- Artificial Intelligence Can Detect Diabetic Eye. Diabetic Retinopathy can be detected easily and accurately using AI.
- EyeArt is a system that helps prevent Diabetic Retinopathy.
- It is 95.5 percent accurate.
Researchers have now found out that an AI screening system is capable of discovering diabetic retinopathy, with 95.5 percent accuracy. It is a well-known fact that high blood sugar damages the tiny blood vessels at the back of the eye. There is a possibility of tiny bulges protruding from the blood vessels, which can at times leak fluid and blood into the retina. A new advancement suggests that an automated; artificial intelligence (AI) screening system can accurately discover diabetic retinopathy now.
Artificial Intelligence Can Detect Diabetic Eye
To add to that, the story suggests that the system does not need any inputs from an expert ophthalmologist. The most brilliant part about it is that it can provide a reading in a time period of just 60 seconds. These findings were presented at AAO 2019, the 123rd Annual Meeting of the American Academy of Ophthalmology.
Diabetic retinopathy develops over time in people who are diagnosed with diabetes and have poor control over their blood sugar levels. The fluid that gets leaked into the retina causes swelling or edema, which affects vision.
Read: Facts About Sugar Substitutes
Initially there are no symptoms, but only mild vision problems. As time passes, it can cause blindness. Ophthalmologists do have effective treatments for the problem, but they only work best when the condition is detected early. Due to this, ophthalmologists recommend the patients suffering from this disease, to be screened every year.
A system known as EyeArt has shown some promise in earlier studies. Previously, it was used to screen 893 patients, who suffered from the disease at 15 different medical locations. Post that, the results were reviewed for clinical accuracy by certified graders. Using only undilated images, the system’s sensitivity was 95.5 percent, along with a specificity of 86 per cent.
Very few patients required dilation for an image good enough to be graded to be captured. When these additional patients were included in the overall analysis, the sensitivity did not change, but the specificity improved to 86.5 percent. Further, the gradability improved to as high as 97.4 percent.
Out of all the patients that were screened, more than 90 percent identified as positive by the EyeArt system for diabetic retinopathy or other eye diseases, as per the reference standard.
“Accurate, real-time diagnosis holds great promise for the millions of patients living with diabetes. In addition to increased accessibility, a prompt diagnosis made possible with AI means identifying those at risk of blindness and getting them in front of an ophthalmologist for treatment before it is too late,”. – Srinivas Sadda, MD of the Doheny Eye Institute/UCLA
This could be a huge development in the field of medicine, which can be used to detect this disease way ahead of time. ‘Prevention is better than cure’ and this could be one instance where a lot of people can be saved from a horrific side-effect of an already terrifying disease. Hopefully, the use of such technology can be used for making people’s lives better in the long run.
