Salivary Raman spectroscopy: Understanding alterations in saliva of tobacco habitués and oral cancer subjects

Abstract

ObjectiveSaliva a non-invasive biofluid, offers diagnostic potential due to its vicinity with tumour along with systemic pooling of biomarkers. The present study evaluated the feasibility of Raman spectroscopy to identify biochemical alterations in saliva samples of healthy volunteers, habitués and oral cancer subjects. MethodsThe saliva samples collected from healthy volunteers (HV, n = 27), oral cancer subjects (T, n = 59) and tobacco habitués (HC, n = 62) were subjected to Raman spectroscopy analysis. ResultsThe 3-model and 2-model Principal Component Linear Discriminant Analysis (PC-LDA) clearly distinguished HV spectra from HC and T spectra while misclassifications were observed among HC and T groups. The sensitivity and misclassifications in 3- and 2-model analysis for HC and T groups was seen to be 55.93 % and 65.52 % and 35 % and 40 % respectively. This could be ascribed to similar tobacco habits and prolonged tobacco exposure possibly leading to cancer field effects and the shedding of clinically undetectable micro-tumours in tobacco habitués. The 3-model approach for classification between HC, HV and T groups, yielded distinct stratification of HV (93 %), while the 2-model approach using HV and HC and HV and T distinctly classified HV with 93 % and 96 % efficiency. The higher classification efficiency highlights the utility of Raman spectroscopy in stratifying the healthy volunteers (HV) from tobacco habitués (HC) and tumour (T) groups. ConclusionsDespite the low stratification accuracy of HC and T groups, distinct classification of HV demonstrates the utility of Salivary Raman spectroscopy as an early diagnostic screening technique to identify high risk individuals.