(50-5) 10 * << * >> * Russian * English * Content * All Issues

Machine learning-based algorithms for identifying age-related macular degeneration biomarkers in OCT images
A.Yu. Ionov1, N.Yu. Ilyasova1, 2, D.V. Kirsh1, 2, K.R. Dotcenko3, A.V. Zolotarev3, 4, N.S. Demin1, 2, Е.Yu. Zubkova4

1 Samara National Research University, Moskovskoye Shosse 34, Samara, 443086, Russia;
2 Image Processing Systems Institute, NRC "Kurchatov Institute",
Molodogvardeyskaya St. 151, Samara, 443001, Russia;
3 Samara State Medical University, Chapaevskaya St. 89, Samara, 443099, Russia;
4 Samara Regional Clinical Ophthalmological Hospital named after T.I. Eroshevky,
Novo-Sadovaya St. 158, Samara, 443068, Russia

  Full text (PDF)

DOI: 10.18287/COJ2001

Article ID: 2001

Language: Russian

Abstract:
The aim of this study was to develop algorithms for the automatic detection of age-related macular degeneration biomarkers in retinal OCT images, aimed at optimizing the patient selection process for nanosecond retinal reparative therapy. Machine learning methods were used for retinal layer segmentation and soft drusen detection. Segmentation was performed using a U-Net convolutional network, which demonstrated high accuracy: the average IoU on the test set was 0.835. Four models were tested to classify atrophic areas, reticular drusen, and drusenoid detachments: QDA, KNN, logistic regression, and Random Forest. Random Forest demonstrated the best results, with an accuracy of 0.88 and an F1-score of up to 0.93, while precision and recall values remained stable across all classes. The algorithms developed in this study automatically identify retinal layers, atrophic zones, and soft drusen, classify pathological changes, and calculate their quantitative parameters: width, height, and area. This approach accelerates OCT image analysis, reduces the workload for physicians, and helps minimize diagnostic errors. Furthermore, the results open up opportunities for personalized therapy selection: using precise quantitative data, it is possible to evaluate the effectiveness of 2RT for a specific patient and predict disease progression.

Keywords:
neural networks, convolutional networks, segmentation, machine learning, age-related macular degeneration.

Acknowledgements:
The work was carried out within the state assignment of NRC “Kurchatov Institute” (theoretical part) and within the state assignment of Samara University, theme FSSS-2026-0003 (experiments and software implementation).

Citation:
Ionov AYu, Ilyasova NYu, Kirsh DV, Dotcenko KR, Zolotarev AV, Demin NS, Zubkova EYu. Machine learning-based algorithms for identifying age-related macular degeneration biomarkers in OCT images. Computer Optics 2026; 50(5): 2001. doi: 10.18287/COJ2001.

References:

  1. Fleckenstein M, Keenan TDL, Guymer RH, Chakravarthy U, Schmitz-Valckenberg S, Klaver CC, Wong WT, Chew EY. Age-related macular degeneration. Nature Reviews Disease Primers. 2021; 7(1): 31. DOI: 10.1038/s41572-021-00265-2.
  2. Lim LS, Mitchell P, Seddon JM, Holz FG, Wong TY. Age-related macular degeneration. The Lancet. 2012; 379(9827): 1728–1738. DOI: 10.1016/S0140-6736(12)60282-7.
  3. Budzinskaya MV. Age-related macular degeneration [In Russian]. Vestnik oftalmologii. 2014; 130(6): 56–61.
  4. Kizi BAN. Age-related macular degeneration of the retina. Academy. 2018; 8(35): 57–62.
  5. Deng Y, Qiao L, Du M, Qu C, Wan K, Li J, Jin Z, Yao X, Shen Y. Age-related macular degeneration: Epidemiology, genetics, pathophysiology, diagnosis, and targeted therapy. Genes & Diseases. 2022; 9(1): 62–79. DOI: 10.1016/j.gendis.2021.02.009.
  6. Russian Association of Ophthalmologists. Age-related macular degeneration. Clinical guidelines [In Russian]. 2024.
  7. Zolotarev AV. Diagnosis and timely treatment of AMD: 2RT nanosecond laser. Field of view [In Russian]. 2021; (5): 10-11.
  8. Querques G, Cicinelli MV, Rabiolo A, Corbelli E, Sacconi R, Bandello F. Laser photocoagulation as treatment of non-exudative age-related macular degeneration: state-of-the-art and future perspectives. Graefes Archive for Clinical and Experimental Ophthalmology. 2018; 256(1): 1-9. DOI: 10.1007/s00417-017-3848-x.
  9. Guymer RH, Wu Z, Hodgson LAB, Caruso E, Brassington KH, Tindill N, Aumann S, McGuinness MB, Fletcher EL, Finnegan S, Foubert E, Harper CA, Li H, Lee G, Matsumoto H, Tan Z, Essex RW, Robman LD. Subthreshold nanosecond laser intervention in age-related macular degeneration: The LEAD randomized controlled clinical trial. Ophthalmology. 2019; 126(6): 829-838. DOI: 10.1016/j.ophtha.2018.09.015.
  10. Mares V, Nehemy MB, Bogunovic H, Frank S, Reiter GS, Schmidt-Erfurth U. AI-based support for optical coherence tomography in age-related macular degeneration. International Journal of Retina and Vitreous. 2024; 10(1): 31. DOI: 10.1186/s40942-024-00549-1.
  11. Muntean GA, Marginean A, Groza A, Militaru M, Olteanu IA, Munteanu A, Balta F. The predictive capabilities of artificial intelligence-based OCT analysis for age-related macular degeneration progression – a systematic review. Diagnostics. 2023; 13(14): 2464. DOI: 10.3390/diagnostics13071234.
  12. Tanthanathewin R, Wongrattanapipat W, Khaing TT, Panjaphongse R, Moongfangklang N, Chinjarernpan P. Automatic exudate and aneurysm segmentation in OCT images using UNET++ and hyperreflective-foci feature based bagged tree ensemble. PLOS ONE. 2024; 19(5): e0304146.
  13. Karn PK, Abdulla WH. Advancing ocular imaging: a hybrid attention mechanism-based U-Net model for precise segmentation of sub-retinal layers in OCT images. Bioengineering. 2024; 11(3): 240.
  14. Sousa JA, Paiva A, Silva A, Costa M, Alves R, Alves J, Azevedo E, Alves S. Automatic segmentation of retinal layers in OCT images with intermediate age-related macular degeneration using U-Net and DexiNed. PLOS ONE. 2021; 16(5): e0251591.
  15. Wang B, Wei W, Qiu S, Wang S, Li D, He H. Boundary aware U-Net for retinal layers segmentation in optical coherence tomography images. IEEE Journal of Biomedical and Health Informatics. 2021; 25(8): 3029–3040.
  16. Du G, Cao X, Liang J, Chen X, Zhan Y, Li K, Zhang L. Medical image segmentation based on U-Net: a review. J Imaging Sci Technol. 2020; 64(2): 20508-1–20508-15.

151, Molodogvardeiskaya str., Samara, 443001, Russia; E-mail: journal@computeroptics.ru; Tel: +7 (846) 242-41-24 (Executive secretary), +7 (846) 332-56-22 (Issuing editor), Fax: +7 (846) 332-56-20