NLM DIR Seminar Schedule

Abstract:

Age-related macular degeneration (AMD) is a progressive irreversible neurodegenerative disease that affects the macula and causes central vision loss in the elderly in developed countries. AMD is predicted to affect more than 288 million people worldwide by 2040. Therefore, early detection of AMD is very crucial for early intervention to slow down the progression of AMD. Optical coherence tomography (OCT) has become an established diagnostic technology in the clinical management of eye diseases, as it provides details about the retinal layers and choroid. OCT can be used to detect AMD features at different stages, so it is very important for clinical management of AMD. However, OCT provides volumetric images of the eye, so 3D processing of the data is necessary to capture contextual information. For this purpose, we have successfully developed state-of-the-art artificial intelligence (AI) models for detecting different features of AMD, including geographic atrophy (GA); the primary lesion in late atrophic AMD, and reticular pseudodrusen (RPD); lesions that happen frequently at intermediate AMD. The developed models comprised 3D convolutional neural networks to process the volumetric OCT data.