Abstract
Introduction: Mitral valve transcatheter edge-to-edge repair (M-TEER) has emerged as a minimally invasive option for high-risk surgical candidates with severe and symptomatic mitral regurgitation (MR), but post-procedure residual mitral valve (MV) dysfunction remains a significant concern. This study evaluates the clinical utility of combining artificial intelligence applied to electrocardiograms (ECG-AI) for diastolic dysfunction (DD) grading and the echocardiography-derived velocity time integral of the MV and left ventricular outflow tract ratio (VTI(MV)/(LVOT)) in predicting prognosis in patients post-M-TEER. Methods: A retrospective analysis of patients who underwent M-TEER between 2014 and 2021 was conducted. Patients were categorized based on VTI(MV/LVOT) and ECG-AI scores into three groups: both normal parameters, either abnormal parameter, or both abnormal parameters to compare outcomes (mortality, major adverse cardiovascular events [MACE], and the need for subsequent MV reintervention) using Kaplan-Meier analysis, multivariable Cox regression models, and net reclassification improvement. Results: Overall, 250 patients were included; the median age was 79.5 (IQR: 73.1, 84.6) and 66.4% were male. The combined abnormal VTI(MV/LVOT) (≥2.5) and ECG-AI score for DD (>1) was associated with higher risk of one-year mortality (adjusted HR: 4.56 [1.04-19.89], p = 0.044) and MACE (adjusted HR: 3.72 [1.09-12.72], p = 0.037) compared to patients with both normal parameters. Conclusions: This study highlights the potential additive value of integrating VTI(MV/LVOT) and ECG-AI scores as a prognostic tool for a personalized approach to the post-operative evaluation and risk stratification in M-TEER patients.