Glioma follow white matter tracts: a multiscale DTI-based model (Q493077)

Starting from a mesoscopic transport equation for the cell density and under some appropriate assumptions, an advection-diffusion equation for tumor cell density is deduced via parabolic scaling. This corresponds to changing the time and space scales from minutes and micrometers (for cell movement) to weeks and centimeters (for glioma). The scaling allows the characterization of the evolution of the macroscopic quantity of interest (i.e., cancer cell density) with the aid of an anisotropic tumor diffusion tensor and tumor drift velocity. Their concrete forms follow from the underlying mesoscopic model, which renders this approach more precise than modeling directly on the macroscopic level. Using formal computations similar to those in the work of \textit{K. J. Painter} and the second author [J. Theor. Biol. 323, 25--39 (2013; Zbl 1314.92083)], the authors show that the tumor diffusion tensor depends on the structure of the underlying fiber network, while the drift velocity contains two components, a directional one due to the underlying fiber network and a supplementary one, which is due to the binding of glioma to extracellular matrix (ECM) components that surrounds the fiber tracts. The modeling approach in this paper is, however, more evolved, as it accounts for three spatiotemporal levels: the microscale for the subcellular dynamics, the mesoscale for migratory events on the cellular level, and the macroscale, on which the whole cell population, i.e., the tumor, is observed. Combining mathematical multiscale modeling with DTI data and using numerical simulations, the anisotropic tumor spread is predicted. The new model predicts finger-like invasion patterns for glioma, as observed in the clinical practice. Hence, these invasion patterns are more realistic than round-shaped predictions from isotropic diffusion models. This model bridges the scales from the subcellular level of receptor binding (which can be seen as the onset of all subsequent processes related to cell motility) over the mesoscale of individual cell behavior in interaction with the tissue, and up to the macroscale of the tumor population. Several open problems and questions remain: many biological relevant processes are still missing or are highly simplified. Besides some modeling issues, there are also some interesting problems with respect to mathematical analysis: the deduction of the macroscopic limit would need a rigorous proof, which is a nontrivial issue. The existence and uniqueness of a positive solution to the kinetic transport equation coupled with the subcellular dynamics can be done by relying on the result by \textit{J. Kelkel} and the last author [Math. Models Methods Appl. Sci. 22, No. 3, 1150017, 25 p. (2012; Zbl 1241.92041)], where an even more complex multiscale model for tumor cell migration has been proposed and analyzed.