Inter-laboratory reproduction and sensitivity study of a finite element model to quantify human femur failure load: Case of metastases.

INTRODUCTION: Metastases increase the risk of fracture when affecting the femur. Consequently, clinicians need to know if the patient's femur can withstand the stress of daily activities. The current tools used in clinics are not sufficiently precise. A new method, the CT-scan-based finite element analysis, gives good predictive results. However, none of the existing models were tested for reproducibility. This is a critical issue to address in order to apply the technique on a large cohort around the world to help evaluate bone metastatic fracture risk in patients. The aim of this study is then to evaluate 1) the reproducibility 2) the transposition of the reproduced model to another dataset and 3) the global sensitivity of one of the most promising models of the literature (original model). METHODS: The model was reproduced based on the paper describing it and discussion with authors to avoid reproduction errors. The reproducibility was evaluated by comparing the results given in the original model by the original first team (Leuven, Belgium) and the reproduced model made by another team (Lyon, France) on the same dataset of CT-scans of ex vivo femurs. The transposition of the model was evaluated by comparing the results of the reproduced model on two different datasets. The global sensitivity analysis was done by using the Morris method and evaluates the influence of the density calibration coefficient, the segmentation, the orientations and the length of the femur. RESULTS: The original and reproduced models are highly correlated (r2 = 0.95), even though the reproduced model gives systematically higher failure loads. When using the reproduced model on another dataset, predictions are less accurate (r2 with the experimental failure load decreases, errors increase). The global sensitivity analysis showed high influence of the density calibration coefficient (mean variation of failure load of 84 %) and non-negligible influence of the segmentation, orientation and length of the femur (mean variation of failure load between 7 and 10 %). CONCLUSION: This study showed that, although being validated, the reproduced model underperformed when using another dataset. The difference in performance depending on the dataset is commonly the cause of overfitting when creating the model. However, the dataset used in the original paper (Sas et al., 2020a) and the Leuven's dataset gave similar performance, which indicates a lesser probability for the overfitting cause. Also, the model is highly sensitive to density parameters and automation of measurement may minimize the uncertainty on failure load. An uncertainty propagation analysis would give the actual precision of such model and improve our understanding of its behavior and is part of future work.

Finite element models with automatic computed tomography bone segmentation for failure load computation.

Bone segmentation is an important step to perform biomechanical failure load simulations on in-vivo CT data of patients with bone metastasis, as it is a mandatory operation to obtain meshes needed for numerical simulations. Segmentation can be a tedious and time consuming task when done manually, and expert segmentations are subject to intra- and inter-operator variability. Deep learning methods are increasingly employed to automatically carry out image segmentation tasks. These networks usually need to be trained on a large image dataset along with the manual segmentations to maximize generalization to new images, but it is not always possible to have access to a multitude of CT-scans with the associated ground truth. It then becomes necessary to use training techniques to make the best use of the limited available data. In this paper, we propose a dedicated pipeline of preprocessing, deep learning based segmentation method and post-processing for in-vivo human femurs and vertebrae segmentation from CT-scans volumes. We experimented with three U-Net architectures and showed that out-of-the-box models enable automatic and high-quality volume segmentation if carefully trained. We compared the failure load simulation results obtained on femurs and vertebrae using either automatic or manual segmentations and studied the sensitivity of the simulations on small variations of the automatic segmentation. The failure loads obtained using automatic segmentations were comparable to those obtained using manual expert segmentations for all the femurs and vertebrae tested, demonstrating the effectiveness of the automated segmentation approach for failure load simulations.

Bone matrix quality in paired iliac bone biopsies from postmenopausal women treated for 12 months with strontium ranelate or alendronate.

Bone quality is altered mainly by osteoporosis, which is treated with modulators of bone quality. Knowledge of their mechanisms of action is crucial to understand their effects on bone quality. The goal of our study was to compare the action of alendronate (ALN) and strontium ranelate (SrRan) on the determinants of bone quality. The investigation was performed on over 60 paired human iliac biopsies. Paired samples correspond to biopsies obtained from the same patient, one before treatment (baseline) and one after 12 months of treatment, in postmenopausal women with osteoporosis. Vibrational spectroscopy (Raman and FTIRM) and nanoindentation were used to evaluate the effect of both drugs on bone quality at the ultrastructural level. Outcomes measured by vibrational spectroscopy and nanoindentation are sensitive to bone age. New bone packets are distinguished from old bone packets. Thus, the effect of bone age is distinguished from the treatment effect. Both drugs modify the mineral and organic composition in new and old bone in different fashions after 12 months of administration. The new bone formed during ALN administration is characterized by an increased mineral content, carbonation and apatite crystal size/perfection compared to baseline. Post-translational modifications of collagen are observed through an increase in the hydroxyproline/proline ratio in new bone. The proteoglycan content is also increased in new bone. SrRan directly modulates bone quality through its physicochemical actions, independent of an effect on bone remodeling. Strontium cations are captured by the hydrated layer of the mineral matrix. The mineral matrix formed during SrRan administration has a lower carbonate content and crystallinity after 12 months than at baseline. Strontium might create bonds (crosslinks) with collagen and noncollagenous proteins in new and old bone. The nanomechanical properties of bone were not modified with either ALN or SrRan, probably due to the short duration of administration. Our results show that ALN and SrRan have differential effects on bone quality in relation to their mechanism of action.

Anisotropic elastic properties of human femoral cortical bone and relationships with composition and microstructure in elderly.

The strong dependence between cortical bone elasticity at the millimetre-scale (mesoscale) and cortical porosity has been evidenced by previous studies. However, bone is an anisotropic composite material made by mineral, proteins and water assembled in a hierarchical structure. Whether the variations of structural and compositional properties of bone affect the different elastic coefficients at the mesoscale is not clear. Aiming to understand the relationships between bone elastic properties and compositions and microstructure, we applied state-of-the-art experimental modalities to assess these aspects of bone characteristics. All elastic coefficients (stiffness tensor of the transverse isotropic bone material), structure of the vascular pore network, collagen and mineral properties were measured in 52 specimens from the femoral diaphysis of 26 elderly donors. Statistical analyses and micromechanical modeling showed that vascular pore volume fraction and the degree of mineralization of bone are the most important determinants of cortical bone anisotropic mesoscopic elasticity. Though significant correlations were observed between collagen properties and elasticity, their effects in bone mesoscopic elasticity were minor in our data. This work also provides a unique set of data exhibiting a range of variations of compositional and microstructural cortical bone properties in the elderly and gives strong experimental evidence and basis for further development of biomechanical models for human cortical bone. STATEMENT OF SIGNIFICANCE: This study reports the relationships between microstructure, composition and the mesoscale anisotropic elastic properties of human femoral cortical bone in elderly. For the first time, we provide data covering the complete anisotropic elastic tensor, the microstructure of cortical vascular porosity, mineral and collagen characteristics obtained from the same or adjacent samples in each donor. The results revealed that cortical vascular porosity and degree of mineralization of bone are the most important determinants of bone anisotropic stiffness at the mesoscale. The presented data gives strong experimental evidence and basis for further development of biomechanical models for human cortical bone.