BACKGROUND:Vertebral metastases compromise bone integrity and may lead to fractures with substantial clinical and economic consequences. Finite element (FE) modeling based on computed tomography (CT) scans enables individualized biomechanical assessment but remains underused clinically due to workflow limitations. OBJECTIVE:To evaluate the sensitivity and accuracy of a patient-specific finite element (FE) simulation pipeline that incorporates automated vertebral segmentation for predicting vertebral strength in metastatic cases. METHODS:Thirty vertebrae (18 metastatic, 12 healthy) from 12 patients were analyzed using two finite element (FE) models (perfectly elasto-plastic vs. linear elastic) within an automated segmentation pipeline. Intra- and inter-operator reproducibility, sensitivity to boundary condition definition, and model-type comparisons were performed. Failure loads were normalized to derive equivalent stresses. RESULTS:Automated segmentation demonstrated excellent agreement with manual reference (R2 = 0.996 p < 0.001; no systematic bias). Sensitivity to operator-defined boundary conditions was minimal (<1%). The elasto-plastic model provided significantly lower failure loads than the linear elastic model (-25% on average). CONCLUSION:This study supports the feasibility of a fully-automated CT-based finite-element (FE) pipeline for vertebral fracture risk assessment in both healthy and metastatic vertebrae.

Spinal metastases are a major cause of pain, vertebral fracture, instability, and neurological compromise. The Spinal Instability Neoplastic Score (SINS) has substantially improved the standardized assessment of neoplastic spinal instability and remains the reference clinical tool for evaluating the mechanical component of metastatic spinal disease. However, it remains a semi-quantitative score, does not directly quantify vertebral strength, and is particularly limited in the intermediate SINS range (7-12), where uncertainty regarding true mechanical risk and therapeutic strategy is often greatest. In this context, vertebral mechanical failure is better understood as the consequence of an imbalance between applied spinal load and residual vertebral strength rather than as a purely morphological imaging abnormality. This review examines the current clinical approach to vertebral mechanical failure in spinal metastases and discusses the potential contribution of patient-specific biomechanical modeling, especially CT-based finite element modeling, to a more quantitative and individualized assessment of mechanical failure risk. Finite element models offer a mechanistically grounded framework that integrates vertebral geometry, lesion characteristics, bone density, and loading conditions to estimate vertebral strength more directly than current clinical scores. Experimental and ex vivo studies support their biomechanical relevance, while recent translational developments suggest increasing feasibility for clinical implementation. Rather than replacing existing clinical frameworks, finite element modeling may become a valuable extension of current assessment, particularly in patients with intermediate-risk lesions in whom decision-making remains uncertain.

Computed tomography (CT) based finite element (FE) models for bone strength prediction of metastasised femurs require calibration of Hounsfield Units (HU) to bone mineral density, to then derive mechanical parameters. Calibration is generally performed using a calibration phantom, which is a limitation for clinical application. The purpose of this study was to evaluate a phantomless calibration method for linear and non-linear FE models of femoral strength prediction. The constitutive laws of the linear and non-linear models, respectively, are defined by two and six density-dependent parameters, respectively. Femurs of 54 patients with bone metastases were CT-scanned with a calibration phantom. Linear and non-linear FE models were created for each femur, following calibration using (i) the phantom, and (ii) the air-fat-muscle (AFM) calibration method, where HU peaks of air, fat, and muscle are extracted and linearly fitted to reference density values. Apparent density and failure loads were compared between each calibration method, and the intra- and inter- operator variability was computed, for each model. A very good reproducibility of the AFM method was found for both models, and a significant correlation with the gold standard phantom-based calibration in terms of apparent density (R2 = 0.78, p<0.01) and femoral strength (R2 = 0.85, p<0.01). Limits of agreement between each method were narrower for the linear model than for the non-linear model, suggesting that the linear model is less sensitive to density calibration. The phantomless calibration method is a promising approach to broaden routine use of bone strength numerical simulation in bone metastatic patients.

Highlights• Automated segmentation of vertebral body showed excellent agreement with expert• Sensitivity to operator-defined boundary conditions was minimal• Discrepancies between elastic and elastoplastic models

Bone permeability is a key parameter that drives osteocyte-based mechanobiological modelling and remodelling. While previous experimental and numerical studies have estimated bone permeability based on the morphology of the lacuno-canalicular network, these studies often relied on simplified geometries. In the current study, bone permeability was characterized using more realistic canalicular geometry for the morphological data. Bone samples harvested from 27 human femoral bones were investigated using synchrotron radiation-based nano-computed tomography with a voxel size of 100 nm. After segmenting the canaliculi and lacunae, each canaliculus was investigated individually by applying a distance map and watershed algorithms. Bone permeability based on canalicular morphology was then assessed using the Kozeny relation, which defines the permeability of a porous medium with capillary-like pores. An averaged intrinsic permeability value of 8.8 10-18 m2 was obtained. It should be noted that this study considered an empty canalicular network, however in vivo, both cellular and peri-cellular matrices decrease space for interstitial fluid flow and thus permeability. Furthermore, a voxel size of 100 nm does not allow for the detection of smaller canaliculi, which may also modify average permeability. With the current data set and the analytic process applied, the results showed a heterogeneous permeability distribution within bone tissue, both when comparing osteonal and interstitial tissues and within an individual osteon. A difference was observed between male and female samples, and permeability appeared to significantly decrease with age. Finally, a significant correlation was found between permeability and canalicular length density, defined as canalicular length per unit bone volume. This study proposes a new form of the Kozeny law to express bone canalicular permeability as a proportional relationship with the canalicular length density. Importantly, this parameter can be directly quantified through confocal fluorescence microcopy, which is more convenient than synchrotron radiation-based nano-computed tomography. In conclusion, the current study confirms that confocal microscopy can be serve as a reliable tool to estimate bone permeability. However, the permeability values calculated here are solely based on canalicular morphology and do not consider cellular and peri-cellular intra-canalicular features.

D’importants progrès thérapeutiques ont été réalisés dans la prise en charge des patients atteints de cancer, avec une amélioration notable de la survie, y compris chez les patients porteurs de métastases osseuses de stade IV. Dans ce contexte, la santé locomotrice devient un enjeu central pour préserver l’autonomie et la qualité de vie des patients. L’évaluation du risque fracturaire des métastases osseuses constitue l’élément pivot de la stratégie décisionnelle, guidant les prescriptions d’agents anti-résorptifs, les recommandations d’activité physique et les interventions locales telles que la radiothérapie, la chirurgie ou la radiologie interventionnelle. Une étape incontournable, du fait du caractère souvent asymptomatique et diffus des métastases osseuses, est la cartographie systématique des localisations osseuses, notamment au sein des os porteurs. Pour chaque localisation identifiée, l’évaluation du risque fracturaire repose actuellement sur des approches qualitatives par imagerie, appuyées par des scores comme ceux de Mirels ou SINS. Ces outils présentent néanmoins des limites importantes, justifiant le développement de nouvelles méthodes d’évaluation du risque fracturaire dans les métastases osseuses et le myélome. La simulation numérique à partir d’images quantitatives scanographiques qCT constitue l’un des outils émergents permettant d’estimer, de façon personnalisée, la résistance osseuse tumorale et de mieux évaluer le risque de fracture pathologique au niveau fémoral et vertébral. Les générations futures de simulations numériques, enrichies par l’intelligence artificielle, intégreront des conditions de charge multiples liées au mouvement afin de refléter au mieux les contraintes de la vie réelle. Cette approche vise à guider la réhabilitation et l’activité physique des patients dans une logique de médecine personnalisée.

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.

Osteopontin (OPN) and Bone Sialoprotein (BSP), abundantly expressed by osteoblasts and osteoclasts, appear to have important, partly overlapping functions in bone. In gene-knockout (KO, -/-) models of either protein and their double (D)KO in the same CD1/129sv genetic background, we analyzed the morphology, matrix characteristics, and biomechanical properties of femur bone in 2 and 4 month old, male and female mice. OPN−/− mice display inconsistent, perhaps localized hypermineralization, while the BSP−/− are hypomineralized throughout ages and sexes, and the low mineralization of young DKO mice recovers with age. The higher contribution of primary bone remnants in OPN−/− shafts suggests a slow turnover, while their lower percentage in BSP−/− indicates rapid remodeling, despite FTIR-based evidence in this genotype of a high maturity of the mineralized matrix. In 3-point bending assays, OPN−/− bones consistently display higher Maximal Load, Work to Max. Load and in young mice Ultimate Stress, an intrinsic characteristic of the matrix. Young male and old female BSP−/− also display high Work to Max. Load along with low Ultimate Stress. Principal Component Analysis confirms the major role of morphological traits in mechanical competence, and evidences a grouping of the WT phenotype with the OPN−/− and of BSP−/− with DKO, driven by both structural and matrix parameters, suggesting that the presence or absence of BSP has the most profound effects on skeletal properties. Single or double gene KO of OPN and BSP thus have multiple distinct effects on skeletal phenotypes, confirming their importance in bone biology and their interplay in its regulation.

Osteocytes are the major actors in bone mechanobiology. Within bone matrix, they are trapped close together in a submicrometric interconnected network: the lacunocanalicular network (LCN). The interstitial fluid circulating within the LCN transmits the mechanical information to the osteocytes that convert it into a biochemical signal. Understanding the interstitial fluid dynamics is necessary to better understand the bone mechanobiology. Due to the submicrometric dimensions of the LCN, making it difficult to experimentally investigate fluid dynamics, numerical models appear as a relevant tool for such investigation. To develop such models, there is a need for geometrical and morphological data on the human LCN. This study aims at providing morphological data on the human LCN from measurement of 27 human femoral diaphysis bone samples using synchrotron radiation nanocomputed tomography with an isotropic voxel size of 100 nm. Except from the canalicular diameter, the canalicular morphological parameters presented a high variability within one sample. Some differences in terms of both lacunar and canalicular morphology were observed between the male and female populations. But it has to be highlighted that all the canaliculi cannot be detected with a voxel size of 100 nm. Hence, in the current study, only a specific population of large canaliculi that could be characterize. Still, to the authors knowledge, this is the first time such a data set was introduced to the community. Further processing will be achieved in order to provide new insight on the LCN permeability.

Purpose: Bone metastasis have a major impact on the quality of life of patients and they are diverse in terms of size and location, making their segmentation complex. Manual segmentation is time-consuming, and expert segmentations are subject to operator variability, which makes obtaining accurate and reproducible segmentations of bone metastasis on CT-scans a challenging yet important task to achieve. Materials and Methods: Deep learning methods tackle segmentation tasks efficiently but require large datasets along with expert manual segmentations to generalize on new images. We propose an automated data synthesis pipeline using 3D Denoising Diffusion Probabilistic Models (DDPM) to enchance the segmentation of femoral metastasis from CT-scan volumes of patients. We used 29 existing lesions along with 26 healthy femurs to create new realistic synthetic metastatic images, and trained a DDPM to improve the diversity and realism of the simulated volumes. We also investigated the operator variability on manual segmentation. Results: We created 5675 new volumes, then trained 3D U-Net segmentation models on real and synthetic data to compare segmentation performance, and we evaluated the performance of the models depending on the amount of synthetic data used in training. Conclusion: Our results showed that segmentation models trained with synthetic data outperformed those trained on real volumes only, and that those models perform especially well when considering operator variability.

Metastases increase the risk of fracture when affecting the femur. Consequently, clinicians need to know if the patients 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. Please see pdf file

Clinical use of finite element analysis requires validation and reproducibility studies. The current study compared two models of vertebral bodies including endplates, on the same experimental dataset and evaluated the influence of the operator on the failure load. Models used were strongly correlated (R2=0.91). The intra-operator reproducibility was 6.4% and 3.5 % for each model. Both simulated results were close to experimental results. The differences in performance could be associated to the differences in segmentation process, mesh (hexahedral vs tetrahedral), material representation and failure criteria. Linear analysis did not decrease model accuracy. Comparison with literature for accuracy and precision shows a wide range of values partly related to the different experimental datasets and the different modelling approaches. Models benchmark using the same experimental dataset are needed to go towards clinical applications.

As far as their mechanical properties are concerned, cancerous lesions can be confused with healthy surrounding tissues in elastography protocols if only the magnitude of moduli is considered. We show that the frequency dependence of the tissue's mechanical properties allows for discriminating the tumor from other tissues, obtaining a good contrast even when healthy and tumor tissues have shear moduli of comparable magnitude. We measured the shear modulus G*(ω) of xenograft subcutaneous tumors developed in mice using breast human cancer cells, compared with that of fat, skin and muscle harvested from the same mice. As the absolute shear modulus |G*(ω)| of tumors increases by 42% (from 5.2 to 7.4 kPa) between 0.25 and 63 Hz, it varies over the same frequency range by 77% (from 0.53 to 0.94 kPa) for the fat, by 103% (from 3.4 to 6.9 kPa) for the skin and by 120% (from 4.4 to 9.7 kPa) for the muscle. These measurements fit well to the fractional model G*(ω)=K(iω)n, yielding a coefficient K and a power-law exponent n for each sample. Tumor, skin and muscle have comparable K parameter values, that of fat being significantly lower; the p-values given by a Mann-Whitney test are above 0.14 when comparing tumor, skin and muscle between themselves, but below 0.001 when comparing fat with tumor, skin or muscle. With regards the n parameter, tumor and fat are comparable, with p-values above 0.43, whereas tumor differs from both skin and muscle, with p-values below 0.001. Tumor tissues thus significantly differs from fat, skin and muscle on account of either the K or the n parameter, i.e. of either the magnitude or the frequency-dependence of the shear modulus.

Supplementary Figure 3 from Dual Function of ERRα in Breast Cancer and Bone Metastasis Formation: Implication of VEGF and Osteoprotegerin