A major clinical challenge for cancer therapies is to obtain an effective treatment strategy for each patient or at least identify a subset of patients who could beneﬁt from a particular treatment. Since each cancer has its own unique features, it is very important to obtain personalized cancer treatments and ﬁnd a way to tailor treatment strategies for each patient. Recently, mathematical models have been commonly used to discover, validate, and test drugs. Since these models are a complex system of nonlinear equations with many unknown parameters, estimating the values of the model's parameters is extremely difﬁcult. Existing parameter estimation methods for these models often use assembled data from various sources rather than a single curated dataset. These datasets are usually obtained through various biological experiments, in vitro and in vivo animal studies. To arrive at personalized treatments, we need to obtain values of parameters of the model for each patient separately. Since the set of variables of the model includes relative amount of each cell type and cytokines in the tumor, we developed a tumor deconvolution software, which is a combination of recently developed methods, to predict the relative amount of these variables from the gene expression profile of the tumor. The output of the tumor deconvolution software can be used to predict the values of the parameters for each patient. In other words, we propose to use patients’ gene expression data of primary tumor to estimate the values of parameters of the mathematical model for each patient separately, instead of the common approach of assuming these parameters have the same values across all patients and using animal studies to estimate them. This new approach provides us with a unique opportunity to suggest the optimal treatment strategy for each patient and predict the efﬁcacy of each treatment for each patient.