Please login first
Comparative Evaluation of MIR and NIR Spectroscopy for Monitoring Kidney Disease Biomarkers
* 1 , 1, 2 , 2 , 1
1  iBB – Institute for Bioengineering and Biosciences, i4HB – The Associate Laboratory Institute for Health and Bioeconomy, IST – Instituto Superior Técnico, Universidade de Lisboa, Lisbon, Portugal
2  ISEL- Instituto Superior de Engenharia de Lisboa, Instituto Politécnico de Lisboa, Lisbon, Portugal
Academic Editor: Shuhai Lin

Abstract:

Introduction: Chronic kidney disease (CKD) represents a major global health challenge, often remaining undetected until advanced stages due to limitations of current biomarkers, such as blood creatinine, urea, and albumin. Fourier-transform infrared spectroscopy (FTIR), particularly in the mid-infrared (MIR) and near-infrared (NIR) regions, has emerged as a promising alternative for monitoring biochemical alterations associated with CKD, since it can be performed with a simple drop of blood. This work compares the predictive potential of MIR and NIR spectroscopy to assess serum levels of creatinine, urea, and albumin in simulated CKD conditions. Methods: A total of 54 serum-based solutions were prepared, reflecting different CKD stages (0–5). The MIR and NIR spectra of these solutions were pre-processed with baseline correction, normalization, and Savitzky–Golay filters, followed by the development of partial least squares (PLS) regression models to predict the target analyte concentration and discriminant analysis (DA) models to classify “healthy” versus “diseased” and to differentiate CKD stages. Results: PLS regression models showed strong predictive ability for urea concentration using both MIR (R2 = 0.97, RMSE = 10 mg/dL) and NIR spectra (R2 = 0.91, RMSE = 17 mg/dL). Albumin concentration was more reliably predicted by NIR spectra (R2 = 0.93, RMSE = 0.11 g/dL), while creatinine prediction had the best results obtained using the third Overtone NIR spectral sub-region (R2 = 0.91, RMSE = 0.48 mg/dL). PCA-DA models confirmed the capacity of MIR and NIR to classify samples as “healthy” versus “diseased” with accuracies > 90%, although stage classification remained more challenging. Conclusions: Both MIR and NIR spectroscopies present strong potential as economic, rapid and minimally invasive tools for monitoring kidney function-related biomarkers like creatinine, urea and albumin. These findings support the integration of FTIR-based platforms in point-of-care diagnostics for the early detection and monitoring of CKD progression.

Keywords: Kidney Disease; FTIR-Spectroscopy; Biomarkers; Monitoring
Top