| Objective:To conduct a scoping review on the application of large language models (LLMs) in diabetic retinopathy (DR) management,and to provide references for clinical practice.Methods:A systematic search was performed in Chinese and English databases for studies related to the application of LLMs in DR management.The search timeframe was from the establishment of each database to January 12, 2026.Two researchers independently screened and extracted data, and summarized and analyzed the included literature.Results? A total of 16 articles were included. The auxiliary application of LLMs in DR management has formed a diverse model ecosystem,encompassing general text models, general multimodal models,and ophthalmology-specific domain models.These models have demonstrated potential in auxiliary diagnosis and risk prediction, patient health education,clinical workflow optimization,and research assistance.Conclusion:LLMs provide novel technological pathways for the screening,management,and education of DR. However, their clinical translation still faces challenges such as a lack of high-level evidence-based evidence and insufficient stability in model output.Future efforts should focus on conducting rigorous effectiveness studies,systematic safety evaluations,and establishing industry standards to promote their safe and effective application in clinical practice. |