Utilization of spatial derivative measurements in circular diffuse optical tomographic imaging to improve image resolution and contrast

Abstract

The resolution of images recovered using diffuse optical tomography without spatial prior information is inherently limited due to the diffusive nature of light transport in scattering dominated biological tissue. Several studies have previously reported an improvement in depth sensitivity in near infra-red spectroscopy by considering the subtraction of two neighbouring measurements. A nearest neighbour, spatial derivative based algorithm is presented, in which the difference between neighbouring source-sharing source-detector pairs is considered for tomographic reconstruction. Using a circular, inward-facing optode geometry, it is shown that deeper abnormalities in oxyhaemoglobin, deoxyhaemoglobin and water can be recovered with better localisation and contrast as compared to reconstructing using absolute values of multi-wavelength data. Relative improvements of up to 25% in quantitative accuracy of recovered physiological values are demonstrated at better resolution for cancerous tissue in a heterogeneous breast model, when using this proposed spatial derivative based parameter reconstruction.