hospitals from mainland Chinapatients enrolled between December 2021 and December 2023 · external validation centers 239
- Participants
- 1,782 patients
- Exposure
- XGBoost model integrating clinical characteristics with ablation quality metrics vs clinical features alone, predicting 1-year AF recurrence
Main finding
As reported in the abstract
AF recurrence · within 1 year
Patients with AF recurrence
294
16.5%
Observational data: associations, not effects.
Further findings
Discrimination in internal validation
- AUC
- 0.918, 95% CI: 0.874-0.962
Feature-combination analyses
- AUC using clinical features alone
- 0.588
- AUC after incorporating ablation quality metrics
- 0.918
Discrimination in 2 independent external validation centers
- AUCs
- 0.992 and 0.951
- Precision-recall AUCs
- 0.960 and 0.815
Original publication
Machine Learning Integration of Ablation Quality Metrics for Predicting Atrial Fibrillation Recurrence. JACC Clin Electrophysiol. Published 3 October 2026. doi:10.
Summarised from the published abstract; every number is checked against it