The latest global medical statics have been expected an ongoing increase in the prevalence of glaucoma as it is rapidly becoming an ageing society. World Health Organisation (WHO) also mentioned glaucoma as one of the three major eye diseases, causing blindness. It is vital to have proper treatment through early medical check-up since glaucoma does not have any obvious symptoms until the late-stage.
As a most eye-catching method, the fundus image enables the ophthalmologist to observe human’s vascular, nerve, retina, etc and detect ophthalmologic diseases in a non-invasive, simple, inexpensive way.
However, the real problem with reading the fundus image is an insufficient number of highly skilled specialists who can detect minute lesions accurately, causing various difficulties in the accuracy of medical check and result. In particular, patients with chronic diseases such as diabetes and high blood pressure have to have a medical check-up of the ophthalmology as well as internal medicine or family medicine which is an inconvenience.
Additionally, since it is found that there is a gap in the diagnosis results of the fundus image by each ophthalmologist, the report on the introduction of AI (Artificial Intelligent) technology for the improvement of the accuracy and consistency in reading the fundus images has recently been announced.
Glaucoma is a disease that can lead to blindness by damaging the Retinal Ganglion Cell (RGC) and its axon and makes patients hardly feel the symptoms such as vision loss until the late-stage. However, due to chronic and irreversible characteristics of glaucoma, if the glaucoma is early discovered, the risk can be slow down through treatment or surgery. In other words, the early detection of glaucoma is important.
Thus, the Korean R&D institute has developed the technology that automatically reads safe images using Convolutional Neural Network for glaucoma diagnosis in an inexpensive, easy way. It is a low-cost diagnostic test with fundus images, accessible to medical devices.
The technology is also able to distinguish between early-stage glaucoma and pre-perimetric open-angle glaucoma and provides deep learning model which can easily check the progress of glaucoma through comparison with the past check-up history.
Its algorithm distinguishes between normal or glaucoma through the model that learned the data obtained by fundus images using CNN, which is the key point.
If it is feasible to read fundus images automatically using AI, it will be able to detect various diseases in early-stage, helping in the decrease in medical expenses and the improvement of the patients’ satisfaction. Plus, it will contribute to reduce the rate of misdiagnosis and improve the quality of medical services due to the decrease in healthcare labour.
There are also more economical opportunities as the development of the new concept of medical devices such as automated glaucoma reading software connected with a digital fundus imaging system or medical device embedded cloud-based automated glaucoma reading, leading to new business development.
The Korean R&D institute is now looking for a global partner interested in deploying this technology on its own platform or developing a new type of glaucoma diagnosis medical device under a license agreement, technical cooperation and commercial agreement with technical assistance.
