Detection of trabecular bone microdamage by micro-computed tomography

Abstract

Microdamage is an important component of bone quality and affects bone remodeling. Improved techniques to assess microdamage without the need for histological sectioning would provide insight into the role of microdamage in trabecular bone strength by allowing the spatial distribution of damage within the trabecular microstructure to be measured. Nineteen cylindrical trabecular bone specimens were prepared and assigned to two groups. The specimens in group I were damaged to 3% compressive strain and labeled with BaSO(4). Group II was not loaded, but was labeled with BaSO(4). Micro-computed tomography (Micro-CT) images of the specimens were obtained at 10 microm resolution. The median intensity of the treated bone tissue was compared between groups. Thresholding was also used to measure the damaged area fraction in the micro-CT scans. The histologically measured damaged area fraction, the median CT intensity, and the micro-CT measured damaged area fraction were all higher in the loaded group than in the unloaded group, indicating that the micro-CT images could differentiate the damaged specimen group from the unloaded specimens. The histologically measured damaged area fraction was positively correlated with the micro-CT measured damaged area fraction and with the median CT intensity of the bone, indicating that the micro-CT images can detect microdamage in trabecular bone with sufficient accuracy to differentiate damage levels between samples. This technique provides a means to non-invasively assess the three-dimensional distribution of microdamage within trabecular bone test specimens and could be used to gain insight into the role of trabecular architecture in microdamage formation.

Figure 1

At every scanning position (1–5), groups of ten images spaced 10 μm apart were obtained. One group was located at the specimen axis, two were centered 1 mm above and below the axis, and two were centered 2.5 mm above and below the axis.

Figure 2

Filtered (left) and thresholded micro-CT images from a loaded and an unloaded sample that were labeled with BaSO₄. White regions in the thresholded images represent regions with higher attenuation labeled by BaSO₄.

Figure 3

A micrograph showing damaged areas (arrows). The damaged area fraction was calculated as the ratio of squares that contained microcracks to the total number of squares containing bone.

Figure 4

(a) The mean damaged areas measured by both histology and by micro-CT were lower for the undamaged samples than for the damaged samples. (b) The BaSO₄ treated samples had higher median intensities than unlabeled control samples, and the loaded samples had a higher median intensity than the unloaded samples. (* indicates significant differences across groups p < 0.02). The boxes in the plots bound the 25th and 75th percentiles of the data, while the bar indicates the range.

Figure 5

(a) The CT measured damage area fraction, Dx.Ar._CT, (p < 0.005, A) and the median image intensity of the bone tissue (p < 0.04, B) were both correlated with the histologically measured damaged area fraction (Dx.Ar._H). The open circles represent the specimens that were mechanically loaded, and the filled circles are specimens that were not loaded.

Figure 6

A backscattered SEM image of a loaded and treated specimen showed several microcracks surrounded by bright regions. The square surrounds a region with several adjacent cracks less than 5 μm in length, which would appear histologically as diffuse damage. EDS was used to measure the elemental composition, which verified that the bright region labeled A contained a higher level of barium than the region labeled B. Cracks that are not labeled with BaSO₄ were considered to be artifacts caused by drying and embedding the sample for SEM imaging.