Abstract

Several Raman spectroscopy studies, over the past few decades have demonstrated, its utility as an adjunct screening tool in oral cancer diagnosis. A screening/diagnostic test should mandatorily have high sensitivity to detect minuscule tumour load. Therefore, in this study, role of tumour load on efficacy of Serum Raman Spectroscopy (SRS) in the Hamster Buccal Pouch (HBP) model was evaluated. Serum samples were collected in a longitudinal manner over 14 weeks from DMBA induced oral carcinogenesis in HBP model. The tumour load i.e., number of tumours on the treated HBP, ranged from 1-8. Raman spectra of sera samples were recorded using Confocal Raman microscope, WITec alpha300R with 532 nm excitation laser (30 mW power) and 600 g/mm grating, in the spectral range 400-4000 cm^-1. Multivariate analyses of averaged serum Raman spectra, generated from pre-processed spectra of each sample which were grouped on basis of tumour load (low and high), in a 5-model system, stratified the distinct phases of oral carcinogenesis (week 0: healthy, week 1-3: inflammation, week 4-7: hyperplasia, week 8-11: dysplasia, and week 12-14: moderate to well-differentiated squamous cell carcinoma). As expected, week 0 symbolising healthy condition clearly distinguished from all other DMBA treated intervals. Misclassifications (biochemical homogeneity) were highest at week interval 1-4 irrespective of the tumour load; majorly with week 0. Another evident observation is after a drop in classification accuracy at weeks 1-4, a gradual increase in classification is noted at later intervals, attributable to progressive DMBA treatments and corresponding development of oral tumours. The findings suggested near equivalent sensitivity for both low and high tumour loads and demonstrate that tumour load has no effect on efficacy of SRS, emphasizing the clinical utility of the technique in oral cancer screening/diagnosis.

© 2023 SPIE