For this purpose, using our model it is possible to investigate at a population level a fine-tuning of model parameters which leads cancer into an extinction condition. An encouraging example of how a computational model combined with experimental data can help to verify how the therapy response influences cell population dynamics is reported in a recent paper by Tyson et al.. The authors have showed that erlotinib – an epidermal growth factor receptor inhibitor – is not able to kill tumor cells, but it leads them into a quiescent state or decreases their proliferation rate. Therefore, expressions of possible system behaviors as mathematical equations can give us the possibility to explore both how different drugs work and against which targets, in term of cell events, therapies must be addressed. Note that we further investigated those parameter combinations where CSC reproduction rate is positive, as reported in Table 1. Previous papers report the crucial role of CSCs to cancer progression, but a connection among some of CSC features, i.e. strong self-renewal, resistance to apoptosis, differentiation abilities, and cancer progression, has not been established. Our results suggested which of these features mostly determine cancer growth dynamics, namely those responsible for global CSC and PC variation. Moreover, PF-4217903 analyzing parameter values obtained from all runs of the MLS algorithm we discovered some interesting linear correlations among CSC differentiation, CSC death, and CSC symmetrical proliferation probability. This is in accordance with what is observed in many solid tumors or mammosphere models, in which both intrinsic and extrinsic mechanisms known to directly affect CSC symmetric division probability and differentiation or apoptotis have been discovered. These mechanisms, which include p53 mutation or depletion in CSCs, and the availability of certain host growth factors – such as EGF and growth-factor-rich niches – can skew division modes in favor of symmetric production of CSCs for up to 85%. Both clones derived from LI, D2 and F11, show the same characteristics.