SCH: Advancing Clinical Decision Support for Glaucoma Detection and Progression Using Multi-Modal AI

Modified Project Summary/Abstract Section This project addresses the growing clinical and public health burden of glaucoma-related vision loss. Glaucoma is a complex, progressive disease that often goes undetected until substantial vision impairment has occurred. Current diagnostic approaches are limited by the complexity of the disease, the variability in patient presentation, and the fragmented nature of healthcare data. This research aims to develop a clinical decision support tool leveraging artificial intelligence (AI) to enhance diagnostic precision and facilitate earlier and more accurate intervention. Three core challenges will be addressed. First, the development of a multi-modal AI framework capable of integrating diverse data types, such as structural retinal imaging and visual field measurements, collected longitudinally across patient visits. This integration is expected to improve the accuracy and reliability of glaucoma detection and functional loss prediction. Second, the research will optimize the interface between clinicians and the AI-based tool to support informed clinical decision-making. Model interpretability and performance transparency will be emphasized to ensure safe use in a range of clinical scenarios. The incorporation of advanced uncertainty estimation methods will provide insights into prediction confidence and support appropriate usage. Third, the study will evaluate model performance across diverse data distributions to ensure consistency and generalizability in clinical applications. By systematically analyzing the impact of data heterogeneity, the project will inform model development strategies that improve clinical utility across populations and care settings.