1158-P: Type 2 Diabetes Patients Subtype Analysis: Unbiased Machine Learning for Future Personalized T2D Care

Abstract

Dorzagliatin (HMS5552) that achieved POC in a recent Ph II trial, demonstrated its potential to treat type 2 diabetes (T2D) patients through a personalized medicine approach in a biomarker guided patient selection study. We here by conducted a data-driven unsupervised clustering analysis on the full analysis set (221 patients) of Dorzagliatin Ph II trail using clinically relevant biomarkers, which generated six clusters. An ANOVA test was done for those variables between clusters, and 40 of them are significantly different (P < 0.05). The study showed that each cluster possesses unique characters related to disease stage of T2D and the stage is progressing to one direction which starts from mild to severe loss of pancreas β-cell function accompany with progressively increase of insulin resistance. Cluster 1 shared similar feature as those in previous Ph Ic study, in which, QD dose of Dorzagliatin offered better response rate of 100%, while in Ph II based cluster, the 75mg QD dose group achieved 1.88% reduction in HbA1c from baseline with the 75mg BID offered HbA1c reduction of 0.95%. However, in Cluster 2 and 4, there are clear trends that BID is more effective than QD in glycemic control as well as improving β-cell function and reduction of insulin resistance. In Cluster 4, it is observed FPG reduction of 1.57mmol/L, PPG reduced of 2.73mmol/L, HOMA-β improvement of 12.22, and HbA1c reduction of 1.69% in 75mg BID dose group, with corresponding changes in 75 QD group of 0.35 mmol/L, 1.7mmol/L, 9.24 and 0.26% respectively adjusted by placebo. The results suggested that the AI methodology is able to differentiate T2D patients into different subclass according to the severity of the disease state, and 75mg BID Dorzagliatin seems offered a minimum therapeutic effective dose (MTED) to all patient population, and 75mg QD regiment is a preferred option for the early stage T2D condition. Disclosure L. Chen: Employee; Self; Hua Medicine. L. Feng: Employee; Self; Hua Medicine. C. Tang: Employee; Self; Hua Medicine. Y. Zhang: Employee; Self; Hua Medicine. Funding Hua Medicine