Stochastic models of tumor growth and the probability of elimination by cytotoxic cells (Q799598)

In a deterministic model, the growth of tumor cell populations under the action of cytotoxic cells can be described by the following differential equation: \[ (*)\quad dT(t)/dt=\gamma T-kT/(K+T), \] where \(T\geq 0\) is the number of tumor cells, while \(\gamma\), k and K are positive constants. Dependent on the sign of an appropriate eigenvalue, T(t) may tend to zero or to infinity. However, for a finite tumor cell population, the stochastic nature of the process cannot be neglected and there exists a nonzero probability that the tumor becomes extinct, for any parameter values. The author considers two stochastic analogues of (*). First is a Markov branching process, with or without density dependence, valid for very small tumors. The second model (for larger tumors) is a density dependent Markov jump process. The density dependent cases are studied by means of their diffusion approximations. Results, collected at the end of the paper, indicate that various stochastic approaches yield qualitatively similar dependence of the extinction probability on the initial tumor size.