Bayesian calibration, validation and uncertainty quantification for predictive modelling of tumour growth: a tutorial

Collis, Joe; Connor, Anthony J.; Paczkowski, Marcin; Kannan, Pavitra; Pitt-Francis, Joe; Byrne, Helen M.; Hubbard, Matthew E.

Abstract

In this work we present a pedagogical tumour growth example, in which we apply calibration and validation techniques to an uncertain, Gompertzian model of tumour spheroid growth. The key contribution of this article is the discussion and application of these methods (that are not commonly employed in the field of cancer modelling) in the context of a simple model, whose deterministic analogue is widely known within the community. In the course of the example we calibrate the model against experimental data that is subject to measurement errors, and then validate the resulting uncertain model predictions. We then analyse the sensitivity of the model predictions to the underlying measurement model. Finally, we propose an elementary learning approach for tuning a threshold parameter in the validation procedure in order to maximize predictive accuracy of our validated model.

Citation

Collis, J., Connor, A. J., Paczkowski, M., Kannan, P., Pitt-Francis, J., Byrne, H. M., & Hubbard, M. E. (2017). Bayesian calibration, validation and uncertainty quantification for predictive modelling of tumour growth: a tutorial. Bulletin of Mathematical Biology, 79(4), 939-974. https://doi.org/10.1007/s11538-017-0258-5

• Journal Article Type: Article
• Acceptance Date: Feb 5, 2017
• Online Publication Date: Mar 13, 2017
• Publication Date: Apr 1, 2017
• Deposit Date: Feb 24, 2017
• Publicly Available Date: Mar 13, 2017
• Journal: Bulletin of Mathematical Biology
• Print ISSN: 0092-8240
• Electronic ISSN: 1522-9602
• Publisher: Springer Verlag
• Peer Reviewed: Peer Reviewed
• Volume: 79
• Issue: 4
• Pages: 939-974
• DOI: https://doi.org/10.1007/s11538-017-0258-5
• Keywords: Bayesian Calibration, Tumour Growth, Model Validation
• Public URL: https://nottingham-repository.worktribe.com/output/969875
• Publisher URL: https://link.springer.com/article/10.1007%2Fs11538-017-0258-5
• Additional Information: The final publication is available at Springer via http://dx.doi.org/10.1007/s11538-017-0258-5