Smartphone-Based AI System for Detecting Multiple Anterior Segment Eye Diseases Without Internet Compared to Cloud-Based AI

What this Trial Is About

Researchers are evaluating a smartphone-based artificial intelligence AI system designed to screen for 16 important anterior segment eye diseases, such as cataract and conjunctivitis. This multi-center, cross-sectional study compares the performance of the AI system, which operates entirely on the smartphone without internet connection, against certified ophthalmologists and a cloud-based AI deployment. The study aims to assess whether the on-device AI can provide accurate and practical screening in low-resource settings where internet access is limited. Participants will be screened using either the on-device AI system or a cloud-based version of the same AI model. The on-device system analyzes images captured by a standard Android smartphone directly on the device, while the cloud-based system sends images to a remote server via internet for analysis. The study includes five components diagnostic accuracy benchmarking, usability testing with non-medical users, direct comparison of deployment methods in the field, population-level screening in communities, and health-economic and environmental impact analyses. During the study, participants will undergo anterior segment eye image capture with a smartphone. Researchers will evaluate diagnostic accuracy compared to ophthalmologists, usability, screening time, user acceptability, and economic and environmental factors such as electricity consumption and cost-effectiveness. The reference standard for diagnosis is slit-lamp examination by certified ophthalmologists. The study is planned to complete by December 2028 and follows established guidelines for AI evaluation in clinical settings.

Development of a Mobile Terminal-Based Intelligent Detection System for Multiple Anterior Segment Diseases of the Eye