Jasper Tromp1, Paul J Seekings2, Chung-Lieh Hung3, Mathias Bøtcher Iversen4, Matthew James Frost4, Wouter Ouwerkerk5, Zhubo Jiang4, Frank Eisenhaber6, Rick S M Goh7, Heng Zhao7, Weimin Huang8, Lieng-Hsi Ling9, David Sim10, Patrick Cozzone11, A Mark Richards12, Hwee Kuan Lee13, Scott D Solomon14, Carolyn S P Lam15, Justin A Ezekowitz16. 1. National Heart Centre Singapore, Singapore; Duke-NUS Medical School, Singapore; Saw Swee Hock School of Public Health, National University of Singapore & National University Health System, Singapore. 2. Bioinformatics Institute, Agency for Science, Technology and Research (A*STAR), Singapore; Us2.ai, Singapore. 3. Department of Medicine and Institute of Biomedical Sciences, Mackay Medical College, Taipei, Taiwan; Cardiovascular Division, Department of Internal Medicine, Mackay Memorial Hospital, Taipei, Taiwan. 4. Us2.ai, Singapore. 5. National Heart Centre Singapore, Singapore; Department of Dermatology, Amsterdam UMC, University of Amsterdam, Amsterdam Infection and Immunity Institute, Amsterdam, Netherlands. 6. Bioinformatics Institute, Agency for Science, Technology and Research (A*STAR), Singapore; Genome Institute of Singapore, Agency for Science, Technology and Research (A*STAR), Singapore; School of Biological Science, Nanyang Technological University, Singapore. 7. Institute of High Performance Computing, Agency for Science, Technology and Research (A*STAR), Singapore. 8. Institute for Infocomm Research, Agency for Science, Technology and Research (A*STAR), Singapore. 9. National University Heart Centre, Singapore; Yong Loo Lin School of Medicine, National University of Singapore, Singapore. 10. National Heart Centre Singapore, Singapore. 11. Singapore Bioimaging Consortium, Biomedical Sciences Institutes, Agency for Science, Technology and Research (A*STAR), Singapore. 12. National University Heart Centre, Singapore; Cardiovascular Research Institute, National University Health System, Singapore; Christchurch Heart Institute, University of Otago, Christchurch, New Zealand. 13. Bioinformatics Institute, Agency for Science, Technology and Research (A*STAR), Singapore; Image and Pervasive Access Lab, CNRS UMI 2955, Singapore; Singapore Eye Research Institute, Singapore. 14. Cardiovascular Division, Brigham and Women's Hospital, Harvard Medical School, Boston, MA, USA. 15. National Heart Centre Singapore, Singapore; Duke-NUS Medical School, Singapore; Department of Cardiology, University of Groningen, University Medical Center Groningen, Groningen, Netherlands. 16. Canadian VIGOUR Centre, University of Alberta, Edmonton, AB, Canada. Electronic address: jae2@ualberta.ca.
Abstract
BACKGROUND: Echocardiography is the diagnostic modality for assessing cardiac systolic and diastolic function to diagnose and manage heart failure. However, manual interpretation of echocardiograms can be time consuming and subject to human error. Therefore, we developed a fully automated deep learning workflow to classify, segment, and annotate two-dimensional (2D) videos and Doppler modalities in echocardiograms. METHODS: We developed the workflow using a training dataset of 1145 echocardiograms and an internal test set of 406 echocardiograms from the prospective heart failure research platform (Asian Network for Translational Research and Cardiovascular Trials; ATTRaCT) in Asia, with previous manual tracings by expert sonographers. We validated the workflow against manual measurements in a curated dataset from Canada (Alberta Heart Failure Etiology and Analysis Research Team; HEART; n=1029 echocardiograms), a real-world dataset from Taiwan (n=31 241), the US-based EchoNet-Dynamic dataset (n=10 030), and in an independent prospective assessment of the Asian (ATTRaCT) and Canadian (Alberta HEART) datasets (n=142) with repeated independent measurements by two expert sonographers. FINDINGS: In the ATTRaCT test set, the automated workflow classified 2D videos and Doppler modalities with accuracies (number of correct predictions divided by the total number of predictions) ranging from 0·91 to 0·99. Segmentations of the left ventricle and left atrium were accurate, with a mean Dice similarity coefficient greater than 93% for all. In the external datasets (n=1029 to 10 030 echocardiograms used as input), automated measurements showed good agreement with locally measured values, with a mean absolute error range of 9-25 mL for left ventricular volumes, 6-10% for left ventricular ejection fraction (LVEF), and 1·8-2·2 for the ratio of the mitral inflow E wave to the tissue Doppler e' wave (E/e' ratio); and reliably classified systolic dysfunction (LVEF <40%, area under the receiver operating characteristic curve [AUC] range 0·90-0·92) and diastolic dysfunction (E/e' ratio ≥13, AUC range 0·91-0·91), with narrow 95% CIs for AUC values. Independent prospective evaluation confirmed less variance of automated compared with human expert measurements, with all individual equivalence coefficients being less than 0 for all measurements. INTERPRETATION: Deep learning algorithms can automatically annotate 2D videos and Doppler modalities with similar accuracy to manual measurements by expert sonographers. Use of an automated workflow might accelerate access, improve quality, and reduce costs in diagnosing and managing heart failure globally. FUNDING: A*STAR Biomedical Research Council and A*STAR Exploit Technologies.
BACKGROUND: Echocardiography is the diagnostic modality for assessing cardiac systolic and diastolic function to diagnose and manage heart failure. However, manual interpretation of echocardiograms can be time consuming and subject to human error. Therefore, we developed a fully automated deep learning workflow to classify, segment, and annotate two-dimensional (2D) videos and Doppler modalities in echocardiograms. METHODS: We developed the workflow using a training dataset of 1145 echocardiograms and an internal test set of 406 echocardiograms from the prospective heart failure research platform (Asian Network for Translational Research and Cardiovascular Trials; ATTRaCT) in Asia, with previous manual tracings by expert sonographers. We validated the workflow against manual measurements in a curated dataset from Canada (Alberta Heart Failure Etiology and Analysis Research Team; HEART; n=1029 echocardiograms), a real-world dataset from Taiwan (n=31 241), the US-based EchoNet-Dynamic dataset (n=10 030), and in an independent prospective assessment of the Asian (ATTRaCT) and Canadian (Alberta HEART) datasets (n=142) with repeated independent measurements by two expert sonographers. FINDINGS: In the ATTRaCT test set, the automated workflow classified 2D videos and Doppler modalities with accuracies (number of correct predictions divided by the total number of predictions) ranging from 0·91 to 0·99. Segmentations of the left ventricle and left atrium were accurate, with a mean Dice similarity coefficient greater than 93% for all. In the external datasets (n=1029 to 10 030 echocardiograms used as input), automated measurements showed good agreement with locally measured values, with a mean absolute error range of 9-25 mL for left ventricular volumes, 6-10% for left ventricular ejection fraction (LVEF), and 1·8-2·2 for the ratio of the mitral inflow E wave to the tissue Doppler e' wave (E/e' ratio); and reliably classified systolic dysfunction (LVEF <40%, area under the receiver operating characteristic curve [AUC] range 0·90-0·92) and diastolic dysfunction (E/e' ratio ≥13, AUC range 0·91-0·91), with narrow 95% CIs for AUC values. Independent prospective evaluation confirmed less variance of automated compared with human expert measurements, with all individual equivalence coefficients being less than 0 for all measurements. INTERPRETATION: Deep learning algorithms can automatically annotate 2D videos and Doppler modalities with similar accuracy to manual measurements by expert sonographers. Use of an automated workflow might accelerate access, improve quality, and reduce costs in diagnosing and managing heart failure globally. FUNDING: A*STAR Biomedical Research Council and A*STAR Exploit Technologies.
Authors: Mariam Haji-Hassan; Tudor Călinici; Tudor Drugan; Sorana D Bolboacă Journal: Int J Environ Res Public Health Date: 2022-03-04 Impact factor: 3.390