Quantitative analysis of spectroscopy data for skin oximetry

Abstract

This work presented the use of Modified Lambert Beer law (MLBL), Extended Modified Lambert Beer model (EMLB) and cumulant based forward model (CM) for the estimation of mean blood oxygen saturation (SmO2) in skin of healthy individuals via an iterative fitting routine. The proposed strategy required a priori knowledge of extinction coefficients of hemoglobin components in the wavelength range of 520--600 nm to give the best estimation of SmO2. This study conducted spectroscopy measurement on left index finger of four Asians at rest condition. Quantitative analysis of the collected data using the employed strategy revealed SmO2 with mean value of 71.5 ± 1.15 %, 95.1 ± 3 % and 96.3 ± 3.8 % given by MLBL, EMLB and CM, respectively. The consistent and high SmO2 value estimated by EMLB and CM agreed considerably well with the value reported in the literature. The underestimation of SmO2 given by the MLBL could be a result of insufficiency of the model at describing light propagation in skin. This work concluded that both EMLB and CM, and the proposed strategy can suitably be used in clinical setting for skin oximetry.