Predicting the future development of diabetic retinopathy using a deep learning algorithm for the analysis of non-invasive retinal imaging

Abstract Diabetic retinopathy (DR) is the most common cause of vision loss in the working age. While over 90% of sight-threatening cases may be treated if detected early, prevalence of yearly detective screening is low until advanced presentation of the disease. We developed a machine learning algorithm for the prediction of future DR development using fundus photography of otherwise healthy eyes. Our algorithm achieves 0.81 Area Under Receiver Operating Curve (AUC) averaging scores from multiple images on the task of predicting development of referrable DR, and 0.76 AUC when using a single image. In conclusion, risk of DR may be predicted from fundus photography alone. Prediction of personalized risk of DR may become key in treatment and contribute to patient compliance across the board. Further prospective research is necessary.