Combining the diffusion approximation and Monte Carlo modeling in analysis of diffuse reflectance spectra from human skin

Abstract

Light propagation in highly scattering biological tissues is often treated in the so-called diffusion approximation (DA). Although the analytical solutions derived within the DA are known to be inaccurate near tissue boundaries and absorbing layers, their use in quantitative analysis of diffuse reflectance spectra (DRS) is quite common. We analyze the artifacts in assessed tissue properties which occur in fitting of numerically simulated DRS with the DA solutions for a three-layer skin model. In addition, we introduce an original procedure which significantly improves the accuracy of such an inverse analysis of DRS. This procedure involves a single comparison run of a Monte Carlo (MC) numerical model, yet avoids the need to implement and run an inverse MC. This approach is tested also in analysis of experimental DRS from human skin.