Bayesian Calibration, Validation and Uncertainty Quantification for Predictive Modelling of Tumour Growth: A...

Area of Science:

• Mathematical Biology
• Computational Oncology
• Biophysics

Background:

• Tumor growth modeling is crucial for understanding cancer progression.
• Gompertzian models are widely used but often lack robust uncertainty quantification.
• Calibration and validation techniques are not commonly applied in cancer modeling.

Purpose of the Study:

• To present a pedagogical example of applying calibration and validation to an uncertain Gompertzian tumor spheroid growth model.
• To demonstrate the utility of these methods in the context of cancer modeling.
• To propose a learning approach for optimizing model validation.

Main Methods:

• Calibration of an uncertain Gompertzian model using experimental data with measurement errors.
• Validation of the calibrated uncertain model predictions.
• Sensitivity analysis of model predictions to the measurement model.
• Development of an elementary learning approach for threshold parameter tuning.

Main Results:

• Successful application of calibration and validation techniques to an uncertain tumor growth model.
• Demonstrated sensitivity of model predictions to measurement error models.
• Proposed a method to enhance predictive accuracy through parameter tuning in validation.

Conclusions:

• Calibration and validation are valuable, underutilized tools for uncertain cancer models.
• Understanding measurement error is critical for accurate tumor growth prediction.
• An adaptive learning approach can improve the reliability of validated tumor growth models.