Filtering Raman spectral features of glial tumor sites based on biochemical correlates

Raman spectroscopy is a sensitive and fast tool for analyzing the molecular composition of tissues, so we used it to differentiate glial tumor and normal tissues during surgery. The structure of Raman spectra motivated us to implement the feature selection technique, based on the filtering of those points of the spectrum for which statistically significant differences were found between the groups. Finally they were matched to the main biochemical components of the studied tissues, and new vectors of Raman features of a smaller dimension were formed.