Elias P, Poterucha TJ, Rajaram V, Moller LM, Rodriguez V, Bhave S, Hahn RT, Tison G, Abreau SA, Barrios J, Torres JN, Hughes JW, Perez MV, Finer J, Kodali S, Khalique O, Hamid N, Schwartz A, Homma S, Kumaraiah D, Cohen DJ, Maurer MS, Einstein AJ, Nazif T, Leon MB, and Perotte AJ
Background: Valvular heart disease is an important contributor to cardiovascular morbidity and mortality and remains underdiagnosed. Deep learning analysis of electrocardiography (ECG) may be useful in detecting aortic stenosis (AS), aortic regurgitation (AR), and mitral regurgitation (MR)., Objectives: This study aimed to develop ECG deep learning algorithms to identify moderate or severe AS, AR, and MR alone and in combination., Methods: A total of 77,163 patients undergoing ECG within 1 year before echocardiography from 2005-2021 were identified and split into train (n = 43,165), validation (n = 12,950), and test sets (n = 21,048; 7.8% with any of AS, AR, or MR). Model performance was assessed using area under the receiver-operating characteristic (AU-ROC) and precision-recall curves. Outside validation was conducted on an independent data set. Test accuracy was modeled using different disease prevalence levels to simulate screening efficacy using the deep learning model., Results: The deep learning algorithm model accuracy was as follows: AS (AU-ROC: 0.88), AR (AU-ROC: 0.77), MR (AU-ROC: 0.83), and any of AS, AR, or MR (AU-ROC: 0.84; sensitivity 78%, specificity 73%) with similar accuracy in external validation. In screening program modeling, test characteristics were dependent on underlying prevalence and selected sensitivity levels. At a prevalence of 7.8%, the positive and negative predictive values were 20% and 97.6%, respectively., Conclusions: Deep learning analysis of the ECG can accurately detect AS, AR, and MR in this multicenter cohort and may serve as the basis for the development of a valvular heart disease screening program., Competing Interests: Funding Support and Author Disclosures This study was supported by a National Institutes of Health Institutional Training Grant (2T32HL007854-21) and National Institutes of Health/National Heart, Lung, and Blood Institute Award (R01HL148248) (to Dr Elias and Dr Perotte’s institution). Dr Elias has received research grant support provided to his institution from Pfizer, Eidos Therapeutics, Google, and Edwards LifeSciences; and is an inventor on a pending patent related to the ValveNet algorithm described in this paper. Dr Poterucha owns stock in Abbott Laboratories and Baxter International with research support provided to his institution from the Amyloidosis Foundation, Eidos Therapeutics, Pfizer, Edwards Lifesciences, and the Glorney-Raisbeck Fellowship Award from the New York Academy of Medicine; and is an inventor on a pending patent related to the ValveNet algorithm described in this paper. Dr Hahn has received speaker fees from Abbott Structural, Edwards Lifesciences, and Philips Healthcare; has institutional consulting contracts with Abbott Structural, Boston Scientific, Edwards Lifesciences, and Gore and Associates; has equity with Navigate; and is Chief Scientific Officer for the Echocardiography Core Laboratory at the Cardiovascular Research Foundation for multiple industry-sponsored trials, for which she receives no direct industry compensation. Dr Tison has received research grants from General Electric, Janssen Pharmaceuticals, and Myokardia; has received personal fees from Myokardia Digital Health as an advisory group member; and has served as an unpaid advisor for Cardiogram. Dr Kodali has served as a consultant for or received honoraria from Admedus, Meril Lifesciences, JenaValve, and Abbott Vascular; has served on the scientific advisory board for and owns equity in Dura Biotech, MicroInterventional Devices, Thubrikar Aortic Valve, Supira, and Admedus; and has received institutional funding to Columbia University and/or the Cardiovascular Research Foundation from Edwards Lifesciences, Medtronic, Abbott Vascular, Boston Scientific, and JenaValve. Dr Khalique has served as a consultant for Boston Scientific, Edwards Lifesciences, and Abbott Structural. Dr Cohen has received research grant support and consulting income from Edwards Lifesciences, Abbott, Boston Scientific, and Medtronic. Dr Einstein has received speaker fees from Ionetix; has received consulting fees from W. L. Gore and Associates; has received authorship fees from Wolters Kluwer Healthcare–UpToDate; and his institution has grants/grants pending from Attralus, Canon Medical Systems, Eidos Therapeutics, GE Healthcare, Pfizer, Roche Medical Systems, W. L. Gore and Associates, and XyloCor Therapeutics. Dr Nazif has served as a consultant for or received honoraria from Edwards Lifesciences, Medtronic, Venus Medtech, and Boston Scientific. Dr Leon has received institutional grants for clinical research from Abbott, Boston Scientific, Edwards, JenaValve, and Medtronic; has received stock options (equity) for advisory board participation in Valve Medical, Picardia, and Venus MedTech; and is an inventor on a pending patent related to the ValveNet algorithm described in this paper. Dr Perotte is Chief Scientific and Innovation Officer at Spiden AG; and is an inventor on a pending patent related to the ValveNet algorithm described in this paper. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose., (Copyright © 2022 The Authors. Published by Elsevier Inc. All rights reserved.)