Near-Infrared (NIR) Tomography for Breast Cancer Imaging

Brian W Pogue, Keith D Paulsen

Thayer School of Engineering, Dartmouth College, Hanover NH 03755, pogue@dartmouth.edu

ABSTRACT

Near-Infrared light can be used to tomographically image tissue in the human body, and several human trials and research studies are ongoing to assess the clinical diagnostic value of this technology in the setting of breast cancer characterization. Near-infrared attenuation provides new information about hemoglobin concentrations, oxygen saturation and water percentage, as well as features of the scattering related to structural elements of the tissue. The challenge in NIR imaging is to provide an accurate tomographic reconstruction algorithm and hardware design that allows separation of the effects of light scattering from light absorption. Since elastic scattering dominants the interaction in tissue, the light transport is modeled accurately by the diffusion equation. Therefore the image reconstruction process requires an iterative solution to the diffusion equation where the absorption and elastic scattering coefficients are progressively updated at each step to provide a match between the forward simulation and measurements from tissue. In our studies, a frequency- domain light source is used to provide light pathlength, or phase shift, information along with amplitude attenuation data. Tomographic images are generated at multiple wavelengths, and the tissue absorber concentrations are estimated, as well as the scatterer parameters of the tissue. This system is currently operational in a clinical trial to determine the efficacy for breast cancer tumor characterization.

Keywords Cancer, diffuse, light, scattering