Skip to content
naulon
Get started

No retraction, correction or concern found when last checked on Nov 3, 2026.

High-Throughput Precision Phenotyping of Left Ventricular Hypertrophy With Cardiovascular Deep Learning

Grant Duffy, Paul P. Cheng, Neal Yuan and 15 more

JAMA Cardiology · 2022

doi:10.1001/jamacardio.2021.6059

Abstract

Question. Can deep learning be used to automate measurements of left ventricular dimensions and identify patients who could benefit from screening for underdiagnosed diseases?

Findings. In this cohort study of 23 745 patients that used a deep learning algorithm to automatically measure left ventricular dimensions and to identify patients with increased wall thickness who may benefit from additional screening for hypertrophic cardiomyopathy and cardiac amyloidosis, the algorithm performed consistently across multiple cohorts while also delivering results in a shorter time than required for human assessment.

Meaning. In this study, a deep learning workflow was able to automate wall thickness evaluation while facilitating identification of hypertrophic cardiomyopathy and cardiac amyloidosis.

Importance. Early detection and characterization of increased left ventricular (LV) wall thickness can markedly impact patient care but is limited by under-recognition of hypertrophy, measurement error and variability, and difficulty differentiating causes of increased wall thickness, such as hypertrophy, cardiomyopathy, and cardiac amyloidosis.

Objective. To assess the accuracy of a deep learning workflow in quantifying ventricular hypertrophy and predicting the cause of increased LV wall thickness.

Design, Settings, and Participants. This cohort study included physician-curated cohorts from the Stanford Amyloid Center and Cedars-Sinai Medical Center (CSMC) Advanced Heart Disease Clinic for cardiac amyloidosis and the Stanford Center for Inherited Cardiovascular Disease and the CSMC Hypertrophic Cardiomyopathy Clinic for hypertrophic cardiomyopathy from January 1, 2008, to December 31, 2020. The deep learning algorithm was trained and tested on retrospectively obtained independent echocardiogram videos from Stanford Healthcare, CSMC, and the Unity Imaging Collaborative.

Main Outcomes and Measures. The main outcome was the accuracy of the deep learning algorithm in measuring left ventricular dimensions and identifying patients with increased LV wall thickness diagnosed with hypertrophic cardiomyopathy and cardiac amyloidosis.

Results. The study included 23 745 patients: 12 001 from Stanford Health Care (6509 [54.2%] female; mean [SD] age, 61.6 [17.4] years) and 1309 from CSMC (808 [61.7%] female; mean [SD] age, 62.8 [17.2] years) with parasternal long-axis videos and 8084 from Stanford Health Care (4201 [54.0%] female; mean [SD] age, 69.1 [16.8] years) and 2351 from CSMS (6509 [54.2%] female; mean [SD] age, 69.6 [14.7] years) with apical 4-chamber videos. The deep learning algorithm accurately measured intraventricular wall thickness (mean absolute error [MAE], 1.2 mm; 95% CI, 1.1-1.3 mm), LV diameter (MAE, 2.4 mm; 95% CI, 2.2-2.6 mm), and posterior wall thickness (MAE, 1.4 mm; 95% CI, 1.2-1.5 mm) and classified cardiac amyloidosis (area under the curve [AUC], 0.83) and hypertrophic cardiomyopathy (AUC, 0.98) separately from other causes of LV hypertrophy. In external data sets from independent domestic and international health care systems, the deep learning algorithm accurately quantified ventricular parameters (domestic: R2, 0.96; international: R2, 0.90). For the domestic data set, the MAE was 1.7 mm (95% CI, 1.6-1.8 mm) for intraventricular septum thickness, 3.8 mm (95% CI, 3.5-4.0 mm) for LV internal dimension, and 1.8 mm (95% CI, 1.7-2.0 mm) for LV posterior wall thickness. For the international data set, the MAE was 1.7 mm (95% CI, 1.5-2.0 mm) for intraventricular septum thickness, 2.9 mm (95% CI, 2.4-3.3 mm) for LV internal dimension, and 2.3 mm (95% CI, 1.9-2.7 mm) for LV posterior wall thickness. The deep learning algorithm accurately detected cardiac amyloidosis (AUC, 0.79) and hypertrophic cardiomyopathy (AUC, 0.89) in the domestic external validation site.

Conclusions and Relevance. In this cohort study, the deep learning model accurately identified subtle changes in LV wall geometric measurements and the causes of hypertrophy. Unlike with human experts, the deep learning workflow is fully automated, allowing for reproducible, precise measurements, and may provide a foundation for precision diagnosis of cardiac hypertrophy.

References 35

  1. 1.

    Mardis ER. A decade’s perspective on DNA sequencing technology. Nature. 2011;470(7333):198-203

    Not in our library10.1038/nature09796

  2. 2.

    National Human Genome Research Institute, National Institutes of Health. DNA sequencing costs: data. Accessed June 1, 2021. https://www.genome.gov/about-genomics/fact-sheets/DNA-Sequencing-Costs-Data

    Not in our library

  3. 3.

    Lindpaintner K, Lee M, Larson MG,. Absence of association or genetic linkage between the angiotensin–converting–enzyme gene and left ventricular mass. N Engl J Med. 1996;334(16):1023-1028

    Not in our library10.1056/nejm199604183341604

  4. 4.

    Vasan RS, Glazer NL, Felix JF,. Genetic variants associated with cardiac structure and function: a meta-analysis and replication of genome-wide association data. JAMA. 2009;302(2):168-178

    Not in our library10.1001/jama.2009.978-a

  5. 5.

    Huda A, Castaño A, Niyogi A,. A machine learning model for identifying patients at risk for wild-type transthyretin amyloid cardiomyopathy. Nat Commun. 2021;12(1):2725

    Not in our library10.1038/s41467-021-22876-9

  6. 6.

    Shah SJ, Katz DH, Selvaraj S,. Phenomapping for novel classification of heart failure with preserved ejection fraction. Circulation. 2015;131(3):269-279

    Not in our library10.1161/circulationaha.114.010637

  7. 7.
    Artificial intelligence-enabled fully automated detection of cardiac amyloidosis using electrocardiograms and echocardiograms · 2021
    No notice
  8. 8.

    Zhang J, Gajjala S, Agrawal P,. Fully automated echocardiogram interpretation in clinical practice. Circulation. 2018;138(16):1623-1635

    Not in our library10.1161/circulationaha.118.034338

  9. 9.

    Davies D, Minamisawa M, Scott C, et al. A simple score to predict transthyretin cardiac amyloidosis in heart failure with preserved ejection fraction. J Am Col Cardiol. 2021;77(18)(suppl 1):521.

    Not in our library

  10. 10.

    Elliott PM, Anastasakis A, Borger MA,; Authors/Task Force members. 2014 ESC Guidelines on diagnosis and management of hypertrophic cardiomyopathy: the Task Force for the Diagnosis and Management of Hypertrophic Cardiomyopathy of the European Society of Cardiology (ESC). Eur Heart J. 2014;35(39):2733-2779

    Not in our library10.1093/eurheartj/ehu284

  11. 11.

    Lee GY, Kim K, Choi JO,. Cardiac amyloidosis without increased left ventricular wall thickness. Mayo Clin Proc. 2014;89(6):781-789

    Not in our library10.1016/j.mayocp.2014.01.013

  12. 12.

    Pagourelias ED, Mirea O, Duchenne J,. Echo Parameters for differential diagnosis in cardiac amyloidosis: a head-to-head comparison of deformation and nondeformation parameters. Circ Cardiovasc Imaging. 2017;10(3):e005588

    Not in our library10.1161/circimaging.116.005588

  13. 13.

    Phelan D, Sperry BW, Thavendiranathan P,. Comparison of ventricular septal measurements in hypertrophic cardiomyopathy patients who underwent surgical myectomy using multimodality imaging and implications for diagnosis and management. Am J Cardiol. 2017;119(10):1656-1662

    Not in our library10.1016/j.amjcard.2017.02.009

  14. 14.

    Angeli F, Verdecchia P, Angeli E,. Day-to-day variability of electrocardiographic diagnosis of left ventricular hypertrophy in hypertensive patients: influence of electrode placement. J Cardiovasc Med (Hagerstown). 2006;7(11):812-816

    Not in our library10.2459/01.jcm.0000250869.78777.09

  15. 15.

    Augusto JB, Davies RH, Bhuva AN,. Diagnosis and risk stratification in hypertrophic cardiomyopathy using machine learning wall thickness measurement: a comparison with human test-retest performance. Lancet Digit Health. 2021;3(1):e20-e28

    Not in our library10.1016/s2589-7500(20)30267-3

  16. 16.

    Howard JP, Stowell CC, Cole GD,. Automated left ventricular dimension assessment using artificial intelligence developed and validated by a UK-wide collaborative. Circ Cardiovasc Imaging. 2021;14(5):e011951

    Not in our library10.1161/circimaging.120.011951

  17. 17.

    Douglas PS, Garcia MJ, Haines DE,; American College of Cardiology Foundation Appropriate Use Criteria Task Force; American Society of Echocardiography; American Heart Association; American Society of Nuclear Cardiology; Heart Failure Society of America; Heart Rhythm Society; Society for Cardiovascular Angiography and Interventions; Society of Critical Care Medicine; Society of Cardiovascular Computed Tomography; Society for Cardiovascular Magnetic Resonance; American College of Chest Physicians. ACCF/ASE/AHA/ASNC/HFSA/HRS/SCAI/SCCM/SCCT/SCMR 2011 appropriate use criteria for echocardiography: a report of the American College of Cardiology Foundation Appropriate Use Criteria Task Force, American Society of Echocardiography, American Heart Association, American Society of Nuclear Cardiology, Heart Failure Society of America, Heart Rhythm Society, Society for Cardiovascular Angiography and Interventions, Society of Critical Care Medicine, Society of Cardiovascular Computed Tomography, Society for Cardiovascular Magnetic Resonance American College of Chest Physicians. J Am Soc Echocardiogr. 2011;24(3):229-267

    Not in our library10.1016/j.echo.2010.12.008

  18. 18.

    Popescu BA, Stefanidis A, Nihoyannopoulos P,. Updated standards and processes for accreditation of echocardiographic laboratories from the European Association of Cardiovascular Imaging: an executive summary. Eur Heart J Cardiovasc Imaging. 2014;15(11):1188-1193

    Not in our library10.1093/ehjci/jeu057

  19. 19.

    Leibundgut G, Rohner A, Grize L,. Dynamic assessment of right ventricular volumes and function by real-time three-dimensional echocardiography: a comparison study with magnetic resonance imaging in 100 adult patients. J Am Soc Echocardiogr. 2010;23(2):116-126

    Not in our library10.1016/j.echo.2009.11.016

  20. 20.

    Farsalinos KE, Daraban AM, Ünlü S, Thomas JD, Badano LP, Voigt JU. Head-to-head comparison of global longitudinal strain measurements among nine different vendors: the EACVI/ASE Inter-Vendor Comparison Study. J Am Soc Echocardiogr. 2015;28(10):1171-1181, e2.

    Not in our library10.1016/j.echo.2015.06.011

  21. 21.

    Ouyang D, He B, Ghorbani A,. Video-based AI for beat-to-beat assessment of cardiac function. Nature. 2020;580(7802):252-256

    Not in our library10.1038/s41586-020-2145-8

  22. 22.

    Poplin R, Varadarajan AV, Blumer K,. Prediction of cardiovascular risk factors from retinal fundus photographs via deep learning. Nat Biomed Eng. 2018;2(3):158-164

    Not in our library10.1038/s41551-018-0195-0

  23. 23.
    Deep learning interpretation of echocardiograms · 2020
    No notice
  24. 24.

    Raghunath S, Ulloa Cerna AE, Jing L,. Prediction of mortality from 12-lead electrocardiogram voltage data using a deep neural network. Nat Med. 2020;26(6):886-891

    Not in our library10.1038/s41591-020-0870-z

  25. 25.

    Github. ConvertDICOMToAVI.ipynb at master echonet/dynamic. Accessed November 1, 2021. https://github.com/echonet/dynamic/blob/master/scripts/ConvertDICOMToAVI.ipynb

    Not in our library

  26. 26.

    Yurtkulu SC, Şahin YH, Unal G. Semantic segmentation with extended DeepLabv3 architecture. Paper presented at: 27th Signal Processing and Communications Applications Conference; April 24-26, 2019; Sivas, Turkey.

    Not in our library

  27. 27.

    Tran D, Wang H, Torresani L, Ray J, LeCun Y, Paluri M. A closer look at spatiotemporal convolutions for action recognition. arXiv. Preprint posted online November 30, 2017.

    Not in our library

  28. 28.

    Castaño A, Narotsky DL, Hamid N,. Unveiling transthyretin cardiac amyloidosis and its predictors among elderly patients with severe aortic stenosis undergoing transcatheter aortic valve replacement. Eur Heart J. 2017;38(38):2879-2887

    Not in our library10.1093/eurheartj/ehx350

  29. 29.

    Tanskanen M, Peuralinna T, Polvikoski T,. Senile systemic amyloidosis affects 25% of the very aged and associates with genetic variation in alpha2-macroglobulin and tau: a population-based autopsy study. Ann Med. 2008;40(3):232-239

    Not in our library10.1080/07853890701842988

  30. 30.

    Gillmore JD, Maurer MS, Falk RH,. Nonbiopsy diagnosis of cardiac transthyretin amyloidosis. Circulation. 2016;133(24):2404-2412

    Not in our library10.1161/circulationaha.116.021612

  31. 31.

    Shah SJ, Borlaug BA, Kitzman DW,. Research priorities for heart failure with preserved ejection fraction: National Heart, Lung, and Blood Institute Working Group summary. Circulation. 2020;141(12):1001-1026

    Not in our library10.1161/circulationaha.119.041886

  32. 32.

    Quiñonero-Candela J, Sugiyama M, Schwaighofer A, Lawrence ND, eds. Dataset Shift in Machine Learning. The MIT Press; 2008

    Not in our library10.7551/mitpress/9780262170055.001.0001

  33. 33.

    Biondetti GP, Gauriau R, Bridge CP, Lu C, Andriole KP. “Name that manufacturer”. Relating image acquisition bias with task complexity when training deep learning models: experiments on head CT. arXiv. Preprint posted online August 19, 2020.

    Not in our library

  34. 34.

    Wu E, Wu K, Daneshjou R, Ouyang D, Ho DE, Zou J. How medical AI devices are evaluated: limitations and recommendations from an analysis of FDA approvals. Nat Med. 2021;27(4):582-584

    Not in our library10.1038/s41591-021-01312-x

  35. 35.

    Banerjee I, Bhimireddy AR, Burns JL,. Reading race: AI recognises patient’s racial identity in medical images. arXiv. Preprint posted online July 21, 2021.

    Not in our library

Cited by papers we hold 3

Reads

Read by people on this page

None yet.

Checked or read by AI through naulon

None yet.

A person counts once a day per paper. We keep no cookie and no address: a reader is a hash that changes every day. Any month under five reads shows no number.

Are you an author?

Is this your paper?

Claim it to add your note and be paid when agents cite it. Free.