Progression of early diagnostic markers for diabetic kidney disease: From single-indicator detection to multi-omics integration modeling

Diabetic kidney disease (DKD) is the leading cause of end-stage renal disease (ESRD), but early diagnosis remains challenging. Current reliance on indicators such as urinary albumin and glomerular filtration rate is limited by insufficient sensitivity and susceptibility to interference. The pathogenesis of DKD is complex, involving multiple intersecting pathways and epigenetic modifications in the progression of the disease. Omics biomarkers offer new directions for early risk assessment, while machine learning can integrate multi-omics data to build efficient diagnostic models. This article reviews recent advances in biomarker research, discusses strategies for model construction, and provides theoretical references for early and precise diagnosis and clinical translation of DKD.