Automatic segmentation of lower limb muscles from MR images of post-menopausal women based on deep learning and data augmentation.

Individual muscle segmentation is the process of partitioning medical images into regions representing each muscle. It can be used to isolate spatially structured quantitative muscle characteristics, such as volume, geometry, and the level of fat infiltration. These features are pivotal to measuring the state of muscle functional health and in tracking the response of the body to musculoskeletal and neuromusculoskeletal disorders. The gold standard approach to perform muscle segmentation requires manual processing of large numbers of images and is associated with significant operator repeatability issues and high time requirements. Deep learning-based techniques have been recently suggested to be capable of automating the process, which would catalyse research into the effects of musculoskeletal disorders on the muscular system. In this study, three convolutional neural networks were explored in their capacity to automatically segment twenty-three lower limb muscles from the hips, thigh, and calves from magnetic resonance images. The three neural networks (UNet, Attention UNet, and a novel Spatial Channel UNet) were trained independently with augmented images to segment 6 subjects and were able to segment the muscles with an average Relative Volume Error (RVE) between -8.6% and 2.9%, average Dice Similarity Coefficient (DSC) between 0.70 and 0.84, and average Hausdorff Distance (HD) between 12.2 and 46.5 mm, with performance dependent on both the subject and the network used. The trained convolutional neural networks designed, and data used in this study are openly available for use, either through re-training for other medical images, or application to automatically segment new T1-weighted lower limb magnetic resonance images captured with similar acquisition parameters.

Automatic segmentation of skeletal muscles from MR images using modified U-Net and a novel data augmentation approach.

Rapid and accurate muscle segmentation is essential for the diagnosis and monitoring of many musculoskeletal diseases. As gold standard, manual annotation suffers from intensive labor and high inter-operator reproducibility errors. In this study, deep learning (DL) based automatic muscle segmentation from MR scans is investigated for post-menopausal women, who normally experience a decline in muscle volume. The performance of four Deep Learning (DL) models was evaluated: U-Net and UNet++ and two modified U-Net networks, which combined feature fusion and attention mechanisms (Feature-Fusion-UNet, FFU, and Attention-Feature-Fusion-UNet, AFFU). The models were tested for automatic segmentation of 16-lower limb muscles from MRI scans of two cohorts of post-menopausal women (11 subjects in PMW-1, 8 subjects in PMW-2; from two different studies so considered independent datasets) and 10 obese post-menopausal women (PMW-OB). Furthermore, a novel data augmentation approach is proposed to enlarge the training dataset. The results were assessed and compared by using the Dice similarity coefficient (DSC), relative volume error (RVE), and Hausdorff distance (HD). The best performance among all four DL models was achieved by AFFU (PMW-1: DSC 0.828 ± 0.079, 1-RVE 0.859 ± 0.122, HD 29.9 mm ± 26.5 mm; PMW-2: DSC 0.833 ± 0.065, 1-RVE 0.873 ± 0.105, HD 25.9 mm ± 27.9 mm; PMW-OB: DSC 0.862 ± 0.048, 1-RVE 0.919 ± 0.076, HD 34.8 mm ± 46.8 mm). Furthermore, the augmentation of data significantly improved the DSC scores of U-Net and AFFU for all 16 tested muscles (between 0.23% and 2.17% (DSC), 1.6%-1.93% (1-RVE), and 9.6%-19.8% (HD) improvement). These findings highlight the feasibility of utilizing DL models for automatic segmentation of muscles in post-menopausal women and indicate that the proposed augmentation method can enhance the performance of models trained on small datasets.

The loading direction dramatically affects the mechanical properties of the mouse tibia.

Introduction: The in vivo tibial loading mouse model has been extensively used to evaluate bone adaptation in the tibia after mechanical loading treatment. However, there is a prevailing assumption that the load is applied axially to the tibia. The aim of this in silico study was to evaluate how much the apparent mechanical properties of the mouse tibia are affected by the loading direction, by using a validated micro-finite element (micro-FE) model of mice which have been ovariectomized and exposed to external mechanical loading over a two-week period. Methods: Longitudinal micro-computed tomography (micro-CT) images were taken of the tibiae of eleven ovariectomized mice at ages 18 and 20 weeks. Six of the mice underwent a mechanical loading treatment at age 19 weeks. Micro-FE models were generated, based on the segmented micro-CT images. Three models using unitary loads were linearly combined to simulate a range of loading directions, generated as a function of the angle from the inferior-superior axis (θ, 0°-30° range, 5° steps) and the angle from the anterior-posterior axis (ϕ, 0°: anterior axis, positive anticlockwise, 0°-355° range, 5° steps). The minimum principal strain was calculated and used to estimate the failure load, by linearly scaling the strain until 10% of the nodes reached the critical strain level of -14,420 µÎµ. The apparent bone stiffness was calculated as the ratio between the axial applied force and the average displacement along the longitudinal direction, for the loaded nodes. Results: The results demonstrated a high sensitivity of the mouse tibia to the loading direction across all groups and time points. Higher failure loads were found for several loading directions (θ = 10°, ϕ 205°-210°) than for the nominal axial case (θ = 0°, ϕ = 0°), highlighting adaptation of the bone for loading directions far from the nominal axial one. Conclusion: These results suggest that in studies which use mouse tibia, the loading direction can significantly impact the failure load. Thus, the magnitude and direction of the applied load should be well controlled during the experiments.

Cortical bone adaptation response is region specific, but not peak load dependent: insights from µ CT image analysis and mechanostat simulations of the mouse tibia loading model.

The two major aims of the present study were: (i) quantify localised cortical bone adaptation at the surface level using contralateral endpoint imaging data and image analysis techniques, and (ii) investigate whether cortical bone adaptation responses are universal or region specific and dependent on the respective peak load. For this purpose, we re-analyse previously published µ CT data of the mouse tibia loading model that investigated bone adaptation in response to sciatic neurectomy and various peak load magnitudes (F = 0, 2, 4, 6, 8, 10, 12 N). A beam theory-based approach was developed to simulate cortical bone adaptation in different sections of the tibia, using longitudinal strains as the adaptive stimuli. We developed four mechanostat models: universal, surface-based, strain directional-based, and combined surface and strain direction-based. Rates of bone adaptation in these mechanostat models were computed using an optimisation procedure (131,606 total simulations), performed on a single load case (F = 10 N). Subsequently, the models were validated against the remaining six peak loads. Our findings indicate that local bone adaptation responses are quasi-linear and bone region specific. The mechanostat model which accounted for differences in endosteal and periosteal regions and strain directions (i.e. tensile versus compressive) produced the lowest root mean squared error between simulated and experimental data for all loads, with a combined prediction accuracy of 76.6, 55.0 and 80.7% for periosteal, endosteal, and cortical thickness measurements (in the midshaft of the tibia). The largest root mean squared errors were observed in the transitional loads, i.e. F = 2 to 6 N, where inter-animal variability was highest. Finally, while endpoint imaging studies provide great insights into organ level bone adaptation responses, the between animal and loaded versus control limb variability make simulations of local surface-based adaptation responses challenging.

Combining PTH(1-34) and mechanical loading has increased benefit to tibia bone mechanics in ovariectomised mice.

Combined treatment with PTH(1-34) and mechanical loading confers increased structural benefits to bone than monotherapies. However, it remains unclear how this longitudinal adaptation affects the bone mechanics. This study quantified the individual and combined longitudinal effects of PTH(1-34) and mechanical loading on the bone stiffness and strength evaluated in vivo with validated micro-finite element (microFE) models. C57BL/6 mice were ovariectomised at 14-week-old and treated either with injections of PTH(1-34), compressive tibia loading or both interventions concurrently. Right tibiae were in vivo microCT-scanned every 2 weeks from 14 until 24-week-old. MicroCT images were rigidly registered to reference tibia and the cortical organ level (whole bone) and tissue level (midshaft) morphometric properties and bone mineral content were quantified. MicroCT images were converted into voxel-based homogeneous, linear elastic microFE models to estimate the bone stiffness and strength. This approach allowed us for the first time to quantify the longitudinal changes in mechanical properties induced by combined treatments in a model of accelerated bone resorption. Both changes of stiffness and strength were higher with co-treatment than with individual therapies, consistent with increased benefits with the tibia bone mineral content and cortical area, properties strongly associated with the tibia mechanics. The longitudinal data shows that the two bone anabolics, both individually and combined, had persistent benefit on estimated mechanical properties, and that benefits (increased stiffness and strength) remained after treatment was withdrawn.

Cortical thickness adaptation to combined mechanical loading and parathyroid hormone treatments is site specific and synergistic in the mouse tibia model.

In this study, we aimed to quantify the localised effects of mechanical loading (ML), low (20 µg/kg/day), moderate (40 µg/kg/day) or high (80 µg/kg/day) dosages of parathyroid hormone (PTH), and combined (PTHML) treatments on cortical bone adaptation in healthy 19-week old female C57BL/6 mice. To this end, we utilise a previously reported image analysis algorithm on µCT data of the mouse tibia published by Sugiyama et al. (2008) to measure changes in cortical area, marrow cavity area and local cortical thickness measures (ΔCt.Ar, ΔMa.Ar, ΔCt.Th respectively), evaluated at two cross-sections within the mouse tibia (proximal-middle (37 %) and middle (50 %)), and are compared to a superposed summation (P + M) of individual treatments to determine the effectiveness of combining treatments in vivo. ΔCt.Ar analysis revealed a non-linear, synergistic interactions between PTH and ML in the 37 % cross-section that saturates at higher PTH dosages, whereas the 50 % cross-section experiences an approximately linear, additive adaptation response. This coincided with an increase in ΔMa.Ar (indicating resorption of the endosteal surface), which was only counteracted by combined high dose PTH with ML in the middle cross-section. Regional analysis of ΔCt.Th changes reveal localised cortical thinning in response to low dose PTH treatment in the posteromedial region of the middle cross-section, signifying that PTH does not provide a homogeneous adaptation response around the cortical perimeter. We observe a synergistic response in the proximal-middle cross-section, with regions of compressive strain experiencing the greatest adaptation response to PTHML treatments, (peak ΔCt.Th of 189.32, 213.78 and 239.30 µm for low, moderate and high PTHML groups respectively). In contrast, PTHML treatments in the middle cross-section show a similar response to the superposed P + M group, with the exception of the combined high dose PTHML treatment which shows a synergistic interaction. These analyses suggest that, in mice, adding mechanical loading to PTH treatments leads to region specific bone responses; synergism of PTHML is only achieved in some regions experiencing high loading, while other regions respond additively to this combined treatment.

The role of bone metastases on the mechanical competence of human vertebrae.

Spine is the most common site for bone metastases. The evaluation of the mechanical competence and failure location in metastatic vertebrae is a biomechanical and clinical challenge. Little is known about the failure behaviour of vertebrae with metastatic lesions. The aim of this study was to use combined micro-Computed Tomography (microCT) and time-lapsed mechanical testing to reveal the failure location in metastatic vertebrae. Fifteen spine segments, each including a metastatic and a radiologically healthy vertebra, were tested in compression up to failure within a microCT. Volumetric strains were measured using Digital Volume Correlation. The images of undeformed and deformed specimens were overlapped to identify the failure location. Vertebrae with lytic metastases experienced the largest average compressive strains (median ± standard deviation: -8506 ± 4748microstrain), followed by the vertebrae with mixed metastases (-7035 ± 15605microstrain), the radiologically healthy vertebrae (-5743 ± 5697microstrain), and the vertebrae with blastic metastases (-3150 ± 4641microstrain). Strain peaks were localised within and nearby the lytic lesions or around the blastic tissue. Failure between the endplate and the metastasis was identified in vertebrae with lytic metastases, whereas failure localised around the metastasis in vertebrae with blastic lesions. This study showed for the first time the role of metastases on the vertebral internal deformations. While lytic lesions lead to failure of the metastatic vertebra, vertebrae with blastic metastases are more likely to induce failure in the adjacent vertebrae. Nevertheless, every metastatic lesion affects the vertebral deformation differently, making it essential to assess how the lesion affects the bone microstructure. These results suggest that the properties of the lesion (type, size, location within the vertebral body) should be considered when developing clinical tools to predict the risk of fracture in patients with metastatic lesions.

Bone metastases do not affect the measurement uncertainties of a global digital volume correlation algorithm.

Introduction: Measurement uncertainties of Digital Volume Correlation (DVC) are influenced by several factors, like input images quality, correlation algorithm, bone type, etc. However, it is still unknown if highly heterogeneous trabecular microstructures, typical of lytic and blastic metastases, affect the precision of DVC measurements. Methods: Fifteen metastatic and nine healthy vertebral bodies were scanned twice in zero-strain conditions with a micro-computed tomography (isotropic voxel size = 39 µm). The bone microstructural parameters (Bone Volume Fraction, Structure Thickness, Structure Separation, Structure Number) were calculated. Displacements and strains were evaluated through a global DVC approach (BoneDVC). The relationship between the standard deviation of the error (SDER) and the microstructural parameters was investigated in the entire vertebrae. To evaluate to what extent the measurement uncertainty is influenced by the microstructure, similar relationships were assessed within sub-regions of interest. Results: Higher variability in the SDER was found for metastatic vertebrae compared to the healthy ones (range 91-1030 µÎµ versus 222-599 µÎµ). A weak correlation was found between the SDER and the Structure Separation in metastatic vertebrae and in the sub-regions of interest, highlighting that the heterogenous trabecular microstructure only weakly affects the measurement uncertainties of BoneDVC. No correlation was found for the other microstructural parameters. The spatial distribution of the strain measurement uncertainties seemed to be associated with regions with reduced greyscale gradient variation in the microCT images. Discussion: Measurement uncertainties cannot be taken for granted but need to be assessed in each single application of the DVC to consider the minimum unavoidable measurement uncertainty when interpreting the results.

Deformable image registration based on single or multi-atlas methods for automatic muscle segmentation and the generation of augmented imaging datasets.

Muscle segmentation is a process relied upon to gather medical image-based muscle characterisation, useful in directly assessing muscle volume and geometry, that can be used as inputs to musculoskeletal modelling pipelines. Manual or semi-automatic techniques are typically employed to segment the muscles and quantify their properties, but they require significant manual labour and incur operator repeatability issues. In this study an automatic process is presented, aiming to segment all lower limb muscles from Magnetic Resonance (MR) imaging data simultaneously using three-dimensional (3D) deformable image registration (single inputs or multi-atlas). Twenty-three of the major lower limb skeletal muscles were segmented from five subjects, with an average Dice similarity coefficient of 0.72, and average absolute relative volume error (RVE) of 12.7% (average relative volume error of -2.2%) considering the optimal subject combinations. The multi-atlas approach showed slightly better accuracy (average DSC: 0.73; average RVE: 1.67%). Segmented MR imaging datasets of the lower limb are not widely available in the literature, limiting the potential of new, probabilistic methods such as deep learning to be used in the context of muscle segmentation. In this work, Non-linear deformable image registration is used to generate 69 manually checked, segmented, 3D, artificial datasets, allowing access for future studies to use these new methods, with a large amount of reliable reference data.

Effect of radiation-induced damage of trabecular bone tissue evaluated using indentation and digital volume correlation.

Exposure to X-ray radiation for an extended amount of time can cause damage to the bone tissue and therefore affect its mechanical properties. Specifically, high-resolution X-ray Computed Tomography (XCT), in both synchrotron and lab-based systems, has been employed extensively for evaluating bone micro-to-nano architecture. However, to date, it is still unclear how long exposures to X-ray radiation affect the mechanical properties of trabecular bone, particularly in relation to lab-XCT systems. Indentation has been widely used to identify local mechanical properties such as hardness and elastic modulus of bone and other biological tissues. The purpose of this study is therefore, to use indentation and XCT-based investigative tools such as digital volume correlation (DVC) to assess the microdamage induced by long exposure of trabecular bone tissue to X-ray radiation and how this affects its local mechanical properties. Trabecular bone specimens were indented before and after X-ray exposures of 33 and 66 h, where variation of elastic modulus was evaluated at every stage. The resulting elastic modulus was decreased, and micro-cracks appeared in the specimens after the first long X-ray exposure and crack formation increased after the second exposure. High strain concentration around the damaged tissue exceeding 1% was also observed from DVC analysis. The outcomes of this study show the importance of designing appropriate XCT-based experiments in lab systems to avoid degradation of the bone tissue mechanical properties due to radiation and these results will help to inform future studies that require long X-ray exposure for in situ experiments or generation of reliable subject-specific computational models.

Comparison of linear and nonlinear stepwise µFE displacement predictions to digital volume correlation measurements of trabecular bone biopsies.

Digital volume correlation (DVC) enables to evaluate the ability of µFE models in predicting experimental results on the mesoscale. In this study predicted displacement fields of three different linear and materially nonlinear µFE simulation methods were compared to DVC measured displacement fields at specific load steps in the elastic regime (StepEl) and after yield (StepUlt). Five human trabecular bone biopsies from a previous study were compressed in several displacement steps until failure. At every compression step, µCT images (resolution: 36 µm) were recorded. A global DVC algorithm was applied to compute the displacement fields at all loading steps. The unloaded 3D images were then used to generate homogeneous, isotropic, linear and materially nonlinear µFE models. Three different µFE simulation methods were used: linear (L), nonlinear (NL), and nonlinear stepwise (NLS). Regarding L and NL, the boundary conditions were derived from the interpolated displacement fields at StepEl and StepUlt, while for the NLS method nonlinear changes of the boundary conditions of the experiments were captured using the DVC displacement field of every available load step until StepEl and StepUlt. The predicted displacement fields of all µFE simulation methods were in good agreement with the DVC measured displacement fields (individual specimens: R2>0.83 at StepEl and R2>0.59 at StepUlt; pooled data: R2>0.97 at StepEl and R2>0.92 at StepUlt). At StepEl, all three simulation methods showed similar intercepts, slopes, and coefficients of determination while the nonlinear µFE models improved the prediction of the displacement fields slightly in all Cartesian directions at StepUlt (individual specimens: L: R2>0.59 and NL, NLS: R2>0.68; pooled data: L: R2>0.92 and NL, NLS: R2>0.94). Damaged/overstrained elements in L, NL, and NLS occurred at similar locations but the number of overstrained elements was overestimated when using the L simulation method. Considering the increased solving time of the nonlinear µFE models as well as the acceptable performance in displacement prediction of the linear µFE models, one can conclude that for similar use cases linear µFE models represent the best compromise between computational effort and accuracy of the displacement field predictions.

Digital volume correlation for the characterization of musculoskeletal tissues: Current challenges and future developments.

Biological tissues are complex hierarchical materials, difficult to characterise due to the challenges associated to the separation of scale and heterogeneity of the mechanical properties at different dimensional levels. The Digital Volume Correlation approach is the only image-based experimental approach that can accurately measure internal strain field within biological tissues under complex loading scenarios. In this minireview examples of DVC applications to study the deformation of musculoskeletal tissues at different dimensional scales are reported, highlighting the potential and challenges of this relatively new technique. The manuscript aims at reporting the wide breath of DVC applications in the past 2 decades and discuss future perspective for this unique technique, including fast analysis, applications on soft tissues, high precision approaches, and clinical applications.

Reproducibility of Densitometric and Biomechanical Assessment of the Mouse Tibia From In Vivo Micro-CT Images.

Interventions for bone diseases (e.g. osteoporosis) require testing in animal models before clinical translation and the mouse tibia is among the most common tested anatomical sites. In vivo micro-Computed Tomography (microCT) based measurements of the geometrical and densitometric properties are non-invasive and therefore constitute an important tool in preclinical studies. Moreover, validated micro-Finite Element (microFE) models can be used for predicting the bone mechanical properties non-invasively. However, considering that the image processing pipeline requires operator-dependant steps, the reproducibility of these measurements has to be assessed. The aim of this study was to evaluate the intra- and inter-operator reproducibility of several bone parameters measured from microCT images. Ten in vivo microCT images of the right tibia of five mice (at 18 and 22 weeks of age) were processed. One experienced operator (intra-operator analysis) and three different operators (inter-operator) aligned each image to a reference through a rigid registration and selected a volume of interest below the growth plate. From each image the following parameters were measured: total bone mineral content (BMC) and density (BMD), BMC in 40 subregions (ten longitudinal sections, four quadrants), microFE-based stiffness and failure load. Intra-operator reproducibility was acceptable for all parameters (precision error, PE < 3.71%), with lowest reproducibility for stiffness (3.06% at week 18, 3.71% at week 22). The inter-operator reproducibility was slightly lower (PE < 4.25%), although still acceptable for assessing the properties of most interventions. The lowest reproducibility was found for BMC in the lateral sector at the midshaft (PE = 4.25%). Densitometric parameters were more reproducible than most standard morphometric parameters calculated in the proximal trabecular bone. In conclusion, microCT and microFE models provide reproducible measurements for non-invasive assessment of the mouse tibia properties.

Geroprotectors and Skeletal Health: Beyond the Headlines.

Osteoporosis and osteoarthritis are the most common age-related diseases of the musculoskeletal system. They are responsible for high level of healthcare use and are often associated with comorbidities. Mechanisms of ageing such as senescence, inflammation and autophagy are common drivers for both diseases and molecules targeting those mechanisms (geroprotectors) have potential to prevent both diseases and their co-morbidities. However, studies to test the efficacy of geroprotectors on bone and joints are scant. The limited studies available show promising results to prevent and reverse Osteoporosis-like disease. In contrast, the effects on the development of Osteoarthritis-like disease in ageing mice has been disappointing thus far. Here we review the literature and report novel data on the effect of geroprotectors for Osteoporosis and Osteoarthritis, we challenge the notion that extension of lifespan correlates with extension of healthspan in all tissues and we highlight the need for more thorough studies to test the effects of geroprotectors on skeletal health in ageing organisms.

MicroFE models of porcine vertebrae with induced bone focal lesions: Validation of predicted displacements with digital volume correlation.

The evaluation of the local mechanical behavior as a result of metastatic lesions is fundamental for the characterization of the mechanical competence of metastatic vertebrae. Micro finite element (microFE) models have the potential of addressing this challenge through laboratory studies but their predictions of local deformation due to the complexity of the bone structure compromized by the lesion must be validated against experiments. In this study, the displacements predicted by homogeneous, linear and isotropic microFE models of vertebrae were validated against experimental Digital Volume Correlation (DVC) measurements. Porcine spine segments, with and without mechanically induced focal lesions, were tested in compression within a micro computed tomography (microCT) scanner. The displacement within the bone were measured with an optimized global DVC approach (BoneDVC). MicroFE models of the intact and lesioned vertebrae, including or excluding the growth plates, were developed from the microCT images. The microFE and DVC boundary conditions were matched. The displacements measured by the DVC and predicted by the microFE along each Cartesian direction were compared. The results showed an excellent agreement between the measured and predicted displacements, both for intact and metastatic vertebrae, in the middle of the vertebra, in those cases where the structure was not loaded beyond yield (0.69 < R2 < 1.00). Models with growth plates showed the worst correlations (0.02 < R2 < 0.99), while a clear improvement was observed if the growth plates were excluded (0.56 < R2 < 1.00). In conclusion, these simplified models can predict complex displacement fields in the elastic regime with high reliability, more complex non-linear models should be implemented to predict regions with high deformation, when the bone is loaded beyond yield.

Subchondral Bone Microarchitectural and Mineral Properties and Expression of Key Degradative Proteinases by Chondrocytes in Human Hip Osteoarthritis.

BACKGROUND: The purpose of this study was to investigate the relationship between the expression of key degradative enzymes by chondrocytes and the microarchitectural and mineral properties of subchondral bone across different stages of cartilage degradation in human hip osteoarthritis (OA). METHODS: Osteochondral samples at different stages of cartilage degradation were collected from 16 femoral heads with OA. Osteochondral samples with normal cartilage were collected from seven femoral heads with osteoporosis. Microcomputed tomography was used for the investigation of subchondral bone microarchitecture and mineral densities. Immunohistochemistry was used to study the expression and distribution of MMP13 and ADAMTS4 in cartilage. RESULTS: The microarchitecture and mineral properties of the subchondral plate and trabecular bone in OA varied with the severity of the degradation of the overlying cartilage. Chondrocytes expressing MMP13 and ADAMTS4 are mainly located in the upper zone(s) of cartilage regardless of the histopathological grades. The zonal expression of these enzymes in OA (i.e., the percentage of positive cells in the superficial, middle, and deep zones), rather than their overall expression (the percentage of positive cells in the full thickness of the cartilage), exhibited significant variation in relation to the severity of cartilage degradation. The associations between the subchondral bone properties and zonal and overall expression of these enzymes in the cartilage were generally weak or nonsignificant. CONCLUSIONS: Phenotypic changes in chondrocytes and remodelling of subchondral bone proceed at different rates throughout the process of cartilage degradation. Biological influences are more important for cartilage degradation at early stages, while biomechanical damage to the compromised tissue may outrun the phenotypic change of chondrocytes and is critical in the advanced stages.

Positive interactions of mechanical loading and PTH treatments on spatio-temporal bone remodelling.

Osteoporosis is one of the most common skeletal diseases, but current therapies are limited to generalized antiresorptive or anabolic interventions, which do not target regions that would benefit from improvements to skeletal health. To improve the evaluation of treatment plans, we used a spatio-temporal multiscale approach that combines longitudinal in vivo micro-computed tomography (micro-CT) and in silico subject-specific finite element modeling to quantitatively map bone adaptation changes due to disease and treatment at high resolution. Our findings show time and region-dependent modifications in bone remodelling following one and two sets of mechanical loading and/or pharmacological interventions. The multiscale results highlighted that the distal section was unaffected by mechanical loading alone but the proximal tibia had the greatest gain from positive interactions of combined therapies. Mechanical loading abated the catabolic effect of PTH, but the main benefit of combined treatments occurred from the additive interactions of the two therapies in periosteal apposition. These results provide detailed insight into the efficacy of combined treatments, facilitating the optimisation of dosage and treatment duration in preclinical mouse studies, and the development of novel interventions for skeletal diseases. STATEMENT OF SIGNIFICANCE: Combined mechanical loading and pharmacotherapy have the potential to slow osteoporosis-induced bone loss but current therapies do not target the regions in need of strengthening. We show for the first time spatial region-dependant interactions between PTH and mechanical loading treatment in OVX mouse tibiae, highlighting local regions in the tibia that benefitted from separate and combined treatments. Combined experimental-computational analysis also detailed the lasting period of each treatment per location in the tibia, the extent of positive (or negative) interactions of the combined therapies, and the impact of each treatment on the regulation of bone adaptation spatio-temporally. This approach can be used to create hypothesis about the interactions of different treatments to optimise the design of biomaterials and medical interventions.

The Role of the Loading Condition in Predictions of Bone Adaptation in a Mouse Tibial Loading Model.

The in vivo mouse tibial loading model is used to evaluate the effectiveness of mechanical loading treatment against skeletal diseases. Although studies have correlated bone adaptation with the induced mechanical stimulus, predictions of bone remodeling remained poor, and the interaction between external and physiological loading in engendering bone changes have not been determined. The aim of this study was to determine the effect of passive mechanical loading on the strain distribution in the mouse tibia and its predictions of bone adaptation. Longitudinal micro-computed tomography (micro-CT) imaging was performed over 2 weeks of cyclic loading from weeks 18 to 22 of age, to quantify the shape change, remodeling, and changes in densitometric properties. Micro-CT based finite element analysis coupled with an optimization algorithm for bone remodeling was used to predict bone adaptation under physiological loads, nominal 12N axial load and combined nominal 12N axial load superimposed to the physiological load. The results showed that despite large differences in the strain energy density magnitudes and distributions across the tibial length, the overall accuracy of the model and the spatial match were similar for all evaluated loading conditions. Predictions of densitometric properties were most similar to the experimental data for combined loading, followed closely by physiological loading conditions, despite no significant difference between these two predicted groups. However, all predicted densitometric properties were significantly different for the 12N and the combined loading conditions. The results suggest that computational modeling of bone's adaptive response to passive mechanical loading should include the contribution of daily physiological load.

A novel approach to evaluate the effects of artificial bone focal lesion on the three-dimensional strain distributions within the vertebral body.

The spine is the first site for incidence of bone metastasis. Thus, the vertebrae have a high potential risk of being weakened by metastatic tissues. The evaluation of strength of the bone affected by the presence of metastases is fundamental to assess the fracture risk. This work proposes a robust method to evaluate the variations of strain distributions due to artificial lesions within the vertebral body, based on in situ mechanical testing and digital volume correlation. Five porcine vertebrae were tested in compression up to 6500N inside a micro computed tomography scanner. For each specimen, images were acquired before and after the application of the load, before and after the introduction of the artificial lesions. Principal strains were computed within the bone by means of digital volume correlation (DVC). All intact specimens showed a consistent strain distribution, with peak minimum principal strain in the range -1.8% to -0.7% in the middle of the vertebra, demonstrating the robustness of the method. Similar distributions of strains were found for the intact vertebrae in the different regions. The artificial lesion generally doubled the strain in the middle portion of the specimen, probably due to stress concentrations close to the defect. In conclusion, a robust method to evaluate the redistribution of the strain due to artificial lesions within the vertebral body was developed and will be used in the future to improve current clinical assessment of fracture risk in metastatic spines.