Calibration of agent based models for monophasic and biphasic tumour growth using approximate Bayesian computation

Area of Science:

• Computational Biology
• Mathematical Oncology
• Cancer Research

Background:

• Agent-based models (ABMs) simulate tumor evolution but face validation challenges due to complexity.
• The Voronoi cell-based model (VCBM) represents individual cell shapes and dynamics.
• Tumor growth can exhibit biphasic patterns, necessitating models that capture these distinct phases.

Purpose of the Study:

• To extend the VCBM to incorporate biphasic tumor growth.
• To employ approximate Bayesian computation for model calibration and uncertainty quantification.
• To validate the enhanced model against in vivo measurements for breast, ovarian, and pancreatic cancers.

Main Methods:

• Extension of the Voronoi cell-based model (VCBM) to include two distinct growth phases.
• Application of approximate Bayesian computation (ABC) for parameter estimation and uncertainty quantification.
• Calibration of the model using in vivo experimental data from multiple cancer types.

Main Results:

• The biphasic VCBM successfully captures tumor growth dynamics observed in experimental data.
• The approach quantifies uncertainty in the switching time between growth phases.
• Precise estimates for cell maturation time were achieved, enabling model refinement.

Conclusions:

• The enhanced VCBM provides valuable insights into tumor evolution and biphasic growth.
• The methodology allows for robust parameter estimation and uncertainty assessment in cancer models.
• This work facilitates improved accuracy in cancer growth modeling and understanding of cell characteristics.