Detection of urinary SERPINA4 by electrochemiluminescence immunoassay and development of a diagnostic model for diabetic nephropathy.

This article focuses on the development and evaluation of an electrochemiluminescence immunoassay (ECLIA) for detecting urinary SERPINA4, a protein investigated as a diagnostic biomarker for diabetic nephropathy (DN). The study established that the ECLIA method for SERPINA4 measurement exhibits high precision and a broad linear detection range, outperforming traditional ELISA techniques. Using data from 98 patients with diabetic kidney disease, a diagnostic model incorporating SERPINA4 normalized to urinary creatinine (SERPINA4/UCr) alongside clinical indicators was developed via the Random Forest algorithm, achieving an area under the curve (AUC) of 0.89, 90% accuracy, 100% sensitivity, and 70% specificity in differentiating DN from non-diabetic renal disease (NDRD). Key variables influencing the model included serum creatinine, microalbuminuria, SERPINA4/UCr ratio, systolic blood pressure, and total urine protein. The study suggests that ECLIA-based SERPINA4 detection combined with machine learning models holds promise for improving non-invasive differential diagnosis of DN, though further validation with larger, multi-center cohorts is needed.