Abstract
Presentation Description :
Clinical trials particularly in slowly progressive disease or pathologies with a wide variability represent a major challenge for investigators as well as sponsors. Artificial intelligence (AI) has the capacity to identify risk profiles and prognostic biomarkers which can be used to characterize study populations in a reliable and stringent manner a priori. Multimodal imaging including high-resolution three-dimensional OCT offers a wealth of morphological features which can be exploited comprehensively using deep learning. Such analyses have been successfully applied to understand the risk of conversion from early to advanced age-related macular degeneration (AMD), prognostic outcomes in neovascular AMD and diabetic macular edema (DME) as well as geographic atrophy (GA). AI provides defined and quantified markers reflecting disease activity beyond classic clinical features such as drusen volume and hyperreflective foci thereby largely expanding the conventional spectrum of pathognomonic markers. This quality enables the search for novel therapeutic targets and furthers insights into the pathophysiology of disease.
This abstract was presented at the 2019 ARVO Annual Meeting, held in Vancouver, Canada, April 28 - May 2, 2019.