Local 3D scaling properties for the analysis of trabecular bone extracted from high-resolution magnetic resonance imaging of human trabecular bone: Comparison with bone mineral density in the prediction of biomechanical strength in vitro

Rationale and Objectives. A novel, nonlinear morphologic measure [ΔP(α)] based on local 3D scaling properties was applied to high-resolution magnetic resonance images (HR-MRI) of human trabecular bone to predict biomechanical strength in vitro. Methods. We extracted ΔP(α) and traditional morphologic parameters (apparent trabecular volume fraction, apparent trabecular separation) from HR-MR images of 32 femoral and 13 spinal bone specimens. Furthermore, bone mineral density (BMD) and maximum compressive strength (MCS) were determined. The morphologic measures were compared with BMD in predicting the biomechanical strength. Results. In the vertebral (femoral) specimens, R² for MCS versus ΔP(α) was 0.87 (0.61) (P < 0.001). Correlation between BMD and MCS was 0.53 (P = 0.05) (0.79 [P < 0.001]) for the vertebral (femoral) specimens. For the femoral specimens, prediction of MCS could be improved further by combining BMD and morphologic parameters by multiple regression (R² = 0.88). Conclusions. Morphologic measures extracted from HR-MRI considering local 3D-scaling properties can be used to predict biomechanical properties of bone in vitro. They are superior to 2-dimensional standard linear morphometric measures and, depending on the anatomic location, more reliably predict bone strength as measured by MCS than does BMD.