Abstract:
In this thesis, an efficient and robust computational framework is developed to predict cortical bone remodelling, capable of including biologically relevant data from across the spatial scales ranging from whole organs to the scale of an osteon. The development of this framework was motivated by gaps in the literature on 3D models of cortical bone, despite its majority role in the load-bearing capabilities of human bones. To this end, the industry-standard methods of Finite Elements and Partial Least Squares Regression, combined with anatomically and physiologically realistic equine and human data, were used to inform novel 3D cortical bone remodelling algorithms and integrated into an accurate and rapid bone remodelling framework.The capabilities of the framework were demonstrated through prediction of bone remodelling in the human femoral neck, emphasising the framework’s applicability to a clinically important scenario.The framework development was iteratively progressed in two computational modelling studies. In the first study, I built a foundation of cortical bone remodelling through the construction of a 3D microscale cortical computational bone remodelling model, validated using equine data, and used Partial Least Squares Regression to rapidly predict remodelling phenomena. In the second study, I shifted the model towards human data, and made large improvements on the bone remodelling algorithm, and finally integrated these developments into a multiscale framework. I demonstrated that it is possible to capture intricate cortical bone remodelling behaviour through this multiscale remodelling framework, robustly accounting for a range of anatomical and physiological parameters to inform a clinically realistic remodelling response over time. I further confirmed that, through a large initial investment into computational resources, a highly efficient surrogate model can be produced to accurately predict this behaviour. This framework can be immediately and easily adapted to investigate complex cortical bone remodelling responses in other joints in the body, which also receive similar loading conditions from surrounding muscles and employ osteon-like remodelling behaviour at the microscale. It is expected that this multiscale approach will allow for detailed investigations into the bone remodelling response from the introduction of drug or disease effects at any scale.