Machine learning for the diagnosis of early-stage diabetes using temporal glucose profiles

Abstract

Machine learning shows remarkable success for recognizing patterns in data. Here, we apply machine learning (ML) for the diagnosis of early-stage diabetes, which is known as a challenging task in medicine. Blood glucose levels are tightly regulated by two counter-regulatory hormones, insulin and glucagon, and the failure of glucose homeostasis leads to a common metabolic disease, diabetes mellitus. It is a chronic disease that has a long latent period that complicates detection of the disease at an early stage. The vast majority of diabetes cases result from that diminished effectiveness of insulin action, and that insulin resistance modifies the temporal profile of blood glucose. Thus, we propose to use ML to detect subtle changes in the temporal pattern of the glucose concentration. Time series data on blood glucose with sufficient resolution is currently unavailable, so we confirm the proposal by using synthetic glucose profiles produced using a biophysical model that considers glucose regulation and hormone action. Multi-layered perceptrons, convolutional neural networks, and recurrent neural networks all identified the degree of insulin resistance with high accuracy above 85%.