Utilizing diagnostic imaging and big data,The Barry Family Center for Ophthalmic Artificial Intelligence and Human Health (BFCOAIHH) will build on the unfolding science of AI. The Center will develop innovative clinical applications that the New York Eye and Ear Infirmary of Mount Sinai (NYEE) and the Icahn School of Medicine at Mount Sinai have initiated in recent years to advance not just visual care, but the much broader field of population health.
Our Approach
The researchers at NYEE and Icahn Mount Sinai have already made great strides in successfully integrating AI and advanced mathematical modeling to enhance the understanding of glaucoma's risk factors and pathophysiology. For example, after analyzing new-onset visual field loss patterns from more than 2,500 affected eyes using AI, researchers gained novel insights into why specific populations are at greater risk for glaucoma blindness. Another research team developed unique models for detecting adult macular degeneration (AMD) at early, intermediate, and advanced stages using an ensemble of deep-learning screening methods and AMD-specific algorithms. Our scientists are now exploring the novel use of the popular AI-driven ChatGPT to provide counseling to glaucoma patients.
Ophthalmic imaging is unique within medicine as it provides methods for the direct visualization and assessment of the effects of ocular diseases like glaucoma, diabetic retinopathy, and macular degeneration, as well as systemic conditions that impact entire populations, such as cardiovascular, neurologic, and vascular diseases. The ability to detect these diseases at their early stage through retinal imaging is expected to reach new heights in the years ahead with the help of sophisticated AI and machine learning algorithms. That strategy is already taking shape through a trio of innovative programs:
- Tele-retinal imaging is a robust program designed to capture patients at high risk of diabetic retinopathy while they undergo routine exams at their primary care physician’s office. Currently in place at seven sites within the Mount Sinai Health System, this initiative uses fundus cameras to provide non-mydriatic images of the back of the eye, which are sent securely to a retinal specialist at NYEE for later reading. Some of these photographs are read in real-time by AI, providing immediate feedback on potential disorders. Once transitioned to a more advanced AI algorithm, tele-retinal imaging is seen as a population-based strategy that will facilitate the early diagnosis of a vast array of ocular and systemic conditions, particularly cardiovascular disorders (whose risk factors recent research has shown can be predicted from standard retinal fundus photographs and deep learning algorithms).
- Tele-consultation streamlines emergent care at Mount Sinai hospitals by linking emergency room physicians via video or telephone to off-site ophthalmologists from NYEE. The consultant logs onto a high-magnification camera in the ER to view and zoom in on external portions of the patient’s eye and surrounding structures in real-time from their laptop or cell phone. Incorporating AI into the image interpretation process could give tele-consults a bold new dimension in terms of fast and comprehensive diagnosis that exceeds the human eye's capabilities.
- The Eye Stroke Program allows patients experiencing a central retinal artery occlusion (CRAO), the ocular equivalent of a cerebral stroke, to get a diagnosis more quickly than ever so they can begin potentially eye-saving treatment with endovascular injection of an infusion of tissue plasminogen activator (tPA). Hospital stroke team physicians upload images from an on-site optical coherence tomography (OCT) machine to a retinal specialist at NYEE for immediate review and consultation. AI could make the CRAO program even more expeditious through an ability to instantly detect central artery occlusion or some other ophthalmic event that might be triggering the patient’s sudden loss of sight.