Automated detection of retinal artery occlusion in fundus photography via self-supervised deep learning and multimodal interpretability using a multimodal AI chatbot
Medical & Biological Engineering & Computing · 2025-09 · 63 (9) · 2679-2691
DOI: 10.1007/s11517-025-03353-7 ↗Kurz gefasst
Fachartikel in der Zeitschrift Medical & Biological Engineering & Computing, online erschienen am 31.03.2025. Verfasst von Sun Young Ryu und 2 weiteren Autor:innen. Beteiligte Einrichtung: B&VIIT Eye Center, Seoul, South Korea. Hauptthemen laut MEDLINE: Deep Learning, Retinal Artery Occlusion, Fundus Oculi, Photography. Die Arbeit wurde bisher 11-mal zitiert (Crossref) und nennt 48 Literaturangaben.
Abstract (Originalsprache)
Retinal artery occlusion (RAO) is a sight-threatening condition that requires prompt diagnosis to prevent irreversible vision loss. This study presents an innovative AI-driven approach for RAO detection from fundus images, marking the first application of deep learning for this purpose. Using a self-supervised learning (SSL) framework with SimCLR, our model addresses the challenge of limited labeled RAO data. The ResNet50 model pretrained with SimCLR demonstrated high diagnostic accuracy, achieving areas under the receiver operating characteristic curve (AUC) of 0.924 and 0.988 on two external validation datasets, highlighting its robustness and generalizability in RAO detection. To enhance transparency in clinical AI, we incorporated a multimodal interpretability approach using a ChatGPT-4-based AI chatbot. This chatbot, combined with Grad-CAM visualizations, provides detailed clinical explanations of the model's predictions, emphasizing key RAO features such as retinal whitening and cherry-red spots. This multimodal interpretability framework improves clinicians' understanding of the model's decision-making process, facilitating clinical adoption and trust. By automating RAO detection, this AI model serves as a valuable tool for the early identification of ocular and systemic vascular risks, enabling timely intervention. These findings highlight the potential of fundus imaging for RAO detection and broader cardiovascular risk assessment, advancing AI's role in predictive healthcare.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Autor:innen und Institutionen
- B&VIIT Eye Center, Seoul, South Korea.
- Department of Biomedical Engineering, Yonsei University, Wonju, South Korea.
- Department of Ophthalmology, Hangil Eye Hospital, Incheon, South Korea. eyetaekeunyoo@gmail.com.
ORCID-Profile: Tae Keun Yoo
Zitieren
APA
Ryu, S. Y., Choi, J. Y., & Yoo, T. K. (2025). Automated detection of retinal artery occlusion in fundus photography via self-supervised deep learning and multimodal interpretability using a multimodal AI chatbot. Medical & Biological Engineering & Computing, 63(9), 2679-2691. https://doi.org/10.1007/s11517-025-03353-7
Vancouver
Ryu SY, Choi JY, Yoo TK. Automated detection of retinal artery occlusion in fundus photography via self-supervised deep learning and multimodal interpretability using a multimodal AI chatbot. Medical & Biological Engineering & Computing. 2025;63(9):2679-2691. doi:10.1007/s11517-025-03353-7
BibTeX
@article{Ryu2025,
author = {Ryu, Sun Young and Choi, Joon Yul and Yoo, Tae Keun},
title = {Automated detection of retinal artery occlusion in fundus photography via self-supervised deep learning and multimodal interpretability using a multimodal AI chatbot},
journal = {Medical & Biological Engineering & Computing},
year = {2025},
volume = {63},
number = {9},
pages = {2679--2691},
doi = {10.1007/s11517-025-03353-7}
}
EUVIMED