Mathematical Modeling in Cancer Metabolism: Tools for Translational Applications in Metabolism-Based Therapy

Cancer metabolism is characterized by extensive reprogramming of biochemical pathways, enabling malignant cells to sustain proliferation, adapt to fluctuating environments, and resist therapeutic stress. While the Warburg effect has long been considered a hallmark of cancer, recent evidence highlights the dynamic metabolic plasticity of tumor cells, which flexibly engage glycolysis, oxidative phosphorylation, glutaminolysis, and lipid biosynthesis depending on nutrient availability and microenvironmental conditions. These adaptations not only promote tumor survival but also generate exploitable metabolic vulnerabilities. Mathematical and computational modeling have become a powerful strategy for unraveling this complexity and translating biological insights into clinical applications. Kinetic models offer a mechanistic resolution of enzymatic flux control, while constraint-based frameworks such as flux balance analysis enable genome-scale prediction of steady-state flux distributions and identification of metabolic liabilities. Agent-based models extend this analysis to capture spatial heterogeneity, tumor-immune interactions, and emergent behaviors within the tumor microenvironment. More recently, machine learning and hybrid data-driven approaches have complemented mechanistic modeling by integrating high-dimensional multi-omics datasets to reveal biomarker patterns, predict therapeutic response, and stratify patients according to metabolic phenotype. Personalized genome-scale metabolic models, constructed from patient-specific omics data, have demonstrated the ability to predict individual vulnerabilities and guide the selection of metabolism-based therapies. Hybrid frameworks such as physics-informed neural networks and neural ordinary differential equations further extend predictive capacity to capture tumor-immune-metabolism dynamics. Collectively, these approaches bridge preclinical experimentation and translational oncology by enabling virtual hypothesis testing, biomarker discovery, and rational design of adaptive therapeutic strategies. By uniting mechanistic insights with predictive modeling, mathematical frameworks are poised to become integral to precision oncology. Their integration into clinical pipelines will accelerate the identification of metabolic targets, improve patient stratification, and advance the development of effective, personalized metabolism-based cancer therapies.