{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,23]],"date-time":"2026-10-23T22:21:58Z","timestamp":1792794118548,"version":"4.5.0"},"reference-count":39,"publisher":"Oxford University Press (OUP)","issue":"3","license":[{"start":{"date-parts":[[2024,2,19]],"date-time":"2024-02-19T00:00:00Z","timestamp":1708300800000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"name":"Pandemic Sciences Institute at the University of Oxford","id":[{"id":"https:\/\/ror.org\/052gg0110","id-type":"ROR","asserted-by":"crossref"}]},{"DOI":"10.13039\/501100000272","name":"National Institute for Health Research","doi-asserted-by":"publisher","id":[{"id":"10.13039\/501100000272","id-type":"DOI","asserted-by":"publisher"}]},{"name":"Oxford Biomedical Research Centre","id":[{"id":"https:\/\/ror.org\/00k27zs44","id-type":"ROR","asserted-by":"crossref"}]},{"name":"Royal Academy of Engineering Research Chair","id":[{"id":"https:\/\/ror.org\/0526snb40","id-type":"ROR","asserted-by":"crossref"}]},{"name":"InnoHK Hong Kong Centre for Centre for Cerebro-cardiovascular Engineering"},{"DOI":"10.13039\/100030827","name":"NHS","doi-asserted-by":"publisher","id":[{"id":"10.13039\/100030827","id-type":"DOI","asserted-by":"publisher"}]},{"name":"RAEng Engineering for Development Research","award":["310790\/2021-2 and 465518\/2014-1"],"award-info":[{"award-number":["310790\/2021-2 and 465518\/2014-1"]}]},{"name":"RAEng Engineering for Development Research","award":["PPM-00428-17 and RED-00081-16"],"award-info":[{"award-number":["PPM-00428-17 and RED-00081-16"]}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2024,5,20]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:sec>\n                    <jats:title>Aims<\/jats:title>\n                    <jats:p>Electrocardiogram (ECG) is widely considered the primary test for evaluating cardiovascular diseases. However, the use of artificial intelligence (AI) to advance these medical practices and learn new clinical insights from ECGs remains largely unexplored. We hypothesize that AI models with a specific design can provide fine-grained interpretation of ECGs to advance cardiovascular diagnosis, stratify mortality risks, and identify new clinically useful information.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Methods and results<\/jats:title>\n                    <jats:p>Utilizing a data set of 2 322 513 ECGs collected from 1 558 772 patients with 7 years follow-up, we developed a deep-learning model with state-of-the-art granularity for the interpretable diagnosis of cardiac abnormalities, gender identification, and hypertension screening solely from ECGs, which are then used to stratify the risk of mortality. The model achieved the area under the receiver operating characteristic curve (AUC) scores of 0.998 (95% confidence interval (CI), 0.995\u20130.999), 0.964 (95% CI, 0.963\u20130.965), and 0.839 (95% CI, 0.837\u20130.841) for the three diagnostic tasks separately. Using ECG-predicted results, we find high risks of mortality for subjects with sinus tachycardia (adjusted hazard ratio (HR) of 2.24, 1.96\u20132.57), and atrial fibrillation (adjusted HR of 2.22, 1.99\u20132.48). We further use salient morphologies produced by the deep-learning model to identify key ECG leads that achieved similar performance for the three diagnoses, and we find that the V1 ECG lead is important for hypertension screening and mortality risk stratification of hypertensive cohorts, with an AUC of 0.816 (0.814\u20130.818) and a univariate HR of 1.70 (1.61\u20131.79) for the two tasks separately.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Conclusion<\/jats:title>\n                    <jats:p>Using ECGs alone, our developed model showed cardiologist-level accuracy in interpretable cardiac diagnosis and the advancement in mortality risk stratification. In addition, it demonstrated the potential to facilitate clinical knowledge discovery for gender and hypertension detection which are not readily available.<\/jats:p>\n                  <\/jats:sec>","DOI":"10.1093\/ehjdh\/ztae014","type":"journal-article","created":{"date-parts":[[2024,2,16]],"date-time":"2024-02-16T17:27:37Z","timestamp":1708104457000},"page":"247-259","source":"Crossref","is-referenced-by-count":33,"title":["Decoding 2.3 million ECGs: interpretable deep learning for advancing cardiovascular diagnosis and mortality risk stratification"],"prefix":"10.1093","volume":"5","author":[{"ORCID":"https:\/\/orcid.org\/0000-0001-9139-8955","authenticated-orcid":false,"given":"Lei","family":"Lu","sequence":"first","affiliation":[{"name":"Institute of Biomedical Engineering, Department of Engineering Science, University of Oxford , Oxford, OX3 7DQ , UK"},{"name":"School of Life Course and Population Sciences, King\u2019s College London , London, SE1 1UL , UK"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Tingting","family":"Zhu","sequence":"additional","affiliation":[{"name":"Institute of Biomedical Engineering, Department of Engineering Science, University of Oxford , Oxford, OX3 7DQ , UK"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Antonio H","family":"Ribeiro","sequence":"additional","affiliation":[{"name":"Department of Information Technology, Uppsala University , Uppsala , Sweden"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Lei","family":"Clifton","sequence":"additional","affiliation":[{"name":"Nuffield Department of Population Health, University of Oxford Big Data Institute , Oxford, OX3 7LF , UK"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Erying","family":"Zhao","sequence":"additional","affiliation":[{"name":"Psychological Science and Health Management Center, Harbin Medical University , Harbin, 150076 , China"},{"name":"Department of Psychiatry, University of Oxford, Oxford, OX3 7JX, UK"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jiandong","family":"Zhou","sequence":"additional","affiliation":[{"name":"Department of Family Medicine and Primary Care, Li Ka Shing Faculty of Medicine, The University of Hong Kong , Hong Kong SAR , China"},{"name":"Division of Health Science, Warwick Medical School, University of Warwick , Coventry, CV4 7AL, UK"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Antonio Luiz P","family":"Ribeiro","sequence":"additional","affiliation":[{"name":"Department of Internal Medicine, Faculdade de Medicina, and Telehealth Center and Cardiology Service, Hospital das Cl\u00ednicas, Universidade Federal de Minas Gerais , Belo Horizonte, Brazil"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yuan-Ting","family":"Zhang","sequence":"additional","affiliation":[{"name":"Department of Electronic Engineering, Chinese University of Hong Kong , Hong Kong SAR , China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"David A","family":"Clifton","sequence":"additional","affiliation":[{"name":"Institute of Biomedical Engineering, Department of Engineering Science, University of Oxford , Oxford, OX3 7DQ , UK"},{"name":"Oxford Suzhou Centre for Advanced Research , Suzhou, 215123 , China"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"286","published-online":{"date-parts":[[2024,2,19]]},"reference":[{"key":"2024052015221920600_ztae014-B1","doi-asserted-by":"crossref","first-page":"e348","DOI":"10.1016\/S2589-7500(20)30107-2","article-title":"Automatic multilabel electrocardiogram diagnosis of heart rhythm or conduction abnormalities with deep learning: a cohort study","volume":"2","author":"Zhu","year":"2020","journal-title":"Lancet Digit Health"},{"key":"2024052015221920600_ztae014-B2","doi-asserted-by":"crossref","first-page":"e11","DOI":"10.1056\/NEJMvcm1400705","article-title":"Electrocardiographic monitoring in adults","volume":"372","author":"Ortega","year":"2015","journal-title":"N Engl J Med"},{"key":"2024052015221920600_ztae014-B3","doi-asserted-by":"crossref","first-page":"4163","DOI":"10.1109\/JBHI.2021.3100425","article-title":"Interactive effects of hrv and p-qrs-t on the power density spectra of ecg signals","volume":"25","author":"Xiang","year":"2021","journal-title":"IEEE J Biomed Health Inform"},{"key":"2024052015221920600_ztae014-B4","doi-asserted-by":"crossref","first-page":"20170821","DOI":"10.1098\/rsif.2017.0821","article-title":"Computational techniques for ECG analysis and interpretation in light of their contribution to medical advances","volume":"15","author":"Lyon","year":"2018","journal-title":"J R Soc Interface"},{"key":"2024052015221920600_ztae014-B5","doi-asserted-by":"crossref","first-page":"226","DOI":"10.1038\/s41746-023-00952-2","article-title":"A medical multimodal large language model for future pandemics","volume":"6","author":"Liu","year":"2023","journal-title":"NPJ Digit Med"},{"key":"2024052015221920600_ztae014-B6","first-page":"1","article-title":"Deep reinforcement learning for multi-class imbalanced training: applications in healthcare","author":"Yang","year":"2023","journal-title":"Mach Learn"},{"key":"2024052015221920600_ztae014-B7","doi-asserted-by":"crossref","first-page":"327","DOI":"10.1161\/CIRCULATIONAHA.122.063651","article-title":"Prediction of shock-refractory ventricular fibrillation during resuscitation of out-of-hospital cardiac arrest","volume":"148","author":"Coult","year":"2023","journal-title":"Circulation"},{"key":"2024052015221920600_ztae014-B8","doi-asserted-by":"crossref","first-page":"70","DOI":"10.1038\/s41591-018-0240-2","article-title":"Screening for cardiac contractile dysfunction using an artificial intelligence\u2013enabled electrocardiogram","volume":"25","author":"Attia","year":"2019","journal-title":"Nat Med"},{"key":"2024052015221920600_ztae014-B9","first-page":"1","article-title":"Semi- supervised learning for multi-label cardiovascular diseases prediction: a multi-dataset study","author":"Zhou","year":"2023","journal-title":"IEEE Trans Pattern Anal Mach Intell"},{"key":"2024052015221920600_ztae014-B10","doi-asserted-by":"crossref","first-page":"532","DOI":"10.1093\/ehjacc\/zuaa021","article-title":"Mortality risk stratification using artificial intelligence-augmented electrocardiogram in cardiac intensive care unit patients","volume":"10","author":"Jentzer","year":"2021","journal-title":"Eur Heart J Acute Cardiovasc Care"},{"key":"2024052015221920600_ztae014-B11","doi-asserted-by":"crossref","first-page":"14","DOI":"10.1038\/s44259-023-00015-2","article-title":"Interpretable machine learning-based decision support for prediction of antibiotic resistance for complicated urinary tract infections","volume":"1","author":"Yang","year":"2023","journal-title":"NPJ Antimicro Resist"},{"key":"2024052015221920600_ztae014-B12","first-page":"4211","article-title":"How can I explain this to you? An empirical study of deep neural network explanation methods","volume":"33","author":"Jeyakumar","year":"2020","journal-title":"Adv Neural Inf Process Sys"},{"key":"2024052015221920600_ztae014-B13","doi-asserted-by":"crossref","first-page":"886","DOI":"10.1038\/s41591-020-0870-z","article-title":"Prediction of mortality from 12-lead electrocardiogram voltage data using a deep neural network","volume":"26","author":"Raghunath","year":"2020","journal-title":"Nat Med"},{"key":"2024052015221920600_ztae014-B14","doi-asserted-by":"crossref","first-page":"e009056","DOI":"10.1161\/CIRCEP.120.009056","article-title":"Discovering and visualizing disease-specific electrocardiogram features using deep learning: proof-of-concept in phospholamban gene mutation carriers","volume":"14","author":"van de Leur","year":"2021","journal-title":"Circ Arrhythm Electrophysiol"},{"key":"2024052015221920600_ztae014-B15","first-page":"1","article-title":"Will two do? Varying dimensions in electrocardiography: the physioNet\/computing in cardiology challenge 2021","volume":"48","author":"Reyna","year":"2021","journal-title":"Comput Cardiol"},{"key":"2024052015221920600_ztae014-B16","author":"Desai"},{"key":"2024052015221920600_ztae014-B17","doi-asserted-by":"crossref","first-page":"217","DOI":"10.1177\/1747493019897870","article-title":"Global epidemiology of atrial fibrillation: an increasing epidemic and public health challenge","volume":"16","author":"Lippi","year":"2021","journal-title":"Int J Stroke"},{"key":"2024052015221920600_ztae014-B18","doi-asserted-by":"crossref","first-page":"e015479","DOI":"10.1161\/JAHA.120.015479","article-title":"Matters of the heart and mind: interpersonal violence and cardiovascular disease in women","volume":"9","author":"El-Serag","year":"2020","journal-title":"J Am Heart Assoc"},{"key":"2024052015221920600_ztae014-B19","doi-asserted-by":"crossref","first-page":"223","DOI":"10.1038\/s41581-019-0244-2","article-title":"The global epidemiology of hypertension","volume":"16","author":"Mills","year":"2020","journal-title":"Nat Rev Nephrol"},{"key":"2024052015221920600_ztae014-B20","doi-asserted-by":"crossref","first-page":"S75","DOI":"10.1016\/j.jelectrocard.2019.09.008","article-title":"Tele-electrocardiography and bigdata: the CODE (clinical outcomes in digital electrocardiography) study","volume":"57","author":"Ribeiro","year":"2019","journal-title":"J Electrocardiol"},{"key":"2024052015221920600_ztae014-B21","doi-asserted-by":"crossref","first-page":"1760","DOI":"10.1038\/s41467-020-15432-4","article-title":"Automatic diagnosis of the 12-lead ECG using a deep neural network","volume":"11","author":"Ribeiro","year":"2020","journal-title":"Nat Commun"},{"key":"2024052015221920600_ztae014-B22","doi-asserted-by":"crossref","first-page":"1206","DOI":"10.1016\/S0140-6736(22)01637-3","article-title":"Artificial intelligence-guided screening for atrial fibrillation using electrocardiogram during sinus rhythm: a prospective non-randomised interventional trial","volume":"400","author":"Noseworthy","year":"2022","journal-title":"Lancet"},{"key":"2024052015221920600_ztae014-B23","doi-asserted-by":"crossref","first-page":"599","DOI":"10.1161\/01.CIR.33.4.599","article-title":"Relationship of fibrillatory waves and P waves in the electrocardio- gram","volume":"33","author":"Peter","year":"1966","journal-title":"Circulation"},{"key":"2024052015221920600_ztae014-B24","first-page":"1","article-title":"AF Classification from a short single lead ECG recording: the PhysioNet\/computing in cardiology challenge 2017","author":"Clifford","year":"2017","journal-title":"Comput Cardiol"},{"key":"2024052015221920600_ztae014-B25","doi-asserted-by":"crossref","first-page":"36","DOI":"10.1186\/s12911-022-01775-z","article-title":"Diabetes mellitus risk prediction in the presence of class imbalance using flexible machine learning methods","volume":"22","author":"Sadeghi","year":"2022","journal-title":"BMC Med Inform Decis Mak"},{"key":"2024052015221920600_ztae014-B26","doi-asserted-by":"crossref","first-page":"e007284","DOI":"10.1161\/CIRCEP.119.007284","article-title":"Age and sex estimation using artificial intelligence from standard 12-lead ECGs","volume":"12","author":"Attia","year":"2019","journal-title":"Circ Arrhythm Electrophysiol"},{"key":"2024052015221920600_ztae014-B27","doi-asserted-by":"crossref","first-page":"2955","DOI":"10.1093\/eurheartj\/ehs230","article-title":"Coronary artery calcium for the prediction of mortality in young adults &lt; 45 years old and elderly adults &gt; 75 years old","volume":"33","author":"Tota-Maharaj","year":"2012","journal-title":"Eur Heart J"},{"key":"2024052015221920600_ztae014-B28","doi-asserted-by":"crossref","first-page":"2398","DOI":"10.1109\/JBHI.2021.3060997","article-title":"A comprehensive explanation framework for biomedical time series classification","volume":"25","author":"Ivaturi","year":"2021","journal-title":"IEEE J Biomed Health Inform"},{"key":"2024052015221920600_ztae014-B29","doi-asserted-by":"crossref","first-page":"10949","DOI":"10.1038\/s41598-021-90285-5","article-title":"Explaining deep neural networks for knowledge discovery in electrocardiogram analysis","volume":"11","author":"Hicks","year":"2021","journal-title":"Sci Reps"},{"key":"2024052015221920600_ztae014-B30","first-page":"618","article-title":"Grad- CAM: visual explanations from deep networks via gradient-based localization","author":"Selvaraju","year":"2017","journal-title":"Proc IEEE Int Conf Comput Vis"},{"key":"2024052015221920600_ztae014-B31","doi-asserted-by":"crossref","first-page":"1520","DOI":"10.1016\/S0140-6736(12)61413-5","article-title":"Ventricular arrhythmias and sudden cardiac death","volume":"380","author":"John","year":"2012","journal-title":"Lancet"},{"key":"2024052015221920600_ztae014-B32","volume-title":"The ECG in Practice","author":"Hampton","year":"2003"},{"key":"2024052015221920600_ztae014-B33","volume-title":"Chou\u2019s Electrocardiography in Clinical Practice: Adult and Pediatric","author":"Surawicz","year":"2008"},{"key":"2024052015221920600_ztae014-B34","doi-asserted-by":"crossref","first-page":"309","DOI":"10.1161\/HYPERTENSIONAHA.119.13708","article-title":"Blood pressure variability and incidence of new-onset atrial fibrillation: a nationwide population-based study","volume":"75","author":"Lee","year":"2020","journal-title":"Hypertension"},{"key":"2024052015221920600_ztae014-B35","doi-asserted-by":"crossref","DOI":"10.1177\/2045894018776496","article-title":"Change in R wave in lead V1 predicts survival of patients with pulmonary arterial hypertension","volume":"8","author":"Sato","year":"2018","journal-title":"Pulm Circ"},{"key":"2024052015221920600_ztae014-B36","doi-asserted-by":"crossref","first-page":"113","DOI":"10.1080\/08037051.2019.1673149","article-title":"Blood pressure-related electrocardiographic findings in healthy young individuals","volume":"29","author":"Hassing","year":"2020","journal-title":"Blood Press"},{"key":"2024052015221920600_ztae014-B37","doi-asserted-by":"crossref","first-page":"199","DOI":"10.1016\/j.ijcard.2020.12.016","article-title":"Electrocardiographic T-wave morphology and risk of mortality","volume":"328","author":"Isaksen","year":"2021","journal-title":"Int J Cardiol"},{"key":"2024052015221920600_ztae014-B38","article-title":"Mortality risk prediction of the electrocardiogram as an informative indicator of cardiovascular diseases","volume":"9","author":"Tsai","year":"2023","journal-title":"Digit Health"},{"key":"2024052015221920600_ztae014-B39","doi-asserted-by":"crossref","first-page":"188","DOI":"10.1038\/s41746-022-00720-8","article-title":"Confounders mediate AI prediction of demographics in medical imaging","volume":"5","author":"Duffy","year":"2022","journal-title":"NPJ Digit Med"}],"container-title":["European Heart Journal - Digital Health"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/academic.oup.com\/ehjdh\/advance-article-pdf\/doi\/10.1093\/ehjdh\/ztae014\/57018032\/ztae014.pdf","content-type":"application\/pdf","content-version":"am","intended-application":"syndication"},{"URL":"https:\/\/academic.oup.com\/ehjdh\/article-pdf\/5\/3\/247\/57765023\/ztae014.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/academic.oup.com\/ehjdh\/article-pdf\/5\/3\/247\/57765023\/ztae014.pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2024,5,20]],"date-time":"2024-05-20T11:23:30Z","timestamp":1716204210000},"score":1,"resource":{"primary":{"URL":"https:\/\/academic.oup.com\/ehjdh\/article\/5\/3\/247\/7610307"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2024,2,19]]},"references-count":39,"journal-issue":{"issue":"3","published-online":{"date-parts":[[2024,2,19]]},"published-print":{"date-parts":[[2024,5,20]]}},"URL":"https:\/\/doi.org\/10.1093\/ehjdh\/ztae014","relation":{"has-preprint":[{"id-type":"doi","id":"10.1101\/2022.11.01.22281722","asserted-by":"object"}]},"ISSN":["2634-3916"],"issn-type":[{"value":"2634-3916","type":"electronic"}],"subject":[],"published-other":{"date-parts":[[2024,5]]},"published":{"date-parts":[[2024,2,19]]}}}