{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"institution":[{"name":"medRxiv"}],"indexed":{"date-parts":[[2026,7,17]],"date-time":"2026-07-17T05:13:07Z","timestamp":1784265187108,"version":"3.55.0"},"posted":{"date-parts":[[2025,11,6]]},"group-title":"Cardiovascular Medicine","reference-count":41,"publisher":"openRxiv","license":[{"start":{"date-parts":[[2025,11,6]],"date-time":"2025-11-06T00:00:00Z","timestamp":1762387200000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/creativecommons.org\/licenses\/by-nc-nd\/4.0\/"}],"funder":[{"DOI":"10.13039\/100000050","name":"National Heart, Lung, and Blood Institute","doi-asserted-by":"crossref","award":["R01HL167858"],"award-info":[{"award-number":["R01HL167858"]}],"id":[{"id":"10.13039\/100000050","id-type":"DOI","asserted-by":"crossref"}]},{"DOI":"10.13039\/100000050","name":"National Heart, Lung, and Blood Institute","doi-asserted-by":"crossref","award":["K23HL153775"],"award-info":[{"award-number":["K23HL153775"]}],"id":[{"id":"10.13039\/100000050","id-type":"DOI","asserted-by":"crossref"}]},{"DOI":"10.13039\/100000050","name":"National Heart, Lung, and Blood Institute","doi-asserted-by":"crossref","award":["F32HL170592"],"award-info":[{"award-number":["F32HL170592"]}],"id":[{"id":"10.13039\/100000050","id-type":"DOI","asserted-by":"crossref"}]},{"DOI":"10.13039\/100000049","name":"National Institute on Aging","doi-asserted-by":"crossref","award":["R01AG089981"],"award-info":[{"award-number":["R01AG089981"]}],"id":[{"id":"10.13039\/100000049","id-type":"DOI","asserted-by":"crossref"}]},{"DOI":"10.13039\/100000862","name":"Doris Duke Foundation","doi-asserted-by":"crossref","award":["2022060"],"award-info":[{"award-number":["2022060"]}],"id":[{"id":"10.13039\/100000862","id-type":"DOI","asserted-by":"crossref"}]},{"DOI":"10.13039\/100000968","name":"American Heart Association","doi-asserted-by":"crossref","award":["26CDA1612298"],"award-info":[{"award-number":["26CDA1612298"]}],"id":[{"id":"10.13039\/100000968","id-type":"DOI","asserted-by":"crossref"}]},{"name":"Robert A. Winn Excellence in Clinical Trials Career Development Award"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"accepted":{"date-parts":[[2026,5,25]]},"abstract":"<jats:title>ABSTRACT<\/jats:title>\n                <jats:sec>\n                  <jats:title>Background<\/jats:title>\n                  <jats:p>Artificial intelligence-enhanced electrocardiography (AI-ECG) enables structural heart disease (SHD) detection. However, its utility as a clinical assay requires consistency of outputs when repeated under similar conditions.<\/jats:p>\n                <\/jats:sec>\n                <jats:sec>\n                  <jats:title>Objectives<\/jats:title>\n                  <jats:p>To characterize test-retest reliability of contemporary AI-ECG models across diverse health systems, identify factors associated with discordance, and determine the predictive significance of screen status changes.<\/jats:p>\n                <\/jats:sec>\n                <jats:sec>\n                  <jats:title>Methods<\/jats:title>\n                  <jats:p>We identified ECG pairs recorded 1-30 days apart in the same individual at the Yale-New Haven Health System (YNHHS), Massachusetts General Hospital (MGH), and Emory University Hospital (EUH). We evaluated internally developed ECG signal- and image-based SHD models, and the EchoNext-Mini model, including disease-specific components and ensemble composites. Reliability was quantified with concordance correlation coefficients (CCCs) and categorical concordance percentages. In patients without prevalent heart failure (HF) with serial ECGs 30-90 days apart, we evaluated association of screen status discordance with new-onset HF risk.<\/jats:p>\n                <\/jats:sec>\n                <jats:sec>\n                  <jats:title>Results<\/jats:title>\n                  <jats:p>We included 731,466 ECG pairs (median interval 5-6 days). At YNHHS, disease-specific model CCCs ranged from 0.77-0.86 across the signal- and image-based model families and 0.50-0.97 for EchoNext-Mini output nodes. Composite SHD models had CCCs of 0.90 (signal-based), 0.90 (image-based), and 0.81 (EchoNext-Mini). The image-based ensemble model achieved categorical screen-status concordance of 87-89% across 3 health systems (versus 80-84% for disease-specific models). Younger age (&lt; 65 years) was the dominant correlate of discordance (OR 1.69 [95% CI, 1.65-1.74]), along with higher odds for inpatient ECGs (ORs 1.30-1.41). Among 65,838 individuals in the predictive cohort, a negative-to-positive change in screen status was associated with elevated risk of new-onset HF (adjusted hazard ratio 1.67-2.37 across sites).<\/jats:p>\n                <\/jats:sec>\n                <jats:sec>\n                  <jats:title>Conclusions<\/jats:title>\n                  <jats:p>Contemporary AI-ECG models had high test-retest reliability, with composite ensembles showing the greatest stability. Serial changes in AI-ECG screen status carry predictive information for new-onset HF.<\/jats:p>\n                <\/jats:sec>\n                <jats:sec>\n                  <jats:title>CONDENSED ABSTRACT<\/jats:title>\n                  <jats:p>Clinical utility of artificial intelligence-enhanced electrocardiography (AI-ECG) requires consistency when repeated under similar conditions. In 731,466 ECG pairs from 3 health systems, AI-ECG models for structural heart disease (SHD) had high test-retest reliability, with composite SHD models showing the greatest stability (concordance correlation coefficient 0.87-0.90). Signal- and image-based models had comparable reliability. Younger age and inpatient acquisition were the primary correlates of screen-status discordance. Among 65,838 individuals at risk, a negative-to-positive change in screen status independently predicted new-onset heart failure (adjusted hazard ratio 1.67-2.37 across sites). Test-retest reliability and serial screen-status changes present meaningful clinical benchmarks for AI-ECG implementation.<\/jats:p>\n                <\/jats:sec>","DOI":"10.1101\/2025.11.04.25339526","type":"posted-content","created":{"date-parts":[[2025,11,6]],"date-time":"2025-11-06T20:35:13Z","timestamp":1762461313000},"source":"Crossref","is-referenced-by-count":2,"title":["Test-Retest Reliability of Artificial Intelligence-Enhanced Electrocardiography: A Multi-Center Study"],"prefix":"10.64898","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-5664-4126","authenticated-orcid":false,"given":"Lovedeep S","family":"Dhingra","sequence":"first","affiliation":[{"name":"Cardiovascular Data Science (CarDS) Lab, Yale School of Medicine, New Haven, CT, USA"},{"name":"Section of Cardiovascular Medicine, Department of Internal Medicine, Yale School of Medicine, New Haven, CT, USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Philip M","family":"Croon","sequence":"additional","affiliation":[{"name":"Cardiovascular Data Science (CarDS) Lab, Yale School of Medicine, New Haven, CT, USA"},{"name":"Section of Cardiovascular Medicine, Department of Internal Medicine, Yale School of Medicine, New Haven, CT, USA"},{"name":"Department of Cardiology, Amsterdam Cardiovascular Sciences, Amsterdam University Medical Center, University of Amsterdam, Amsterdam, The Netherlands"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Bruno","family":"Batinica","sequence":"additional","affiliation":[{"name":"Cardiovascular Data Science (CarDS) Lab, Yale School of Medicine, New Haven, CT, USA"},{"name":"Section of Cardiovascular Medicine, Department of Internal Medicine, Yale School of Medicine, New Haven, CT, USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-4362-0720","authenticated-orcid":false,"given":"Evangelos K","family":"Oikonomou","sequence":"additional","affiliation":[{"name":"Cardiovascular Data Science (CarDS) Lab, Yale School of Medicine, New Haven, CT, USA"},{"name":"Section of Cardiovascular Medicine, Department of Internal Medicine, Yale School of Medicine, New Haven, CT, USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-9467-6199","authenticated-orcid":false,"given":"Rohan","family":"Khera","sequence":"additional","affiliation":[{"name":"Cardiovascular Data Science (CarDS) Lab, Yale School of Medicine, New Haven, CT, USA"},{"name":"Section of Cardiovascular Medicine, Department of Internal Medicine, Yale School of Medicine, New Haven, CT, USA"},{"name":"Center for Outcomes Research and Evaluation (CORE), Yale New Haven Hospital, New Haven, CT, USA"},{"name":"Department of Biomedical Informatics and Data Science, Yale School of Medicine, New Haven, CT, USA"},{"name":"Section of Health Informatics, Department of Biostatistics, Yale School of Public Health, New Haven, CT, USA"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"54368","reference":[{"key":"2026052707100519000_2025.11.04.25339526v2.1","doi-asserted-by":"publisher","DOI":"10.1093\/eurheartj\/ehab649"},{"key":"2026052707100519000_2025.11.04.25339526v2.2","doi-asserted-by":"crossref","first-page":"465","DOI":"10.1038\/s41569-020-00503-2","article-title":"Artificial intelligence-enhanced electrocardiography in cardiovascular disease management","volume":"18","year":"2021","journal-title":"Nat Rev Cardiol"},{"key":"2026052707100519000_2025.11.04.25339526v2.3","doi-asserted-by":"crossref","unstructured":"Picano E . Economic and biological costs of cardiac imaging. Cardiovasc Ultrasound. 2005;3.","DOI":"10.1186\/1476-7120-3-13"},{"key":"2026052707100519000_2025.11.04.25339526v2.4","doi-asserted-by":"publisher","DOI":"10.1007\/s12350-009-9104-y"},{"key":"2026052707100519000_2025.11.04.25339526v2.5","doi-asserted-by":"publisher","DOI":"10.1016\/j.jacc.2025.01.030"},{"key":"2026052707100519000_2025.11.04.25339526v2.6","doi-asserted-by":"publisher","DOI":"10.1093\/eurheartj\/ehae914"},{"key":"2026052707100519000_2025.11.04.25339526v2.7","doi-asserted-by":"publisher","DOI":"10.1161\/CIRCULATIONAHA.122.062646"},{"key":"2026052707100519000_2025.11.04.25339526v2.8","doi-asserted-by":"publisher","DOI":"10.1161\/CIRCULATIONAHA.121.057869"},{"key":"2026052707100519000_2025.11.04.25339526v2.9","unstructured":"Poterucha TJ , Jing L , Ricart RP , et al. Detecting structural heart disease from electrocardiograms using AI. Nature. 2025:1\u201310."},{"key":"2026052707100519000_2025.11.04.25339526v2.10","doi-asserted-by":"crossref","unstructured":"Hughes JW , Jing L , Finer J , et al. EchoNext-mini: A dataset and baseline AI model for detecting structural heart disease from electrocardiograms. NEJM AI. 2026;3.","DOI":"10.1056\/AIdbp2500516"},{"key":"2026052707100519000_2025.11.04.25339526v2.11","doi-asserted-by":"crossref","unstructured":"Pedroso AF , Nascimento BR , Dhingra LS , et al. Artificial intelligence-enabled electrocardiography to triage echocardiography for structural heart disease diagnosis in a low-resource setting. Am J Prev Cardiol. 2026:101539.","DOI":"10.1016\/j.ajpc.2026.101539"},{"key":"2026052707100519000_2025.11.04.25339526v2.12","doi-asserted-by":"crossref","first-page":"117","DOI":"10.1001\/jamacardio.2025.4591","article-title":"Detecting transthyretin cardiac amyloidosis with artificial intelligence: A nonrandomized clinical trial: A nonrandomized clinical trial","volume":"11","year":"2026","journal-title":"JAMA Cardiol"},{"key":"2026052707100519000_2025.11.04.25339526v2.13","doi-asserted-by":"crossref","first-page":"991","DOI":"10.1038\/s44161-025-00685-3","article-title":"Identification of hypertrophic cardiomyopathy on electrocardiographic images with deep learning","volume":"4","year":"2025","journal-title":"Nat Cardiovasc Res"},{"key":"2026052707100519000_2025.11.04.25339526v2.14","doi-asserted-by":"crossref","first-page":"35","DOI":"10.1016\/j.amjcard.2024.11.028","article-title":"A multicenter evaluation of the impact of therapies on deep learning-based electrocardiographic hypertrophic cardiomyopathy markers","volume":"237","year":"2025","journal-title":"Am J Cardiol"},{"key":"2026052707100519000_2025.11.04.25339526v2.15","doi-asserted-by":"crossref","first-page":"10","DOI":"10.1007\/s11897-026-00748-x","article-title":"Artificial intelligence-enhanced electrocardiography for heart failure screening and risk stratification","volume":"23","year":"2026","journal-title":"Curr Heart Fail Rep"},{"key":"2026052707100519000_2025.11.04.25339526v2.16","doi-asserted-by":"crossref","first-page":"102572","DOI":"10.1016\/j.jacadv.2025.102572","article-title":"Artificial intelligence-enhanced electrocardiogram models for detection of left ventricular dysfunction: A comparison study","volume":"5","year":"2026","journal-title":"JACC Adv"},{"key":"2026052707100519000_2025.11.04.25339526v2.17","doi-asserted-by":"crossref","first-page":"ztag057","DOI":"10.1093\/ehjdh\/ztag057","article-title":"Open science requires trust and rigour: a framework for responsible evaluation of shared AI-ECG tools","volume":"7","year":"2026","journal-title":"Eur Heart J Digit Health"},{"key":"2026052707100519000_2025.11.04.25339526v2.18","doi-asserted-by":"crossref","first-page":"843","DOI":"10.1016\/j.jacc.2025.08.021","article-title":"Guiding the AI revolution in cardiovascular medicine","volume":"86","year":"2025","journal-title":"J Am Coll Cardiol"},{"key":"2026052707100519000_2025.11.04.25339526v2.19","doi-asserted-by":"crossref","first-page":"691","DOI":"10.1016\/j.jacc.2025.07.040","article-title":"Moving beyond the model","volume":"86","year":"2025","journal-title":"J Am Coll Cardiol"},{"key":"2026052707100519000_2025.11.04.25339526v2.20","doi-asserted-by":"crossref","first-page":"1054","DOI":"10.1093\/eurheartj\/ehae906","article-title":"Using artificial intelligence to spot heart failure from ECGs: is it prime time?","volume":"46","year":"2025","journal-title":"Eur Heart J"},{"key":"2026052707100519000_2025.11.04.25339526v2.21","doi-asserted-by":"crossref","first-page":"1314","DOI":"10.1016\/j.jacc.2025.02.008","article-title":"Implementing ECG-AI in Clinical Care: From Pixel to Practice","volume":"85","year":"2025","journal-title":"J Am Coll Cardiol"},{"key":"2026052707100519000_2025.11.04.25339526v2.22","doi-asserted-by":"publisher","DOI":"10.1007\/s40037-017-0347-z"},{"key":"2026052707100519000_2025.11.04.25339526v2.23","unstructured":"Anon. Anumana ECG-AI LEF Accessed February 11, 2025. https:\/\/anumana.ai\/ecg-ai-lef."},{"key":"2026052707100519000_2025.11.04.25339526v2.24","first-page":"419","article-title":"Artificial intelligence enabled ECG screening for left ventricular systolic dysfunction: a systematic review","volume":"28","year":"2023","journal-title":"Heart Fail Rev"},{"key":"2026052707100519000_2025.11.04.25339526v2.25","doi-asserted-by":"publisher","DOI":"10.1038\/s41591-018-0240-2"},{"key":"2026052707100519000_2025.11.04.25339526v2.26","doi-asserted-by":"crossref","first-page":"2885","DOI":"10.1093\/eurheartj\/ehab153","article-title":"Electrocardiogram screening for aortic valve stenosis using artificial intelligence","volume":"42","year":"2021","journal-title":"Eur Heart J"},{"key":"2026052707100519000_2025.11.04.25339526v2.27","doi-asserted-by":"crossref","unstructured":"Parsa S , Keyes T , Dash D , et al. Designing clinically useful AI: A blueprint for impact. NEJM AI. 2026;3.","DOI":"10.1056\/AIp2500808"},{"key":"2026052707100519000_2025.11.04.25339526v2.28","doi-asserted-by":"crossref","first-page":"101184","DOI":"10.1016\/j.ajogmf.2023.101184","article-title":"Electrocardiogram-based deep learning model to screen peripartum cardiomyopathy","volume":"5","year":"2023","journal-title":"Am J Obstet Gynecol MFM"},{"key":"2026052707100519000_2025.11.04.25339526v2.29","doi-asserted-by":"crossref","first-page":"e791","DOI":"10.1016\/S2589-7500(24)00172-9","article-title":"Artificial intelligence-enabled electrocardiogram for mortality and cardiovascular risk estimation: a model development and validation study","volume":"6","year":"2024","journal-title":"Lancet Digit Health"},{"key":"2026052707100519000_2025.11.04.25339526v2.30","doi-asserted-by":"crossref","first-page":"106329","DOI":"10.1016\/j.ejim.2025.04.036","article-title":"Artificial intelligence and the electrocardiogram: A modern renaissance","volume":"140","year":"2025","journal-title":"Eur J Intern Med"},{"key":"2026052707100519000_2025.11.04.25339526v2.31","doi-asserted-by":"crossref","first-page":"97","DOI":"10.1016\/j.jacc.2024.05.003","article-title":"Transforming cardiovascular care with artificial intelligence: From discovery to practice","volume":"84","year":"2024","journal-title":"J Am Coll Cardiol"},{"key":"2026052707100519000_2025.11.04.25339526v2.32","doi-asserted-by":"crossref","first-page":"3204","DOI":"10.1093\/eurheartj\/ehae415","article-title":"Artificial intelligence-enhanced patient evaluation: bridging art and science","volume":"45","year":"2024","journal-title":"Eur Heart J"},{"key":"2026052707100519000_2025.11.04.25339526v2.33","doi-asserted-by":"crossref","unstructured":"Koscova Z , Li Q , Robichaux C , et al. The Harvard-Emory ECG database. medRxiv. 2025:2024.09.27.24314503.","DOI":"10.1101\/2024.09.27.24314503"},{"key":"2026052707100519000_2025.11.04.25339526v2.34","doi-asserted-by":"crossref","unstructured":"Koscova Z , Moura Junior V , Reyna M , et al. Harvard-Emory ECG database. 2025.","DOI":"10.1101\/2024.09.27.24314503"},{"key":"2026052707100519000_2025.11.04.25339526v2.35","doi-asserted-by":"crossref","first-page":"1282","DOI":"10.1161\/CIRCULATIONAHA.125.076279","article-title":"Phenotypic selectivity of artificial intelligence-enhanced electrocardiography in cardiovascular diagnosis and risk prediction","volume":"152","year":"2025","journal-title":"Circulation"},{"key":"2026052707100519000_2025.11.04.25339526v2.36","doi-asserted-by":"publisher","DOI":"10.2307\/2532051"},{"key":"2026052707100519000_2025.11.04.25339526v2.37","doi-asserted-by":"crossref","first-page":"e253351","DOI":"10.1001\/jamahealthforum.2025.3351","article-title":"Benefit-risk reporting for FDA-cleared artificial intelligence-enabled medical devices","volume":"6","year":"2025","journal-title":"JAMA Health Forum"},{"key":"2026052707100519000_2025.11.04.25339526v2.38","doi-asserted-by":"crossref","first-page":"3005","DOI":"10.3390\/biomedicines13123005","article-title":"Machine learning-enabled medical devices authorized by the US Food and Drug Administration in 2024: Regulatory characteristics, predicate lineage, and transparency reporting","volume":"13","year":"2025","journal-title":"Biomedicines"},{"key":"2026052707100519000_2025.11.04.25339526v2.39","unstructured":"Center for Devices , Radiological Health . Methods and tools for effective postmarket monitoring of artificial intelligence (AI)-enabled medical devices. US Food and Drug Administration. 2024. Accessed May 14, 2026. https:\/\/www.fda.gov\/medical-devices\/medical-device-regulatory-science-research-programs-conducted-osel\/methods-and-tools-effective-postmarket-monitoring-artificial-intelligence-ai-enabled-medical-devices."},{"key":"2026052707100519000_2025.11.04.25339526v2.40","doi-asserted-by":"crossref","first-page":"e253172","DOI":"10.1001\/jamahealthforum.2025.3172","article-title":"Early recalls and clinical validation gaps in artificial intelligence-enabled medical devices","volume":"6","year":"2025","journal-title":"JAMA Health Forum"},{"key":"2026052707100519000_2025.11.04.25339526v2.41","doi-asserted-by":"publisher","DOI":"10.1093\/ehjdh\/ztae090"}],"container-title":[],"original-title":[],"link":[{"URL":"https:\/\/syndication.highwire.org\/content\/doi\/10.1101\/2025.11.04.25339526","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,5,27]],"date-time":"2026-05-27T14:10:26Z","timestamp":1779891026000},"score":1,"resource":{"primary":{"URL":"http:\/\/medrxiv.org\/lookup\/doi\/10.1101\/2025.11.04.25339526"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,11,6]]},"references-count":41,"URL":"https:\/\/doi.org\/10.1101\/2025.11.04.25339526","relation":{},"subject":[],"published":{"date-parts":[[2025,11,6]]},"subtype":"preprint"}}