Slow uptake of AI systems for identification of diabetic eye disease

A new study published in JAMA Ophthalmology has identified barriers to the uptake of artificial intelligence (AI) systems for the detection of diabetic retinopathy.

The research highlighted that two systems, LumineticsCore(formerly IDx-DR) and EyeArt, have received approval from the US Food and Drug Administration for the identification of diabetic eye disease.

The authors highlighted that while vision loss from diabetic retinopathy is “largely preventable,” less than two thirds of diabetes patients undergo an annual eye examination.

In order to examine national trends within the US of diabetic retinopathy detection using AI systems, the researchers searched a national patient database, TriNetX, for the records of patients with diabetes between January 2019 and December 2023.

They identified instances where the AI-based screening reimbursement code was used in patient records.

Within the study cohort, only 4.2% of diabetic patients received ophthalmic imaging for diabetic retinopathy over 5 years.

Among the group that did receive imaging, only 2.2% underwent screening using AI-based systems.

“Despite AI-based imaging leading to more OCT referrals compared to traditional methods, barriers persist, such as cost, awareness, integration, and FDA approval of AI software for imaging devices,” the authors highlighted.

They shared that use of AI-based systems helps to increase the detection of diabetic eye disease in primary care settings, while optimising ophthalmic examinations for those with vision-threatening diabetic retinopathy.