Discriminant analysis for classification of colonic tissue autofluorescence spectra

摘要

This study evaluates the potential of a discriminant analysis to classify colonic mucosa from autofluorescence spectral characteristics. With 337 nm excitation, the autofluorescence spectra of colonic tissues were measured using a FLS920 spectrofluorimeter. Principal component analysis (PCA) combined with Fisher's discriminant analysis was performed for tissue classification. As a result, the sensitivity and specificity of the discriminant analysis is 92.3% and 90.5%, respectively. The results suggest the relative concentrations of collagen and nicotinamide adenine dinucleotide (NADH) are the potential diagnostic biomarkers for colonic tissue classification using autofluorescence spectroscopy, and the discriminant analysis based on PCA is useful to differentiate adenocarcinoma from normal tissue.