Diffuse reflectance spectroscopy for monitoring physiological and morphological changes in oral cancer

Oral cancer is prevalent in India due to addictive habits such as smoking and tobacco chewing. Simple steady-state diffuse reflectance spectroscopic technique may facilitate routine check up and early diagnosis of oral cancer. In this regard, inverse Monte Carlo algorithm is implemented to estimate the optical properties of tissue from diffuse reflectance spectra. A modified tissue model has been proposed in this algorithm which considers predominant Rayleigh scattering in the visible region and absorption of methemoglobin and melanin in tissue. The proposed tissue model improves the accuracy of estimation of optical properties. The root mean square variance between the measured and modeled diffuse reflectance spectra is found to be less than 5%. Further, lookup table based technique adopted in this algorithm enables real-time estimation of optical parameters of tissue in about 5 s. Significant variations in the scattering coefficient, hemoglobin oxygen saturation and fraction of methemoglobin are identified between normal and cancer tissues. Step-wise linear discriminant analysis have revealed that the tissue parameters were statistically significant and the cancer tissue are discriminated from normal with 90% sensitivity and 90% specificity.