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

Abstract

AbstractAimsDiabetic retinopathy (DR) is the most common cause of vision loss in the working age. This research aimed to develop a machine learning model which can predict the development of referrable DR from fundus imagery of otherwise healthy eyes.MethodsOur researchers trained a machine learning algorithm on the EyePacs dataset, consisting of 156,363 fundus images. Referrable DR was defined as any level above mild on the International Clinical Diabetic Retinopathy scale.ResultsThe algorithm achieved 0.81 Area Under Receiver Operating Curve (AUC) when averaging scores from multiple images on the task of predicting development of referrable DR, and 0.76 AUC when using a single image.ConclusionOur results suggest that 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, particularly when supported by further prospective research.SynopsisThe deep learning algorithm our researchers developed was able to predict the development of referrable diabetic retinopathy in diabetic patients with otherwise healthy eyes with 0.81 AUC.