Accuracy of MRI-based finite element assessment of distal tibia compared to mechanical testing

摘要

High-resolution MRI-derived finite element analysis (FEA) has been used in translational research to estimate the mechanical competence of human bone. However, this method has yet to be validated adequately under in vivo imaging spatial resolution or signal-to-noise conditions. We therefore compared MRI-based metrics of bone strength to those obtained from direct, mechanical testing. The study was conducted on tibiae from 17 human donors (12 males and five females, aged 33 to 88 years) with no medical history of conditions affecting bone mineral homeostasis. A 25 mm segment from each distal tibia underwent MR imaging in a clinical 3-Tesla scanner using a fast large-angle spin-echo (FLASE) sequence at 0.137 mm × 0.137 mm × 0.410 mm voxel size, in accordance with in vivo scanning protocol. The resulting high-resolution MR images were processed and used to generate bone volume fraction maps, which served as input for the micro-level FEA model. Simulated compression was applied to compute stiffness, yield strength, ultimate strength, modulus of resilience, and toughness, which were then compared to metrics obtained from mechanical testing. Moderate to strong positive correlations were found between computationally and experimentally derived values of stiffness (R² = 0.77, p < 0.0001), yield strength (R² = 0.38, p = 0.0082), ultimate strength (R² = 0.40, p = 0.0067), and resilience (R² = 0.46, p = 0.0026), but only a weak, albeit significant, correlation was found for toughness (R² = 0.26, p = 0.036). Furthermore, experimentally derived yield strength and ultimate strength were moderately correlated with MRI-derived stiffness (R² = 0.48, p = 0.0022 and R² = 0.58, p = 0.0004, respectively). These results suggest that high-resolution MRI-based finite element (FE) models are effective in assessing mechanical parameters of distal skeletal extremities.