Saliva could be an optimal sample for non-invasive cancer detection, as it contains plenty of proteins and metabolites which can reflect the health status of an individual. Moreover, pairing it with high-sensitivity, label-free detection techniques could prove successful for early cancer diagnosis. In this study, we explore the enhancement of salivary characteristic Raman bands by using label-free, ultrasensitive surface-enhanced Raman scattering (SERS) based on gold nanoparticles. SERS maps were acquired from dry samples of saliva supernatant mixed with Au colloidal nanoparticles, which was then pipetted on clean glass slides. The SERS spectra presented a high variability of signal intensities and frequency shifts. However, several reproducible SERS spectra showing well-resolved bands were obtained at certain locations on the maps, where Au nanoparticles clustered together during the air-drying. The healthy and oral cancer saliva could be differentiated using principal components analysis based on several SERS bands assigned mainly to amino acids and proteins. Moreover, thiocyanate Raman modes were detected in saliva samples of both smoking and non-smoking volunteers and cancer patients. The analysis indicated that the cancer group displayed an overall higher level of the 2126 cm-1 band area assigned to C-N stretching vibrations of thiocyanate.
Keywords: Diagnosis; Oral cancer; Saliva; Surface-enhanced Raman spectroscopy.