AEVR Focuses on Artificial Intelligence in Retinal Disease in the International Age-Related Macular Degeneration (AMD) Awareness Week 2020 Congressional Briefing

Tiarnán D.L. Keenan, BM BCh, PhD (NEI)
Tiarnán D.L. Keenan, BM BCh, PhD (NEI)

As in past years, AEVR began its virtual Sixth Annual EVS Day with the International AMD Awareness Week Congressional Briefing, also held during Healthy Aging Month. Since NEI has a robust portfolio of research into both the “wet” and “dry” forms of AMD through various initiatives, including its Diabetic Retinopathy Clinical Research (DRCR) Retina Network, AEVR featured NEI’s Tiarnán D.L. Keenan, BM BCh, PhD who serves as Staff Clinician in Retinal Diseases in the Division of Epidemiology and Clinical Applications. Since his current research focuses on the diagnosis, treatment, and prevention of AMD using Artificial Intelligence (AI) and ‘Big Data” applications, he spoke about AI applications in medicine, generally, and in ophthalmology, specifically, since the latter is highly image-based and well-suited to automated diagnosis and classification. In fact, ophthalmology has taken a lead in AI, with the first autonomous AI system in use in any field of medicine approved by the Food and Drug Administration (FDA) in 2018 for the automated detection of clinically significant Diabetic Retinopathy [DR, which is treated with the same FDA-approved anti-Vascular Endothelial Growth Factor (VEGF) therapies as in “wet” AMD].

Using a series of both fundus photographs and Optical Coherence Tomography images (OCT, a non-invasive technology that creates three-dimensional images of the retina, the light-sensitive back of the eye), he described how AI can detect both “wet” and “dry” AMD, classify its severity, and predict the risk of progression to the late stage of the disease. Specifically with OCT, he demonstrated graphically how AI can detect and quantify retinal fluid in “wet” AMD and diabetic macular edema—whether intraretinal or subretinal—which can drive treatment decisions. Although in several cases he compared the AI results to those from retina specialists, he concluded that AI is a tool to assist physicians, rather than replace them, in diagnosis, triage, severity classification, risk prediction, and quantification of retinal fluid.

T.Y. Alvin Liu, MD (Wilmer Eye Institute/Johns Hopkins University)
T.Y. Alvin Liu, MD (Wilmer Eye Institute/Johns Hopkins University)

T. Y. Alvin Liu, MD, an Assistant Professor of Ophthalmology at the Wilmer Eye Institute/Johns Hopkins University and a participant in the Sixth Annual EVS Day, spoke about challenges for wide-spread adoption of AI technologies in medicine in general and Johns Hopkins Medicine’s adoption of the FDA-approved autonomous AI system for Diabetic Retinopathy screening. Among the various considerations he described were: taking a patient-centric approach in algorithm development (that is, focusing on meaningful applications that can positively impact patient outcomes); consideration of federal regulatory agency requirements and professional organization recommendations; the complex coordination between institutional departments (e.g., Medical/Legal, Billing, IT, Internal Medicine, and Ophthalmology); and adaptable implementation that includes regular quality improvement.

Dr. Liu noted that the implementation of AI-assisted Diabetic Retinopathy screening across the Johns Hopkins Medicine system was delayed during the COVID-19 pandemic clinical slowdown.

AEVR concluded the Briefing with comments presented by American Macular Degeneration Foundation patient advocate and veteran Jack Stewart. Mr. Stewart, who retired from the Navy early due to vision loss from AMD, spoke about the quality of life challenges which patients face. Despite this, he remains active, currently serving as the Secretary and Treasurer of the Blind Veterans Association of Vermont.  

AEVR Wishes to thank its co-sponsors:
Research to Prevent Blindness
American Macular Degeneration Foundation
ARVO
Macular Degeneration Partnership

AEVR also thanks Novartis for its video management support.