Abstract B030: Decoding cancer evolution: Computational models for tumor dynamics and precision oncology

Abstract Cancer evolution, driven by genetic instability and clonal selection, challenges effective treatment. This review synthesizes computational approaches to model tumor progression, heterogeneity, and resistance, bridging biology with precision medicine. We integrate multi-omics data with mathematical frameworks to uncover adaptive behaviors, guiding therapies that exploit evolutionary vulnerabilities. Our methods span cancer’s multiscale nature. Ordinary differential equations (ODEs) model biochemical networks, like cyclin-dependent kinase dynamics, with cell cycle progression as \(\frac{d[CycD]}{dt} = k_s - k_d [CycD] - k_p [CycD][CDK4/6]\), revealing treatment effects. Probabilistic models, like branching processes, predict lineage evolution with exponential waiting times \(P(t) = \lambda e^{-\lambda t}\). Partial differential equations (PDEs) address tumor microenvironment influences, e.g., hypoxia-driven selection: \(\frac{\partial u}{\partial t} = D \nabla^2 u + f(u) - g(u,v)\), where \(u\) and \(v\) denote cell and nutrient density. Agent-based models (ABMs) capture phenotypic plasticity, while Boolean networks identify regulatory attractors. Dynamical systems theory quantifies chaos via Lyapunov exponents (\(\lambda = \lim_{t \to \infty} \frac{1}{t} \ln \left| \frac{\delta x(t)}{\delta x(0)} \right|\)) and fractal dimensions, signaling relapse risks. Algorithmic information dynamics, using Kolmogorov complexity \(K(s) = \min \{ |p| : U(p) = s \}\), infers causal gene networks via block decomposition. Machine learning, including recurrent neural networks, forecasts omics time-series, while phylogenetic tools like PiCnIc reconstruct clonal trajectories. Analysis of breast and pancreatic cancer datasets shows chaotic attractors linked to chemoresistance, with multifractal spectra (Hurst index \(H\)) quantifying growth irregularity. Adaptive dosing, modeled via optimal control, reduces resistance by 40-60%, with PDE-ABM hybrids predicting clonal expansion under therapies. In conclusion, these computational tools reveal cancer as an evolving ecosystem, enabling personalized, evolutionary-guided therapies to outmaneuver tumor adaptation for improved outcomes. Citation Format: Peter Oloche David. Decoding cancer evolution: Computational models for tumor dynamics and precision oncology [abstract]. In: Proceedings of the AACR Special Conference in Cancer Research: Cancer Evolution: The Dynamics of Progression and Persistence; 2025 Dec 4-6; Albuquerque, NM. Philadelphia (PA): AACR; Cancer Res 2025;85(23_Suppl):Abstract nr B030.