Creation and Validation of an Algorithm Predicting Recurrence of Atrial Fibrillation Following Pulmonary Vein Isolation Utilizing Ensemble Modeling Techniques

Abstract Catheter ablation (CA) of atrial fibrillation (AF) represents a mainstay of treatment of this prevalent arrhythmia. Clinical trials investigating the efficacy of CA may poorly represent real-world patient populations. However, many real-world data sets possess missing data, which may impede their applicability in research. We sought to use ensemble modeling to address missing data and develop a model to estimate the probability of AF recurrence following CA. We retrospectively analyzed clinical variables in 476 patients who underwent an initial CA of AF. Univariate and multivariate logistic regression was performed to determine those variables predictive of AF recurrence. A multivariate logistic model was created to estimate the probability of AF recurrence after CA. Missing data was addressed using ensemble modeling and variable selection was performed using the aggregate of multiple models. After analysis, six variables remained in the model. Predictive modeling was performed using these variables for 1000 randomly partitioned datasets (80% training, 20% testing) and 1000 random imputations for each partitioned dataset. The model predicted AF recurrence with an accuracy of 74.34 ± 3.99%. Application of this model to patients undergoing CA may help identify those at risk of AF recurrence.