AI Retinal Screening Achieves Clinical Adoption

Scobleizer · x · 2026-07-12

This repost highlights the medical value and clinical progress of **AI retinal image diagnosis**. Key points include: - The retina is one of the few windows to directly observe the central nervous system non-invasively. Diabetes silently damages retinal blood vessels, and patients often miss the optimal window for intervention by the time symptoms appear. - In 2018, the FDA approved **IDx-DR**, the first medical AI capable of providing autonomous diagnoses without a physician reading the images. This was followed by approvals for systems like **EyeArt** and **AEYE-DS**. - However, about one-third of diabetic patients in the US still do not receive screening. - China's **CARE** system was trained on **207,228** retinal photos, can simultaneously identify **14** types of retinal abnormalities, and matched ophthalmologists' performance in tests across **35** clinics. - The post concludes by noting that the same camera hardware already exists in US clinics, but the application simply isn't widespread yet.

Related event: AI Retinal Screening Deployed in Primary Care(2 posts)→

Original post →

More from Apps

Apps channel →