The role of artificial intelligence and machine learning technology in the diagnosis of age-related macular degeneration

Background. According to the Beaver Dam Eye Study, the prevalence of advanced age-related macular degeneration (AMD) reaches 1.6 % among patients aged 43 to 86 years. In this context, automating the diagnostic process through the use of artificial intelligence (AI) and machine learning is a promising avenue. This study explores the potential of AI models for analyzing fundus images to detect drusen in AMD. The purpose was to investigate the feasibility and effectiveness of using artificial intelligence, specifically the YOLOv10 model, to analyze fundus camera images for AMD diagnosis. Materials and methods. The study was conducted on 200 fundus images from the publicly available RFMiD dataset. The YOLOv10 model was adapted to detect drusen and changes characteristic of AMD. Results. The YOLOv10 model demonstrated high accuracy and specificity in detecting drusen and degenerative changes, such as pigmentary abnormalities and geographic atrophy. The study provides statistical performance metrics of the model. Conclusions. The application of YOLOv10 in AMD diagnosis opens new possibilities for automation and improved diagnostic accuracy, representing a promising direction for implementation in ophthalmic practice.