Identification of hypertrophic cardiomyopathy on electrocardiographic images with deep learning

Veer Sangha^#^{1

2}, Lovedeep Singh Dhingra^#¹, Arya Aminorroaya¹, Philip M Croon³, Nikhil V Sikand¹, Sounok Sen¹, Matthew W Martinez^{4

5}, Martin S Maron⁶, Harlan M Krumholz^{1

7}, Folkert W Asselbergs^{3

8

9}, Evangelos K Oikonomou¹, Rohan Khera^{10

11

12}

Affiliations

¹ Section of Cardiovascular Medicine, Department of Internal Medicine, Yale School of Medicine, New Haven, CT, USA.
² Department of Engineering Science, Oxford University, Oxford, UK.
³ Department of Cardiology, Amsterdam Cardiovascular Sciences, Amsterdam University Medical Centre, University of Amsterdam, Amsterdam, The Netherlands.
⁴ Department of Cardiovascular Medicine, Atlantic Health, Morristown Medical Center, Morristown, NJ, USA.
⁵ Sports Cardiology and Hypertrophic Cardiomyopathy, Morristown Medical Center, Morristown, NJ, USA.
⁶ HCM Institute, Tufts Medical Center, Boston, MA, USA.
⁷ Center for Outcomes Research and Evaluation (CORE), Yale New Haven Hospital, New Haven, CT, USA.
⁸ Institute of Health Informatics, University College London, London, UK.
⁹ National Institute for Health Research, University College London Hospitals, Biomedical Research Centre, University College London, London, UK.
¹⁰ Section of Cardiovascular Medicine, Department of Internal Medicine, Yale School of Medicine, New Haven, CT, USA. rohan.khera@yale.edu.
¹¹ Center for Outcomes Research and Evaluation (CORE), Yale New Haven Hospital, New Haven, CT, USA. rohan.khera@yale.edu.
¹² Section of Health Informatics, Department of Biostatistics, Yale School of Public Health, New Haven, CT, USA. rohan.khera@yale.edu.

^# Contributed equally.

PMID: 40696040
DOI: 10.1038/s44161-025-00685-3

Identification of hypertrophic cardiomyopathy on electrocardiographic images with deep learning

Veer Sangha et al. Nat Cardiovasc Res. 2025 Aug.

. 2025 Aug;4(8):991-1000.

doi: 10.1038/s44161-025-00685-3. Epub 2025 Jul 22.

Authors

Affiliations

¹ Section of Cardiovascular Medicine, Department of Internal Medicine, Yale School of Medicine, New Haven, CT, USA.
² Department of Engineering Science, Oxford University, Oxford, UK.
³ Department of Cardiology, Amsterdam Cardiovascular Sciences, Amsterdam University Medical Centre, University of Amsterdam, Amsterdam, The Netherlands.
⁴ Department of Cardiovascular Medicine, Atlantic Health, Morristown Medical Center, Morristown, NJ, USA.
⁵ Sports Cardiology and Hypertrophic Cardiomyopathy, Morristown Medical Center, Morristown, NJ, USA.
⁶ HCM Institute, Tufts Medical Center, Boston, MA, USA.
⁷ Center for Outcomes Research and Evaluation (CORE), Yale New Haven Hospital, New Haven, CT, USA.
⁸ Institute of Health Informatics, University College London, London, UK.
⁹ National Institute for Health Research, University College London Hospitals, Biomedical Research Centre, University College London, London, UK.
¹⁰ Section of Cardiovascular Medicine, Department of Internal Medicine, Yale School of Medicine, New Haven, CT, USA. rohan.khera@yale.edu.
¹¹ Center for Outcomes Research and Evaluation (CORE), Yale New Haven Hospital, New Haven, CT, USA. rohan.khera@yale.edu.
¹² Section of Health Informatics, Department of Biostatistics, Yale School of Public Health, New Haven, CT, USA. rohan.khera@yale.edu.

^# Contributed equally.

PMID: 40696040
DOI: 10.1038/s44161-025-00685-3

Abstract

Hypertrophic cardiomyopathy (HCM) is frequently underdiagnosed. Although deep learning (DL) models using raw electrocardiographic (ECG) voltage data can enhance detection, their use at the point of care is limited. Here we report the development and validation of a DL model that detects HCM from images of 12-lead ECGs across layouts. The model was developed using 124,553 ECGs from 66,987 individuals at the Yale New Haven Hospital (YNHH), with HCM features determined by concurrent imaging (cardiac magnetic resonance (CMR) or echocardiography). External validation included ECG images from MIMIC-IV, the Amsterdam University Medical Center (AUMC) and the UK Biobank (UKB), where HCM was defined by CMR (YNHH, MIMIC-IV and AUMC) and diagnosis codes (UKB). The model demonstrated robust performance across image formats and sites (areas under the receiver operating characteristic curve (AUROCs): 0.95 internal testing; 0.94 MIMIC-IV; 0.92 AUMC; 0.91 UKB). Discriminative features localized to anterior/lateral leads (V4 and V5) regardless of layout. This approach enables scalable, image-based screening for HCM across clinical settings.

PubMed Disclaimer

Conflict of interest statement

Competing interests: V.S. and R.K. are co-inventors of US pending patent applications WO2023230345A1, ‘Articles and methods for format-independent detection of hidden cardiovascular disease from printed electrocardiographic images using deep learning’, and 63/484,426, ‘Biometric contrastive learning for data-efficient deep learning from electrocardiographic images’, filed by Yale University. These patents cover the methods of training an AI model to detect structural heart disease in a format-independent manner using varied image formats as inputs and the biocontrastive pretraining used in the present study. They are also co-founders of Ensight-AI, along with H.M.K. P.M.C. is the founder and CEO of DGTL Health BV. R.K. is an Associate Editor of JAMA. In addition to awards from the National Heart, Lung, and Blood Institute and the Doris Duke Charitable Foundation, he receives research support, through Yale University, from Bristol Myers Squibb, Novo Nordisk and BridgeBio. R.K. and E.K.O. are co-inventors of US pending patent applications 63/562,335, ‘Artificial intelligence-guided screening of under-recognized cardiomyopathies adapted for point-of-care cardiac ultrasound’; 18/813,882, ‘Multimodality artificial intelligence systems to track the progression of pre-clinical amyloid cardiomyopathy’; 63/619,241, ‘A multi-modal video-based progression score for aortic stenosis using artificial intelligence’; US20220336048A1, ‘Methods for neighborhood phenomapping for clinical trials’; 63/508,315, ‘Machine learning method for adaptive trial enrichment’; and 63/606,203, ‘Methods of generating digital twin-based datasets’, filed by Yale University and unrelated to this work. R.K. and E.K.O. are co-founders of Evidence2Health, a precision health platform to improve evidence-based cardiovascular care. E.K.O. has been a consultant for Caristo Diagnostics, Ltd. and Ensight-AI, Inc. and has received royalty fees from technology licensed through the University of Oxford. H.M.K. works under contract with the Centers for Medicare & Medicaid Services to support quality measurement programs; was a recipient of a research grant from Johnson & Johnson, through Yale University, to support clinical trial data sharing; was a recipient of a research agreement, through Yale University, from the Shenzhen Center for Health Information for work to advance intelligent disease prevention and health promotion; collaborates with the National Center for Cardiovascular Diseases in Beijing; receives payment from the Arnold & Porter Law Firm for work related to the Sanofi clopidogrel litigation, from the Martin Baughman Law Firm for work related to the Cook Celect IVC filter litigation and from the Siegfried and Jensen Law Firm for work related to Vioxx litigation; chairs a Cardiac Scientific Advisory Board for UnitedHealth; was a member of the IBM Watson Health Life Sciences Board; is a member of the Advisory Board for Element Science, the Advisory Board for Facebook and the Physician Advisory Board for Aetna; and is the co-founder of Hugo Health, a personal health information platform, and the co-founder of Refactor Health, a healthcare AI-augmented data management company. The remaining authors declare no competing interests.

Update of

Identification of Hypertrophic Cardiomyopathy on Electrocardiographic Images with Deep Learning.
Sangha V, Dhingra LS, Aminorroaya A, Croon PM, Sikand NV, Sen S, Martinez MW, Maron MS, Krumholz HM, Asselbergs FW, Oikonomou EK, Khera R. Sangha V, et al. medRxiv [Preprint]. 2025 May 28:2023.12.23.23300490. doi: 10.1101/2023.12.23.23300490. medRxiv. 2025. Update in: Nat Cardiovasc Res. 2025 Aug;4(8):991-1000. doi: 10.1038/s44161-025-00685-3. PMID: 38234746 Free PMC article. Updated. Preprint.

References

1. Maron, B. J. Hypertrophic cardiomyopathy. JAMA 287, 1308–1320 (2002). - PubMed
1. Maron, B. J. & Maron, M. S. Hypertrophic cardiomyopathy. Lancet 381, 242–255 (2013). - PubMed
1. Semsarian, C., Ingles, J., Maron, M. S. & Maron, B. J. New perspectives on the prevalence of hypertrophic cardiomyopathy. J. Am. Coll. Cardiol. 65, 1249–1254 (2015). - PubMed
1. Sabater-Molina, M., Pérez-Sánchez, I., Hernández del Rincón, J. P. & Gimeno, J. R. Genetics of hypertrophic cardiomyopathy: a review of current state. Clin. Genet. 93, 3–14 (2018). - PubMed
1. Geske, J. B., Ommen, S. R. & Gersh, B. J. Hypertrophic cardiomyopathy. JACC Heart Fail. 6, 364–375 (2018). - PubMed

Publication types

Actions

MeSH terms

Actions
Actions
Actions
Actions
Actions
Actions
Actions
Actions
Actions
Actions
Actions
Actions
Actions
Actions
Actions

Grants and funding

LinkOut - more resources

Full Text Sources
- Nature Publishing Group

Save citation to file

Email citation

Add to Collections

Add to My Bibliography

Your saved search

Create a file for external citation management software

Your RSS Feed

Identification of hypertrophic cardiomyopathy on electrocardiographic images with deep learning

Affiliations

Identification of hypertrophic cardiomyopathy on electrocardiographic images with deep learning

Authors

Affiliations

Abstract

Conflict of interest statement

Update of

References

Publication types

MeSH terms

Grants and funding

LinkOut - more resources

Full Text Sources