AI brings clearer results to diabetic eye disease

Artificial Intelligence is helping doctors to accurately diagnose eye disease in 96 per cent of patients at the Dubai Diabetes Centre.

Thousands of routine eye scans taken during a trial this month – UAE Innovation Month – were fed into an AI programme to identify retinal damage caused by diabetes.

The successful programme could now be permanently adopted at the centre, because it is helping doctors to treat more patients at an earlier stage.

“To diagnose diabetic retinopathy, regular retinal imaging is required and in some cases optical coherence tomography is also used,” said director Dr M Hamed Farooqi.

“This imaging test provides cross-sectional images of the retina that show the thickness of the retina, which will help to determine whether fluid has leaked into retinal tissue.”

Diabetic retinopathy is a diabetes complication caused by damage to the blood vessels of light-sensitive tissue at the back of the eye. At first, there may be no symptoms or only mild vision problems, but it can develop into blindness.

The second phase of the Dubai Diabetes Survey 2017 revealed that the prevalence of diabetes among Emiratis in the city is 19 per cent; the total rate of undiagnosed cases is 11 per cent; and the rate of pre-diabetic patients is 18.6 per cent.

Usually, seven pictures of each eye are needed to diagnose retinal damage, which means doctors must go through thousands of scans. The AI programme, developed by Dubai company Artelus, makes that process faster and more accurate.

The images fed into the programme were assessed by retina experts at Dubai Hospital and then compared with the programme’s assessment.

The accuracy of its diagnosis reached 96 per cent of all referrable cases. The more scans fed into the programme, the more accurate the diagnoses will ultimately be.

“These positive results will be considered in possibly adopting this programme at the centre,” Dr Farooqi said.