The National Eye Institute of the National Institutes of Health has awarded Mingquan Lin, PhD—an assistant professor in the Division of Computational Health Sciences at the University of Minnesota Medical School—a five-year, $3.7 million grant to develop next-generation artificial intelligence for primary open-angle glaucoma prognosis.
Glaucoma remains one of the leading causes of irreversible blindness worldwide, largely because disease progression can be difficult to anticipate. Clinicians currently rely on longitudinal changes across tests such as fundus photography, optical coherence tomography (OCT), and visual field measurements, a process that can be time-consuming and sensitive to variability.
Lin’s project, titled “Developing Robust Multimodal AI for Primary Open-Angle Glaucoma Prognosis,” is designed to build trustworthy, clinically usable models that integrate multiple data streams. The approach combines imaging features from retinal fundus photographs and OCT scans with functional information from visual field testing, while also incorporating relevant clinical variables.
A central goal is to improve early identification of patients at high risk for vision loss. Rather than maximizing accuracy alone, the work emphasizes robustness across real-world conditions—such as differences in imaging quality, patient demographics, and measurement noise—so that AI outputs remain reliable in routine care.
The team aims to create multimodal predictive systems that can estimate the likelihood of glaucoma progression and support decision-making. This includes designing models that can generalize beyond specific datasets, mitigating risks like overfitting and dataset bias that can undermine performance when deployed.
Trustworthiness is treated as an engineering requirement, not an afterthought. The project will focus on making predictions more interpretable and dependable for clinical workflows, aligning computational outputs with the needs of ophthalmologists.
Collaborations for the study span the University of Minnesota, Weill Cornell Medicine, and Washington University in St. Louis, bringing together complementary expertise in AI modeling and ophthalmic research.
As an outcome, the grant supports the development of clinically deployable technology intended to help clinicians act earlier and more precisely, ultimately improving patient outcomes while maintaining the safeguards required for medical AI.
Keywords
Glaucoma
Primary open-angle glaucoma
Multimodal AI
Trustworthy machine learning
Fundus photography
OCT
Visual field testing
Clinical decision support
Robust prognosis models
Subject of Research: Primary open-angle glaucoma prognosis using multimodal, trustworthy AI.
Article Title: Developing Robust Multimodal AI for Primary Open-Angle Glaucoma Prognosis
News Publication Date:
Web References: https://reporter.nih.gov/search/CgbzV_CDeEq9b9SGOUz5vQ/project-details/11365407
References:
Image Credits:
Tags: addressing variability in ophthalmic imagingadvancing glaucoma management with machine learningAI for real-world ophthalmology applicationsAI glaucoma prognosisearly detection of glaucoma progressionintegrating clinical variables in AI diagnosticsmultimodal artificial intelligence in eye careNIH grant for glaucoma researchpredictive modeling for primary open-angle glaucomaretinal imaging and OCT analysisrobust AI models for clinical usevisual field testing in glaucoma diagnosis
