Infrared spectroscopy in clinical chemistry, using chemometric calibration techniques

The aim of our studies was to further develop the assays of fecal fat and urinary calculi. The development consisted of the investigation of the applicability of new infrared spectroscopic methods for routine use in the clinical laboratory. Because most of these assays made use of authentic sample material, quantification of the analyte concentrations was often hampered, because of the complex sample matrices. Therefore, we also investigated the application of chemometrical methods for quantification of the analyte concentrations from the spectral results. We applied artificial neural networks and partial least-squares regression analysis for both calibration and prediction of the outcome of both kinds of assays. Furthermore, we gave some information about the pathophysiology background of our studies.