Postmortem imaging allows for validation of the origins of image contrast through comparisons with histology. However, the inclusion of formalin fixative substantially reduces the T2. This reduction is (approximately) linear with concentration. Prior to scanning, samples are often placed in a fluid that has more favourable properties for imaging (e.g. perfluorocarbon fluids). This may lead to an outflow of fixative and an increase in T2 at the tissue surface. Here we propose to correct for T2 inhomogeneity by modelling the outflow of fixative within whole, human brains using a novel kinetic tensor model, which incorporates the effects of diffusion anisotropy.