Related Experiment Video
Updated: Sep 18, 2025
Analyses of Proteinuria, Renal Infiltration of Leukocytes, and Renal Deposition of Proteins in Lupus-prone MRL/lpr Mice
Published on: June 8, 2022
A deep learning system for detecting systemic lupus erythematosus from retinal images
Tingyao Li1, Shiqun Lin2, Zhouyu Guan3
1Shanghai International Joint Laboratory of Intelligent Prevention and Treatment for Metabolic Diseases, Department of Computer Science and Engineering, School of Electronic, Information, and Electrical Engineering, Shanghai Jiao Tong University, Department of Endocrinology and Metabolism, Shanghai Sixth People's Hospital Affiliated to Shanghai Jiao Tong University School of Medicine, Shanghai Diabetes Institute, Shanghai Clinical Center for Diabetes, Shanghai, China; MOE Key Laboratory of AI, School of Electronic, Information, and Electrical Engineering, Shanghai Jiao Tong University, Shanghai, China; Institute for Proactive Healthcare, Shanghai Jiao Tong University, Shanghai, China.
A new AI system, DeepSLE, can detect systemic lupus erythematosus (SLE) and its complications from retinal images. This deep learning tool shows high accuracy and potential for early disease detection in diverse populations.
Area of Science:
- Ophthalmology
- Nephrology
- Rheumatology
- Artificial Intelligence
Background:
- Systemic lupus erythematosus (SLE) is a complex autoimmune disease primarily affecting women, with diagnostic challenges due to complex criteria and low public awareness.
- Retinal manifestations in SLE can indicate disease activity and severity, offering a potential avenue for monitoring and diagnosis.
- Current screening methods for SLE and its complications face global challenges, necessitating innovative diagnostic approaches.
Purpose of the Study:
- To develop and validate a deep learning system (DeepSLE) for detecting SLE and its associated retinal and kidney complications using retinal images.
- To assess the performance of DeepSLE across diverse ethnic and demographic groups.
- To evaluate the clinical utility of DeepSLE compared to primary care physicians.
Main Methods:
- Development of a deep learning system (DeepSLE) trained on a large dataset of retinal images.
- Multi-ethnic validation using datasets from China and the UK, comprising 247,718 images.
- Qualitative and quantitative analyses for explainability and a prospective reader study comparing DeepSLE to primary care physicians.
Main Results:
- DeepSLE achieved high diagnostic performance for SLE, with areas under the receiver operating characteristic curve ranging from 0.822 to 0.969.
- The system demonstrated robust performance across various subgroups, including different genders, ages, ethnicities, and socioeconomic statuses.
- In a prospective reader study, DeepSLE exhibited higher sensitivity in detecting SLE complications compared to primary care physicians.
Conclusions:
- DeepSLE provides a digital solution for detecting SLE and its complications from retinal images, leveraging AI for enhanced diagnostic capabilities.
- The system's performance across diverse populations and its superior sensitivity compared to physicians suggest significant potential for clinical deployment.
- DeepSLE holds promise for improving early detection and management of systemic lupus erythematosus and its related organ damage.
