Failure of radiative transport theory in homogeneous scattering media to predict the Point Spread Function (PSF) of transillumination images through some biologic tissues

Abstract

Monte Carlo simulations or, to a lesser degree of accuracy, the resolution of the diffusion are generally considered as giving reasonable evaluations of the photon irradiance in tissues, globally described by the radiative transport theory. They are frequently used to calculate the Point Spread Function (PSF) describing the sharpness of transillumination images, resolved either in time or in frequency domain. The PSF is itself completely determined, for an homogenous slab of tissue and in the approximation of the diffusion equation, by the reduced scattering and absorption coefficients: µs' and µa respectively. For some tissue preparations, precise measurements of the PSF show, however, significant discrepancies between the predictions from µs′ and µa and the measured PSF. The propagation of light in biological tissues has been studied both theoretically by Monte Carlo simulations and experimentally in vitro on bovine and porcine liver, fat emulsions, human breast and brain tissues. Absorption and reduced scattering coefficients have been obtained and assessed by matching both the spatial and temporal profiles of a pulsed, collimated light beam to the experimental data. Measurements on fat emulsions and liver exhibit an excellent agreement between the predictions of the radiative transport theory for both spatial and temporal profiles. On the contrary, the spatial profile measured on adipose, breast and brain tissues are systematically too large to be predicted by the radiative transport theory in homogenous tissues. The tissue structure, in particular, the tissue micro and macroheterogeneities and possible site percolation in the model of light transport, not yet evidenced experimentally, could explain the unexpected broadening of the PSF in transillumination images through some tissues like adipose and brain .