‘Will soon emerge as core application technology of AI-based medical service’

At a time when smart healthcare technology is catching on as a new growth business, two local institutions have joined hands to develop “AI-aided early diagnosis technology of glaucoma,” officials said Wednesday.

The operators of the national health examination program have used funduscopic tests to sort out diabetic retinopathy. If they can analyze the slight changes in photographed images, however, early detection of glaucoma is also possible, they said.

Konyang University Hospital in Nonsan, South Chungcheong Province, will develop AI-aided diagnosis technology of glaucoma, jointly with Electronic and Telecommunications Research Institute.

It is against this backdrop Konyang University Hospital and Electronics and Telecommunications Research Institute (ETRI) decided to jointly develop a medical device loaded with an artificial intelligence function that allows for the “screening test of early-phase glaucoma,” the university said in a press release.

The Ministry of Science, ICT and Future Planning has designated it as one of its primary research tasks and decided to provide 840 million won ($74,800) to help finance the project over the next two years.

Ocular fundus, or the eyeground, is the only body part whose blood vessels can be observed without irradiation or invasive treatment. That means doctors may be able to diagnose chronic diseases, such as hypertension and diabetes if only they can analyze fundus image through deep neural networks, university officials said.

The university hospital will be responsible for establishing medical image big data related to glaucoma diagnosis for machine learning, preliminary clinical experiment, and the evaluation of model performance. The state agency will develop fundus image processing algorithm and deep neural network model for glaucoma diagnosis. Afterward, they will evaluate the reading model’s performance by using funduscopic pictures of glaucoma patients and normal people.

If and when the research proves successful, the institutions expect to upgrade the national health screening program through diagnosing glaucoma by funduscopic tests, while applying the technology to developing technology related to detecting not only various eye problems but systemic diseases.

“We expect the funduscopic camera loaded with artificial intelligence function to replace the existing filming equipment not just in Korea but overseas markets as well,” said Cho Hyun-sung Cho, a senior researcher at ETRI.

Professor Kim Jong-yeop of Konyang University Hospital also emphasized the importance of the project. “Glaucoma is one of the eye diseases that can cause blindness as well as cataracts and macular deafness,” he said. “We hope the research will help to reduce indirect social costs by detecting it earlier and easier.”

Other experts expected a more advanced image preprocessing technology would be necessary for the diagnosis of glaucoma than the diagnosis of diabetic retinopathy in fundus image. If the research team can accumulate the know-how of image preprocessing, they will be able to present additional ways of utilizing the technology, the experts added.

Copyright © KBR Unauthorized reproduction, redistribution prohibited