Fine-tuning an ECG Foundation Model to Predict Coronary CT Angiography Outcomes

Fine-tuning an ECG Foundation Model to Predict Coronary CT Angiography Outcomes | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Fine-tuning an ECG Foundation Model to Predict Coronary CT Angiography Outcomes Yujie Xiao, Gongzheng Tang, Deyun Zhang, Jun Li, Guangkun Nie, and 6 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8234977/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 12 You are reading this latest preprint version Abstract Coronary artery disease (CAD) remains a major global health burden. Accurate identification of the culprit vessel and assessment of stenosis severity are essential for guiding individualized therapy. Although coronary CT angiography (CCTA) is the first-line non-invasive modality for CAD diagnosis, its dependence on high-end equipment, radiation exposure, and strict patient cooperation limits large-scale use. With advances in artificial intelligence (AI) and the widespread availability of electrocardiography (ECG), AI-ECG offers a promising alternative for CAD screening. In this study, we developed an interpretable AI-ECG model to predict severe or complete stenosis of the four major coronary arteries on CCTA. On the internal validation set, the model's AUCs for the right coronary artery (RCA), left main coronary artery (LM), left anterior descending artery (LAD), and left circumflex artery (LCX) were 0.794, 0.818, 0.744, and 0.755, respectively; on the external validation set, the AUCs reached 0.749, 0.971, 0.667, and 0.727, respectively. Performance remained stable in a clinically normal-ECG subset, indicating robustness beyond overt ECG abnormalities. Subgroup analyses across demographic and acquisition-time strata further confirmed model stability. Risk stratification based on vessel-specific incidence thresholds showed consistent separation on calibration and cumulative event curves. Interpretability analyses revealed distinct waveform differences between high- and low-risk groups, highlighting key electrophysiological regions contributing to model decisions and offering new insights into the ECG correlates of coronary stenosis. Health sciences/Cardiology Biological sciences/Computational biology and bioinformatics Health sciences/Diseases Electrocardiogram Coronary CT Angiography Artificial Intelligence Deep Learning Coronary Heart Disease Full Text Additional Declarations No competing interests reported. Cite Share Download PDF Status: Under Review Version 1 posted Editorial decision: Revision requested 31 Dec, 2025 Reviews received at journal 16 Dec, 2025 Reviews received at journal 16 Dec, 2025 Reviewers agreed at journal 12 Dec, 2025 Reviews received at journal 05 Dec, 2025 Reviewers agreed at journal 05 Dec, 2025 Reviewers agreed at journal 05 Dec, 2025 Reviewers agreed at journal 05 Dec, 2025 Reviewers invited by journal 05 Dec, 2025 Editor assigned by journal 03 Dec, 2025 Submission checks completed at journal 03 Dec, 2025 First submitted to journal 29 Nov, 2025 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-8234977","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":556071185,"identity":"9a038c92-8242-4519-8f52-ff7a86db0568","order_by":0,"name":"Yujie Xiao","email":"","orcid":"","institution":"Peking University","correspondingAuthor":false,"prefix":"","firstName":"Yujie","middleName":"","lastName":"Xiao","suffix":""},{"id":556071186,"identity":"df9a0f10-af7b-4308-b3c3-57a005f9f067","order_by":1,"name":"Gongzheng Tang","email":"","orcid":"","institution":"Peking University","correspondingAuthor":false,"prefix":"","firstName":"Gongzheng","middleName":"","lastName":"Tang","suffix":""},{"id":556071187,"identity":"dd5376a3-81fd-47cf-b277-2d34433bdca2","order_by":2,"name":"Deyun Zhang","email":"","orcid":"","institution":"Heart Voice Medical Technology","correspondingAuthor":false,"prefix":"","firstName":"Deyun","middleName":"","lastName":"Zhang","suffix":""},{"id":556071188,"identity":"6781c5ff-5278-4337-83ac-b3e20255654c","order_by":3,"name":"Jun Li","email":"","orcid":"","institution":"Peking University","correspondingAuthor":false,"prefix":"","firstName":"Jun","middleName":"","lastName":"Li","suffix":""},{"id":556071189,"identity":"1b25a256-e012-4e0a-92b8-34ef694efb11","order_by":4,"name":"Guangkun Nie","email":"","orcid":"","institution":"Peking University","correspondingAuthor":false,"prefix":"","firstName":"Guangkun","middleName":"","lastName":"Nie","suffix":""},{"id":556071190,"identity":"583616f1-fd0c-4b20-b846-fae2cbd58073","order_by":5,"name":"Haoyu Wang","email":"","orcid":"","institution":"University of Chinese Academy of Sciences","correspondingAuthor":false,"prefix":"","firstName":"Haoyu","middleName":"","lastName":"Wang","suffix":""},{"id":556071191,"identity":"acca4549-7034-470e-a7d6-c04573a3b013","order_by":6,"name":"Shun Huang","email":"","orcid":"","institution":"Peking University","correspondingAuthor":false,"prefix":"","firstName":"Shun","middleName":"","lastName":"Huang","suffix":""},{"id":556071192,"identity":"265be780-678f-4aee-809f-fbb0cdbbf334","order_by":7,"name":"Tong Liu","email":"","orcid":"","institution":"Tianjin Key Laboratory of Ionic-Molecular Function of Cardiovascular Disease, Department of Cardiology, Tianjin Institute of Cardiology, The Second Hospital of Tianjin Medical University","correspondingAuthor":false,"prefix":"","firstName":"Tong","middleName":"","lastName":"Liu","suffix":""},{"id":556071193,"identity":"3e66a94b-8f92-4cdf-a4d5-4edfda7806ec","order_by":8,"name":"Qinghao Zhao","email":"","orcid":"","institution":"Peking University People's Hospital","correspondingAuthor":false,"prefix":"","firstName":"Qinghao","middleName":"","lastName":"Zhao","suffix":""},{"id":556071194,"identity":"a61d2e95-3218-4b40-9317-c66baeab1b7f","order_by":9,"name":"Kangyin Chen","email":"","orcid":"","institution":"Tianjin Key Laboratory of Ionic-Molecular Function of Cardiovascular Disease, Department of Cardiology, Tianjin Institute of Cardiology, The Second Hospital of Tianjin Medical University","correspondingAuthor":false,"prefix":"","firstName":"Kangyin","middleName":"","lastName":"Chen","suffix":""},{"id":556071195,"identity":"0035195b-c85e-4944-81fb-a95fcbe30bf0","order_by":10,"name":"Shenda Hong","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA5klEQVRIiWNgGAWjYBACAxDB2AAkDjAfANMMB4jXwpZAshYeA+K0mLP3Hn75c4dNHt/xM98ezmxjkOO7kcD4uQCPFsuec2nWvGfSiiXP5G433NjGYCx5I4FZegY+h93IMTNmbDucuOFA7jbJh20MiRtuJLAx8xDQYviz7X/ihvNvnoG01BOjxfgBb9sBoOE5bJJAhyUYENRy5owZM29bcuLMG8/MDWeckzCceeZhszReLcd7jD/+bLNL7Duf/OxhT5mNPN/x5IOf8WkBAjYJGAOIQWxI9OADzB+QtIyCUTAKRsEowAQAl1FXreTma4kAAAAASUVORK5CYII=","orcid":"","institution":"Peking University","correspondingAuthor":true,"prefix":"","firstName":"Shenda","middleName":"","lastName":"Hong","suffix":""}],"badges":[],"createdAt":"2025-11-29 07:08:05","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8234977/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8234977/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":97661777,"identity":"99cdb126-fd1a-4397-acb5-88b57919e8d6","added_by":"auto","created_at":"2025-12-08 07:58:48","extension":"json","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":11884,"visible":true,"origin":"","legend":"","description":"","filename":"f51b7b90c0fa4708b687f4cc3b9d8cef.json","url":"https://assets-eu.researchsquare.com/files/rs-8234977/v1/f6df034a01a94d2c5490debe.json"},{"id":97674709,"identity":"55ce336a-aff0-4161-80b2-676451889154","added_by":"auto","created_at":"2025-12-08 09:43:55","extension":"png","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":136344,"visible":true,"origin":"","legend":"","description":"","filename":"Roca.png","url":"https://assets-eu.researchsquare.com/files/rs-8234977/v1/6278b25f4f1b2a281fef5cb3.png"},{"id":97661779,"identity":"a774df0c-f2ba-4e72-b29c-eba342580adb","added_by":"auto","created_at":"2025-12-08 07:58:49","extension":"png","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":119377,"visible":true,"origin":"","legend":"","description":"","filename":"Rocb.png","url":"https://assets-eu.researchsquare.com/files/rs-8234977/v1/ea8d93c42190c372edcc7b9c.png"},{"id":97661782,"identity":"080ee978-fddd-4ae2-8a12-bdf0c792448e","added_by":"auto","created_at":"2025-12-08 07:58:49","extension":"png","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":122989,"visible":true,"origin":"","legend":"","description":"","filename":"Rocc.png","url":"https://assets-eu.researchsquare.com/files/rs-8234977/v1/f3ad3d43450a65e721af8a2b.png"},{"id":97661780,"identity":"7b2ed1cc-0b0b-41bf-a9a8-9bac95f4dc0a","added_by":"auto","created_at":"2025-12-08 07:58:49","extension":"png","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":93857,"visible":true,"origin":"","legend":"","description":"","filename":"box.png","url":"https://assets-eu.researchsquare.com/files/rs-8234977/v1/aca69e1fcb649f6828d45e53.png"},{"id":97674838,"identity":"81b89e00-8aac-4fa3-a85c-1f224dbd3921","added_by":"auto","created_at":"2025-12-08 09:44:25","extension":"png","order_by":5,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":269554,"visible":true,"origin":"","legend":"","description":"","filename":"calibration.png","url":"https://assets-eu.researchsquare.com/files/rs-8234977/v1/24ae42d7625e530e3dbca714.png"},{"id":97675041,"identity":"e35b013b-7216-4b1f-82bb-fa9e2401c8b9","added_by":"auto","created_at":"2025-12-08 09:45:45","extension":"pdf","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":4297902,"visible":true,"origin":"","legend":"","description":"","filename":"fdkbykxxypkynvrsbkcrgzjdvfsjqdpt.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8234977/v1/4e25f4859b62c5b94d2df12c.pdf"},{"id":97674999,"identity":"6eef0b44-aec0-4712-86b0-311d0cd3dec0","added_by":"auto","created_at":"2025-12-08 09:45:15","extension":"png","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":126313,"visible":true,"origin":"","legend":"","description":"","filename":"kmcurve.png","url":"https://assets-eu.researchsquare.com/files/rs-8234977/v1/215e1586411adb00d3c9af97.png"},{"id":97661783,"identity":"624fd94a-94a4-496d-b976-20122b95cadd","added_by":"auto","created_at":"2025-12-08 07:58:49","extension":"bbl","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":22126,"visible":true,"origin":"","legend":"","description":"","filename":"main.bbl","url":"https://assets-eu.researchsquare.com/files/rs-8234977/v1/c276099dbe9ed4e9939e7610.bbl"},{"id":97661784,"identity":"a9b7cfeb-3adb-4666-8cdc-87cdcab52d5d","added_by":"auto","created_at":"2025-12-08 07:58:49","extension":"bst","order_by":9,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":30606,"visible":true,"origin":"","legend":"","description":"","filename":"numbered.bst","url":"https://assets-eu.researchsquare.com/files/rs-8234977/v1/9419e1bc58b06034c4bad17d.bst"},{"id":97661785,"identity":"02a44127-01d4-48de-8baf-1b6fafcd2e15","added_by":"auto","created_at":"2025-12-08 07:58:49","extension":"sty","order_by":10,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":6172,"visible":true,"origin":"","legend":"","description":"","filename":"numcompress.sty","url":"https://assets-eu.researchsquare.com/files/rs-8234977/v1/0179a00fa83dd2d632996f48.sty"},{"id":97661788,"identity":"ab44cc0a-d4f0-49de-8740-d723d4cb1f18","added_by":"auto","created_at":"2025-12-08 07:58:49","extension":"png","order_by":11,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":99120,"visible":true,"origin":"","legend":"","description":"","filename":"subgroup.png","url":"https://assets-eu.researchsquare.com/files/rs-8234977/v1/1c6fa05adc02349aef76def7.png"},{"id":97675010,"identity":"4cd37f0b-892a-49c0-a583-bd6cb655cc46","added_by":"auto","created_at":"2025-12-08 09:45:20","extension":"png","order_by":12,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":2435124,"visible":true,"origin":"","legend":"","description":"","filename":"subgroupforest.png","url":"https://assets-eu.researchsquare.com/files/rs-8234977/v1/871d3ea1e41178f72a9126a0.png"},{"id":97661791,"identity":"693c5e63-b129-4398-bb29-40ec8d73f637","added_by":"auto","created_at":"2025-12-08 07:58:49","extension":"png","order_by":13,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1035829,"visible":true,"origin":"","legend":"","description":"","filename":"visualize.png","url":"https://assets-eu.researchsquare.com/files/rs-8234977/v1/d676106bca2f6221f1a5ddf3.png"},{"id":97661789,"identity":"77825d72-ce2c-4eb6-9f4b-77495d1d4bbf","added_by":"auto","created_at":"2025-12-08 07:58:49","extension":"xml","order_by":14,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":66358,"visible":true,"origin":"","legend":"","description":"","filename":"f51b7b90c0fa4708b687f4cc3b9d8cef1structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-8234977/v1/133d45b7a5915aa3f8b0ebfc.xml"},{"id":97678903,"identity":"25126126-70d6-4888-9f5a-8ff3dc3a7619","added_by":"auto","created_at":"2025-12-08 09:56:37","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":4540962,"visible":true,"origin":"","legend":"","description":"","filename":"fdkbykxxypkynvrsbkcrgzjdvfsjqdpt.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8234977/v1_covered_4f2d3236-463c-49db-b5e9-2933a3ad15ac.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Fine-tuning an ECG Foundation Model to Predict Coronary CT Angiography Outcomes","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":true,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":true,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"npj-digital-medicine","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"npjdigitalmed","sideBox":"Learn more about [npj Digital Medicine](http://www.nature.com/npjdigitalmed/)","snPcode":"41746","submissionUrl":"https://submission.springernature.com/new-submission/41746/3","title":"npj Digital Medicine","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"NPJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Electrocardiogram, Coronary CT Angiography, Artificial Intelligence, Deep Learning, Coronary Heart Disease","lastPublishedDoi":"10.21203/rs.3.rs-8234977/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8234977/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"Coronary artery disease (CAD) remains a major global health burden. Accurate identification of the culprit vessel and assessment of stenosis severity are essential for guiding individualized therapy. Although coronary CT angiography (CCTA) is the first-line non-invasive modality for CAD diagnosis, its dependence on high-end equipment, radiation exposure, and strict patient cooperation limits large-scale use. With advances in artificial intelligence (AI) and the widespread availability of electrocardiography (ECG), AI-ECG offers a promising alternative for CAD screening. In this study, we developed an interpretable AI-ECG model to predict severe or complete stenosis of the four major coronary arteries on CCTA. On the internal validation set, the model's AUCs for the right coronary artery (RCA), left main coronary artery (LM), left anterior descending artery (LAD), and left circumflex artery (LCX) were 0.794, 0.818, 0.744, and 0.755, respectively; on the external validation set, the AUCs reached 0.749, 0.971, 0.667, and 0.727, respectively. Performance remained stable in a clinically normal-ECG subset, indicating robustness beyond overt ECG abnormalities. Subgroup analyses across demographic and acquisition-time strata further confirmed model stability. Risk stratification based on vessel-specific incidence thresholds showed consistent separation on calibration and cumulative event curves. Interpretability analyses revealed distinct waveform differences between high- and low-risk groups, highlighting key electrophysiological regions contributing to model decisions and offering new insights into the ECG correlates of coronary stenosis.","manuscriptTitle":"Fine-tuning an ECG Foundation Model to Predict Coronary CT Angiography Outcomes","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-12-08 07:58:44","doi":"10.21203/rs.3.rs-8234977/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2025-12-31T16:17:27+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-12-16T21:40:21+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-12-16T14:27:51+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"118326094863625122294230889794569648469","date":"2025-12-12T08:36:31+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-12-06T03:14:07+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"295830624552428702944649408914652826924","date":"2025-12-06T01:38:33+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"36014143736712074667810059427897607663","date":"2025-12-05T14:16:38+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"225766478194460863788876306899594417109","date":"2025-12-05T10:08:31+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-12-05T06:07:48+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-12-04T00:56:27+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-12-03T12:19:58+00:00","index":"","fulltext":""},{"type":"submitted","content":"npj Digital Medicine","date":"2025-11-29T06:53:45+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"npj-digital-medicine","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"npjdigitalmed","sideBox":"Learn more about [npj Digital Medicine](http://www.nature.com/npjdigitalmed/)","snPcode":"41746","submissionUrl":"https://submission.springernature.com/new-submission/41746/3","title":"npj Digital Medicine","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"NPJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"588422b8-6456-4fb5-8f71-e0d5f022ab4c","owner":[],"postedDate":"December 8th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[{"id":59187169,"name":"Health sciences/Cardiology"},{"id":59187170,"name":"Biological sciences/Computational biology and bioinformatics"},{"id":59187171,"name":"Health sciences/Diseases"}],"tags":[],"updatedAt":"2026-05-20T03:09:04+00:00","versionOfRecord":[],"versionCreatedAt":"2025-12-08 07:58:44","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8234977","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8234977","identity":"rs-8234977","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}