Clinical Report: Tongue Phenotyping Reveals Variability in Diabetic Kidney Disease
Overview
This study identifies two distinct tongue-phenotype subtypes in patients with diabetic kidney disease (DKD) through unsupervised clustering of tongue features.
Background
Diabetic kidney disease (DKD) is a prevalent complication of diabetes and a leading cause of end-stage renal disease. The heterogeneity of DKD complicates treatment and management, necessitating refined stratification methods. Recent advances in machine learning and clustering techniques offer potential for identifying clinically relevant subtypes in DKD.
Data Highlights
Cluster
Number of Patients
Key Features
1
108 (32.6%)
Darker tongue color, higher saturation, more yellow coating
2
223 (67.4%)
Higher brightness, lower saturation, thicker coating, more tooth marks
Key Findings
Two clusters of tongue phenotypes were identified in DKD patients: Cluster 1 and Cluster 2.
Cluster 2 exhibited higher brightness and lower saturation compared to Cluster 1.
Significant differences in coating thickness and tooth marks were observed between the two clusters.
No significant differences in laboratory or composite indices were found after correction for multiple testing.
The clustering methods showed high concordance, with an adjusted Rand index of 0.715.
Clinical Implications
The identification of tongue-phenotype subtypes in DKD may provide additional data for understanding patient variability.
Conclusion
Objective tongue phenotyping reveals significant heterogeneity in DKD that is not captured by traditional laboratory profiles.
In a phase 2b trial, the highest dose produced an estimated 12% mean weight reduction vs 1% with placebo, although gastrointestinal adverse events were common.
