This dissertation reports on a study that explored whether the three-dimensional geometry of pore networks in bones is morphologically optimized to resist local mechanical strain.
The research presented in this dissertation had the purpose of determining whether cortical porosity morphologically adapts to localized mechanical conditions, to reduce the probability of microcrack propagation and spontaneous fracture. The author’s research tested the hypothesis that porosity in bones is morphologically optimized for localized mechanical loading environment, i.e., the author asked if pore geometry is less prone to microdamage in high-strain regions, where microscopic damage is more likely to occur. The author hypothesized an overarching structure-chain model in which two things happen: high-strain regions accumulate smaller, more isolated, longitudinally oriented pores due to more frequent remodeling; and low-strain regions accumulate larger, more highly connected, obliquely oriented pores due to more frequent disuse-related resorption. The author extracted regions from human bones, including the right-side femoral neck, for being common sites of osteoporotic fracture; and the right-side midshaft fourth rib, which served as a relatively unloaded control; and sample sources included one male and female per age decade, from 20s to 80s. The author used high-resolution micro-CT imaging to reconstruct three-dimensional pore networks from 10mm thick cross-sections, and processed them with custom routines that extracted and characterized porosity by bone type, pore type, and cross-section region. The author notes that structure-strain relationship deteriorates with age, as pores become significantly more convergent in the femoral neck and rib, and that decay is accelerated in females due to the loss of endosteum-preserving estrogen at menopause. The author also notes that there are significant interactions with cortical thickness, and with broader demographic co-variates that represent body size, bone mineral density, femoral neck gross geometry, and cross-sectional mass and shape. The research produced broadly applicable, open-source code for many applications in processing micro-CT images of bone tissue for morphometric analysis, including brightness-contrast thresholding, automated pore extraction, and batch process lateral merging of large cut examples.