RESUMO
Pulsed laser-induced-fluorescence studies of pathologically certified oral tissues are carried out at different excitations and time delays. Among the several excitations used, 325 nm produced noticeably different spectral profile for normal and malignant tissues. Extensive curve analysis was carried out in order to understand changes in biochemical composition of tissue based on spectral profiles. Curve resolution and principal component analysis (PCA) show that the fluorescence intensity changes from normal to malignant tissue samples are not completely explained in terms of simple collagen and NAD(P)H intensity changes. The spectra require at least five components to be fully accounted for. Several discrimination methodologies based on PCA and intensity differences between different emission peaks (resultant peaks of curve analysis) were also evaluated. The results obtained indicate PCA using Mahalanobis distance and spectral residual as discrimination parameters provides best discrimination and can be used for matching unknown samples to standard calibration sets. Intensity ratio of bound NAD(P)H to collagen seems to be more suitable for discrimination between normal and malignant oral tissue, compared to ratio of collagen to total intensity of all the other components together.
