Detection of Oral Cancer Using the Fluorescence Spectroscopy and Classification of Different Stages of Cancer by Multivariate Analysis

Abstract

Fluorescence spectroscopy is used as a diagnostic tool for the detection of oral cancer in the present study. For the fluorescence measurement, an excitation wavelength of 350 nm is used. Two diagnostic media, namely human tissue and saliva are incorporated in this study. Measurements are accomplished on oral squamous cell carcinoma (OSCC), dysplastic, and normal tissue and saliva samples. Fluorescence spectra obtained from human oral tissue consists of major bands of collagen and NADH near 390 and 445 nm. However, saliva shows only one major band of NADH near 440 nm. Multivariate analysis has been employed on the fluorescence data of human oral tissue and saliva for the classification of different stages of cancer. In the multivariate analysis, principal component analysis (PCA), Mahalanobis distance (MD) model, and receiver operating characteristic (ROC) analysis are utilized. Fluorescence spectroscopy on human oral tissue and saliva is competent to differentiate OSCC to normal, dysplasia to normal and OSCC to dysplasia with overall accuracies of 90%, 83%, 80% and 100%, 86%, 85% respectively. Obtained results using the fluorescence spectroscopy on human saliva are comparable to human tissue samples. Results imply that we may make use of saliva as a non-invasive diagnostic medium for the detection of oral cancer and multivariate analysis can be employed as a classification tool.