Related Experiment Video
Updated: Jan 10, 2026
09:17
Using Retinal Imaging to Study Dementia
Published on: November 6, 2017
22.2K
Noninvasive Coronary Artery Disease Detection Using Retinal Images: A Multimodal Study
Xiaohui Li1, Xiaoyu Dong2, Leilei Chen2
1School of Computer Science and Engineering, Nanjing University of Science and Technology, Nanjing, Jiangsu, China.
JACC. Advances
|November 20, 2025
Summary
Deep learning using retinal images shows promise for detecting coronary artery disease (CAD). This AI-based approach offers a noninvasive method for risk stratification, potentially improving patient outcomes.
Area of Science:
- Ophthalmology and Cardiology
- Artificial Intelligence in Healthcare
- Medical Imaging Analysis
Background:
- Coronary artery disease (CAD) is a major global health concern.
- Current CAD detection methods have limitations regarding safety and applicability.
- There is a critical need for safe, noninvasive diagnostic tools for CAD.
Purpose of the Study:
- To develop a deep learning framework for CAD detection using retinal images.
- To create a visual and multimodal detection system for coronary artery disease.
- To explore the utility of retinal vasculature in CAD diagnosis.
Main Methods:
- A multicenter cross-sectional study involved 383 patients undergoing coronary angiography.
- Developed three AI models: retinal image CNN, clinical indicator MLP, and multimodal fusion model.
- Utilized a cross-modal attention mechanism for integrating visual and clinical data.
Main Results:
- The retinal image-only model achieved an AUC of 0.80, outperforming clinical scores in intermediate-risk patients.
- Multimodal models integrating retinal images and clinical data showed superior performance.
- The best multimodal model achieved an AUC of 0.91, with 87.0% accuracy and 92.1% sensitivity.
Conclusions:
- Retinal images and vasculature contain valuable information for noninvasive CAD detection.
- Retinal imaging presents a promising, safe clinical risk factor for new diagnostic tools.
- AI-based fundus imaging can aid in risk stratification for coronary heart disease.
