Automated interpretation of systolic and diastolic function on the echocardiogram: a multicohort study.

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.

Informations de copyright

Copyright © 2022 The Author(s). Published by Elsevier Ltd. This is an Open Access article under the CC BY-NC-ND 4.0 license. Published by Elsevier Ltd.. All rights reserved.

Déclaration de conflit d'intérêts

Declaration of interests JT is supported by a National University of Singapore Startup Grant and has received personal grants and speaker fees from Olink Proteomics, Roche Diagnostics, Daiichi Sankyo, and Us2.ai. CSPL is supported by a Clinician Scientist Award from the National Medical Research Council of Singapore; has received research support from Boston Scientific, Bayer, Roche Diagnostics, AstraZeneca, Medtronic, and Vifor Pharma; has served as consultant or on the advisory board, steering committee, or executive committee for Boston Scientific, Bayer, Roche Diagnostics, AstraZeneca, Medtronic, Vifor Pharma, Novartis, Amgen, Merck, Janssen Research & Development, Menarini, Boehringer Ingelheim, Novo Nordisk, Abbott Diagnostics, Corvia, Stealth BioTherapeutics, JanaCare, Biofourmis, Darma, Applied Therapeutics, MyoKardia, Cytokinetics, WebMD Global, Radcliffe Group, and Corpus; and is co-founder and non-executive director of Us2.ai. MBI, MJF, and ZJ are employees of Us2.ai and hold equity in the company. PJS was an employee of Us2.ai and held equity in the company during the study. JT, PJS, WO, and CSPL own a patent entitled “Automatic clinical workflow that recognises and analyses 2D and doppler modality echocardiogram images for automated cardiac measurements and the diagnosis, prediction and prognosis of heart disease” related to the present work. JT, WO, and JAE are scientific advisors of Us2.ai and hold equity in the company. AMR is supported by a Senior Translational Research award from the National Medical Research Council of Singapore; holds the New Zealand Heart Foundation Chair of Cardiovascular Studies; has received research support from Boston Scientific, Bayer, AstraZeneca, Medtronic, Roche Diagnostics, Abbott Laboratories, Thermo Fisher Scientific, and Critical Diagnostics; and has consulted for Bayer, Novartis, Merck, AstraZeneca, and Roche Diagnostics. SDS reports grants from Actelion, Alnylam, Amgen, AstraZeneca, Bellerophon, Bayer, Bristol Myers Squibb, Celladon, Cytokinetics, Eidos, Gilead Sciences, GlaxoSmithKline, Ionis Pharmaceuticals, LoneStar Heart, Mesoblast, MyoKardia, NeuroTronik, NHLBI, Novartis, Novo Nordisk, Respicardia, Sanofi Pasteur, and Theracos; and personal fees from Abbott, Actelion, Akros, Alnylam Pharmaceuticals, Amgen, Arena, AstraZeneca, Bayer, Boehringer Ingelheim, Bristol Myers Squibb, Cardior Pharmaceuticals, Cardurion Pharmaceuticals, Corvia Medical, Cytokinetics, Daiichi Sankyo, Gilead Sciences, GlaxoSmithKline, Ironwood Pharmaceuticals, Eli Lilly, Merck, MyoKardia, Novartis, Roche, Takeda, Theracos, Quantum Genetics, Cardurion Pharmaceuticals, AOBiome, Janssen, Cardiac Dimensions, Sanofi Pasteur, Tenaya, DiNAQOR, Tremeau Pharmaceuticals, CellProthera, and Moderna. All other authors declare no competing interests.