Test of label-free Nasopharyngeal carinoma tissue at different stages by Raman spectroscopy

Abstract

Raman spectroscopy (RS) of Nasopharyngeal carcinoma (NPC) tissue contained various biomedicine features. These features indicated molecular-level information of tissue at different carcinoma development-level. This study suggested an automatic and quick method for the NPC Raman spectra classification at different stages by multivariate statistical analysis. In the RS measurement, high quality Raman spectra was acquired from each NPC tissue sample in two groups: one group consisted of 30 NPC patients at the early stages (I-II), another group was 46 NPC patients at the advanced stages (III-IV). Moreover, tentative diagnostic algorithms based on principle components analysis (PCA) and support vector machine (SVM) were employed to classify the multivariate data of Raman spectra effectively. The classification performance (sensitivities and specificities were 70% (21/30) and 91% (42/46)) was achieved by the PCA-SVM in conjunction with leave-one-out cross validation method. In this beneficial study, the RS technique in conjunction with PCA-SVM provided a great clinical potential for rapid NPC stage diagnosis.