Cortical porosity assessment in the distal radius: A comparison of HR-pQCT measures with Synchrotron-Radiation micro-CT-based measures.

Abstract

To determine the agreement between cortical porosity derived from high resolution peripheral quantitative computed tomography (HR-pQCT) (via standard threshold, mean density and density inhomogeneity methods) and synchrotron radiation micro-CT (SR-μCT) derived porosity at the distal radius. We scanned 10 cadaveric radii (mean donor age: 79, SD 11 years) at the standard distal region using HR-pQCT and SR-μCT at voxel sizes of 82 μm and 17.7 μm, respectively. Common cortical regions were delineated for each specimen in both imaging modalities. HR-pQCT images were analyzed for cortical porosity using the following methods: Standard threshold, mean density, and density inhomogeneity (via recommended and optimized equations). We assessed agreement in porosity measures between HR-pQCT methods and SR-μCT by reporting predicted variance from linear regression and mean bias with limits of agreement (LOA). The standard threshold and mean density methods predicted 85% and 89% of variance and indicated underestimation (mean bias -9.1%, LOA -15.9% to -2.2%) and overestimation (10.4%, 4.6% to 16.2%) of porosity, respectively. The density inhomogeneity method with recommended equation predicted 89% of variance and mean bias of 14.9% (-4.3 to 34.2) with systematic over-estimation of porosity in more porous specimens. The density inhomogeneity method with optimized equation predicted 91% of variance without bias (0.0%, -5.3 to 5.2). HR-pQCT imaged porosity assessed with the density inhomogeneity method with optimized equation indicated the best agreement with SR-μCT derived porosity.