Detecting structural heart disease from electrocardiograms using AI

Early detection of structural heart disease is critical to improving outcomes, but widespread screening remains limited by the cost and accessibility of imaging tools such as echocardiography<sup>1,2</sup>. Recent advances in machine learning applied to heart rhythm recordings have shown promise in identifying disease<sup>3,4</sup>, although previous work has been limited by development in narrow populations or targeting only select heart conditions<sup>5</sup>. Here we introduce a deep learning model, EchoNext, trained on more than 1 million heart rhythm and imaging records across a large and diverse health system to detect many forms of structural heart disease. The model demonstrated high diagnostic accuracy in internal and external validation, outperforming cardiologists in a controlled evaluation and showing consistent performance across different care settings and racial and/or ethnic groups. The models were prospectively evaluated in a clinical trial of patients without previous cardiac imaging, successfully identifying previously undiagnosed heart disease. These findings support the potential of artificial intelligence to expand access to heart disease screening at scale. To enable further development and transparency, we have publicly released model weights and a large, annotated dataset linking heart rhythm data to imaging-based diagnoses.

Authors

Institutions

Publication Details

Journal
Nature
Published
2025-07-16
DOI
https://doi.org/10.1038/s41586-025-09227-0
Citations
47
Primary Topic
Cardiac Valve Diseases and Treatments
Type
article
Field-Weighted Citation Impact
65.24

Funders

Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
article

Detecting structural heart disease from electrocardiograms using AI

Joshua Finer, Polydoros Ν. Kampaktsis, Ashley Beecy, Christopher M. Haggerty et al.
47 citations
Nature
Cardiac Valve Diseases and Treatments
65.24
article

Detecting structural heart disease from electrocardiograms using AI

Joshua Finer, Polydoros Ν. Kampaktsis, Ashley Beecy, Christopher M. Haggerty, Denis Corbin, Joshua P. Barrios, Dustin N. Hartzel, Geoffrey H. Tison, Carol A. Waksmonski, Siqin Ye, José Dizon, Corey Bradley, Prashant Vaishnava, J. Weston Hughes, Aaron S. Long, Deepa Kumaraiah, Allan Schwartz, Andrew Jeffrey Einstein, Alexander Volodarskiy, Pierre Elias, Brock Daniels, Vratika Agarwal, Ersilia M. DeFilippis, Michael Adjei-Mosi, Olivier Tastet, Rebecca Tung Hahn, Sofia Shames, David P. vanMaanen, Mathew S. Maurer, FRANCIS H. OLIVER, Gregg Rosner, Timothy J. Poterucha, Shalmali Joshi, Michelle Castillo, Shah Raj Ali, Jeffrey Ruhl, Marc A. Probst, David Ouyang, Mark Lebehn, Richard Ricart, Robert Avram, Qi Liu, Linyuan Jing, D. Rocha, I-Min Chiu, Shunichi Homma, Martin B. Leon, Heidi Hartman, C. D. Kelsey, Paloma P Malta
article en
47 citations

Abstract

Early detection of structural heart disease is critical to improving outcomes, but widespread screening remains limited by the cost and accessibility of imaging tools such as echocardiography<sup>1,2</sup>. Recent advances in machine learning applied to heart rhythm recordings have shown promise in identifying disease<sup>3,4</sup>, although previous work has been limited by development in narrow populations or targeting only select heart conditions<sup>5</sup>. Here we introduce a deep learning model, EchoNext, trained on more than 1 million heart rhythm and imaging records across a large and diverse health system to detect many forms of structural heart disease. The model demonstrated high diagnostic accuracy in internal and external validation, outperforming cardiologists in a controlled evaluation and showing consistent performance across different care settings and racial and/or ethnic groups. The models were prospectively evaluated in a clinical trial of patients without previous cardiac imaging, successfully identifying previously undiagnosed heart disease. These findings support the potential of artificial intelligence to expand access to heart disease screening at scale. To enable further development and transparency, we have publicly released model weights and a large, annotated dataset linking heart rhythm data to imaging-based diagnoses.

NatureVol. 644(8075)
Cedars-Sinai Medical Center (US), Cornell University (US), Columbia University Irving Medical Center (US), Montreal Heart Institute (CA), New York Hospital Queens (US), Kaohsiung Chang Gung Memorial Hospital (TW), Cardiovascular Research Foundation (US), Weill Cornell Medicine (US), Columbia University (US)
American Heart Association, Patient-Centered Outcomes Research Institute, Amyloidosis Foundation, Fondation Institut de Cardiologie de Montréal, Institut de Cardiologie de Montréal, National Institutes of Health, National Heart, Lung, and Blood Institute
Good health and well-being
Openalex Percentile: Top 0%
Cardiac Valve Diseases and Treatments
65.24
DOIOpen Access PDFOpenAlex
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.