Left atrial wall stress and the outcome of catheter ablation for atrial fibrillation: artificial intelligence-based prediction of clinical outcome. (24th May 2021)

• Record Type: Journal Article
• Authors: Lee, J; Kwon, OS; Lee, JS; Yu, HT; Kim, TH; Uhm, JS; Joung, BY; Lee, MH; Pak, HN
• Is Part Of: Europace. Volume 23:Supplement 3(2021)
• Publication Date: 2021-05-24
• Subjects: Arrhythmia -- Treatment -- Periodicals; Cardiac pacing -- Periodicals; Catheter ablation -- Periodicals; Heart -- Physiology -- Periodicals; Electrophysiology -- Periodicals; 617.4120645
• DOI: 10.1093/europace/euab116.296
• Languages: English
• ISSNs: 1099-5129

Abstract

Abstract: Funding Acknowledgements: Type of funding sources: None. Background/Introduction: Left atrial (LA) wall stress (LAW-str) may contribute to the mechanism of atrial fibrillation (AF). Purpose: We explored the clinical characteristics and the rhythm outcome of AF depending on LAW-str among the patients who underwent AF catheter ablation (AFCA). We also tested whether artificial intelligence (AI) properly estimate LAW-str without invasive parameters. Methods: We included 2223 patients (72.8% male, 59.0 [52.0–67.0] years old, 28.7% with persistent AF [PeAF]) who underwent radiofrequency catheter ablation (RFCA). LAW-str was calculated by the Law of Laplace using LA diameter (echocardiogram), peak LA pressure, and mean LA wall thickness (computed tomography) measured by customized software. Based on the quartile (Q1–4) or AI-estimated values of LAW-str, we compared clinical characteristics and rhythm outcome. Results: LAW-str was independently associated with PeAF (p < 0.001), diabetes (p = 0.012) and vascular disease (p = 0.002), body mass index (p < 0.001), E/Em (p < 0.001), and mean LA voltage (p < 0.001). During 26.0 (12.0–52.0) months follow-up, clinical recurrence of AF was significantly higher in the Q4 of LAW-str group (log rank p = 0.001). LAW-str was independently associated with clinical recurrence after AFCA (HR 1.001 [1.000–1.002], p = 0.013). AI-based model using non-invasive parameters predicted Q4 of LAW-str with area under the curve (AUC) 0.734, which was similar to logistic regression based predictive model using all data including invasive parameters (AUC 0.726). Patients in Q4 of LAW-str showed consistently worse rhythm outcome regardless of the type of AF, sex, or AI-based prediction (p = 0.039). Conclusions: The LAW-str seems to be associated with rhythm outcome of AFCA and AI can predict this complex parameter with moderate accuracy.