Interpretation of Sectoral Standard Artificial Intelligence Medical Device—Specific Requirement for Datasets: Color Fundus Images of Diabetic Retinopathy

In recent years, with the rapid development of the artificial intelligence(AI) medical device industry, regulatory authorities have been continuously strengthening quality control over datasets. In 2024, the National Medical Products Administration (NMPA) issued the significant industry standard YY/T 1949-2024 Artificial Intelligence Medical Devices—Particular Requirements for Datasets: Color Fundus Photographs of Diabetic Retinopathy, which is China's first specialized standard for evaluating dataset quality in AI medical devices targeting a specific disease area.Under the existing general standards framework for AI medical devices, this standard focuses specifically on diabetic retinopathy fundus photograph datasets, systematically proposing specialized quality requirements and testing methods. Its implementation will provide clear technical guidance for dataset manufacturers, help improve the development quality of AI-assisted diagnostic software for diabeticretinopathy, and promote standardized industry advancement.