Savefor later Page saved! You can go back to this later in your Diabetes and Me Close

Artificial intelligence to keep an eye on retinopathy

Project summary

Diabetic retinopathy is a common complication of diabetes that can lead to loss of vision. People living with diabetes aged over 12 years have access to annual eye screenings, where their retina is photographed to identify signs of retinopathy. Dr Paul Nderitu wants to use a type of artificial intelligence to develop a way to predict who is likely to develop retinopathy as early as possible, so people can get the treatment they need before damage is done.

Background to research

Diabetic retinopathy affects about a third of people living with diabetes. It happens when high blood sugar levels cause the blood vessels in the eye to become damaged and leaky. If it's not caught and treated in time it can lead to sight loss. People with diabetes aged over 12 years have annual retinal screening appointments, to check for the development of retinopathy before it becomes severe. But there's no way to predict the risk of retinopathy getting worse. 

Research aims

Dr Nderitu wants to use deep learning, a type of artificial intelligence (AI), to develop an improved way to predict when someone retinopathy is likely to worsen and identify the features that are linked with progression of retinopathy. Dr Nderitu’s will use millions of retinal photographs from the diabetes eye screening programme to train AI models and fine tune the prediction tool.

Potential benefit to people with diabetes

Spotting signs of diabetic retinopathy early is key to preventing devastating sight loss. If Dr Nderitu’s model works, it could help identify eye damage even earlier than is currently possible, so that people with diabetes can get the right treatment and advice in time to prevent eye damage getting worse.

Brand Icons/Telephonecheck - FontAwesomeicons/tickicons/uk