FT-IR spectrometry has proved to be a useful tool for determining a series of plasma molecular concentrations. Dedicated experiments were first performed to test the analytical performance that could be obtained by FT-IR spectrometry using a synthesized N3-peptide exhibiting a -N3 absorption centered at 2110 cm(-1), a spectral region where no organic material of biological samples absorbs. Further, we investigated whether this technology was able to allow quantification of metabolic parameters (glucose and lactic acid) within plasma, cells, and tissues as an alternative method to the "classical" biochemical approaches, which require sophisticated biological material treatment and expensive reagents. For this purpose we used a series of plasma samples to determine glucose and lactic acid concentrations, which are common markers of cancer growth. We compared the results of the main spectral data treatments commonly achieved for FT-IR data analysis, such as univariate (Beer-Lambert) or multivariate (PLS) calibrations, as well as the deconvolution of the spectral interval of interest (1200-900 cm(-1)). No significant differences were found regarding the analytical performances of these methods. Spectral deconvolution was finally undertaken on cultured and on xenografted cells (U87 glial cells implied in human gliomas) to determine glucose and lactic acid concentrations. In this case, qualification was allowed by FT-IR imaging on the cellular models since biochemical approaches are not efficient to reach metabolic concentrations at the cellular level while keeping tissue organization.