Abstract
Purpose
Based on a structure function relationship, bone density distribution has been described as being representative of skeletal loading. As such, computed tomography (CT) may be used to visualize the structure of femoral head subchondral bone to allow in vivo quantification of joint mechanics without the need for implanted hardware. This study aims to characterize the distribution of subchondral bone density in the femoral head. We hypothesize that a non-uniform distribution of bone density will be observed, with correlation between left and right sides for a given patient.
Method
Femoral head surfaces were created bilaterally for thirty patients through semi-automatic segmentation of reconstructed CT data and used to map bone density, by shrinking them into the subchondral bone and averaging the grey values (linearly related to bone density) within five millimeters of the articular surface. Density maps were then oriented with the center of the head at the origin, the femoral mechanical axis (FMA) aligned with the vertical, and the posterior condylar axis (PCA) aligned with the horizontal. Twelve regions were created by dividing the density maps into three concentric rings at increments of thirty degrees from the horizontal, then splitting into four quadrants along the anterior-posterior and medial-lateral axes. Average bone density within each region was then calculated using histogram analysis. All analysis was performed with AmriaDEV 5.2.2 image analysis software (Visage Imaging, Carlsbad USA).
Results
The regions representing the most-superior medial portion of the femoral head (one and four) were found to have significantly higher densities compared to all other regions (p<0.05). One exception to this was on the left side where region five, on the posteromedial side (adjacent but inferior to region one) was not found to be significantly different from region four (p=0.595). Significant side-to-side correlations were found for all regions (r=0.90 to r=0.40), with very strong correlations for the highest density regions (r=0.85 for region one, r=0.84 for region four). Side-to-side differences in measured bone density were seen for two of the low-density regions in the anterolateral portion of the femoral head (p<0.05).
Conclusion
The locations of highest bone density on the femoral head correspond with principal contact areas found in previous investigations of loading in the human hip joint. Regions of the femoral head found to have highest average bone density, correspond well with high-density regions found previously for the acetabulum. The high correlation found between the left and right sides indicates that this tool may be used to detect early differences in bone density caused by unilateral hip pathologies, such as osteonecrosis or osteoarthritis of the femoral head, prior to the presentation of clinical symptoms.
