Prediction of Atrial Fibrillation Recurrence after Thoracoscopic Surgical Ablation Using Machine Learning Techniques.

Sarah W E Baalman,Joris R De Groot,Ricardo R Lopes,Jolien Neefs,Antoine H G Driessen,Bas A J M De Mol,Lucas A Ramos,Wimjan P Van Boven,Henk A Marquering

doi:10.3390/diagnostics11101787

Sarah W E Baalman, Joris R De Groot + Show 7 more

Open Access

https://doi.org/10.3390/diagnostics11101787

Copy DOI

Abstract

Thoracoscopic surgical ablation (SA) for atrial fibrillation (AF) has shown to be an effective treatment to restore sinus rhythm in patients with advanced AF. Identifying patients who will not benefit from this procedure would be valuable to improve personalized AF therapy. Machine learning (ML) techniques may assist in the improvement of clinical prediction models for patient selection. The aim of this study is to investigate how available baseline characteristics predict AF recurrence after SA using ML techniques. One-hundred-sixty clinical baseline variables were collected from 446 AF patients undergoing SA in our tertiary referral center. Multiple ML models were trained on five outcome measurements, including either all or a number of key variables selected by using the least absolute shrinkage and selection operator (LASSO). There was no difference in model performance between different ML techniques or outcome measurements. Variable selection significantly improved model performance (AUC: 0.73, 95% CI: 0.68–0.77). Subgroup analysis showed a higher model performance in younger patients (<55 years, AUC: 0.82 vs. >55 years, AUC 0.66). Recurrences of AF after SA can be predicted best when using a selection of baseline characteristics, particularly in young patients.

Highlights

In patients with advanced atrial fibrillation (AF), thoracoscopic surgical ablation (SA)is effective to restore sinus rhythm (SR) [1]
Despite our knowledge of risk factors that are associated with lower efficacy and more recurrences, there are no risk scores or prediction models available that consider all the available pre-procedural clinical data that may affect the outcome of SA
The aim of this study was (I) to evaluate the proportion of baseline characteristics that are causal risk factors for AF recurrence after SA using different Machine learning (ML) techniques; (II) to investigate the differential performance of ML models on multiple conventional and modified definitions of AF recurrence; and (III) to analyze whether the accuracy of the ML models is pertinent for clinically relevant subgroups

Summary

Introduction

In patients with advanced atrial fibrillation (AF), thoracoscopic surgical ablation (SA)is effective to restore sinus rhythm (SR) [1]. In patients with advanced atrial fibrillation (AF), thoracoscopic surgical ablation (SA). Invasive SA for AF using videoassisted thoracoscopic surgery has increasingly been performed and has a success rate of. Several clinical variables predicting AF recurrence after catheter ablation (CA) have been identified. These variables are currently being applied for patient selection for both. Despite our knowledge of risk factors that are associated with lower efficacy and more recurrences, there are no risk scores or prediction models available that consider all the available pre-procedural clinical data that may affect the outcome of SA. A systematic analysis tool to assess the risk of any ablation failure could potentially lead to enhanced identification of patients who may benefit from SA versus those in whom SA therapy would be futile

Objectives

Methods

Conclusion