Diffuse reflectance spectroscopy as a tool to evaluate liver tissue

Sammanfattning: The shape of a diffuse reflectance spectra can provide knowledge about tissue composition. By analysing this spectra one can, with the right evaluation methods, find out information about the tissue. With known absorption properties of the present tissue chromophores it is possible, with a curve fit algorithm, to find their volume fractions in any measured tissue. In this work, the usability of diffuse reflectance spectroscopy (DRS) as a tool in tissue diagnostics was investigated. A model based on diffusion theory with a semi-infinite boundary condition and known tissue chromophore data was developed. It was used to evaluate DRS signals in tissue phantoms and on murinae, porcine and human livers. To ease the evaluation, an interface was developed to run the instruments (one light source and two spectrometers) from the computer and to compute the tissue composition from the provided diffusion model. The tissue composition was computed through a Levenberg-Marquardt least-squares algorithm in MATLAB. Validation of the diffusion model was performed on tissue phantoms together with Monte Carlo (MC) simulations and a Time-of-Flight Spectroscopy (TOFS) measurement. The results from the validation measurements were varying. Many different kinds of phantoms were created. Commonly used phantoms, containing water, blood and intralipid gave results systematically underestimating the absorption while phantoms with greater fractions of lipid, mixed with agar or Triton-X100, gave results with less good correlation between signals and concentrations. The poor phantom results were discovered to be due to a low haemoglobin concentration in the used blood whilst there was also limitations in the diffusion criteria for these phantoms. The results from liver measurements gave better correlation between signals and concentrations, and when investigating the criteria for diffusion these were also fulfilled in a greater wavelength region. The liver results also showed an evident difference between healthy and malignant tissue.