Background Artificial intelligence (AI) applied to routine electrocardiograms (ECGs) offers promise for screening of structural heart disease (SHD), yet broad clinical integration remains limited by high false positive rates and the lack of tailored deployment strategies.

Methods We developed TARGET-AI, a multimodal AI-enabled pipeline that integrates longitudinal electronic health record (EHR) data with ECG images to identify optimal intersections of healthcare encounters and patient phenotypes for targeted AI-ECG screening. The approach is built on (1) a foundation model pretrained on 118 million coded EHR events from 159,322 individuals to generate temporal patient embeddings and identify high-risk screening candidates, followed by (2) a contrastive vision-language model trained on 754,533 ECG-echocardiogram pairs to detect SHD with tunable performance characteristics. We evaluated this joint strategy in a temporally distinct cohort of 5,198 individuals referred for their first transthoracic echocardiogram (TTE) within 90 days of an ECG and externally in 33,518 participants from the UK Biobank undergoing ECG and cardiac magnetic resonance imaging.

Results Our pre-trained AI-ECG image foundation model discriminated 27 SHD subtypes, from left ventricular systolic dysfunction (AUROC of 0.90) to severe aortic stenosis (AUROC of 0.85) and elevated right ventricular systolic pressure (AUROC of 0.82). Compared with untargeted AI-ECG screening, EHR-informed TARGET-AI-guided screening significantly reduced false positive predictions across SHD labels (median reduction: 87.8%; interquartile range [IQR], 82.4%-98.2%) and improved F1 score (median increase: 0.25; IQR, 0.19-0.41). In the UK Biobank, targeted screening reduced false positives by 61.7% (IQR, 50.4%-89.1%) while preserving sensitivity.

Conclusions TARGET-AI enables the context-aware deployment of AI-ECG screening by leveraging key longitudinal EHR phenotypes and multimodal ECG-echocardiogram representations, thereby defining an interoperable, data-driven strategy for the more precise deployment of AI screening tools across health systems.

E.K.O. is an Associate Editor for European Heart Journal, a co-founder of Evidence2Health LLC, a co-inventor in patent applications 18/813,882, 17/720,068, 63/508,315, 63/580,137, 63/619,241, 63/562,335, and granted patents US12067714B2, US11948230B2, has been an ad hoc consultant for Caristo Diagnostics Ltd and Ensight-AI Inc, and has received royalty fees from technology licensed through the University of Oxford. R.K. is an Associate Editor of JAMA and receives research support, through Yale, from the Blavatnik Foundation, Bristol-Myers Squibb, Novo Nordisk, and BridgeBio. He is a coinventor of Pending Patent Applications WO2023230345A1, US20220336048A1, 63/346,610, 63/484,426, 63/508,315, 63/580,137, 63/606,203, 63/619,241, and 63/562,335, and a co-founder of Ensight-AI, Inc and Evidence2Health, LLC. The remaining authors have nothing to disclose.

The authors acknowledge support by the National Heart, Lung, And Blood Institute of the National Institutes of Health (under award numbers R01HL167858 and K23HL153775 to RK, and F32HL170592 to EKO), the National Institute on Aging of the National Institutes of Health (under award number R01AG089981 to RK), and the Doris Duke Charitable Foundation (under award number 2022060 to RK). The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

I confirm all relevant ethical guidelines have been followed, and any necessary IRB and/or ethics committee approvals have been obtained.

Yes

The details of the IRB/oversight body that provided approval or exemption for the research described are given below:

This retrospective study was approved by the Yale Institutional Review Board (IRB) with a waiver of informed consent. UK Biobank analyses were conducted under project #71033.

I confirm that all necessary patient/participant consent has been obtained and the appropriate institutional forms have been archived, and that any patient/participant/sample identifiers included were not known to anyone (e.g., hospital staff, patients or participants themselves) outside the research group so cannot be used to identify individuals.

Yes

I understand that all clinical trials and any other prospective interventional studies must be registered with an ICMJE-approved registry, such as ClinicalTrials.gov. I confirm that any such study reported in the manuscript has been registered and the trial registration ID is provided (note: if posting a prospective study registered retrospectively, please provide a statement in the trial ID field explaining why the study was not registered in advance).

Yes

I have followed all appropriate research reporting guidelines, such as any relevant EQUATOR Network research reporting checklist(s) and other pertinent material, if applicable.

Yes

Funding: The authors acknowledge support by the National Heart, Lung, And Blood Institute of the National Institutes of Health (under award numbers R01HL167858 and K23HL153775 to RK, and F32HL170592 to EKO), the National Institute on Aging of the National Institutes of Health (under award number R01AG089981 to RK), and the Doris Duke Charitable Foundation (under award number 2022060 to RK). The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

The sample code for training and inference, including a pipeline for embedding extraction from new ECG images, as well as reference embeddings for case and control centroids across 27 representative echocardiographic labels from the YNHHS training set, will be released through our Lab’s GitHub repository upon formal acceptance of the peer-reviewed manuscript. Model weights for the ECG image vision transformer will also be made available through our Lab’s HuggingFace repository.