OphtAI uses AI to screen for eye disease in people with diabetes

[ad_1]

OphtAI won the innovation prize on December 4, 2019 at the 12th Congress of the French Society of Digital Health. The French joint venture, which has developed diabetic retinopathy (RD) screening software, will benefit from a support program delivered by the National Union of the Medical Technology Industry and the Medicen Paris Region competitiveness cluster. Objective: detect eye damage on the retina caused by excess blood sugar that causes permanent loss of vision, through the use of artificial intelligence (AI).

Two photos are enough to detect lesions

With only two photos of each eye taken by an orthoptist, the tool is able to tell if a diabetic person has diabetic retinopathy by detecting eye damage on the retina. If the machine detects a large number of lesions, the person must go to an ophthalmologist for treatment, because moderate diabetic retinopathy can be treated. This device makes it possible to cope with the shortage of ophthalmologists in France to which is added the constant increase in people with diabetes.

The fruit of a long collaboration

The creation of this software is the result of a long collaboration. Doctors and researchers from the Assistance Publique-Hôpitaux com Paris (AP-HP) created a telemedicine network called "OphDiaT" in 2004. Orthoptists took pictures of the retina with a retinograph and then ophthalmologists remotely analyzed cliches. "But there were not enough doctors to perform this task", says retinologist Ali Erginay of Lariboisière hospital in Paris, contacted by The Digital Factory.

From there came the idea to delegate this task to a machine. After several projects, the software is put in place. And it is Evolucare, software publisher, and ADCIS, SME specializing in image processing, who are coming together to create a joint venture in May 2019 and thus market the system.

Detect a large number of eye diseases

"A first version of the software was CE marked in May 2019", specifies Alexandre Le Guilcher, CEO of OphtAI, interviewed by The Digital Factory. But the development of this tool is still in its infancy. "Thanks to deep learning, we can machine learn to detect other diseases"explains Dr. Ali Erginay. A new version of the software should make it possible to detect the level of severity of diabetic retinopathy and thus its level of urgency."We also added two diseases to screen for: glaucoma and age-related macular degeneration (AMD).", specifies Alexandre Le Guilcher. Finally, a third version of the device is in preparation. It will be able, this time, to detect"41 different risks" and so "cover 90% of patients at risk".

"Imagine that in a station, we replace the photo booths with cameras capable of taking pictures of the retina and you send the results by email", projects Dr. Ali Erginay. So only sick people will have to see a specialist. This system lowers the costs associated with a consultation with an ophthalmologist and allows as many people as possible to be followed regularly. OphDiaT telemedicine network still exists within the AP-HP and has not yet been replaced by software. "40 centers participate in this screening and treat around 2,000 patients per month ", says retinologist Ali Erginay.

In France, diabetic retinopathy (RD) is the leading cause of blindness before age 65. This eye disease, which affects 50% of type 2 diabetic patients.

[ad_2]