AI-Enabled Diabetic Retinopathy Screening Programs in Primary Care

Juan Ding, OD, PhD

Assistant Professor,

UMass Chan Medical School

What is your presentation about?
We set out to evaluate the accuracy of artificial intelligence (AI)-enabled diabetic retinopathy (DR) screening in our rural community, and how this improves the DR screening rate in this population.

How do you hope your presentation will impact diabetes research or care?
I hope that this research allows more communities to be able to apply AI for DR screening and to use this technology successfully without the pitfalls. I look forward to more real-world data on the efficacy of using AI for DR screening in various settings.

How did you become involved with this area of diabetes research or care?
I have always been passionate about improving compliance of DR screening and have worked on this area of research since I joined UMass Medical Center. AI has become the most promising approach to improve DR screening, and through my collaborative research I am convinced this is true both now and for the future.

---

AI-Enabled Diabetic Retinopathy Screening in a National Screening Program

Colin S. Tan, MD

Associate Professor of Ophthalmology,
Duke-NUS Medical School, National Healthcare Group Eye Institute

What is your presentation about?
Diabetic retinopathy and diabetic macular edema are one of the commonest causes of visual impairment and blindness throughout the world and have a significant impact on patients’ lives as well as health care costs. Screening for diabetic retinopathy is important for early detection of the disease and to reduce the burden of sight-threatening complications. This presentation will describe the Singapore Integrated Diabetic Retinopathy Program, which is a national screening program for diabetic retinopathy in Singapore, and our experiences in integrating artificial intelligence (AI) as an integral part of the system. The performance of the AI model and its impact on cost and time savings will be discussed.

How do you hope your presentation will impact diabetes research or care?
This presentation will share our experiences in designing and implementing a national screening program for diabetic retinopathy and the subsequent introduction of AI into the screening algorithm. It will illustrate how AI can improve efficiency and reduce the workload for screening programs. The validation of the AI system and its accuracy compared to human graders will enable comparisons with other existing systems in use. This will help everyone involved in treating diabetes and its complications evaluate the most appropriate screening programs for their communities.

How did you become involved with this area of diabetes research or care?
I have been involved in the treatment of diabetic retinopathy since my residency training and have worked on the various screening programs available at that time. I was part of the Singapore Integrated Diabetic Retinopathy Program since its inception and currently serve as the Chair of the Manpower, Training and Development Sub-Committee, responsible for the screening protocols as well as the training program for our graders. I have also been involved in grading of images used in multi-center, randomized controlled clinical trials and am the Head of the Fundus Image Reading Center, which grades images from various ophthalmic clinical trials, including those involving diabetic retinopathy and diabetic macular edema.