The use of artificial intelligence in the diagnosis and monitoring of retinal diseases: analysis of clinical effectiveness and prospects for implementation into domestic protocols (a review)
Ophthalmology Reports · 22.10.2026 · 19 (3) · 75-85
DOI: 10.17816/ov706112 ↗Kurz gefasst
Fachartikel in der Zeitschrift Ophthalmology Reports, online erschienen am 22.10.2026. Verfasst von Anastasia N. Koltakova und 11 weiteren Autor:innen. Die Arbeit nennt 54 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC-ND 4.0).
Abstract (Originalsprache)
Age-related macular degeneration and diabetic retinopathy remain the leading causes of preventable vision loss worldwide. In Russia, the availability of ophthalmologists is 18 specialists per 100,000 population, which is substantially below the World Health Organization recommendation of 55 per 100,000 population. Given the rising prevalence of diabetes mellitus, this circumstance requires the introduction of automated screening systems. Results of meta-analyses indicate high diagnostic accuracy of artificial intelligence models: sensitivity for detecting diabetic retinopathy ranges from 87% to 95%, specificity from 87% to 94%, and the area under the ROC curve reaches 0.98. The U.S. Food and Drug Administration has approved autonomous systems for diabetic retinopathy screening (IDx-DR, EyeArt, AEYE-DS), as well as platforms for remote monitoring of age-related macular degeneration based on home optical coherence tomography—SCANLY Home OCT. The use of AI enables the selection of only 10%–15% of images for manual assessment by specialists, thereby significantly saving specialists’ working time. The application of home optical coherence tomography for remote monitoring of patients with age-related macular degeneration increases the interval between injections from 8.0 to 15.3 weeks (p = 0.004). The main barriers to the implementation of artificial intelligence into clinical practice include insufficient personnel training (20.5%), high implementation costs (16.5%), and the risk of algorithmic bias (44.2%). Successful integration of these technologies into domestic clinical protocols requires the development of educational programs for physicians, validation of artificial intelligence systems on Russian population-based data, the formation of a regulatory framework, the creation of insurance reimbursement mechanisms, as well as the inclusion of issues of AI application in the curricula of medical universities.
Zitieren
APA
Koltakova, A. N., Raminskaya, A. A., Grebenyuk, A. I., Bagomedova, Z. R., Plotko, A. I., Gubadova, G. C., Gnezdilova, M. V., Gartvikh, E. V., Kikova, A. A., Klycheva, A. R., Uzdenova, L. A., & Mukha, I. V. (2026). The use of artificial intelligence in the diagnosis and monitoring of retinal diseases: analysis of clinical effectiveness and prospects for implementation into domestic protocols (a review). Ophthalmology Reports, 19(3), 75-85. https://doi.org/10.17816/ov706112
Vancouver
Koltakova AN, Raminskaya AA, Grebenyuk AI, Bagomedova ZR, Plotko AI, Gubadova GC, et al. The use of artificial intelligence in the diagnosis and monitoring of retinal diseases: analysis of clinical effectiveness and prospects for implementation into domestic protocols (a review). Ophthalmology Reports. 2026;19(3):75-85. doi:10.17816/ov706112
BibTeX
@article{Koltakova2026,
author = {Koltakova, Anastasia N. and Raminskaya, Anna A. and Grebenyuk, Anna I. and Bagomedova, Zaira R. and Plotko, Anastasia I. and Gubadova, Giuler Ch. and Gnezdilova, Maria V. and Gartvikh, Ekaterina V. and Kikova, Aza A. and Klycheva, Amina R. and Uzdenova, Larisa A. and Mukha, Ivan V.},
title = {The use of artificial intelligence in the diagnosis and monitoring of retinal diseases: analysis of clinical effectiveness and prospects for implementation into domestic protocols (a review)},
journal = {Ophthalmology Reports},
year = {2026},
volume = {19},
number = {3},
pages = {75--85},
doi = {10.17816/ov706112}
}
EUVIMED