Abstract
IntroductionWe developed a prediction model for atrial fibrillation (AF) progression and tested whether machine learning (ML) could reproduce the prediction power in an independent cohort using pre-procedural non-invasive variables alone.MethodsCohort 1 included 1,214 patients and cohort 2, 658, and all underwent AF catheter ablation (AFCA). AF progression to permanent AF was defined as sustained AF despite repeat AFCA or cardioversion under antiarrhythmic drugs. We developed a risk stratification model for AF progression (STAAR score) and stratified cohort 1 into three groups. We also developed an ML-prediction model to classify three STAAR risk groups without invasive parameters and validated the risk score in cohort 2.ResultsThe STAAR score consisted of a stroke (2 points, p = 0.003), persistent AF (1 point, p = 0.049), left atrial (LA) dimension ≥43 mm (1 point, p = 0.010), LA voltage <1.109 mV (2 points, p = 0.004), and PR interval ≥196 ms (1 point, p = 0.001), based on multivariate Cox analyses, and it had a good discriminative power for progression to permanent AF [area under curve (AUC) 0.796, 95% confidence interval (CI): 0.753–0.838]. The ML prediction model calculated the risk for AF progression without invasive variables and achieved excellent risk stratification: AUC 0.935 for low-risk groups (score = 0), AUC 0.855 for intermediate-risk groups (score 1–3), and AUC 0.965 for high-risk groups (score ≥ 4) in cohort 1. The ML model successfully predicted the high-risk group for AF progression in cohort 2 (log-rank p < 0.001).ConclusionsThe ML-prediction model successfully classified the high-risk patients who will progress to permanent AF after AFCA without invasive variables but has a limited discrimination power for the intermediate-risk group.
Highlights
We developed a prediction model for atrial fibrillation (AF) progression and tested whether machine learning (ML) could reproduce the prediction power in an independent cohort using pre-procedural non-invasive variables alone
The cohort was adjusted based on the following exclusion criteria: [1] permanent AF refractory to electrical cardioversion, [2] AF with rheumatic valvular disease, or any mechanical or bioprosthetic heart valve, [3] prior cardiac surgery with concomitant AF surgery or AF catheter ablation, and [4] the absence of a mean left atrial (LA) voltage that could be obtained during the de novo ablation procedure
We considered the progression to permanent AF, which is the endpoint of this study, as sustaining AF detected on an electrocardiogram (ECG) or Holter/event-monitor after last ablation procedures with antiarrhythmic drugs (AAD) or electrical cardioversion
Summary
We developed a prediction model for atrial fibrillation (AF) progression and tested whether machine learning (ML) could reproduce the prediction power in an independent cohort using pre-procedural non-invasive variables alone. Recurrence after AFCA is defined as AF or atrial tachycardia (AT) of 30 s or more regardless of symptoms [1] At this point, it is important to reduce a recurrence rate of atrial arrhythmias lasting 30 s, it might be very important to identify progression to permanent AF in which it is difficult to control sustained AF even after repeated procedures or with use of antiarrhythmic drugs (AADs). Pre-discovery of patients who are likely to progress to permanent AF using pre-procedural parameters may contribute to improvements in AFCA rhythm and clinical outcomes. Variable studies reported or validated the risk model for AF recurrence in the patients who underwent repeat ablations [5]. Those studies showed the range of ACU from 0.487 to 0.833. AI has been suggested for predicting invasive parameters or invasive cardiovascular outcomes [13]
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
