Jennifer Lim, MD: Can AI Fill the Gap in Diabetic Retinopathy Screening?

Currently 8.5% of the world’s population has diabetes, and this number is only growing said Jennifer I. Lim, MD, Marion H. Schenk Esq. Chair, Professor of Ophthalmology, and Director of Retina Service at the University of Illinois at Chicago.

Lim spoke with MD Magazine® about a clinical trial she and her team conducted to test an artificial intelligence’s ability to detect referable diabetic retinopathy. The EyeArt system was developed as an effort to begin closing the gap between patients with diabetes who need screening and the limited numbers of health care providers who are able to screen for diabetic retinopathy.

“In the world today only about 50% at best undergo diabetic retinopathy screening and we can't scale it,” said Lim. She added that as the population of people with diabetes continues to grow, the gap will continue to grow.

The trial results were presented at the 2019 Association for Research in Vision and Ophthalmology (ARVO) Imaging in the Eye Conference in Vancouver, BC.

Why are you turning to artificial intelligence for diabetic retinopathy screening?

There's a burden for screening right now in order to examine all patients with diabetes for diabetic retinopathy and ensure that they don't have diabetic eye disease. And that's important because diabetic eye disease is a leading cause of preventable blindness in the working age population. In order to prevent it, you have to adequately diagnose it and then do timely treatment to prevent devastating eye disease.

In the world today only about 50% at best undergo diabetic retinopathy screening and we can't scale it. What we have available today, with the physicians and other health care workers who can screen for diabetic retinopathy—there's not enough to supply going forward in, say, 20 years. Right now, we have 415 million people with diabetes—that's 8.5% of the world's population. And it's estimated in 20 years that's going to increase the 642 million people. So, that's a huge burden of screening that has to be taken care of.

So, some groups have approached this by saying, perhaps we could do digital imaging and screen patients who for one reason or another are not getting to the ophthalmologist's office or the optometrist's office and maybe we can screen these pictures and look for diabetic retinopathy. And these have actually been validated, they actually work just as well with proper screening programs.

More recently groups have looked to artificial intelligence and have validated artificial intelligence as being able to grade photographs and determine whether there is diabetic retinopathy present that requires a referral to an ophthalmologist. So, we too sought to use artificial intelligence and we've applied a system known as the EyeArt system.