Clinical cone beam computed tomography compared to high-resolution peripheral computed tomography in the assessment of distal radius bone.

UNLABELLED: Clinical cone beam computed tomography (CBCT) was compared to high-resolution peripheral quantitative computed tomography (HR-pQCT) for the assessment of ex vivo radii. Strong correlations were found for geometry, volumetric density, and trabecular structure. Using CBCT, bone architecture assessment was feasible but compared to HR-pQCT, trabecular parameters were overestimated whereas cortical ones were underestimated. INTRODUCTION: HR-pQCT is the most widely used technique to assess bone microarchitecture in vivo. Yet, this technology has been only applicable at peripheral sites, in only few research centers. Clinical CBCT is more widely available but quantitative assessment of the bone structure is usually not performed. We aimed to compare the assessment of bone structure with CBCT (NewTom 5G, QR, Verona, Italy) and HR-pQCT (XtremeCT, Scanco Medical AG, Brüttisellen, Switzerland). METHODS: Twenty-four distal radius specimens were scanned with these two devices with a reconstructed voxel size of 75 µm for Newtom 5G and 82 µm for XtremeCT, respectively. A rescaling-registration scheme was used to define the common volume of interest. Cortical and trabecular compartments were separated using a semiautomated double contouring method. Density and microstructure were assessed with the HR-pQCT software on both modality images. RESULTS: Strong correlations were found for geometry parameters (r = 0.98-0.99), volumetric density (r = 0.91-0.99), and trabecular structure (r = 0.94-0.99), all p < 0.001. Correlations were lower for cortical microstructure (r = 0.80-0.89), p < 0.001. However, absolute differences were observed between modalities for all parameters, with an overestimation of the trabecular structure (trabecular number, 1.62 ± 0.37 vs. 1.47 ± 0.36 mm(-1)) and an underestimation of the cortical microstructure (cortical porosity, 3.3 ± 1.3 vs. 4.4 ± 1.4 %) assessed on CBCT images compared to HR-pQCT images. CONCLUSIONS: Clinical CBCT devices are able to analyze large portions of distal bones with good spatial resolution and limited irradiation. However, compared to dedicated HR-pQCT, the assessment of microarchitecture by NewTom 5G dental CBCT showed some discrepancies, for density measurements mainly. Further technical developments are required to reach optimal assessment of bone characteristics.

Multi-frequency shear modulus measurements discriminate tumorous from healthy tissues.

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.

Influence of loading conditions in finite element analysis assessed by HR-pQCT on ex vivo fracture prediction.

Many fractures occur in individuals with normal areal Bone Mineral Density (aBMD) measured by Dual X-ray Absorptiometry (DXA). High Resolution peripheral Quantitative Computed Tomography (HR-pQCT) allows for non-invasive evaluation of bone stiffness and strength through micro finite element (µFE) analysis at the tibia and radius. These µFE outcomes are strongly associated with fragility fractures but do not provide clear enhancement compared with DXA measurements. The objective of this study was to establish whether a change in loading conditions in standard µFE analysis assessed by HR-pQCT enhance the discrimination of low-trauma fractured radii (n = 11) from non-fractured radii (n = 16) obtained experimentally throughout a mechanical test reproducing a forward fall. Micro finite element models were created using HR-pQCT images, and linear analyses were performed using four different types of loading conditions (axial, non-axial with two orientations and torsion). No significant differences were found between the failure load assessed with the axial and non-axial models. The different loading conditions tested presented the same area under the receiver operating characteristic (ROC) curves of 0.79 when classifying radius fractures with an accuracy of 81.5%. In comparison, the area under the curve (AUC) is 0.77 from DXA-derived ultra-distal aBMD of the forearm with an accuracy of 85.2%. These results suggest that the restricted HR-pQCT scanned region seems not sensitive to loading conditions for the prediction of radius fracture risk based on ex vivo experiments (n = 27).

A credible homogenized finite element model to predict radius fracture in the case of a forward fall.

Fragility fractures that occur after a fall from a standing height or less are almost always due to osteoporosis, which remains underdiagnosed and untreated. Patient-specific finite element (FE) models have been introduced to predict bone strength and strain. This approach, based on structure mechanics, is derived from Quantitative Computed Tomography (QCT), and element mechanical properties are computed from bone mineral densities. In this study, we developed a credible finite element model of the radius to discriminate low-trauma-fractured radii from non-fractured radii obtained experimentally. Thirty cadaveric radii were impacted with the same loading condition at 2 m/s, and experimental surface strain was retrieved by stereo-correlation in addition to failure loads in fracture cases. Finite element models of the distal radius were created from clinical computed tomography. Different density-elasticity relationships and failure criteria were tested. The strongest agreement (simulations-experiments) for average strain showed a Spearman's rank correlation (ρ) between 0.75 and 0.82, p < 0.0001, with a root mean square error between 0.14 and 0.19%. The experimental mean strain was 0.55%. Predicted failure load error (23%) was minimized for derived Pistoia's failure criterion. Numerical failure demonstrated area under the receiver operating characteristic (ROC) curves of 0.76 when classifying radius fractures with an accuracy of 82%. These results suggest that a credible FE modelling method in a large region of interest (distal radius) is a suitable technique to predict radius fractures after a forward fall.

Variabilities in µQCT-based FEA of a tumoral bone mice model.

A finite element analysis based on Micro-Quantitative Computed Tomography (µQCT) is a method with high potential to improve fracture risk prediction. However, the segmentation process and model generation are generally not automatized in their entirety. Even with a rigorous protocol, the operator might add uncertainties during the creation of the model. The aim of this study was to evaluate a µQCT-based model of mice tumoral and sham tibias in terms of the variabilities induced by the operator and sensitivity to operator-dependent variables (such as model orientation or length). Two different operators generated finite element (FE) models from µCT images of 8 female Balb/c nude mice tibias aged 10 weeks old with bone tumors induced in the right tibia and with sham injection in the left. From these models, predicted failure load was determined for two different boundary conditions: fixed support and spherical joints. The difference between the predicted and experimental failure load of both operators was large (-122% to 93%). The difference in the predicted failure load between operators was less for the spherical joints boundary conditions (9.8%) than for the fixed support (58.3%), p < 0.001, whereas varying the orientation of bone tibia caused more variability for the fixed support boundary condition (44.7%) than for the spherical joints (9.1%), p < 0.002. Varying tibia length had no significant effect, regardless of boundary conditions (<4%). When using the same mesh and same orientation, the difference between operators is non-significant (<6%) for each model. This study showed that the operator influences the failure load assessed by a µQCT-based finite element model of the tumoral and sham mice tibias. The results suggest that automation is needed for better reproducibility.

Intrinsic properties of osteomalacia bone evaluated by nanoindentation and FTIRM analysis.

Osteomalacia is a pathological bone condition consisting in a deficient primary mineralization of the matrix, leading to an accumulation of osteoid tissue and reduced bone mechanical strength. The amounts, properties and organization of bone constituents at tissue level, are known to influence its mechanical properties. It is then important to investigate the relationship between mechanical behavior and tissue composition at this scale in order to provide a better understanding of bone fragility mechanisms associates with this pathology. Our purpose was to analyze the links between ultra-structural properties and the mechanical behavior of this pathological bone tissue (osteomalacia) at tissue level (mineral and osteoid separately, or global). Four bone biopsies were taken from patients with osteomalacia, and subsequently embedded, sectioned, and polished. Then nanoindentation tests were performed to determine local elastic modulus E, contact hardness Hc and true hardness H for both mineralized and organic bone phases and for the global bone. The creep of the bone was also studied using a special indentation procedure in order to assess visco-elasto-plastic (creep) bone behavior. This allowed a detailed study of the rheological models adapted to the bone and to calculate the parameters associated to a Burgers model. Ultra-structural parameters were measured by Fourier Transform InfraRed Microspectroscopy (FTIRM) on the same position as the indents. The use of rheological models confirmed a significant contribution from the organic phase on the viscous character of bone tissue. The elastic E and the elasto-plastic Hc deformation were correlated to both collagen maturity and Mineral/Matrix. The pure plastic deformation H was only correlated to the mineral phase. Our data show that mineral phase greatly affects mechanical variables (moduli and viscosities) and that organic phase (as illustrated in osteoid tissue) may play an important role in the creep behavior of bone. In conclusion, this study brings mechanical and physicochemical values for osteoid and mineral phases.

Deletion of OPN in BSP knockout mice does not correct bone hypomineralization but results in high bone turnover.

The two SIBLING (Small Integrin Binding Ligand N-linked Glycoproteins), bone sialoprotein (BSP) and osteopontin (OPN) are expressed in osteoblasts and osteoclasts. In mature BSP knockout (KO, -/-) mice, both bone formation and resorption as well as mineralization are impaired. OPN-/- mice display impaired resorption, and OPN is described as an inhibitor of mineralization. However, OPN is overexpressed in BSP-/- mice, complicating the understanding of their phenotype. We have generated and characterized mice with a double KO (DKO) of OPN and BSP, to try and unravel their respective contributions. Despite the absence of OPN, DKO bones are still hypomineralized. The SIBLING, matrix extracellular phosphoglycoprotein with ASARM motif (MEPE) is highly overexpressed in both BSP-/- and DKO and may impair mineralization through liberation of its ASARM (Acidic Serine-Aspartate Rich MEPE associated) peptides. DKO mice also display evidence of active formation of trabecular, secondary bone as well as primary bone in the marrow-ablation repair model. A higher number of osteoclasts form in DKO marrow cultures, with higher resorption activity, and DKO long bones display a localized and conspicuous cortical macroporosity. High bone formation and resorption parameters, and high cortical porosity in DKO mice suggest an active bone modeling/remodeling, in the absence of two key regulators of bone cell performance. This first double KO of SIBLING proteins thus results in a singular, non-trivial phenotype leading to reconsider the interpretation of each single KO, concerning in particular matrix mineralization and the regulation of bone cell activity.

Bone cortical thickness and porosity assessment using ultrasound guided waves: An ex vivo validation study.

Several studies showed the ability of the cortex of long bones such as the radius and tibia to guide mechanical waves. Such experimental evidence has given rise to the emergence of a category of quantitative ultrasound techniques, referred to as the axial transmission, specifically developed to measure the propagation of ultrasound guided waves in the cortical shell along the axis of long bones. An ultrasound axial transmission technique, with an automated approach to quantify cortical thickness and porosity is described. The guided modes propagating in the cortex are recorded with a 1-MHz custom made linear transducer array. Measurement of the dispersion curves is achieved using a two-dimensional spatio-temporal Fourier transform combined with singular value decomposition. Automatic parameters identification is obtained through the solution of an inverse problem in which the dispersion curves are predicted with a two-dimensional transverse isotropic free plate model. Thirty-one radii and fifteen tibiae harvested from human cadavers underwent axial transmission measurements. Estimates of cortical thickness and porosity were obtained on 40 samples out of 46. The reproducibility, given by the root mean square error of the standard deviation of estimates, was 0.11 mm for thickness and 1.9% for porosity. To assess accuracy, site-matched micro-computed tomography images of the bone specimens imaged at 9 µm voxel size served as the gold standard. Agreement between micro-computed tomography and axial transmission for quantification of thickness and porosity at the radius and tibia ranged from R2=0.63 for porosity (root mean square error RMSE=1.8%) to 0.89 for thickness (RMSE=0.3 mm). Despite an overall good agreement for porosity, the method performs less well for porosities lower than 10%. The heterogeneity and general complexity of cortical bone structure, which are not fully accounted for by our model, are suspected to weaken the model approximation. This study presents the first validation study for assessing cortical thickness and porosity using the axial transmission technique. The automatic signal processing minimizes operator-dependent errors for parameters determination. Recovering the waveguide characteristics, that is to say cortical thickness and porosity, could provide reliable information about skeletal status and future fracture risk.

Intrinsic mechanical properties of trabecular calcaneus determined by finite-element models using 3D synchrotron microtomography.

To determine intrinsic mechanical properties (elastic and failure) of trabecular calcaneus bone, chosen as a good predictor of hip fracture, we looked for the influence of image's size on a numerical simulation. One cubic sample of cancellous bone (9 x 9 x 9 mm(3)) was removed from the body of the calcaneus (6 females, 6 males, 79+/-9 yr). These samples were tested under compressive loading. Before compressive testing, these samples were imaged at 10.13 microm resolution using a 3D microcomputed tomography (muCT) (ESRF, France). The muCT images were converted to finite-element models. Depending on the bone density values (BV/TV), we compared two different finite element models: a linear hexahedral and a linear beam finite element models. Apparent experimental Young's modulus (E(app)(exp)) and maximum apparent experimental compressive stress (sigma(max)(exp)) were significantly correlated with bone density obtained by Archimedes's test (E(app)(exp)=236+/-231 MPa [19-742 MPa], sigma(max)(exp)=2.61+/-1.97 MPa [0.28-5.81 MPa], r>0.80, p<0.001). Under threshold at 40 microm, the size of the numerical samples (5.18(3) and 6.68(3)mm(3)) seems to be an important parameter on the accuracy of the results. The numerical trabecular Young's modulus was widely higher (E(trabecular)(num)=34,182+/-22,830 MPa [9700-87,211 MPa]) for the larger numerical samples and high BV/TV than those found classically by other techniques (4700-15,000 MPa). For rod-like bone samples (BV/TV<12%, n=7), Young's modulus, using linear beam element (E(trabecular)(num-skeleton): 10,305+/-5500 MPa), were closer to the Young's modulus found by other techniques. Those results show the limitation of hexahedral finite elements at 40 microm, mostly used, for thin trabecular structures.

An ex vivo experiment to reproduce a forward fall leading to fractured and non-fractured radii.

Forward falls represent a risk of injury for the elderly. The risk is increased in elderly persons with bone diseases, such as osteoporosis. However, half of the patients with fracture were not considered at risk based on bone density measurement (current clinical technique). We assume that loading conditions are of high importance and should be considered. Real loading conditions in a fall can reach a loading speed of 2m/s on average. The current study aimed to apply more realistic loading conditions that simulate a forward fall on the radius ex vivo. Thirty radii from elderly donors (79y.o.±12y.o., 15 males, 15 females) were loaded at 2m/s using a servo-hydraulic testing machine to mimic impact that corresponds to a fall. Among the 30 radii, 14 had a fracture after the impact, leading to two groups (fractured and non-fractured). Surfacic strain fields were measured using stereovision and allow for visualization of fracture patterns. The average maximum load was 2963±1274N. These experimental data will be useful for assessing the predictive capability of fracture risk prediction methods such as finite element models.

Genetic algorithms-based inversion of multimode guided waves for cortical bone characterization.

Recent progress in quantitative ultrasound has exploited the multimode waveguide response of long bones. Measurements of the guided modes, along with suitable waveguide modeling, have the potential to infer strength-related factors such as stiffness (mainly determined by cortical porosity) and cortical thickness. However, the development of such model-based approaches is challenging, in particular because of the multiparametric nature of the inverse problem. Current estimation methods in the bone field rely on a number of assumptions for pairing the incomplete experimental data with the theoretical guided modes (e.g. semi-automatic selection and classification of the data). The availability of an alternative inversion scheme that is user-independent is highly desirable. Thus, this paper introduces an efficient inversion method based on genetic algorithms using multimode guided waves, in which the mode-order is kept blind. Prior to its evaluation on bone, our proposal is validated using laboratory-controlled measurements on isotropic plates and bone-mimicking phantoms. The results show that the model parameters (i.e. cortical thickness and porosity) estimated from measurements on a few ex vivo human radii are in good agreement with the reference values derived from x-ray micro-computed tomography. Further, the cortical thickness estimated from in vivo measurements at the third from the distal end of the radius is in good agreement with the values delivered by site-matched high-resolution x-ray peripheral computed tomography.

Relationship between compressive properties of human os calcis cancellous bone and microarchitecture assessed from 2D and 3D synchrotron microtomography.

The aim of this study was to determine the contribution of 2D and 3D microarchitectural characteristics in the assessment of the mechanical strength of os calcis cancellous bone. A sample of cancellous bone was removed in a medio-lateral direction from the posterior body of calcaneus, taken at autopsy in 17 subjects aged 61-91 years. The sample was first used for the assessment of morphological parameters from 2D morphometry and 3D synchrotron microtomography (microCT) (spatial resolution=10 microm). The 2D morphometry was obtained from three slices extracted from the 3D microCT images. Very good concordance was shown between 3D microCT slices and the corresponding physical histologic slices. In 2D, the standard histomorphometric parameters, fractal dimension, mean intercept length, and connectivity were computed. In 3D, histomorphometric parameters were computed using both the 3D mean intercept length method and model-independent techniques. The 3D fractal dimension and the 3D connectivity, assessed by Euler density, were also evaluated. The cubic samples were subjected to elastic compressive tests in three orthogonal directions (X, Y, Z) close to the main natural trabecular network directions. A test was performed until collapse of trabecular network in the main direction (Z). The mechanical properties were significantly correlated to most morphological parameters resulting from 2D and 3D analysis. In 2D, the correlation between the mechanical strength and bone volume/tissue volume was not significantly improved by adding structural parameters or connectivity parameter (nodes number/tissue volume). In 3D, one architectural parameter (the trabecular thickness, Tb.Th) permitted to improve the estimation of the compressive strength from the bone volume/tissue volume alone. However, this improvement was minor since the correlation with the BV/TV alone was high (r=0.96). In conclusion, which is in agreement with the statistic's rules, we found, in this study, that the determination of the os calcis bone compressive strength using the 3D bone volume fraction cannot be improved by adding 3D architectural parameters.

The degree of mineralization is a determinant of bone strength: a study on human calcanei.

Strength of bones depends on bone matrix volume (BMV), bone microarchitecture, and also on the degree of mineralization of bone (DMB). We have recently shown in osteoporotic patients treated with alendronate that fracture risk decreased and bone mineral density increased with a parallel increase of the DMB due to prolonged secondary mineralization but without modifications of BMV or bone microarchitecture. DMB and strength were both measured at the tissue level in calcaneus bone samples taken at autopsy from 20 subjects (aged 78 +/- 8 years, 8 women, 12 men) who died suddenly without apparent bone disease. DMB parameters measured on microradiographs (mean DMB, distribution of DMB, most frequent maximum DMB value, and width at half maximum, an index reflecting the homogeneity of DMB) were compared with those reported in iliac cancellous bone samples of 43 human bones. Histomorphometric measurements of microarchitectural parameters (TbTh, TbN, and TbSp) were also measured. Compression tests were performed on contiguous samples of the same calcaneus on a universal screw-driven machine (Schenck RSA 250). A 5000-N load cell (TME, F 501 TC) measured the compressive load. The displacement was measured directly on the sample using a specific displacement transducer developed by the <> The apparent Young's modulus (E), the maximal strength (sigma(max)), and the work (W) until failure were measured. In human cancellous bone tissue, mean DMB (+/- SD) was higher in calcaneus (1.135 +/- 0.147 g/cm(3)) than in iliac crest (1.098 +/- 0.077 g/cm(3)). The mean most frequent maximum DMB values (mean DMB freq. max.) were 1.118 +/- 0.175 g/cm(3) in calcaneus and 1.108 +/- 0.095 g/cm(3) in iliac samples, and DMB was more heterogeneous in calcaneus than in iliac samples (mean width at half maximum were 0.270 +/- 0.127 versus 0.227 +/- 0.056 g/cm(3), respectively). Compression tests revealed significant positive linear correlations between DMB and both elastic modulus (r(2) = 0.69) and maximal strength (r(2) = 0.69). Correlations with DMB persisted (P < 0.003) even after adjustment for both calcified bone volume, for the Young's modulus (E), the maximal strength (sigma(max)) (r(2) = 0.44 and 0.41, respectively), and microarchitectural parameters (0.50 < r(2) < 0.56, P < 0.001). The same results were obtained with the work to fracture (W) (0.23 < r(2) < 0.46, P < 0.045). We conclude that the more the cancellous tissue was mineralized, the higher was its stiffness and compressive strength. This may explain the increase in bone strength when DMB is modified in a physiological range without necessary changes of BMV and bone microarchitecture. The impact of such modifications on fracture risk and the therapeutic implications of these data remain to be analyzed.

Finite element dependence of stress evaluation for human trabecular bone.

Numerical simulation using finite element models (FEM) has become more and more suitable to estimate the mechanical properties of trabecular bone. The size and kind of elements involved in the models, however, may influence the results. The purpose of this study is to analyze the influence of hexahedral elements formulation on the evaluation of mechanical stress applied to trabeculae bone during a compression test simulation. Trabecular bone cores were extracted from 18 L2 vertebrae (12 women and 6 men, mean age: 76 ± 11, BV/TV=7.5 ± 1.9%). Samples were micro-CT scanned at 20 µm isotropic voxel size. Micro-CT images have been sub-sampled (20, 40 and 80 µm) to create 5.6 mm cubic FEM. For each sample, a compression test FEM has been created, using either 8-nodes linear hexahedral elements with full or reduced integration or 20-nodes quadratic hexahedral elements fully integrated, resulting in nine models per samples. Bone mechanical properties have been assumed isotropic, homogenous and to follow a linear elastic behavior law (Young modulus: 8 GPa, Poisson ratio: 0.3). Despite micro-architecture modifications (loss of connectivity, trabeculae thickening) due to voxel size increase, apparent mechanical properties calculated with low resolution models are significantly correlated with high resolution results, no matter the element formulation. However, stress distributions are more sensitive to both resolution and element formulation modifications. With linear elements, increasing voxel size leads to an alteration of stress concentration areas due to stiffening errors. On the opposite, the use of reduced integration induces severe smoothing and underestimation of stress fields resulting in stress raisers loss. Notwithstanding their high computational cost, quadratic elements are most appropriate for stress prediction in low resolution trabecular bone FEM. These observations are dependent on trabecular bone micro-architecture, and are more significant for low density sample displaying low trabecular thickness. In conclusion, we found that element formulation is almost important as element size when evaluating trabecular bone mechanical behavior at trabeculae scale. Therefore, element type should be chosen carefully when evaluating trabecular bone behavior using FEM.

Computed tomography, histological and ultrasonic measurements of adolescent scoliotic rib hump geometrical and material properties.

In Adolescent Idiopathic Scoliosis (AIS), numerical models can enhance orthopaedic or surgical treatments and provide reliable insights into the mechanism of progression. Computational methods require knowledge of relevant parameters, such as the specific geometrical or material properties of the AIS rib, about which there is currently a lack of information. The aim of our study was to determine the geometrical and material properties (Young's modulus [E] and Poisson's ratio [ν]) for AIS rib bones. Twelve ribs extracted during gibbectomy on 15 and 17 year old girls were tested using computed tomography (CT) scanner, histology and ultrasonic scanner. The mean porosity (± standard deviation (SD)) is 1.35 (±0.52)% and the mean (±SD) bone mineral density is 2188 (±19)mmHA/cc. The cortical part of the AIS rib hump is found to be thicker than physiological values in the literature. To mimic the rib hump for an AIS girl, our results suggest that ribs should be modeled as hollow circular cylinders with a 10.40 (±1.02)mm external radius and 7.56mm (±0.75) internal radius, and material properties with a mean E of 14.9GPa (±2.6) and a mean ν of 0.26 (±0.08).

Respective roles of organic and mineral components of human cortical bone matrix in micromechanical behavior: an instrumented indentation study.

Bone is a multiscale composite material made of both a type I collagen matrix and a poorly crystalline apatite mineral phase. Due to remodeling activity, cortical bone is made of Bone Structural Units (BSUs) called osteons. Since osteon represents a fundamental level of structural hierarchy, it is important to investigate the relationship between mechanical behavior and tissue composition at this scale for a better understanding of the mechanisms of bone fragility. The aim of this study is to analyze the links between ultrastructural properties and the mechanical behavior of bone tissue at the scale of osteon. Iliac bone biopsies were taken from untreated postmenopausal osteoporotic women, embedded, sectioned and microradiographed to assess the degree of mineralization of bone (DMB). On each section, BSUs of known DMB were indented with relatively high load (~500 mN) to determine local elastic modulus (E), contact hardness (H(c)) and true hardness (H) of several bone lamellae. Crystallinity and collagen maturity were measured by Fourier Transform InfraRed Microspectroscopy (FTIRM) on the same BSUs. Inter-relationships between mechanical properties and ultrastructural components were analyzed using multiple regression analysis. This study showed that elastic deformation was only explained by DMB whereas plastic deformation was more correlated with collagen maturity. Contact hardness, reflecting both elastic and plastic behaviors, was correlated with both DMB and collagen maturity. No relationship was found between crystallinity and mechanical properties at the osteon level.