Correlation analysis of cortical geometry of tibia and humerus of white leghorns using clinical quantitative computed tomography and microcomputed tomography scans1
Peripheral quantitative computed tomography (QCT) has been used in poultry bone research in recent years to analyze cortical and cross-sectional geometry. For QCT to be used as a standard research tool for analysis of bones of laying hens (cortical thickness <2 mm), the accuracy of the scans must be assessed. The primary difference between the QCT and micro-computed tomography (micro-CT) is image resolution. Image resolution is inversely related to the pixel size. The aim of the current study was to correlate the cortical parameters measured using clinical CT scans with the measurements from micro-CT, the current gold standard. A total of 15 tibiae and 14 humeri of Lohmann White hens was scanned using clinical CT and micro-CT. Reconstruction of the scans generated images with final voxel resolution of 195 μm for clinical CT scans and 46 μm for micro-CT scans. Cortical and total area were measured using MIMICS® software at proximal, middle, and distal locations of 20 mm sections of humerus diaphysis and 30 mm sections of tibia diaphysis. The total area for proximal and middle locations as well as proximal cortical area measurements for humeri produced strong correlation coefficients (R ≥ 0.70). Moderate strength correlation coefficients (R = 0.40 to 0.60) in humeri were seen in middle and distal cortical areas. Distal total area in humeri displayed a weak correlation coefficient (R ≤ 0.3; P = 0.25). Overall, tibiae demonstrated a weaker correlation. Proximal and middle cortical areas indicated moderate correlation coefficients (R = 0.40 to 0.60), while proximal and middle total areas accompanied by distal cortical and total area displayed weak correlation coefficients (R ≤ 0.3). Only the middle cortical area measurement for tibiae was significant (P = 0.03). These results indicate stronger correlation for humeri measurements among the scans than tibia. Overall, cross-sectional area measurements were only low to moderately correlated between clinical and micro-CT scans.