{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"institution":[{"name":"medRxiv"}],"indexed":{"date-parts":[[2026,10,13]],"date-time":"2026-10-13T20:25:28Z","timestamp":1791923128604,"version":"4.3.4"},"posted":{"date-parts":[[2024,11,18]]},"group-title":"Cardiovascular Medicine","reference-count":40,"publisher":"openRxiv","license":[{"start":{"date-parts":[[2024,11,18]],"date-time":"2024-11-18T00:00:00Z","timestamp":1731888000000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/creativecommons.org\/licenses\/by-nc\/4.0\/"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"accepted":{"date-parts":[[2025,4,16]]},"abstract":"<jats:title>ABSTRACT<\/jats:title>\n                <jats:sec>\n                  <jats:title>Importance<\/jats:title>\n                  <jats:p>Echocardiography is a cornerstone of cardiovascular care but relies on expert interpretation and manual reporting from a series of videos. We propose an artificial intelligence (AI) system, PanEcho, to automate echocardiogram interpretation with multi-task deep learning.<\/jats:p>\n                <\/jats:sec>\n                <jats:sec>\n                  <jats:title>Objective<\/jats:title>\n                  <jats:p>To develop and evaluate the accuracy of PanEcho on a comprehensive set of 39 echocardiographic labels and measurements on transthoracic echocardiography (TTE).<\/jats:p>\n                <\/jats:sec>\n                <jats:sec>\n                  <jats:title>Design, Setting, and Participants<\/jats:title>\n                  <jats:p>This study represents the development and retrospective, multi-site validation of an AI system. PanEcho was developed using a sample of TTE studies conducted at Yale-New Haven Health System (YNHHS) hospitals and clinics from January 2016-June 2022 during routine care. The trained model was internally validated in a temporally distinct YNHHS cohort from July-December 2022, externally validated across four diverse external cohorts, and made publicly available.<\/jats:p>\n                <\/jats:sec>\n                <jats:sec>\n                  <jats:title>Main Outcomes and Measures<\/jats:title>\n                  <jats:p>The primary outcome was the area under the receiver operating characteristic curve (AUC) for diagnostic classification tasks and mean absolute error (MAE) for parameter estimation tasks, comparing AI predictions with the assessment of the interpreting cardiologist.<\/jats:p>\n                <\/jats:sec>\n                <jats:sec>\n                  <jats:title>Results<\/jats:title>\n                  <jats:p>This study included 1.2 million echocardiographic videos from 32,265 TTE studies of 24,405 patients across YNHHS hospitals and clinics. PanEcho performed 18 diagnostic classification tasks with a median AUC of 0.91 (IQR: 0.88-0.93) and estimated 21 echocardiographic parameters with a median normalized MAE of 0.13 (0.10-0.18) in internal validation. For instance, the model accurately estimated left ventricular (LV) ejection fraction (MAE: 4.2% internal; 4.5% external) and detected moderate or higher LV systolic dysfunction (AUC: 0.98 internal; 0.99 external), RV systolic dysfunction (0.93 internal; 0.94 external), and severe aortic stenosis (0.98 internal; 1.00 external). PanEcho maintained excellent performance in limited imaging protocols, performing 15 diagnosis tasks with 0.91 median AUC (IQR: 0.87-0.94) in an abbreviated TTE cohort and 14 tasks with 0.85 median AUC (0.77-0.87) on real-world point-of-care ultrasound acquisitions by non-experts from YNHHS emergency departments.<\/jats:p>\n                <\/jats:sec>\n                <jats:sec>\n                  <jats:title>Conclusions and Relevance<\/jats:title>\n                  <jats:p>We report an AI system that automatically interprets echocardiograms, maintaining high accuracy across geography and time from complete and limited studies. PanEcho may be used as an adjunct reader in echocardiography labs or rapid AI-enabled screening tool in point-of-care settings.<\/jats:p>\n                <\/jats:sec>\n                <jats:sec>\n                  <jats:title>KEY POINTS<\/jats:title>\n                  <jats:sec>\n                    <jats:title>Question<\/jats:title>\n                    <jats:p>Can artificial intelligence (AI) fully automate echocardiogram interpretation?<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Findings<\/jats:title>\n                    <jats:p>\n                      We report the development and validation of an automated AI system for echocardiogram analysis, called\n                      <jats:italic>PanEcho<\/jats:italic>\n                      , that performed 18 diagnostic classification tasks with a median area under the receiver operating characteristic curve (AUC) of 0.91 and 21 echocardiographic parameter estimation tasks with a median normalized mean absolute error (MAE) of 0.14.\n                    <\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Meaning<\/jats:title>\n                    <jats:p>An AI system can automate complete echocardiogram interpretation with high accuracy, potentially accelerating workflows and enabling rapid cardiovascular health screening in point-of-care settings with limited access to trained experts.<\/jats:p>\n                  <\/jats:sec>\n                <\/jats:sec>","DOI":"10.1101\/2024.11.16.24317431","type":"posted-content","created":{"date-parts":[[2024,11,18]],"date-time":"2024-11-18T14:01:08Z","timestamp":1731938468000},"source":"Crossref","is-referenced-by-count":21,"title":["PanEcho: Complete AI-enabled echocardiography interpretation with multi-task deep learning"],"prefix":"10.64898","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-5657-3081","authenticated-orcid":false,"given":"Gregory","family":"Holste","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-4362-0720","authenticated-orcid":false,"given":"Evangelos K.","family":"Oikonomou","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-3036-4131","authenticated-orcid":false,"given":"M\u00e1rton","family":"Tokodi","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-2320-6434","authenticated-orcid":false,"given":"Attila","family":"Kov\u00e1cs","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Zhangyang","family":"Wang","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-9467-6199","authenticated-orcid":false,"given":"Rohan","family":"Khera","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"54368","reference":[{"key":"2025041804300562000_2024.11.16.24317431v2.1","doi-asserted-by":"publisher","DOI":"10.1016\/j.echo.2018.06.004"},{"issue":"6","key":"2025041804300562000_2024.11.16.24317431v2.2","doi-asserted-by":"crossref","first-page":"569","DOI":"10.1016\/j.echo.2023.01.002","article-title":"Variation in cost of echocardiography within and across United States hospitals","volume":"36","year":"2023","journal-title":"J Am Soc Echocardiogr"},{"key":"2025041804300562000_2024.11.16.24317431v2.3","unstructured":"Virnig BA , Shippee ND , O\u2019Donnell B , Zeglin J , Parashuram S . Trends in the Use of Echocardiography, 2007 to 2011. Agency for Healthcare Research and Quality (US); 2014."},{"issue":"5","key":"2025041804300562000_2024.11.16.24317431v2.4","doi-asserted-by":"crossref","first-page":"562","DOI":"10.1016\/j.echo.2024.01.001","article-title":"Precision of echocardiographic measurements","volume":"37","year":"2024","journal-title":"J Am Soc Echocardiogr"},{"key":"2025041804300562000_2024.11.16.24317431v2.5","doi-asserted-by":"crossref","first-page":"520","DOI":"10.1038\/s41586-023-05947-3","article-title":"Blinded, randomized trial of sonographer versus AI cardiac function assessment","volume":"616","year":"2023","journal-title":"Nature"},{"key":"2025041804300562000_2024.11.16.24317431v2.6","doi-asserted-by":"publisher","DOI":"10.1056\/NEJMra1916062"},{"issue":"6","key":"2025041804300562000_2024.11.16.24317431v2.7","doi-asserted-by":"crossref","first-page":"624","DOI":"10.1001\/jamacardio.2021.0185","article-title":"Utility of a Deep-Learning Algorithm to Guide Novices to Acquire Echocardiograms for Limited Diagnostic Use","volume":"6","year":"2021","journal-title":"JAMA Cardiol"},{"key":"2025041804300562000_2024.11.16.24317431v2.8","doi-asserted-by":"publisher","DOI":"10.1093\/eurheartj\/ehad456"},{"issue":"7","key":"2025041804300562000_2024.11.16.24317431v2.9","doi-asserted-by":"crossref","first-page":"769","DOI":"10.1016\/j.echo.2023.03.008","article-title":"Fully Automated Artificial Intelligence Assessment of Aortic Stenosis by Echocardiography","volume":"36","year":"2023","journal-title":"J Am Soc Echocardiogr"},{"key":"2025041804300562000_2024.11.16.24317431v2.10","first-page":"614","volume-title":"Proceedings of the 6th Machine Learning for Healthcare Conference","volume":"149","year":"2021"},{"issue":"6","key":"2025041804300562000_2024.11.16.24317431v2.11","doi-asserted-by":"crossref","first-page":"534","DOI":"10.1001\/jamacardio.2024.0595","article-title":"A Multimodality Video-Based AI Biomarker For Aortic Stenosis Development And Progression","volume":"9","year":"2024","journal-title":"JAMA Card"},{"key":"2025041804300562000_2024.11.16.24317431v2.12","doi-asserted-by":"publisher","DOI":"10.1161\/CIRCULATIONAHA.118.034338"},{"key":"2025041804300562000_2024.11.16.24317431v2.13","doi-asserted-by":"publisher","DOI":"10.1038\/s41586-020-2145-8"},{"key":"2025041804300562000_2024.11.16.24317431v2.14","doi-asserted-by":"crossref","unstructured":"Ghorbani A , Ouyang D , Abid A , et al. Deep learning interpretation of echocardiograms. NPJ Digit Med. 2020;3:10.","DOI":"10.1038\/s41746-019-0216-8"},{"issue":"5","key":"2025041804300562000_2024.11.16.24317431v2.15","doi-asserted-by":"crossref","first-page":"482","DOI":"10.1016\/j.echo.2023.01.015","article-title":"Video-Based Deep Learning for Automated Assessment of Left Ventricular Ejection Fraction in Pediatric Patients","volume":"36","year":"2023","journal-title":"J Am Soc Echocardiogr"},{"issue":"1","key":"2025041804300562000_2024.11.16.24317431v2.16","doi-asserted-by":"crossref","first-page":"6776","DOI":"10.1038\/s41467-022-34245-1","article-title":"A formal validation of a deep learning-based automated workflow for the interpretation of the echocardiogram","volume":"13","year":"2022","journal-title":"Nat Commun"},{"key":"2025041804300562000_2024.11.16.24317431v2.17","doi-asserted-by":"crossref","unstructured":"Zeng Y , Tsui PH , Pang K , et al. MAEF-Net: Multi-attention efficient feature fusion network for left ventricular segmentation and quantitative analysis in two-dimensional echocardiography. Ultrasonics. 2023;127:106855.","DOI":"10.1016\/j.ultras.2022.106855"},{"issue":"1","key":"2025041804300562000_2024.11.16.24317431v2.18","doi-asserted-by":"crossref","first-page":"97","DOI":"10.1016\/j.jacc.2024.05.003","article-title":"Transforming Cardiovascular Care With Artificial Intelligence: From Discovery to Practice: JACC State-of-the-Art Review","volume":"84","year":"2024","journal-title":"J Am Coll Cardiol"},{"issue":"1","key":"2025041804300562000_2024.11.16.24317431v2.19","doi-asserted-by":"crossref","first-page":"2726","DOI":"10.1038\/s41467-021-22877-8","article-title":"Artificial intelligence-enabled fully automated detection of cardiac amyloidosis using electrocardiograms and echocardiograms","volume":"12","year":"2021","journal-title":"Nat Commun"},{"key":"2025041804300562000_2024.11.16.24317431v2.20","doi-asserted-by":"publisher","DOI":"10.1161\/CIRCULATIONAHA.124.068996"},{"key":"2025041804300562000_2024.11.16.24317431v2.21","unstructured":"Ferreira DL , Salaymang Z , Arnaout R. Label-free segmentation from cardiac ultrasound using self-supervised learning. arXiv [eessIV]. Published online October 10, 2022. http:\/\/arxiv.org\/abs\/2210.04979"},{"key":"2025041804300562000_2024.11.16.24317431v2.22","doi-asserted-by":"crossref","unstructured":"Vrudhula A , Duffy G , Vukadinovic M , Liang D , Cheng S. High Throughput Deep Learning Detection of Mitral Regurgitation. medRxiv. Published online 2024. https:\/\/www.medrxiv.org\/content\/10.1101\/2024.02.08.24302547.abstract","DOI":"10.1101\/2024.02.08.24302547"},{"key":"2025041804300562000_2024.11.16.24317431v2.23","doi-asserted-by":"publisher","DOI":"10.1101\/2024.06.22.24309332"},{"issue":"4","key":"2025041804300562000_2024.11.16.24317431v2.24","doi-asserted-by":"crossref","first-page":"386","DOI":"10.1001\/jamacardio.2021.6059","article-title":"High-Throughput Precision Phenotyping of Left Ventricular Hypertrophy With Cardiovascular Deep Learning","volume":"7","year":"2022","journal-title":"JAMA Cardiol"},{"key":"2025041804300562000_2024.11.16.24317431v2.25","doi-asserted-by":"publisher","DOI":"10.1038\/s41591-024-02959-y"},{"issue":"1","key":"2025041804300562000_2024.11.16.24317431v2.26","doi-asserted-by":"crossref","first-page":"133","DOI":"10.1038\/s43856-024-00538-3","article-title":"Efficient deep learning-based automated diagnosis from echocardiography with contrastive self-supervised learning","volume":"4","year":"2024","journal-title":"Commun Med"},{"key":"2025041804300562000_2024.11.16.24317431v2.27","doi-asserted-by":"crossref","unstructured":"Magyar B , Tokodi M , So\u00f3s A , et al. RVENet: A large echocardiographic dataset for the deep learning-based assessment of right ventricular function. In: Lecture Notes in Computer Science. Lecture notes in computer science. Springer Nature Switzerland; 2023:569\u2013583.","DOI":"10.1007\/978-3-031-25066-8_33"},{"issue":"8","key":"2025041804300562000_2024.11.16.24317431v2.28","doi-asserted-by":"crossref","first-page":"1005","DOI":"10.1016\/j.jcmg.2023.02.017","article-title":"Deep Learning-Based Prediction of Right Ventricular Ejection Fraction Using 2D Echocardiograms","volume":"16","year":"2023","journal-title":"JACC Cardiovasc Imaging"},{"issue":"2","key":"2025041804300562000_2024.11.16.24317431v2.29","doi-asserted-by":"crossref","first-page":"e113","DOI":"10.1016\/S2589-7500(24)00249-8","article-title":"Artificial intelligence-guided detection of under-recognised cardiomyopathies on point-of-care cardiac ultrasonography: a multicentre study","volume":"7","year":"2025","journal-title":"Lancet Digit Health"},{"key":"2025041804300562000_2024.11.16.24317431v2.30","unstructured":"Vaswani A , Shazeer NM , Parmar N , et al. Attention is All you Need. Adv Neural Inf Process Syst. Published online June 12, 2017:5998\u20136008."},{"issue":"1","key":"2025041804300562000_2024.11.16.24317431v2.31","doi-asserted-by":"crossref","first-page":"e46","DOI":"10.1016\/S2589-7500(21)00235-1","article-title":"Automated interpretation of systolic and diastolic function on the echocardiogram: a multicohort study","volume":"4","year":"2022","journal-title":"Lancet Digit Health"},{"key":"2025041804300562000_2024.11.16.24317431v2.32","unstructured":"Vukadinovic M , Tang X , Yuan N , et al. EchoPrime: A multi-video view-informed vision-language model for comprehensive echocardiography interpretation. arXiv [csCV]. Published online October 12, 2024. http:\/\/arxiv.org\/abs\/2410.09704"},{"key":"2025041804300562000_2024.11.16.24317431v2.33","doi-asserted-by":"publisher","DOI":"10.1016\/j.jacc.2016.06.070"},{"issue":"4","key":"2025041804300562000_2024.11.16.24317431v2.34","first-page":"36","article-title":"Future-proofing UK echocardiography","volume":"30","year":"2023","journal-title":"Br J Cardiol"},{"key":"2025041804300562000_2024.11.16.24317431v2.35","doi-asserted-by":"publisher","DOI":"10.1038\/s41598-023-33011-7"},{"key":"2025041804300562000_2024.11.16.24317431v2.36","doi-asserted-by":"publisher","DOI":"10.1016\/j.jacc.2012.01.037"},{"key":"2025041804300562000_2024.11.16.24317431v2.37","doi-asserted-by":"publisher","DOI":"10.1001\/jama.2020.27205"},{"issue":"12","key":"2025041804300562000_2024.11.16.24317431v2.38","doi-asserted-by":"crossref","first-page":"1932","DOI":"10.1111\/echo.14167","article-title":"Diagnostic accuracy of bicuspid aortic valve by echocardiography","volume":"35","year":"2018","journal-title":"Echocardiography"},{"issue":"100274","key":"2025041804300562000_2024.11.16.24317431v2.39","first-page":"100274","article-title":"Echocardiographic evaluation of the right atrial size and function: Relevance for clinical practice","volume":"27","year":"2023","journal-title":"Am Heart J Plus"},{"key":"2025041804300562000_2024.11.16.24317431v2.40","doi-asserted-by":"publisher","DOI":"10.1161\/CIRCIMAGING.112.978031"}],"container-title":[],"original-title":[],"link":[{"URL":"https:\/\/syndication.highwire.org\/content\/doi\/10.1101\/2024.11.16.24317431","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,1,15]],"date-time":"2026-01-15T20:20:30Z","timestamp":1768508430000},"score":1,"resource":{"primary":{"URL":"http:\/\/medrxiv.org\/lookup\/doi\/10.1101\/2024.11.16.24317431"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2024,11,18]]},"references-count":40,"URL":"https:\/\/doi.org\/10.1101\/2024.11.16.24317431","relation":{},"subject":[],"published":{"date-parts":[[2024,11,18]]},"subtype":"preprint"}}