Rapid detection of kidney disease based on urine surface-enhanced Raman spectroscopy and principal components...

Immunoglobulin A nephropathy (IgAN) and idiopathic membranous nephropathy (IMN) are the most prevalent primary glomerulonephritis (PGN) subtypes and can lead to end-stage renal disease. Conventional diagnostic methods in certain aspects are often limited by time-consuming, low-sensitivity, and invasive. Thus, developing a rapid, non-invasive, and high-sensitivity PGN diagnostic method is essential. This study proposes a novel method utilizing urinary SERS spectra for the diagnosis of PGN/HC and the classification of IgAN/IMN. Firstly, the Mann-Whitney U tests confirmed significant spectral differences between groups. Subsequently, PCA-SVM and PCA-RF models were established. Finally, leave-one-out-cross-validation (LOOCV) assessment showed PCA-SVM outperformed PCA-RF in PGN/HC classification in accuracy (98.36 % vs 96.72 %) and specificity (100 % vs 90.16 %), but underperformed in IgAN/IMN classification in accuracy (72.95 % vs 68.85 %) and sensitivity (83.61 % vs 75.41 %). The results indicate that measuring urine with SERS can more effectively distinguish PGN/HC and IgAN/IMN. Our findings are helpful for clinicians to diagnose PGN patients and monitor the progress and recurrence of the disease.