Multiscale predictions of vertebral strength in tissue, microarchitecture, and body levels

Abstract

Osteoporosis is a global health crisis with the fatal consequence of fragility fractures. Almost all bones in the human skeleton could occur fragility fractures, among which vertebral fractures are the most common but the hardest to define. In the past half-century, researchers in the field have put a large amount of effort to understand vertebral strength for predicting vertebral fracture risk. Generally, vertebral strength depends on three scales of mechanical characteristics. They are tissue properties (µm), trabecular microarchitecture (mm), and vertebral body (cm). Even though the previous studies have answered many significant problems, there are still some consequential questions in all three levels. This thesis aims to deepen insights into the multiscale mechanical properties of the vertebra further. Firstly, only a few studies have investigated if the mechanical properties of bone tissue could vary during deformation. I used micro-indentation to examine the indentation modulus and hardness of bone tissue under various indentation depths; and found the nonlinear decrease of stiffness and hardness with the increase of the deformation volumes induced by longitudinal and lateral indentation loads. Secondly, no mechanical parameter could quantify the contribution of microarchitecture to mechanical properties of cancellous bone. Morphological and topological parameters have strong correlations with mechanical properties of cancellous bone, but they are not easy to interpret mechanically. To solve this problem, I put forward the concept of the ineffective bone mass (IneffBM); and built micro finite element (FE) models to directly quantify the effects of microarchitecture on the apparent modulus of the cancellous bone (E). My results showed IneffBM, together with bone volume fraction, could explain about 95% of the variation in E. The third question is at the vertebral level in clinical practice. The clinical screening of patients at high vertebral fracture risk is dependent on the lumbar bone mineral density (BMD) regardless of the type of vertebral fracture. But there are various forms of BMD and three different types of vertebral fractures. Nobody has asked if the predictor of fracture risk is dependent on the type of fracture. I asked the question and designed a novel testing rig to induce compression and wedge types of vertebral fractures. By studying the correlations between various forms of BMD and vertebral strength of compression and wedge types of fractures independently and unitedly, I found the fracture type could influence the selection of BMD predictors of vertebral strength. Finally, to predict the fracture type and strength in different fracture types, I built nonlinear FE models to simulate the experimental work of the third study. At present, my proposed parameter cannot accurately predict the fracture types, but the simulated strength has a stronger correlation with the experimental one than the best BMD predictor. This thesis partly answered the unresolved multiscale (µm-mm-cm) questions about the mechanical properties of the vertebra. Insights from this thesis are of significance in predicting vertebral strength.