CoChat
Sign Up
Cited by 130ECG Monitoring and AnalysisCardiomyopathy and Myosin StudiesArtificial Intelligence in Healthcare

Multinational Federated Learning Approach to Train ECG and Echocardiogram Models for Hypertrophic Cardiomyopathy Detection

Open access
Goto, Shinichi, Solanki, Divyarajsinhji, John, Jenine E., Yagi, Ryuichiro, Homilius, Max, Ichihara, Genki, Katsumata, Yoshinori, Gaggin, Hanna K. +3 more·2022·Circulation
DOI 10.1161/circulationaha.121.058696PMC PMC9439630
Open this paper in CoChat — free accountOpen PDF

Abstract

Background: Novel targeted treatments increase the need for prompt hypertrophic cardiomyopathy (HCM) detection. However, its low prevalence (0.5%) and resemblance to common diseases present challenges that may benefit from automated machine learning–based approaches. We aimed to develop machine learning models to detect HCM and to differentiate it from other cardiac conditions using ECGs and echocardiograms, with robust generalizability across multiple cohorts. Methods: Single-institution HCM ECG models were trained and validated on external data. Multi-institution models for ECG and echocardiogram were trained on data from 3 academic medical centers in the United States and Japan using a federated learning approach, which enables training on distributed data without data sharing. Models were validated on held-out test sets for each institution and from a fourth academic medical center and were further evaluated for discrimination of HCM from aortic stenosis, hypertension, and cardiac amyloidosis. Last, automated detection was compared with manual interpretation by 3 cardiologists on a data set with a realistic HCM prevalence. Results: We identified 74 376 ECGs for 56 129 patients and 8392 echocardiograms for 6825 patients at the 4 academic medical centers. Although ECG models trained on data from each institution displayed excellent discrimination of HCM on internal test data (C statistics, 0.88–0.93), the generalizability was limited, most notably for a model trained in Japan and tested in the United States (C statistic, 0.79–0.82). When trained in a federated manner, discrimination of HCM was excellent across all institutions (C statistics, 0.90–0.96 and 0.90–0.96 for ECG and echocardiogram model, respectively), including for phenotypic subgroups. The models further discriminated HCM from hypertension, aortic stenosis, and cardiac amyloidosis (C statistics, 0.84, 0.83, and 0.88, respectively, for ECG and 0.93, 0.94, 0.85, respectively, for echocardiogram). Analysis of electrocardiography-echocardiography paired data from 11 823 patients from an external institution indicated a higher sensitivity of automated HCM detection at a given positive predictive value compared with cardiologists (0.98 versus 0.81 at a positive predictive value of 0.01 for ECG and 0.78 versus 0.59 at a positive predictive value of 0.24 for echocardiogram). Conclusions: Federated learning improved the generalizability of models that use ECGs and echocardiograms to detect and differentiate HCM from other causes of hypertrophy compared with training within a single institution.

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1161/circulationaha.121.058696

Cites

4 of 4 references verified0 citation errors · 0 not checkable

Statistics

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

  • Hypertrophic cardiomyopathy, cardiac amyloidosis, and pulmonary hypertension AI-ECG applications remain earlier or pathway-incomplete.

    “hypertrophic cardiomyopathy, cardiac amyloidosis, and pulmonary hypertension remain earlier or pathway-incomplete applications.”

    From AI-ECG for Echocardiography Triage in Structural Heart Disease: Evidence, Implementation, and Future Directions (2026) · cites this paper as [27]

    The Discussion (page unavailable) says downstream workflows will inform clinical implementation of the HCM models, supporting the pathway-incomplete characterization for HCM. The paper mentions cardiac amyloidosis only as a condition that can resemble HCM; it does not assess an AI-ECG application for amyloidosis or pulmonary hypertension. (LLM, conf=0.87)

    Experimental — accuracy not yet measured

  • HCM and cardiac amyloidosis should be regarded as clinically important but pathway-incomplete domains rather than deployable screening indications comparable to low-LVEF AI-ECG.

    “Taken together, HCM and CA should therefore be regarded as clinically important but pathway-incomplete domains rather than deployable screening indications comparable to low-LVEF AI-ECG.26–30,43”

    From AI-ECG for Echocardiography Triage in Structural Heart Disease: Evidence, Implementation, and Future Directions (2026) · cites this paper as [27]

    The Clinical Implications section describes HCM as underdiagnosed and clinically important, while the Discussion says downstream workflows will inform how its screening models are implemented. The paper treats cardiac amyloidosis as a condition to distinguish from HCM, but does not assess its screening pathway or compare either condition’s deployment readiness with low-LVEF AI-ECG. (LLM, conf=0.85)

    Experimental — accuracy not yet measured

  • Equitable AI-ECG deployment depends not only on broadly similar performance across demographic subgroups, but also on transportability across institutions, acquisition workflows, and care environments.

    “Equitable AI-ECG deployment depends not only on broadly similar performance across demographic subgroups, but also on transportability across institutions, acquisition workflows, and care environments.”

    From AI-ECG for Echocardiography Triage in Structural Heart Disease: Evidence, Implementation, and Future Directions (2026) · cites this paper as [27]

    The Discussion (page unavailable) reports that single-institution ECG models can perform poorly on external data and suggests that institutional differences may reflect vendor signal processing. The Results, under “Federated Learning Improves the Generalizability of the HCM-ECG Model” (page unavailable), report external validation and an outpatient-setting analysis. The paper does not establish broadly similar performance across demographic subgroups or test transportability across acquisition workflows and care environments generally. (LLM, conf=0.84)

    Experimental — accuracy not yet measured

  • The low prevalence of individual structural heart diseases limits the potential implementation of models for broad screening because of the low PPVs of the proposed models.

    “While these models provide a strong foundation for the role of ECG-based detection of SHDs, the low prevalence of these individual diseases, their potential implementation for broad screening is limited by the low PPVs of the proposed models.10-12,16,32-34”

    From Ensemble Deep Learning Algorithm for Structural Heart Disease Screening Using Electrocardiographic Images (2025) · cites this paper as [33]

    The Methods describe an HCM surveillance cohort with 0.5% prevalence, and the Results report a 1% PPV for the ECG model at a high-sensitivity cutoff. The Discussion says PPV affects which follow-up actions are appropriate, but it does not conclude that low PPV precludes broad screening; it proposes systematic ECG screening followed by echocardiography. The paper addresses HCM, not structural heart diseases generally. (LLM, conf=0.84)

    Experimental — accuracy not yet measured

  • In a multinational Circulation study, ECG and echocardiography models trained across three academic centers in the United States and Japan showed substantially better cross-site generalizability than single-institution models and improved discrimination of HCM from other causes of hypertrophy.

    “In a multinational Circulation study, ECG and echocardiography models trained across three academic centers in the United States and Japan showed substantially better cross-site generalizability than single-institution models and improved discrimination of HCM from other causes of hypertrophy.27”

    From AI-ECG for Echocardiography Triage in Structural Heart Disease: Evidence, Implementation, and Future Directions (2026) · cites this paper as [27]

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Cites

Flagged first, then by citations

Research works printed but not linked: 23

  1. [2]Seidman JG Seidman C. The genetic basis for cardiomyopathy from mutation identification to mechanistic paradigms. Cell. 2001; 104: 557– 567. doi: 10.1016/s0092-8674(01)00242-2 11239412 · DOI
  2. [3]Maron BJ. Clinical course and management of hypertrophic cardiomyopathy. N Engl J Med. 2018; 379: 655– 668. doi: 10.1056/NEJMra1710575 30110588 · DOI · Paper page
  3. [4]Ommen SR Mital S Burke MA Day SM Deswal A Elliott P Evanovich LL Hung J Joglar JA Kantor P. 2020 AHA/ACC guideline for the diagnosis and treatment of patients with hypertrophic cardiomyopathy. Circulation. 2020; 142: e558– e631. doi:10.1161/CIR.0000000000000937 33215931 · DOI
  4. [5]Iacopo O Artur O Roberto B-V Theodore PA Ahmad M Pablo G-P Sara S Neal KL Matthew TW Anjali O. Mavacamten for treatment of symptomatic obstructive hypertrophic cardiomyopathy (EXPLORER-HCM): a randomised, double-blind, placebo-controlled, phase 3 trial. Lancet. 2020; 396: 759– 769. doi: 10.1016/S0140-6736(20)31792-X 32871100 · DOI · Paper page
  5. [6]Ho CY Mealiffe ME Bach RG Bhattacharya M Choudhury L Edelberg JM Hegde SM Jacoby D Lakdawala NK Lester SJ. Evaluation of mavacamten in symptomatic patients with nonobstructive hypertrophic cardiomyopathy. J Am Coll Cardiol. 2020; 75: 2649– 2660. doi: 10.1016/j.jacc.2020.03.064 32466879 · DOI
  6. [7]Maron BJ Desai MY Nishimura RA Spirito P Rakowski H Towbin JA Rowin EJ Maron MS Sherrid MV. Diagnosis and evaluation of hypertrophic cardiomyopathy: JACC state-of-the-art review. J Am Coll Cardiol. 2022; 79: 372– 389. doi: 10.1016/j.jacc.2021.12.002 35086660 · DOI
  7. [8]Hughes SE. The pathology of hypertrophic cardiomyopathy. Histopathology. 2004; 44: 412– 427. doi: 10.1111/j.1365-2559.2004.01835.x 15139989 · DOI
  8. [9]López-Cuenca D Muñoz-Esparza C Peñalver MN Alberola AG Blanes JRG. Hypertrophic or hypertensive cardiomyopathy? Int J Cardiol. 2016; 203: 891– 892. 26609686
  9. [10]Maron MS Rowin EJ Maron BJ. How to image hypertrophic cardiomyopathy. Circ Cardiovasc Imaging. 2017; 10: e005372. doi: 10.1161/CIRCIMAGING.116.005372 28701526 · DOI
  10. [11]Lemery R Kleinebenne A Nihoyannopoulos P Aber V Alfonso F McKenna WJ. Q waves in hypertrophic cardiomyopathy in relation to the distribution and severity of right and left ventricular hypertrophy. J Am Coll Cardiol. 1990; 16: 368– 374. doi: 10.1016/0735-1097(90)90587-f 2373813 · DOI
  11. [12]Maron BJ Wolfson JK Ciró E Spirito P. Relation of electrocardiographic abnormalities and patterns of left ventricular hypertrophy identified by 2-dimensional echocardiography in patients with hypertrophic cardiomyopathy. Am J Cardiol. 1983; 51: 189– 194. doi: 10.1016/s0002-9149(83)80034-4 6217739 · DOI
  12. [13]Usui M Inoue H Suzuki J Watanabe F Sugimoto T Nishikawa J. Relationship between distribution of hypertrophy and electrocardiographic changes in hypertrophic cardiomyopathy. Am Heart J. 1993; 126: 177– 183. doi: 10.1016/s0002-8703(07)80026-3 8322662 · DOI
  13. [14]Tison GH Zhang J Delling FN Deo RC. Automated and interpretable patient ECG profiles for disease detection, tracking, and discovery. Circ Cardiovasc Qual Outcomes. 2019; 12: e005289. doi: 10.1161/CIRCOUTCOMES.118.005289 31525078 PMC6951431 · DOI · Paper page
  14. [15]Morita SX Kusunose K Haga A Sata M Hasegawa K Raita Y Reilly MP Fifer MA Maurer MS Shimada YJ. Deep learning analysis of echocardiographic images to predict positive genotype in patients with hypertrophic cardiomyopathy. Front Cardiovasc Med. 2021; 8: 669860. doi: 10.3389/fcvm.2021.669860 34513940 PMC8429777 · DOI
  15. [18]McMahan HB Moore E Ramage D Hampson S Arcas B. Communication-efficient learning of deep networks from decentralized data. arXiv. Preprint posted online February 17, 2016. doi: 10.48550/arXiv.1602.05629 · DOI · Paper page
  16. [19]Li W Milletarì F Xu D Rieke N Hancox J Zhu W Baust M Cheng Y Ourselin S Cardoso MJ. Machine Learning in Medical Imaging, 10th international workshop, MLMI 2019, held in conjunction with MICCAI 2019, Shenzhen, China, October 13, 2019, Proceedings. Lect Notes Comput Sci. 2019: 133– 141. PMC7616593 39453415
  17. [21]Ho CY Day SM Colan SD Russell MW Towbin JA Sherrid MV Canter CE Jefferies JL Murphy AM Cirino AL. The burden of early phenotypes and the influence of wall thickness in hypertrophic cardiomyopathy mutation carriers: findings from the HCMNet study. JAMA Cardiol. 2017; 2: 419. doi: 10.1001/jamacardio.2016.5670 28241245 PMC5541992 · DOI
  18. [22]McInnes L Healy J Melville J. UMAP: Uniform Manifold Approximation and Projection for dimension reduction. arXiv. Preprint posted online December 6, 2018. doi: 10.48550/arXiv.1802.03426 · DOI · Paper page
  19. [23]Selvaraju RR Cogswell M Das A Vedantam R Parikh D Batra D. Grad-CAM: visual explanations from deep networks via gradient-based localization. arXiv. Preprint posted online December 3, 2019. doi: 10.48550/arXiv.1610.02391 · DOI
  20. [24]Goto S Mahara K Beussink-Nelson L Ikura H Katsumata Y Endo J Gaggin HK Shah SJ Itabashi Y MacRae CA. Artificial intelligence-enabled fully automated detection of cardiac amyloidosis using electrocardiograms and echocardiograms. Nat Commun. 2021; 12: 2726. doi: 10.1038/s41467-021-22877-8 33976142 PMC8113484 · DOI · Paper page
  21. [25]Morita SX Kusunose K Haga A Sata M Hasegawa K Raita Y Reilly MP Fifer MA Maurer MS Shimada YJ. Deep learning analysis of echocardiographic images to predict positive genotype in patients with hypertrophic cardiomyopathy. Front Cardiovasc Med. 2021; 8: 669860. doi: 10.3389/fcvm.2021.669860 34513940 PMC8429777 · DOI
  22. [26]Carlini N Tramèr F Wallace E Jagielski M Herbert-Voss A Lee K Roberts A Brown TB Song DX Erlingsson Ú Oprea A Raffel C. Extracting Training Data from Large Language Models. arXiv. Preprint posted online December 14, 2020. doi: 10.48550/arXiv.2012.07805 · DOI · Paper page
  23. [28]Ma C Poor HV. On safeguarding privacy and security in the framework of federated learning. IEEE Network. 2020; 34: 242– 248. doi: 10.1109/MNET.001.1900506 · DOI · Paper page

Cited by

Sorted by citations

Authors
Goto, S., Solanki, D., John, J. E., Yagi, R., Homilius, M., Ichihara, G., Katsumata, Y., Gaggin, H. K., Itabashi, Y., MacRae, C. A., & Deo, R. C.
Affiliations
  • Brigham and Women's Hospital (US)
  • Keio University (JP)
  • Harvard University (US)
Corresponding author
Deo, Rahul C.
ORCID
Venue
Circulation
Publisher
Ovid Technologies (Wolters Kluwer Health)
Published
2022-09-06
Volume
146
Issue
10
Pages
755-769
ISSN
  • 0009-7322 (ISSN-L)
  • 1524-4539
DOI
10.1161/circulationaha.121.058696
PMC
PMC9439630
OpenAlex
W4289525005
Cited by
130
References
4
Citations by year
2026: 32 · 2025: 48 · 2024: 30 · 2023: 19 · 2022: 1
FWCI
17.05
Type
journal article
Language
en
Open access
Licence
Locations
  • Circulation (publishedVersion, cc-by-nc-nd)
  • PubMed Central (publishedVersion, cc-by-nc-nd)
  • PubMed (publishedVersion)
Topics
  • Health Sciences › Medicine › Cardiology and Cardiovascular Medicine › ECG Monitoring and Analysis
  • Health Sciences › Medicine › Cardiology and Cardiovascular Medicine › Cardiomyopathy and Myosin Studies
  • Health Sciences › Health Professions › Health Information Management › Artificial Intelligence in Healthcare
Keywords
hypertrophic cardiomyopathy, automated disease detection, ECG, echocardiogram, federated learning, multi-institutional collaboration, model generalization, aortic stenosis, hypertension, cardiac amyloidosis, phenotypic subgroups, external validation, data privacy, cardiovascular screening, Japan-US comparison
MeSH terms
Amyloidosis, Cardiomyopathy, Hypertrophic, Echocardiography, Electrocardiography, Humans, Hypertension

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review