Speaker
Description
Colorectal cancer (CRC) is the world’s third most common malignancy, and patients with the microsatellite instability–high/mismatch repair–deficient (MSI-H/dMMR) phenotype respond poorly to conventional chemotherapy yet show striking responsiveness to immune checkpoint inhibitors such as pembrolizumab. Mathematical models based on ordinary differential equations (ODEs) provide a powerful framework for analysing the immunobiology underpinning disease dynamics. In contrast, agent-based models (ABMs) explicitly represent individual entities and their interactions; inherent randomness can naturally give rise to heterogeneity—features that deterministic, mean-field ODEs cannot capture. However, ABMs typically encode many interactions and processes, making simulations computationally expensive and parameter calibration to experimental data difficult.
In this talk, we present a minimal ODE model of pembrolizumab therapy in locally advanced MSI-H/dMMR CRC (laMCRC) to reveal the core drivers of the tumour–immune response and the key immune interactions involved. We also outline a practical workflow for converting ODE models to ABMs and apply it to our minimal model, enabling accurate and efficient parameter estimation for the ABM. Finally, we compare ABM trajectories with those from the minimal ODE model, verifying consistency and faithful reproduction of the dynamics, thereby laying a foundation for future ABMs of laMCRC and other cancers.
