Diabetic Retinopathy: Can Artificial Intelligence Provide a Better Way to Detect Disease?

Linda Brookes, MSc


June 07, 2018

The recent approval by the US Food and Drug Administration (FDA) of the first autonomous diagnostic system for the detection of diabetic retinopathy (DR)[1]—the IDx-DR (IDx, LLC; Coralville, Iowa)—appeared to signal the beginning of the inevitable advance of artificial intelligence (AI) toward replacing, rather than merely assisting, physicians in diagnosing diseases. Despite this possible threat, the initial reaction of eye care specialists and primary care physicians has been to welcome it as an addition to, but not a replacement for, the comprehensive eye examination—at least for now.

The rationale behind the development of the device, according to IDx founder and president Michael Abramoff, MD, PhD, was to expand screening to the primary care setting, thus freeing up time for ophthalmologists to concentrate on treating patients already diagnosed with DR.[2]

That is an attractive prospect for such ophthalmologists as Charles Wykoff, MD, PhD, director of clinical research at Retina Consultants of Houston, and deputy chair of ophthalmology for the Blanton Eye Institute at Houston Methodist Hospital. "I think this is a positive evolution of our space, and it will allow patients to get eye care earlier than they do currently," Wykoff said.

Nonetheless, it may not necessarily be practical for many primary care physicians or diabetologists, such as Jay Shubrook, DO, professor in the department of primary care at Touro University California College of Osteopathic Medicine.

The key is to be able to ensure that it facilitates, not disrupts, the normal procedure of care.

"I think AI can be applied very effectively in many cases, and if this tool can give a good interpretation for the back of the eye, then it will be very effective," Shubrook said. "But the key is to be able to ensure that it facilitates, not disrupts, the normal procedure of care. In addition, if you are trying to implement it in primary care, it has to be easy enough and not cause a loss of income."

How It Works

The IDx-DR device is a software program incorporating a deep-learning enhanced algorithm that analyzes two images of the retina (one disc-centered, one macula-centered) taken with a nonmydriatic retinal camera (TRC-NW400; Topcon, Tokyo, Japan).[3,4] If the images are of sufficient quality, the software produces one of two results: (1) negative for more-than-mild DR (mtmDR), with a recommendation to rescreen in 12 months, or (2) positive for mtmDR, with a recommendation to refer to an eye care specialist.

Because the device provides a screening decision without requiring a clinician to interpret the images or the results, the IDx-DR device can be used by healthcare professionals who may not normally be involved in eye care.[1]


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.