AI Diabetic Retinopathy Screening

Diabetic retinopathy is an eye disease associated with diabetes that can damage retinal blood vessels and threaten vision. Screening aims to identify people who may need a comprehensive eye evaluation. AI-enabled systems analyze retinal photographs and return a screening result based on their intended use. FDA’s De Novo summary for IDx-DR documents a specific device and workflow; it is not a general approval for every camera, algorithm, patient group, or online photo. IDx-DR's FDA-authorized intended use is specifically for adults with diabetes who have not previously been diagnosed with diabetic retinopathy. It analyzes retinal images taken with the specified Topcon NW400 camera to detect more-than-mild diabetic retinopathy. This is a screening use under the device's labeling, not authorization to screen children, people with a prior diabetic-retinopathy diagnosis, other diseases, or images from arbitrary cameras. The device does not screen for diabetes or glaucoma. A result indicating diabetic retinopathy requires referral to an eye-care provider. If the system cannot produce a result, the person should be retested or referred according to the labeling. A no-disease screen does not replace all eye care or guarantee future health. The National Eye Institute recommends regular comprehensive dilated eye exams for people with diabetes. Clinics should use the authorized device as labeled, train staff, monitor image quality, and ensure referrals are completed. Patients should ask who will review the result, what happens after an ungradable image, and whether they still need a dilated eye exam. AI screening can extend access in some settings, but it does not treat retinopathy or replace an eye-care professional’s assessment and care plan.