Image and optical property reconstruction from noninvasive measurements of turbid media, in vitro and in vivo: A promising method for optical mammography.

摘要

Due to their multiple scattering, near infrared photons diffuse in human tissues providing interesting physiological information over relatively large volumes. By employing frequency-domain techniques along with a diffusion model for photon transport, one can separate contributions to light attenuation from absorption and scattering processes. Both the absorption and reduced scattering parameters are sensitive to physiological characteristics of tissues. I have employed the diffusion model, which assumes that tissue is homogeneous, to create an efficient reconstruction routine based on the back-projection technique to detect and locate inhomogeneities qualitatively.; Before my work in this field, researchers were uncertain whether the diffusion model would be valid for a medium containing inhomogeneities. I, along with a handful of other researchers demonstrated that the diffusion model is accurate enough to include well behaved heterogeneous tissue models in which a number of macroscopic objects of different size and shape are embedded in a homogeneous background medium. I have employed a fitting of the experimental data to the diffusion model to accurately reconstruct not only the position and size, but also the optical properties including the relative index of refraction, absorption and scattering coefficients of objects immersed in highly scattering media. In this thesis I describe a model for locating and characterizing optical inhomogeneities in a turbid medium. I will go on to discuss limits of the technique and present data collected in vivo from a 55 year old female subject with a malignant breast cancer that was detected and optically characterized by these techniques.