Comparison of methods for assigning the material properties of the distraction callus in computational models
CitationMora‐Macías, J., Giráldez‐Sánchez, M. Á., López, M., Domínguez, J., & Reina‐Romo, M. E. (2019). Comparison of methods for assigning the material properties of the distraction callus in computational models. In International Journal for Numerical Methods in Biomedical Engineering (Vol. 35, Issue 9). Wiley. https://doi.org/10.1002/cnm.3227
AbstractIn silico models of distraction osteogenesis and fracture healing usually assume constant mechanical properties for the new bone tissue generated. In addition, these models do not always account for the porosity of the woven bone and its evolution. In this study, finite element analyses based on computed tomography (CT) are used to predict the stiffness of the callus until 69 weeks after surgery using 15 CT images obtained at different stages of an experiment on bone transport, technique in which distraction osteogenesis is used to correct bone defects. Three different approaches were used to assign the mechanical properties to the new bone tissue. First, constant mechanical properties of the hard callus tissue and no porosity were assumed. Nevertheless, this approach did not show good correlations. Second, random variations in the elastic modulus and porosity of the woven bone were taken from previous experimental studies. Finally, the elastic properties of each element were assigned depending on gray scale in CT images. The numerically predicted callus stiffness was compared with previous in vivo measurements. It was concluded firstly that assignment depending on gray scale is the method that provides the best results and secondly that the method that considers a random distribution of porosity and elastic modulus of the callus is also suitable to predict the callus stiffness from 15 weeks after surgery. This finding provides a method for assigning the material properties of the distraction callus, which does not require CT images and may contribute to improve current in silico models.