cancer density) or toc 3 (in our case this is non-zero corresponding
to appreciable cell cancer density). The quantityc 2 plays a role of a
switch in this dynamics. If cancer cells concentrationcgets higher
thanc 2 , then the fate of the point of tissue is in principle deter-
mined, and a stably dense cancer scenario there arises if diffusion
does not carry away cancer cells fast enough. On the other hand, if
the value ofcis lower, the organism can avoid this pathological
situation. It is clear that in this model thenc 2 (in union with
diffusion which—by construction—lowers cellular population)
characterizes the role of the immune system against cancer
progression.
3 Results
The model previously described can be cast in a non-dimensional
form, a standard procedure in any numerical situation and studied
in different dimensions.
In one dimension (1D), the resulting equation is
∂C
∂T
¼
∂^2 C
∂X^2
þaCðÞ 1 CðÞCα
where Cis non-dimensional concentration, and X and Tare
non-dimensional space and time respectively. Parametersa> 0
and 0<α<1 describe the entire dynamics then. The 1D equation
above can be both studied analytically and numerically. In the
former case, one can find a traveling wave exact solution for an
infinite spatial domain given by
C¼
1
2
1 þtanh
ffiffiffi
a
p
ðÞXþVT
2
ffiffiffi
2
p
with constant non-dimensional velocityV¼
ffiffia
2
p
ðÞ 1 2 α. This is a
traveling wave that transfers high concentrations of cancer cells in
regions that had before almost zero values for this species.
These equations have been numerically investigated in 1D and
2D numerical simulations, showing how a nontrivial initial data for
cancer cells leads to a tumor cells cancellation in some regions but
finally to a growth in the space of tumor mass.
Similar studies can also be performed in 3D, even in the pres-
ence of more complicated situations in which the diffusion coeffi-
cient is not a constant but a function of space and time
(inhomogeneous diffusion) instead, i.e.,D¼Dt;x
⇀
.
This is what can be done for instance in studying tumor diffu-
sion processes in the brain, a standard application of mathematical
modeling extremely useful in Neurosurgery [6, 33]. We point out
that a more realistic (but complicated) scenario would be the one in
which the diffusion coefficient is not a scalar but a matrix (diffusion
Systems Biology Modelling 209