European health authorities have validated a new diabetic retinopathy screening solution that combines retinal fundus cameras with artificial intelligence algorithms in a portable, easy-to-use device, designed to bring this test closer to primary care, ophthalmology clinics and community screening programs.
A study published in the journal Diabetes Technology & Therapeutics evaluated the algorithm's performance using 624 retinal images from 156 people, achieving 94.8% sensitivity, 91.4% specificity and 92.7% diagnostic accuracy for detecting signs of diabetic retinopathy.
As a portable system that doesn't require an ophthalmologist present at the time of testing, this kind of technology aims to widen access to screening in areas with few specialists, catching diabetic retinopathy at early stages, when treatment is most effective.