1. Home
  2. EchoAI: A Multi-View Deep Learning Web Platform for Automated Echocardiographic Assessment

EchoAI: A Multi-View Deep Learning Web Platform for Automated Echocardiographic Assessment

1. Department of Health Information Technology and Management, School of Allied Medical Sciences, Shahid Beheshti University of Medical Sciences, Tehran, Iran
2. Cardiovascular Diseases Research Center, Department of Cardiology, Heshmat Hospital, School of Medicine, Guilan University of Medical Sciences, Rasht, Iran
3. Department of Computer Engineering, Sharif University of Technology, Tehran, Iran
4. Department of Artificial Intelligence in Medicine, Faculty of Advanced Technologies in Medicine, Iran University of Medical Sciences, Tehran, Iran

Published 2025 Article ID:1019

Abstract

Manual interpretation of echocardiography is a labor-intensive process characterized by significant inter-observer variability. Although Deep Learning (DL) has shown expert-level potential, its clinical integration is often hindered by "domain shift" across different ultrasound vendors and a lack of multi-view analysis capabilities. To address these challenges, we developed EchoAI, a secure, vendor-agnostic web-based Clinical Decision Support System (CDSS) that incorporates a multi-task Unsupervised Domain Adaptation (UDA) engine. This platform enables the simultaneous and automated quantification of Left Ventricular Ejection Fraction (LVEF) and cardiac wall thickness across standard A4C, A2C, and PLAX views. EchoAI utilizes a user-centered design and a "human-in-the-loop" workflow, allowing physicians to verify and edit AI-generated segmentation masks in real-time, thereby fostering diagnostic trust. Clinical validation conducted at Guilan University of Medical Sciences (GUMS) demonstrated high operational efficiency with an average processing time of 1.15 seconds per cardiac cycle. The system achieved a strong correlation with expert manual measurements (r=0.95, P<0.001) and an exceptionally low mean bias of -0.17%. Usability assessments yielded a high satisfaction score of 6.18 out of 7, with 86% of the AI outputs being accepted without modification. By providing an interactive and transparent interface, EchoAI effectively bridges the gap between algorithmic potential and routine clinical practice, offering a scalable solution for enhanced cardiac assessment in diverse healthcare settings.

Highlights

EchoAI is a vendor-agnostic web-based clinical decision support platform integrating multi-view deep learning and unsupervised domain adaptation for automated echocardiographic assessment. Its interactive human-in-the-loop workflow enables rapid, reliable, and physician-verified cardiac quantification across diverse ultrasound systems.

References

1. Nazari M, Emami H, Rabiei R, Rabiee HR, Salari A, Sadr H. Enhancing cardiac function assessment: developing and validating a domain adaptive framework for automating the segmentation of echocardiogram videos. Comput Med Imaging Graph. 2025:102627. [DOI:10.1016/j.compmedimag.2025.102627] [PMID:40834844] 
2. Sahashi Y, Ieki H, Yuan V, Christensen M, Vukadinovic M, Binder-Rodriguez C, et al. Artificial intelligence automation of echocardiographic measurements. J Am Coll Cardiol. 2025;86(13):964-78. [PMID:40914895] [DOI:10.1016/j.jacc.2025.07.053]
3. Cacao GF, Du D, Nair N. Unsupervised image segmentation on 2D echocardiogram. Algorithms. 2024;17(11):515. [DOI:10.3390/a17110515]
4. Sadr H, Salari A, Ashoobi MT, Nazari M. Cardiovascular disease diagnosis: a holistic approach using the integration of machine learning and deep learning models. Eur J Med Res. 2024;29(1):455. [PMCID:PMC11389500] [PMID:39261891][DOI:10.1186/s40001-024-02044-7] 
5. Nazari M, Emami H, Rabiei R, Hosseini A, Rahmatizadeh S. Detection of cardiovascular diseases using data mining approaches: application of an ensemble-based model. Cogn Comput. 2024;16(5):2264-78. [DOI:10.1007/s12559-024-10306-z]
6. Khodaverdian Z, Kozegar E, Sadr H, Nazari M, Parsa NA, Mirrazeghi SF, et al. A unified multi-task learning framework for automated assessment of left ventricular structure and its systolic function from echocardiography. Sci Rep. 2025; 16:2220. [PMCID:PMC12816601] [PMID:41387528] [DOI:10.1038/s41598-025-31773-w] 
7. Wang X, Cheng Z. Cross-sectional studies: strengths, weaknesses, and recommendations. Chest. 2020;158(1):S65-71. [DOI:10.1016/j.chest.2020.03.012] [PMID:32658654] [PMCID:PMC12469180]
8. Al‐Hawari F, Alufeishat A, Alshawabkeh M, Barham H, Habahbeh M. The software engineering of a three‐tier web‐based student information system (MyGJU). Comput Appl Eng Educ. 201;25(2):242-63. [DOI:10.1002/cae.21794]
9. Still B, Crane K. Fundamentals of user-centered design: A practical approach. CRC press; 2017. [DOI:10.4324/9781315200927]
10. Lewis JR. Computer System Usability Questionnaire (CSUQ). Int J Hum Comput Interact. 1995. APA PsycTests. [DOI:10.1037/t32698-000]

Files

History

Receive Date: 05 October 2025

Revise Date: 06 November 2025

Accept Date: 10 December 2025

Publish Date: 29 December 2025

Share

How to cite

Statistics

Article View: 89

PDF Download: 8

Keywords