Subsurface sensing of biomedical tissues using a miniaturized Raman probe: Study of thin-layered model samples

Abstract

A ball lens hollow-fiber Raman probe (BHRP) is a powerful tool for in vivo nondestructive subsurface analysis of biomedical tissues in a living body. It has confocal-like optical properties, but its collection volume is rather large in comparison with that of a conventional confocal Raman system. Therefore, the obtained Raman spectra have contributions from the upper and lower layers at different rates depending on the thickness of the upper layer when the measurement point is close to the boundary surface of the two layers. In the present study, we describe a methodology to extract quantitative information about the thickness of the subsurface layer structure by using a BHRP combined with the partial least-square regression (PLSR) analysis. The simulation study indicates that distribution of the collection efficiency in the collection volume of the BHRP is similar to a Gaussian distribution. The empirical study suggests that the PLSR model built with only a principal component (PC) 1 based on the linearized depth data gives good prediction.