Design and evaluation of a reflectance diffuse optical tomography system

We have designed a continuous wave, back-reflection diffuse optical tomography system; and developed a new practical calibration method including both optode efficiency and positional error corrections. System design, data acquisition and calibration protocols are described in detail. Monte Carlo (MC) simulations of photon distribution for tissue phantoms have been used to obtain the weight matrix to be used in the Rytov approximation to the photon diffusion equation. The system has been evaluated by acquiring data from a tissue phantom with a background scattering coefficient ($$\upmu _{\mathrm{s}}^{\prime }$$μs′) of $$10\,\hbox {cm}^{-1}$$10cm-1 and absorption coefficient ($$\upmu _{\mathrm{a}})$$μa) of $$0.04\,\hbox {cm}^{-1}$$0.04cm-1. An inclusion made of 1 % Intralipid and indocyanine green with $$\upmu _{s}^{\prime }=10\,\hbox {cm}^{-1}$$μs′=10cm-1 and $$\upmu _{\mathrm{a}}= 0.16\,\hbox {cm}^{-1}$$μa=0.16cm-1 was placed at a 2 cm depth from the tissue phantom surface. After calibration, the average value of the measurements over source–detector pairs at the same distance for each neighborhood was calculated. Perturbation data were obtained by subtracting the average data from the measurements with the same source–detector separation. In the reconstruction, weight matrixes obtained from MC simulation for $$\upmu _{\mathrm{s}}^{\prime }= 7$$μs′=7 to $$12\,\hbox {cm}^{-1}$$12cm-1 were used. The Depth Compensation Algorithm was used in the Tikhonov regularization to identify the location of the inclusion correctly in the reconstruction.