Optical processing of light-induced autofluorescence for characterization of tissue pathology

We describe an optical processing method for characterizing tissue pathology that is based on principal-component analysis of light-induced autofluorescence. A set of optical spectral filters, which are related to the principal-component loading vectors, is designed to process the autofluorescence signal optically and to generate principal-component scores from the autofluorescence spectra. The scores are then correlated with the tissue pathology. An optical processing system is designed that uses the in vivo fluorescence spectra recorded from nasopharyngeal tissues. We demonstrate that the system can differentiate nasopharyngeal carcinoma from normal tissue with a high degree of sensitivity and specificity and that the optical filters used in the system can be manufactured.