High-resolution MR images of distal skeletal extremities have been shown to be suitable for the generation of finite-element models that allow the estimation of bone strength. However, the resolutions achievable by in vivo MRI are on the order of trabecular bone thickness due to signal-to-noise ratio limitations imposed by scan time. The goal of this study was to validate the MRI assessment of bone strength using a protocol analogous to that used for in vivo imaging by comparing to those obtained by direct mechanical testing of bone specimens.

Human tibia specimens encompassing the distal metaphysis were harvested from cadavers of 13 male and 5 female donors from 33 to 88 years of age. Donors that have a medical history involving conditions affecting bone or bone mineral homeostasis were excluded. From each donor, a 20-mm thick segment of the distal tibia encompassing the distal metaphysis was sectioned perpendicular to the bone’s anatomic axis such that the distal boundary is 15mm proximal to the distal end plate. Residual soft tissue on the periosteal cortex was carefully removed. This segment was selected to correspond to the in vivo MRI scan region.The specimens was imaged distal-end first and anterior-side upward (analogous to feet-first supine position in patient imaging of distal tibia) on a 3T whole body clinical MRI scanner (Siemens Tim Trio) with the same in-house built surface coil used for MR imaging of the ankle in patients. A 3D fast large-angle spin echo (FLASE) pulse sequence was used with the following parameters: flip angle 140O, TR/TE 80/10.5 ms, and voxel size 0.137x0.137x0.410 mm3 with the third dimension representing the axial direction. The imaging parameters were identical to those used for MR imaging of the ankle in human subject studies in our laboratory. Immediately following imaging specimens were stored at -20OC until mechanical testing.

The acquired µMR images were processed first using a local thresholding algorithm to correct intensity variations caused by inhomogeneous sensitivity of the receiving coil. The grayscale voxel values of the image were linearly scaled with pure marrow and pure bone having minimum and maximum values, respectively. The resulting 3D array is referred to as the bone-volume-fraction map with entries representing the percentage of the voxel occupied by bone. Each of the voxels was then directly converted to a hexahedral (brick) finite element with dimensions corresponding to the voxel size. Tissue material properties were chosen as isotropic and linearly elastic with each element’s Young’s modulus set linearly proportional to the respective BVF value of the corresponding voxel. The whole-bone section axial stiffness was computed by custom designed finite element solver. Simulated compression was applied along the bone’s longitudinal axis by applying constant displacement to all vertices of finite elements in the proximal face while keeping those in the distal face fixed. The whole bone section axial stiffness was computed as the ratio of the reaction force on the proximal face to calculated applied displacement. The specimens were then subjected to mechanical testing. Uniaxial compression tests were performed on a servo-hydraulic material testing machine (MTS810, MTS, Minneapolis, MN) by placing the 20-mm whole-segment distal tibia specimens between two steel plates. The displacement was measured using a 25-mm gauge length extensometer attached to loading platens (MTS 634. 11F-24, MTS, Minneapolis, MN) while a 100kN load transducer (MTS 661.20E-03, MTS, Minneapolis, MN) was used to record the load. Specimens were tested destructively under displacement at a rate of 1mm/min until the ultimate load was reached. Stiffness was then calculated as the initial tangent of the force-displacement curve. A graphical illustration of the complete methodology is presented in Figure 1.

We examined the hypothesis that MRI-derived axial stiffness is highly correlated with stiffness and obtained by mechanical testing. Correlation between the computationally-predicted stiffness and experimentally measured values was assessed. All statistical analysis was performed using JMP Discovery Software (Version 7.02, SAS Institute, Inc.). The analysis reviled that there is a strong positive correlation (R2 = 0.84) between the experimental and computational stiffness values (Figure 2).Previous studies have compared MRI-derived mechanical parameters of distal skeletal extremities to those obtained on the basis of high-resolution (~0.025 mm) micro-CT images of cadaveric bone. The present study provided a direct validation of MRI-based strength assessment of bone. Further studies will examine the accuracy of predicting yield strength, ultimate strength, and fracture toughness using MRI-based finite element modeling.The findings from this study support the use of MRI-based finite element analysis to reliably predict the mechanical competence of distal extremities in human subjects.