[home] [Personal Program] [Help]

---

Session: Musculoskeletal System
Session starts: Thursday 24 January, 10:40
Presentation starts: 10:40
Room: Lecture room 558

---

Michele Colloca (Department of Biomedical Engineering, Eindhoven University of Technology)
Christian Hellmich (Institute for Mechanics of Materials and Structures, Vienna University of Technology)
Keita Ito (Department of Biomedical Engineering, Eindhoven University of Technology)
Bert van Rietbergen (Department of Biomedical Engineering, Eindhoven University of Technology)

Abstract:
Load-adaptive bone remodeling algorithms implementing micro-FE models to represent the bone structure and calculate local loading conditions (Huiskes et al., 2000) can explain many features of bone adaptation at the tissue- and cell-level (Ruimermann et al., 2005). The application of these algorithms for patient-specific predictions, however, is limited by the high computational costs and the fact that at most sites (e.g. hip and spine) it is not possible to measure the bone structure in vivo. An alternative method that can reduce the computational time and does not require microstructural measurements in vivo is proposed in this study. We developed a multiscale analytical model to predict changes in bone density due to changes in cell activity or loading by combining, in a rigorous way, a micromechanical formulation of the mechanical stimulus at the tissue and cell level and a non linear differential equation expressing the density evolution of the bone remodeling system at the organ level. We used a representative volume element (RVE) of trabecular bone composed of a two-phase material: bone matrix (modeled as transversely isotropic material, Malandrino et al. 2012) and cylindrical voids. This assumption allowed for finding a closed-form solution for the mechanical stimulus sensed by the osteocytes, that is, the micromechanics-derived strain energy density based on an Eshelby matrix-inclusion problem. Hence, the typical RVE (mm) and pore (µm) scales were linked to predict the stress state on the trabecular surface where the bone remodeling takes place. The analytical evolution of the bone volume fraction was compared to that predicted by a corresponding numerical model, based on the previously validated micro-FE algorithm of bone remodeling. Good agreement was found with a difference of less than 2.4% while the computational time was dramatically reduced by a factor of almost one million. We thus expect that the proposed model can provide an efficient tool for simulating patient-specific bone remodeling at the organ level depending on changes in cell activity, material properties or loading conditions at lower levels.