Role of trabecular microarchitecture and its heterogeneity parameters in the mechanical behavior of ex vivo human L3 vertebrae

Abstract

Low bone mineral density (BMD) is a strong risk factor for vertebral fracture risk in osteoporosis. However, many fractures occur in people with moderately decreased or normal BMD. Our aim was to assess the contributions of trabecular microarchitecture and its heterogeneity to the mechanical behavior of human lumbar vertebrae. Twenty-one human L(3) vertebrae were analyzed for BMD by dual-energy X-ray absorptiometry (DXA) and microarchitecture by high-resolution peripheral quantitative computed tomography (HR-pQCT) and then tested in axial compression. Microarchitecture heterogeneity was assessed using two vertically oriented virtual biopsies--one anterior (Ant) and one posterior (Post)--each divided into three zones (superior, middle, and inferior) and using the whole vertebral trabecular volume for the intraindividual distribution of trabecular separation (Tb.Sp*SD). Heterogeneity parameters were defined as (1) ratios of anterior to posterior microarchitectural parameters and (2) the coefficient of variation of microarchitectural parameters from the superior, middle, and inferior zones. BMD alone explained up to 44% of the variability in vertebral mechanical behavior, bone volume fraction (BV/TV) up to 53%, and trabecular architecture up to 66%. Importantly, bone mass (BMD or BV/TV) in combination with microarchitecture and its heterogeneity improved the prediction of vertebral mechanical behavior, together explaining up to 86% of the variability in vertebral failure load. In conclusion, our data indicate that regional variation of microarchitecture assessment expressed by heterogeneity parameters may enhance prediction of vertebral fracture risk.

Fig. 1

HR-pQCT slice of L₃ vertebra. Trabecular region of interest (ROI) was defined manually in order to exclude cortical component of the vertebral body. Virtual biopsies were positioned using two lines drawn on the vertebral body, one line for the middle anteroposterior axis and one line for the middle mediolateral axis. Each line divided the vertebral body into four quadrants. Biopsies were strictly centered on the middle anteroposterior axis and on both sides of the mediolateral axis to avoid the cortical shell anteriorly and the venus plexus posteriorly by projection in the vertical direction in the HR-pQCT slice stack.

Fig. 2

Whole trabecular volume of L₃ vertebra and the two virtual biopsies (82-µm isotropic voxel size) each divided into three vertical zones (superior, middle, and inferior).