Abstract
Proximal femoral fractures are a serious life-threatening injury with high morbidity and mortality. Magnetic resonance (MR) imaging has potential to non-invasively assess proximal femoral bone strength in vivo through usage of finite element (FE) modelling (a technique referred to as MR-FE). To precisely assess bone strength, knowledge of measurement error associated with different MR-FE outcomes is needed. The objective of this study was to characterize the short-term in vivo precision errors of MR-FE outcomes (e.g., stress, strain, failure loads) of the proximal femur for fall and stance loading configurations using 13 participants (5 males and 8 females; median age: 27 years, range: 21-68), each scanned 3 times. MR-FE models were generated, and mean von Mises stress and strain as well as principal stress and strain were calculated for 3 regions of interest. Similarly, we calculated the failure loads to cause 5% of contiguous elements to fail according to the von Mises yield, Brittle Coulomb-Mohr, normal principal, and Hoffman stress and strain criteria. Precision (root-mean squared coefficient of variation) of the MR-FE outcomes ranged from 3.3% to 11.8% for stress and strain-based mechanical outcomes, and 5.8% to 9.0% for failure loads. These results provide evidence that MR-FE outcomes are a promising non-invasive technique for monitoring femoral strength in vivo.