Simulation-based parameter estimation for complex models: a breast cancer natural history modelling illustration.

Area of Science:

• Computational Biology
• Mathematical Modeling
• Cancer Research

Background:

• Estimating parameters in complex stochastic models is challenging.
• Natural history models are crucial for understanding disease progression, such as breast cancer.
• Existing methods may struggle with the complexity of real-world biological systems.

Purpose of the Study:

• To demonstrate simulation-based parameter estimation for a breast cancer natural history model.
• To apply these methods using data from the Surveillance, Epidemiology, and End Results (SEER) registry.
• To highlight the utility of simulation in computing maximum likelihood estimators for complex models.

Main Methods:

• Developed a natural history model for breast cancer.
• Assumed tumor growth follows a geometric Brownian motion process.
• Employed simulation techniques to compute the maximum likelihood estimator (MLE).

Main Results:

• Successfully estimated model parameters using SEER registry data.
• Demonstrated that simulation provides a computationally feasible approach for parameter estimation.
• Confirmed the effectiveness of simulation for models with substantial complexity.

Conclusions:

• Simulation-based parameter estimation is a powerful and practical tool for complex stochastic models.
• This approach is well-suited for analyzing cancer natural history models.
• The method offers a straightforward way to compute estimators, facilitating deeper insights into disease dynamics.