Our goal is to build personalized computational models of cancer that capture how an individual’s disease progresses and responds to treatment. By simulating the unique characteristics of each tumor and its therapy response, we aim to generate clinically useful insights that support personalized care. At the core of our approach is a mechanistic model based on quantitative systems pharmacology (QSP), which represents biochemical and biomechanical processes within tumors through a large system of nonlinear equations. A major challenge in QSP is parameter estimation: traditional models typically assume patients are homogeneous, which reduces accuracy when fitting models to diverse clinical data. We instead develop a computational framework that estimates parameters from individual patient data, tailoring the QSP model to each person’s tumor biology. We use sensitivity analysis and uncertainty quantification to identify the most influential interactions and to characterize prediction ranges. In addition, we apply machine learning and statistical methods to stratify patients and predict who is most likely to benefit from a given therapy. For example, we developed the Python package TMEImmune to identify patients who may respond to immune checkpoint inhibitors. By integrating mechanistic modeling with data-driven methods in a patient-specific framework, we seek to improve prediction of cancer progression and treatment response and, ultimately, to contribute to more effective personalized cancer therapy. REFERENCES [1] N. Mirzaei, L. Shahriyari, Modeling Cancer Progression: An Integrated Workflow Extending Data-Driven Kinetic Models to Bio-Mechanical PDE Models, Physical Biology, 2024. [2] Q. Zhou , L. Shahriyari, TMEImmune: A Python Package for deriving prognostic tumor micro-environment score, SoftwareX, 2025.