Analysis of a mathematical model of the growth of necrotic tumors (Q5934259)

In this paper a necrotic tumor is schematized as a radially symmetric system in which the central core is necrotic (and mass losing) and the proliferating region is possibly vascularized. Necrosis occurs when the nutrient concentration reaches a given (positive) threshold. After suitable rescaling, the mathematical problem is formulated as follows \[ c\frac{\partial \sigma}{\partial t}-\Delta \sigma +\lambda \sigma H (\sigma - \sigma_0)=0,\quad\text{ a.e.} \text{ for}\quad 0<R<R(t),\;t>0,\tag{1} \] \[ \frac{\partial \sigma}{\partial r} (0,t)=0,\quad t>0,\tag{2} \] \[ \sigma(R(t),t)=\tilde \sigma, \quad t>0, \tag{3} \] \[ \sigma (r,t)=\sigma_0,\quad \text{ for} \quad 0 \geq r \geq \rho(t), \quad t>0,\tag{4} \] \[ \frac{\partial \sigma}{\partial r} (\rho (t), t)=0, \quad t>0, \tag{5} \] \[ R^2 \dot R =\int^R_\rho [\sigma (r,t)-\tilde \sigma]r^2 dr -\mu \rho^3, \quad t>0, \tag{6} \] plus initial conditions for \(\sigma\), \(\rho\), \(R\). Here the unknowns are: \(\sigma\) (nutrient concentration), \(R\) (outer radius), \(\rho (t)\) (radius of the necrotic region), and \(c,\lambda\), \(\bar \sigma\), \(\sigma_0\), \(\tilde \sigma\), \(\mu\), are positive constants. In (1) \(H\) is the Heaviside function. The paper is very long and contains several results. In Section 2 it is shown that the stationary problem has a unique solution, provided the given parameters entering (1)--(6) satisfy some conditions. In Sections 3, 4 the auxiliary problem (1)--(5) is studied with \(R(t)\) given and \(\rho (t)\) as the only free boundary. Existence and regularity are proved. In Section 5 the full problem is studied using a fixed point argument and concluding that (i) there exists a unique solution as long as \(R(t)\) belongs to a suitable interval; (ii) the solution is smooth. More estimates are derived in the next sections, allowing to say that \(R(t)\) must remain in the critical interval for all \(t>0\), thus extending the solution to arbitrarily large time intervals. Thus it makes sense to study the behaviour of the solution as \(t \to \infty\). It is shown that the functions \(\sigma\), \(\rho\), \(R\) converge to the respective stationary values and that the rate of convergence is \(\exp(-Kt^{\frac{1}{2}})\) for some \(K>0\). Among the assumptions in Section 5, the most important one from the physical point of view is that the constant \(c\) in (1) is small enough (i.e. diffusivity is sufficiently large). This is indeed the case e.g. for oxygen (\(c\) is so small that the quasi-steady approximation is usually adopted).