Validation of a voxel-based FE method for prediction of the uniaxial apparent modulus of human trabecular bone using macroscopic mechanical tests and nanoindentation

Assessment of the mechanical properties of trabecular bone is of major biological and clinical importance for the investigation of bone diseases, fractures and their treatments. Finite element (FE) methods are getting increasingly popular for quantifying the elastic and failure properties of trabecular bone. In particular, voxel-based FE methods have been previously used to calculate the effective elastic properties of trabecular microstructures. However, in most studies, bone tissue moduli were assumed or back-calculated to match the apparent elastic moduli from experiments, which often lead to surprisingly low values when compared to nanoindentation results. In this study, voxel-based FE analysis of trabecular bone is combined with physical measures of volume fraction, micro-CT (microCT) reconstructions, uniaxial mechanical tests and specimen-specific nanoindentation tests for proper validation of the method. Cylindrical specimens of cancellous bone were extracted from human femurs and their volume fraction determined with Archimede's method. Uniaxial apparent modulus of the specimens was measured with an improved tension-compression testing protocol that minimizes boundary artefacts. Their microCT reconstructions were segmented to match the measured bone volume fraction and used to create full-size voxel models with 30-45 microm element size. For each specimen, linear isotropic elastic material properties were defined based on specific nanoindentation measurements of its embedded bone tissue. Linear FE analyses were finally performed to simulate the uniaxial mechanical tests. Additional parametric analyses were performed to evaluate the potential errors on the predicted apparent modulus arising from variations in segmentation threshold, tissue modulus, and the use of 125-mm(3) cubic sub-regions. The results demonstrate an excellent correspondence between experimental measures and FE predictions of uniaxial apparent modulus. In conclusion, the adopted voxel-based FE approach is found to be a robust method to predict the linear elastic properties of human cancellous bone, provided segmentation of the microCT reconstructions is carefully calibrated, tissue modulus is known a priori and the entire region of interest is included in the analysis.     

Concept and development of an orthotropic FE model of the proximal femur

PURPOSE: In contrast to many isotropic finite-element (FE) models of the femur in literature, it was the object of our study to develop an orthotropic FE "model femur" to realistically simulate three-dimensional bone remodelling. METHODS: The three-dimensional geometry of the proximal femur was reconstructed by CT scans of a pair of cadaveric femurs at equal distances of 2mm. These three-dimensional CT models were implemented into an FE simulation tool. Well-known "density-determined" bony material properties (Young's modulus; Poisson's ratio; ultimate strength in pressure, tension and torsion; shear modulus) were assigned to each FE of the same "CT-density-characterized" volumetric group.In order to fix the principal directions of stiffness in FE areas with the same "density characterization", the cadaveric femurs were cut in 2mm slices in frontal (left femur) and sagittal plane (right femur). Each femoral slice was scanned into a computer-based image processing system. On these images, the principal directions of stiffness of cancellous and cortical bone were determined manually using the orientation of the trabecular structures and the Haversian system. Finally, these geometric data were matched with the "CT-density characterized" three-dimensional femur model. In addition, the time and density-dependent adaptive behaviour of bone remodelling was taken into account by implementation of Carter's criterion. RESULTS: In the constructed "model femur", each FE is characterized by the principal directions of the stiffness and the "CT-density-determined" material properties of cortical and cancellous bone. Thus, on the basis of anatomic data a three-dimensional FE simulation reference model of the proximal femur was realized considering orthotropic conditions of bone behaviour. CONCLUSIONS: With the orthotropic "model femur", the fundamental basis has been formed to realize realistic simulations of the dynamical processes of bone remodelling under different loading conditions or operative procedures (osteotomies, total hip replacements, etc).     

The dependence of the elastic properties of osteoporotic cancellous bone on volume fraction and fabric

Osteoporosis is a progressive systemic skeletal condition characterized by low bone mass and microarchitectural deterioration, with a consequent increase in susceptibility to fracture. Hence, osteoporosis would be best diagnosed by in vivo measurements of bone strength. As this is not clinically feasible, our goal is to estimate bone strength through the assessment of elastic properties, which are highly correlated to strength. Previously established relations between morphological parameters (volume fraction and fabric) and elastic constants could be applied to estimate cancellous bone stiffness in vivo. However, these relations were determined for normal cancellous bone. Cancellous bone from osteoporotic patients may require different relations. In this study we set out to answer two questions. First, can the elastic properties of osteoporotic cancellous bone be estimated from morphological parameters? Second, do the relations between morphological parameters and elastic constants, determined for normal bone, apply to osteoporotic bone as well? To answer these questions we used cancellous bone cubes from femoral heads of patients with (n=26) and without (n=32) hip fractures. The elastic properties of the cubes were determined using micro-finite element analysis, assuming equal tissue moduli for all specimens. The morphological parameters were determined using microcomputed tomography. Our results showed that, for equal tissue properties, the elastic properties of cancellous bone from fracture patients could indeed be estimated from morphological parameters. The morphology-based relations used to estimate the elastic properties of cancellous bone are not different for women with or without fractures.     

Estimation of bone matrix apparent stiffness variation caused by osteocyte lacunar size and density

The role of osteocyte lacunar size and density on the apparent stiffness of bone matrix was predicted using a mechanical model from the literature. Lacunar size and lacunar density for different bones from different gender and age groups were used to predict the range of matrix apparent stiffness values for human cortical and cancellous tissue. The results suggest that bone matrix apparent stiffness depends on tissue type (cortical versus cancellous), age, and gender, the magnitudes of the effects being significant but small in all cases. Males had a higher predicted matrix apparent stiffness than females for vertebral cancellous bone (p< I0(-7)) and the difference increased with age (p =0.0007). In contrast, matrix apparent stiffness was not different between males and females forfemoral cortical bone and increased with age in both males (p < 0.0001) and females (p < 0.0364). Osteocyte lacunar density and size may cause significant gender and age-related variations in bone matrix apparent stiffness. The magnitude of variations in matrix apparent stiffness was small within the physiological range of lacunar size and density for healthy bone, whereas the variations can be profound in certain pathological cases. It was proposed that the mechanical effects of osteocyte density be uncoupled from their biological effects by controlling lacunar size in normal bone.     

Trabecular shear stress in human vertebral cancellous bone: intra- and inter-individual variations

Correlation of the mean and standard deviation of trabecular stresses has been proposed as a mechanism by which a strong relationship between the apparent strength and stiffness of cancellous bone can be achieved. The current study examined whether the relationship between the mean and standard deviation of trabecular von Mises stresses can be generalized for any group of cancellous bone. Cylindrical human vertebral cancellous bone specimens were cut in the infero-superior direction from T12 of 23 individuals (inter-individual group). Thirty nine additional specimens were prepared similarly from the T4-T12 and L2-L5 vertebrae of a 63 year old male (intra-individual group). The specimens were scanned by micro-computed tomography (microCT) and trabecular von Mises stresses were calculated using finite element modeling. The expected value, standard deviation and coefficient of variation of the von Mises stress were calculated form a three-parameter Weibull function fitted to von Mises stress data from each specimen. It was found that the average and standard deviation of trabecular von Mises shear stress were: (i) correlated with each other, supporting the idea that high correlation between the apparent strength and stiffness of cancellous bone can be achieved through controlling the trabecular level shear stress variations, (ii) dependent on anatomical site and sample group, suggesting that the variation of stresses are correlated to the mean stress to different degrees between vertebrae and individuals, and (iii) dependent on bone volume fraction, consistent with the idea that shear stress is less well controlled in bones with low BV/TV. The conversion of infero-superior loading into trabecular von Mises stresses was maximum for the tissue at the junction of the thoracic and lumbar spine (T12-L1) consistent with this junction being a common site of vertebral fracture.     

Mapping anisotropy of the proximal femur for enhanced image based finite element analysis

Finite element (FE) models of bone derived from quantitative computed tomography (QCT) rely on realistic material properties to accurately predict bone strength. QCT cannot resolve bone microarchitecture, therefore QCT-based FE models lack the anisotropy apparent within the underlying bone tissue. This study proposes a method for mapping femoral anisotropy using high-resolution peripheral quantitative computed tomography (HR-pQCT) scans of human cadaver specimens. Femur HR-pQCT images were sub-divided into numerous overlapping cubic sub-volumes and the local anisotropy was quantified using a ‘direct-mechanics’ method. The resulting directionality reflected all the major stress lines visible within the trabecular lattice, and provided a realistic estimate of the alignment of Harvesian systems within the cortical compartment. QCT-based FE models of the proximal femur were constructed with isotropic and anisotropic material properties, with directionality interpolated from the map of anisotropy. Models were loaded in a sideways fall configuration and the resulting whole bone stiffness was compared to experimental stiffness and ultimate strength. Anisotropic models were consistently less stiff, but no statistically significant differences in correlation were observed between material models against experimental data. The mean difference in whole bone stiffness between model types was approximately 26%, suggesting that anisotropy can still effect considerable change in the mechanics of proximal femur models. The under prediction of whole bone stiffness in anisotropic models suggests that the orthotropic elastic constants require further investigation. The ability to map mechanical anisotropy from high-resolution images and interpolate information into clinical-resolution models will allow testing of new anisotropic material mapping strategies.     

Biomechanical properties of bones from rats treated with sevelamer

Sevelamer hydrochloride is used for ten years in patients on dialysis as a phosphate binder. We have previously shown that oral application of sevelamer prevents the bone loss and increases the bone volume in ovariectomized rats. In this study we further analysed the biomechanical properties of bones from rats treated with sevelamer utilizing a threepoint bending test to determine the mechanical properties of the cortical bone of the mid-shaft femur, while the indentation test was used to determine the mechanical properties of cancellous bone in the marrow cavity of the distal femoral metaphysis. Parameters analyzed included: maximum load (F(u)), stiffness (S), energy absorbed (W), toughness (T) and ultimate strength (sigma). The intrinsic properties, stress, elastic modulus and toughness were determined from measured maximum load, strains, stiffness, energy absorbed, outer and inner diameters, and calculated bone cross-sectional moment of inertia. Sevelamer was given to rats for 25 weeks with a content of 3% of sevelamer in a standard diet, starting immediately following ovariectomy (OVX). Animals were divided to the following groups: (1) Sham; (2) Sham + sevelamer 3%; (3) OVX; (4) OVX + sevelamer 3%. Our results showed that sevelamer particularly influenced the rat trabecular bone by increasing the maximum load for 26.2%, energy absorbed for 24.2% and the ultimate strength for 26.2% in sham animals treated with sevelamer 3%, as compared to sham rats. Sevelamer 3% in OVX rats also increased the maximum load for 71.4%, stiffness for 70.7%, energy absorbed for 55.9% and the ultimate strength for 71.3% as compared to OVX controls. In the three bending test sevelamer had a very little effect on preventing loss of bone strenght in the cortical bone. These results collectively suggest that sevelamer improves bone biomechanical properties, mainly affecting trabecular bone quality in both normal and ovariectomized rats.     

The influence of material property and morphological parameters on specimen-specific finite element models of porcine vertebral bodies

The use of finite element (FE) methods in spinal research is increasing, but there is only limited information available on the influence of different input parameters on the model predictions. The aim of this study was to investigate the role of these parameters in FE models of the vertebral body. Experimental tests were undertaken on porcine lumbar vertebral bodies and scans of the specimens were used to create specimen-specific FE models. Three models were created for each specimen with combinations of generic and specimen-specific parameters. Stiffness and strength predictions were also made directly from the specimen trabecular bone volume fraction (BVF) and cross-sectional area (CSA). The agreement between the experimental results and the FE models with generic morphology was poorer (concordance coefficients = 0.058, 0.125 for stiffness, strength) than those made from the BVF and CSA (concordance coefficients = 0.638, 0.609). The greatest levels of agreement were found with the morphologically specific models including element-specific material properties (concordance coefficients = 0.881, 0.752). This indicates that highly specific models, both in terms of morphology and bone quality, are necessary if the FE tool is to be used effectively for spinal research and clinical practice.     

An approximate model for cancellous bone screw fixation

This paper presents a finite element (FE) model to identify parameters that affect the performance of an improved cancellous bone screw fixation technique, and hence potentially improve fracture treatment. In cancellous bone of low apparent density, it can be difficult to achieve adequate screw fixation and hence provide stable fracture fixation that enables bone healing. Data from predictive FE models indicate that cements can have a significant potential to improve screw holding power in cancellous bone. These FE models are used to demonstrate the key parameters that determine pull-out strength in a variety of screw, bone and cement set-ups, and to compare the effectiveness of different configurations. The paper concludes that significant advantages, up to an order of magnitude, in screw pull-out strength in cancellous bone might be gained by the appropriate use of a currently approved calcium phosphate cement.     

Dependence of mechanical compressive strength on local variations in microarchitecture in cancellous bone of proximal human femur

Human cancellous bone is a heterogeneous material. Despite this, most of the published studies report correlations between mechanical properties and morphometric parameters averaged on the whole specimen. This work investigated whether local variations in morphometric parameters were linked to the localized failure regions of cancellous bone. Additionally, it was examined whether local values of morphometric parameters can predict the ultimate stress better than the average bone volume fraction (BV/TV). Cylindrical cancellous bone specimens extracted along the primary compressive group of human femoral heads were studied. These were microCT-imaged to assess the morphometric parameters, compressed to determine the ultimate stress, and rescanned by microCT to visualize the failure region. Failure involved slightly less than half of the free height of the specimens. Significant differences were found in the morphometric parameters calculated in the failure and in the non-failure regions. The cross-sections containing minimum BV/TV values were those most often located inside the failure region (83%, p<0.001). Regression analysis confirmed that variations in BV/TV best describe variations in ultimate stress (R2=0.84) out of the averaged morphometric parameters. The prediction of ultimate stress increased when minimum or maximum values of the morphometric parameters were taken, with the highest prediction found by considering the minimum BV/TV (R2=0.95). In conclusion, due to the heterogeneity of cancellous bone, there may exist regions characterized by a different microarchitecture, where the bone is weaker and consequently is more likely to fail. These regions mostly contain minimum values in BV/TV, which were found to predict ultimate stress better than average BV/TV.     

The apparent critical isotherm for cryoinsult-induced osteonecrotic lesions in emu femoral heads

Cryoinsult-induced osteonecrosis (ON) in the emu femoral head provides a unique opportunity to systematically explore the pathogenesis of ON in an animal model that progresses to human-like femoral head collapse. Among the various characteristics of cryoinsult, the maximally cold temperature attained is one plausible determinant of tissue necrosis. To identify the critical isotherm required to induce development of ON in the cancellous bone of the emu femoral head, a thermal finite element (FE) model of intraoperative cryoinsults was developed. Thermal material property values of emu cancellous bone were estimated from FE simulations of cryoinsult to emu cadaver femora, by varying model properties until the FE-generated temperatures matched corresponding thermocouple measurements. The resulting FE model, with emu bone-specific thermal properties augmented to include blood flow effects, was then used to study intraoperatively performed in vivo cryoinsults. Comparisons of minimum temperatures attained at FE nodes corresponding to the three-dimensional histologically apparent boundary of the region of ON were made for six experimental cryoinsults. Series-wide, a critical isotherm of 3.5 degrees C best corresponded to the boundary of the osteonecrotic lesions.     

HR-pQCT-based homogenised finite element models provide quantitative predictions of experimental vertebral body stiffness and strength with the same accuracy as μFE models

This study validated two different high-resolution peripheral quantitative computer tomography (HR-pQCT)-based finite element (FE) approaches, enhanced homogenised continuum-level (hFE) and micro-finite element (μFE) models, by comparing them with compression test results of vertebral body sections. Thirty-five vertebral body sections were prepared by removing endplates and posterior elements, scanned with HR-pQCT and tested in compression up to failure. Linear hFE and μFE models were created from segmented and grey-level CT images, and apparent model stiffness values were compared with experimental stiffness as well as strength results. Experimental and numerical apparent elastic properties based on grey-level/segmented CT images (N=35) correlated well for μFE (r2=0.748/0.842) and hFE models (r2=0.741/0.864). Vertebral section stiffness values from the linear μFE/hFE models estimated experimental ultimate apparent strength very well (r2=0.920/0.927). Calibrated hFE models were able to predict quantitatively apparent stiffness with the same accuracy as μFE models. However, hFE models needed no back-calculation of a tissue modulus or any kind of fitting and were computationally much cheaper.     

Compressive axial mechanical properties of rat bone as functions of bone volume fraction,apparent density and micro-ct based mineral density

Mechanical testing has been regarded as the gold standard to investigate the effects of pathologies on the structure–function properties of the skeleton. With recent advances in computing power of personal computers, virtual alternatives to mechanical testing are gaining acceptance and use. We have previously introduced such a technique called structural rigidity analysis to assess mechanical strength of skeletal tissue with defects. The application of this technique is predicated upon the use of relationships defining the strength of bone as a function of its density for a given loading mode. We are to apply this technique in rat models to assess their compressive skeletal response subjected to a host of biological and pharmaceutical stimulations. Therefore, the aim of this study is to derive a relationship expressing axial compressive mechanical properties of rat cortical and cancellous bone as a function of equivalent bone mineral density, bone volume fraction or apparent density over a range of normal and pathologic bones.We used bones from normal, ovariectomized and partially nephrectomized animals. All specimens underwent micro-computed tomographic imaging to assess bone morphometric and densitometric indices and uniaxial compression to failure.We obtained univariate relationships describing 71–78% of the mechanical properties of rat cortical and cancellous bone based on equivalent mineral density, bone volume fraction or apparent density over a wide range of density and common skeletal pathologies. The relationships reported in this study can be used in the structural rigidity analysis introduced by the authors to provide a non-invasive method to assess the compressive strength of bones affected by pathology and/or treatment options.     

Trabecular bone's mechanical properties are affected by its non-uniform mineral distribution

The bone remodeling process takes place at the surface of trabeculae and results in a non-uniform mineral distribution. This will affect the mechanical properties of cancellous bone, because the properties of bone tissue depend on its mineral content. We investigated how large this effect is by simulating several non-uniform mineral distributions in 3D finite element models of human trabecular bone and calculating the apparent stiffness of these models. In the ‘linear model’ we assumed a linear relation between mineral content and Young's modulus of the tissue. In the ‘exponential model’ we included an empirical exponential relation in the model. When the linear model was used the mineral distribution slightly changed the apparent stiffness, the difference varied between an 8% decrease and a 4% increase compared to the uniform model with the same BMD. The exponential model resulted in up to 20% increased apparent stiffness in the main load-bearing direction. A thin less mineralized surface layer (28 μm) and highly mineralized interstitial bone (mimicking mineralization resulting from anti-resorptive treatment) resulted in the highest stiffness. This could explain large reductions in fracture risk resulting from small increases in BMD. The non-uniform mineral distribution could also explain why bone tissue stiffness determined using nano-indentation is usually higher than finite element (FE)-determined stiffness. We conclude that the non-uniform mineral distribution in trabeculae does affect the mechanical properties of cancellous bone and that the tissue stiffness determined using FE-modeling could be improved by including detailed information about mineral distribution in trabeculae in the models.     

Micro-CT and micro-FE analysis of pedicle screw fixation under different loading conditions

Anchorage of pedicle screw instrumentation in the elderly spine with poor bone quality remains challenging. In this study, micro finite element (µFE) models were used to assess the specific influence of screw design and the relative contribution of local bone density to fixation mechanics. These were created from micro computer tomography (µCT) scans of vertebras implanted with two types of pedicle screws, including a full region-or-interest of 10 mm radius around each screw, as well as submodels for the pedicle and inner trabecular bone of the vertebral body. The local bone volume fraction (BV/TV) calculated from the µCT scans around different regions of the screw (pedicle, inner trabecular region of the vertebral body) were then related to the predicted stiffness in simulated pull-out tests as well as to the experimental pull-out and torsional fixation properties mechanically measured on the corresponding specimens. Results show that predicted stiffness correlated excellently with experimental pull-out strength (R² > 0.92, p < .043), better than regional BV/TV alone (R2 = 0.79, p = .003). They also show that correlations between fixation properties and BV/TV were increased when accounting only for the pedicle zone (R² = 0.66–0.94, p ≤ .032), but with weaker correlations for torsional loads (R² < 0.10). Our analyses highlight the role of local density in the pedicle zone on the fixation stiffness and strength of pedicle screws when pull-out loads are involved, but that local apparent bone density alone may not be sufficient to explain resistance in torsion.     

Mechanical testing of cancellous bone from the femoral head: experimental errors due to off-axis measurements

The aim of this study was to verify whether a misalignment between the testing direction and the trabecular main direction has a significant effect on the compressive behaviour of cancellous bone. Ten cylindrical specimens were extracted from femoral heads with a misalignment to the trabecular main direction of approximately 20 degrees. Each specimen was paired with a specimen extracted aligned with the main direction of the trabeculae on the basis of the closest bone volume fraction, obtaining two groups, one 'aligned' and one 'misaligned'. The average off-axis angle was 6.1 degrees and 21.6 degrees for the 'aligned' and 'misaligned' group, respectively. The specimens underwent micro-tomographic analysis, compressive testing, micro-indentation testing and ashing. No significant differences were found in histomorphometric parameters, hardness and ash density between the two groups, whereas significant differences were found in Young's modulus and ultimate stress: both parameters, measured for the 'misaligned' group, were about 40% lower than those measured for the 'aligned' group. These results demonstrate a great effect of the angle between the testing direction and the main direction of the trabecular structure on the compressive behaviour of cancellous bone. This angle should be reduced as much as possible (in the present work the average value was 6.6+/-3.3 degrees), in any case measured, and always reported together with the mechanical parameters of cancellous bone.     

Critical evaluation of known bone material properties to realize anisotropic FE-simulation of the proximal femur

PURPOSE: In a meta-analysis of the literature we evaluated the present knowledge of the material properties of cortical and cancellous bone to answer the question whether the available data are sufficient to realize anisotropic finite element (FE)-models of the proximal femur. MATERIAL AND METHOD: All studies that met the following criteria were analyzed: Young's modulus, tensile, compressive and torsional strengths, Poisson's ratio, the shear modulus and the viscoelastic properties had to be determined experimentally. The experiments had to be carried out in a moist environment and at room temperature with freshly removed and untreated human cadaverous femurs. All material properties had to be determined in defined load directions (axial, transverse) and should have been correlated to apparent density (g/cm(3)), reflecting the individually variable and age-dependent changes of bone material properties. RESULTS: Differences in Young's modulus of cortical [cancellous] bone at a rate of between 33% (58%) (at low apparent density) and 62% (80%) (at high apparent density), are higher in the axial than in the transverse load direction. Similar results have been seen for the compressive strength of femoral bone. For the tensile and torsional strengths, Poisson's ratio and the shear modulus, only ultimate values have been found without a correlation to apparent density. For the viscoelastic behaviour of bone only data of cortical bone and in axial load direction have been described up to now. CONCLUSIONS: Anisotropic FE-models of the femur could be realized for most part with the summarized material properties of bone if characterized by apparent density and load directions. Because several mechanical properties have not been correlated to these main criteria, further experimental investigations will be necessary in future.     

A three-dimensional finite element analysis of the upper tibia

A three-dimensional finite element model of the proximal tibia has been developed to provide a base line for further modeling of prosthetic resurfaced tibiae. The geometry for the model was developed by digitizing coronal and transverse sections made with the milling machine, from one fresh tibia of average size. The load is equally distributed between the medial and lateral compartments over contact areas that were reported in the literature. An indentation test has been used to measure the stiffness and the ultimate strength of cancellous bone in four cadaver tibiae. These values provided the statistical basis for characterising the inhomogeneous distribution of the cancellous bone properties in the proximal tibia. All materials in the model were assumed to be linearly elastic and isotropic. Mechanical properties for the cortical bone and cartilage have been taken from the literature. Results have been compared with strain gage tests and with a two-dimensional axisymmetric finite element model both from the literature. Qualitative comparison between trabecular alignment, and the direction of the principal compressive stresses in the cancellous bone, showed a good relationship. Maximum stresses in the cancellous bone and cortical bone, under a load which occurs near stance phase during normal gait, show safety factors of approximately eight and twelve, respectively. The load sharing between the cancellous bone and the cortical bone has been plotted for the first 40 mm distally from the tibial eminence.     

Anisotropy and inhomogeneity of the trabecular structure can describe the mechanical strength of osteoarthritic cancellous bone

Osteoarthritic cancellous bone was studied to investigate the development of this pathology, and the functional changes it induces in the bone. In order to predict how the morphological alterations of the tissue induced by the pathology can change the mechanical properties of the structure, two different strategies have been used in the literature: (1) emphasising the influence of structural anisotropy; (2) stressing the highly inhomogeneous characteristics of cancellous bone. The aim of the present study was to verify the theory that mechanical strength of osteoarthritic cancellous bone depends both on tissue anisotropy and inhomogeneity.Twenty-five specimens were extracted from osteoarthritic femoral heads, along selected directions, and analysed by means of a microtomograph. The same specimens were mechanically tested in compression to determine the mechanical strength. The most representative structural parameters, confirmed by a stepwise analysis, were used to define four models to describe the measured mechanical strength. The models were applied neglecting (global analysis) or considering (local analysis) tissue inhomogeneities to verify whether the correlation with ultimate stress could be improved.The coefficient of determination increased from 0.53, considering only bone volume fraction, up to 0.88, combining it with off-axis angle and normalised eigenvalue. A further improvement was found performing a local analysis (R²=0.90), which corresponded to a decrease of 17% in the residual error.The proposed approach of considering both tissue anisotropy and inhomogeneity improved the accuracy in predicting the mechanical behaviour of cancellous bone tissue and should be suitable for more general loading conditions.