Solving the Data Challenge in Ophthalmic AI
At ARVO, Cecilia S. Lee, MD, MS, and Aaron Y. Lee, MD, MSCI, discuss barriers to AI deployment in ophthalmology, including interoperability and model development.
By
Cecilia S. Lee, MD, MS
Aaron Y. Lee, MD, MSCI
May 1, 2026
Clinical Scorecard: Solving the Data Challenge in Ophthalmic AI
At a Glance
Category Detail
Condition Ophthalmic AI Deployment Key Mechanisms Standardization of imaging devices, data sharing, privacy protection, community engagement. Target Population Ophthalmology practitioners and researchers. Care Setting Clinical and research environments in ophthalmology.
Key Highlights
Barriers include lack of standardization in imaging devices. DICOM and FHIR standards are being implemented for better data interoperability. Data sharing while protecting participant privacy is a significant challenge. AI-READI project provides high-quality datasets for training AI models. Efforts are underway to safely deploy AI models in clinical settings. Guideline-Based Recommendations
Diagnosis
Utilize standardized imaging devices for consistent data collection. Management
Engage community partners in the development and sharing of datasets. Monitoring & Follow-up
Implement privacy protection mechanisms for participant data. Risks
Potential for reidentification of participants if data is not properly managed. Patient & Prescribing Data
Participants in ophthalmic studies and clinical trials.
Focus on developing AI models that enhance patient care delivery.
Clinical Best Practices
Adopt DICOM and FHIR standards for data interoperability. Ensure community engagement in AI model development. Utilize both publicly accessible and controlled access datasets for research. Related Resources & Content
