Abstract

Digital twins can offer personalised models of human tissue by combining imaging data with advanced computational methods. In this talk, I will describe our efforts to build digital twins of the retina, capturing the structure and function of its vascular networks and how these change in disease. By using deep generative learning, we can personalise these models to individual patients, even when imaging data are limited, creating simulations that reflect patient-specific anatomy and physiology. The retina is particularly valuable as a testbed as pathological angiogenesis in diabetic retinopathy shares many features with the abnormal blood vessel growth seen in cancer. This makes retinal digital twins a powerful prototype for understanding cancer angiogenesis and for developing new approaches to predicting treatment response.