Identify Eye Disease Early at Your Family Doctor’s Office

As an ophthalmologist, Michael D. Abramoff, M.D., had seen the consequences of diabetic retinopathy, glaucoma and other age-related eye diseases — all leading causes of blindness in America. He had an idea: Could autonomous artificial intelligence — similar to the software in IBM’s robot, Watson — make a difference? ​

Abramoff, 58, a University of Iowa professor, developed a breakthrough AI system for performing eye exams on people with diabetes. Called IDx-DR, the system can determine whether a test taker may have diabetic retinopathy or diabetic macular edema. “Diabetic retinopathy is a giant problem,” he says. “We know how to treat it, but only 15 percent of people with diabetes get yearly tests” from an eye doctor. ​

​Approved by the Food and Drug Administration (FDA) in 2018, the testing device is now used in the offices of primary care doctors and even in supermarket and mobile health clinics in Pennsylvania and Delaware. The American Diabetes Association added autonomous AI for eye checks to its 2020 standards of medical care in diabetes, and Medicare and private insurers often cover the cost. In a 2018 study of 900 people with diabetes, IDx-DR identified diabetic retinopathy 87 percent of the time, a greater level of success than that of retina specialists. ​

William Lamm’s primary care doctor offered him the breakthrough screening test in 2019. Lamm, 69, of Johns Creek, Georgia, has type 2 diabetes. When his exam showed that he had early signs of diabetic retinopathy, Lamm moved up his next appointment with his eye doctor to share the results. So far, controlling his blood sugar, along with good eating and exercise habits, has kept his eyes in good shape. “My eye doctor can tell when I’ve been going to the gym and riding my bike,” he says. “The blood vessels in my retinas are healthier.” ​

Abramoff hopes to have AI tests for glaucoma, other eye diseases and perhaps even for skin cancer ready for FDA review in the next few years. ​