AI catches diabetes earlier than human doctors

The news: An AI model from AEYE Health accurately diagnoses patients with diabetes, per HIT consultant.

Here’s how it works: The AI scans a single image per eye, obtained from either a desktop or handheld retinal camera. 

  • Ophthalmologists typically look for signs of diabetic retinopathy, a diabetes complication visible in retinal imaging. Most doctors that manage to detect diabetes usually find that the disease has been present for a long time.
  • AEYE Health’s AI can generally diagnose diabetes from retinal images of otherwise healthy retinas—this early detection can lead to proper treatment and mitigation.

Why this matters: Diabetes is often under diagnosed and when doctors discover it, it’s usually in an advanced state. 

  • An estimated 7.3 million adults ages 18+ have diabetes but are undiagnosed (21.4% of all adults with diabetes), per the National Institute of DIabetes and Digestive and Kidney Diseases (NIDDK).

The opportunity: Healthcare AI is usually designed to supplant some tasks doctors perform, but AEYE Health’s model is performing tasks that human doctors cannot accomplish on their own.

  • Supplemental diagnoses can increase the accuracy of finding and treating medical conditions. In the case of diabetes, if left untreated, it can lead to heart disease, kidney failure, vision loss, and even death.

What’s next? AEYE Health is aiming to make diagnostic screening practical, accurate, and accessible by seeking FDA clearance for its autonomous screening solution for diabetic retinopathy.