State of the art of glaucoma diagnosis:
neural networks and artificial intelligence

A.V. Kuroyedov^1,2, G.A. Ostapenko³, K.V. Mitroshina², A.B. Movsisyan^2,4

¹Mandryka Central Military Clinical Hospital, Moscow, Russian Federation

²Pirogov Russian National Research Medical University, Moscow, Russian Federation

³Voronezh State Technical University, Voronezh, Russian Federation

⁴Hospital for Disabled Veterans No. 2, Moscow, Russian Federation

Predicting of the borderline between health and disease as well as determining of minimum characteristics of disease progression in primary open-angle glaucoma still remain high priority for ophthalmologists worldwide. Machine diagnostic search continues and improves along with the development of computers. Modern trends in this search shift towards artificial intelligence (AI) technologies. This shift is determined by present and future modalities aimed to improve the whole system of glaucoma management. The idea of ​​symbiosis of technology for early detection of a disease and a rational diagnostic algorithm is not new. This concept originated many years ago when studying the processes occurring in the brain and when trying to model these processes. Moreover, the use of the results of traditional diagnostic methods, which became the basis for databases of AI technology, is still in demand. The specificity of modern AI is a new technological level due to the introduction of novel algorithms to automatize eye disease diagnostics including glaucoma screening based on color eye fundus image, optical coherence tomography, and visual field test. Automated data extraction allows AI network to learn difficult functions of direct achievement of final results independently of subjective clinician’s opinion. Ability for independent data extraction is particularly important considering growing data volume and emerging novel areas of machine teaching application.

Keywords: primary open-angle glauco ma, artificial intelligence, neural networks, intraocular pressure, structural diagnostics, static automated perimetry.

For citation: Kuroyedov A.V., Ostapenko G.A., Mitroshina K.V., Movsisyan A.B. State of the art of glaucoma diagnosis: neural networks and artificial intelligence. Russian Journal of Clinical Ophthalmology. 2019;19(4):230–237.