PREDICTION OF ATRIAL FIBRILLATION RECURRENCE AFTER REPEAT ABLATION USING MACHINE LEARNING

BackgroundPulmonary vein isolation is a widely used method to treat Atrial fibrillation (AF). However, many patients experience AF recurrence, leading repeated ablation requirements. In this study, we sought identify whether left atrial (LA) geometry can at risk of recurrence after second ablation.ObjectiveTo predict repeat using machine learning on clinical covariates and LA geometric features calculated from CT scans.MethodsThis retrospective single-center study included 102 who experienced single radiofrequency procedure underwent ablation. 38 had within 12 months Cardiac images taken prior were collected for each patient. From the segmented 3D geometry, 11 which along with 35 analyzed model development. Performance was k-fold cross-validation 76 relevant optimized selected maximize performance (Fig. 1).ResultsThe include sphericity compactness (how much shape resembles that sphere) while CHA2DS2-VASc score, history any cardioversion ARB medication. On held-out test set 26 patients, radial-basis support vector (RBF-SVM) achieved best obtained an accuracy 0.67±0.08, specificity 0.67±0.07, sensitivity 0.67±0.15 NPV 0.77±0.09 2).ConclusionView Large Image Figure ViewerDownload Hi-res image Download (PPT)Receiver operating characteristic (ROC) curve SVM-RBF classifier patients. Pulmonary ObjectiveTo scans. To MethodsThis 1). This ResultsThe 2). The ConclusionReceiver Receiver