Comparison of different Raman spectrometer models in the quantification of blood serum analytes.

Abstract

The analytical performance of three commercial Raman spectrometers in determining the concentrations of analytes associated with diagnosing metabolic, cardiovascular, and renal disorders has been compared. Two portable and one benchtop spectrometers were tested to predict the serum concentrations of triglycerides (TRI), cholesterol (COL), high-density cholesterol, creatinine, urea, and glucose in 193 serum samples using partial least squares (PLS) regression and PLS discriminant analysis, the latter for classifying samples as either altered or within reference values. Strong correlations (r > 0.81) were obtained for TRI and COL analytes using the benchtop and one of the portable spectrometers, and the classification accuracy rates exceeded 90%, suggesting potential for use in clinical screening. Adding a glucose solution improved the correlation and the root mean square error for TRI and COL analytes; however, it did not enhance the correlation or error for the glucose analyte. Raman spectroscopy showed potential to support routine laboratory activities and may have applications in clinical screening.