- Email: [email protected]
89 Derivation of Structure-Function Relations in Cortical Bone
Giuseppe and Toffanin
125
known n and provided experimental n values with a precision between 0.1 and 3%. In bone, n was 0.42 0.01 (osteonal tissue n 5 0.41+/-0.04; interstitial tissue n 5 0.43 0.02; p ! 0.001) and corresponding values of Young’s modulus of 17.9 2.5 GPa (osteonal tissue E 5 16.8 2.4 GPa; interstitial tissue E 5 19.2 1.8; p ! 0.001). Our values of n are at the upper limit of the range of values derived from macroscopic mechanical tests or low frequency ultrasound testing. A direct comparison with our tissue-level measurements requires a careful analysis of the impact of cortical porosity on n. Our preliminary results indicate that combination of high-resolution SAM and nanoindentation may be relevant to determine both Poisson’s ratio and Young modulus of bone tissue.
87 PREDICTION OF HIP FRACTURE LOAD FROM RADIOGRAPHS BY COMBINED ANALYSIS OF TRABECULAR BONE STRUCTURE AND BONE GEOMETRY
86 IN VITRO STUDIES OF INFLUENCE OF INTRINSIC PROPERTIES OF THE PORE-FLUID ON THE ULTRASONIC PARAMETERS IN CANCELLOUS BONES
Since conventional radiography is widely available with low imaging cost, it is of considerable interest to discover how well bone mechanical competence can be determined using this technology. We tested the hypothesis that mechanical strength of femur can be estimated by the combined analysis of bone trabecular structure and geometry. The sample consisted of 62 cadaver femurs (34 females, 28 males). After radiography and DXA, femora were mechanically tested in side impact configuration. Fracture patterns were classified as being cervical or trochanteric. Computerized image analysis was applied to obtain structure-related trabecular parameters (trabecular bone area, Euler number, homogeneity index, and trabecular main orientation), and set of geometrical variables (neck-shaft angle, medial calcar and femoral shaft cortex thicknesses, and femoral neck axis length). Multiple linear regression analysis was performed to identify the variables that best explain variation in failure load between subjects. In cervical fracture cases, trabecular bone area and femoral neck axis length explained 64% of the variability in failure loads, while femoral neck BMD also explained 64%. In trochanteric fracture cases, Euler number and femoral cortex thickness explained 66% of the variability in failure load, while trochanteric BMD explained 72%. As a conclusion, structural parameters of trabecular bone and bone geometry predict in vitro failure loads of the proximal femur with similar accuracy as DXA, when using appropriate image analysis technology.
M. Pakula F. Padilla and P. Laugier, Laboratoire d’Imagerie Parametrique, University Paris 6, Paris, France; The paper is focused on comparative in vitro ultrasonic experiments of cancellous bones filled with different fluids, marrow, water and alcohol with the goal of evaluating the contribution of different fluid physical properties to wave attenuation and dispersion. Thirty trabecular plates were prepared from fresh human condyles obtained from 15 individuals (one plate from both the left and right femur of the each individual). Ultrasonic measurements were performed on intact specimens (i.e., with marrow inside) and subsequently on water and alcohol saturated specimens (i.e. the marrow was removed). Whereas the three fluids have similar mass density, they differ by their viscosity (viscosity in marrow is 2 to 3 orders of magnitude higher than in water and alcohol) and elastic properties (bulk modulus of water and marrow is twice higher than for alcohol). The measurements were performed using two pairs of broadband ultrasonic transducers (center frequencies 0.5 and 1 MHz respectively) using the insertion method by scanning the specimens in spatial steps of 2 mm. Then, the wave parameters (frequency-dependent attenuation and velocity) were calculated at each scan point after deconvolution of the pulse transmitted through the specimens by the reference pulse transmitted through water. The spatial variation of the acoustic properties reflects the strong heterogeneity of cancellous bone. Within one specimen the amplitude of the transmitted pulse may vary within a range of 60 dB. Moreover, overlapped wave pulses (fast and slow Biot waves) were often associated with dense areas precluding further signal analysis without their former separation. To overcome the problem of overlapped pulses, we decided to restrict the spectral analysis for the low bone density areas, where only a single pulse was observed. It was found that phase velocity differs between alcohol saturated specimens (1215 7 m/s at 0.8 MHz) on one hand and water (1502 5 m/s) and marrow (1482 16 m/s) saturated specimens on the other hand, reflecting the impact of elastic properties of the fluid filling pores. Given the low density of tested areas, the values are close to the speed of sound in the free fluid. In contrast, the values of the attenuation coefficient remained similar for the three saturating fluid even for intact specimens, where the marrow viscosity in comparison with the other fluids is very high. This result showed that attenuation is poorly affected by the viscous properties of the saturating fluid, suggesting that attenuation is mainly due to absorption by the bone tissue and/or scattering by the microarchitecture. These findings should guide us in the future for the choice of appropriate propagation models, in order to be able to solve inverse problems and monitor these properties from ultrasound measurements. PhaseV elocity[ m/s]
1550 1500 1450 1400 1350 1300 1250 1200
Attenuation[ dB/cm]
0,2
0,4
0,6
0,8
1,0
9,0 8,5 8,0 7,5 7,0 6,5 6,0 5,5 5,0 4,5 4,0 3,5
1,2
1,4
Alcohol Marrow Water
0,2
0,4
0,6
0,8
1,0
1,2
1,4
Frequency [MHz]
P. Pulkkinen,1 T. Ja¨msa¨,1 E.-M. Lochm€ uller,2 V. Kuhn,3 M. Nieminen,4 and F. Eckstein5, 1University of Oulu, Oulu, Finland, 2Ludwig-MaximiliansUniversita¨t, M€ unchen, Germany, 3Medical University, Innsbruck, Austria, 4Oulu University Hospital, Oulu, Finland, 5Paracelsus Medical University, Salzburg, Austria;
88 BMD AND BONE GEOMETRY IN THE DISCRIMINATION OF FRACTURE CASES IN FEMALES WITHOUT OSTEOPOROSIS P. Pulkkinen,1 J. Partanen,2 P. Jalovaara,1 and T. Ja¨msa¨1, 1University of Oulu, Oulu, Finland, 2Oulu Deaconess Institute, Oulu, Finland; Even though majority of fractures occur among people who are not classified as having osteoporosis, there is still limited attention in the fracture risk assessment in individuals with normal BMD. In order to achieve a more detailed figure on the factors associated with increased fracture risk in this population, we investigated BMD and geometrical risk factors in individuals without osteoporosis. The study subjects consisted of 57 postmenopausal females with non-pathologic cervical (n 5 39) or trochanteric (n 5 18) hip fractures (mean age 74.2 years). The control group consisted of 40 age-matched females (mean age 73.7 years). BMD was measured at the femoral neck (FNBMD) and trochanter (TRBMD) with Lunar DPX scanner using standard measurement routines. T-score -2.5 was used to exclude subjects with osteoporosis. Anterioposterior pelvic roentgenograms were taken, and several dimensions were measured from the digitized X-rays, such as hip axis length (HAL), femoral neck axis length (FNAL), femoral neck and shaft cortex thicknesses (FNC and FSC), and femoral neck-shaft angle (NSA). Results showed that 23 cervical (59 %) and 6 trochanteric (33.3 %) fracture cases were classified not to have osteoporosis. In those without osteoporosis, TRBMD was able to discriminate fracture patients from controls in the cases of trochanteric fractures (p ! 0.001). On the contrary, FNBMD predicted a cervical fracture only in individuals with low BMD (osteoporosis). For subjects without osteoporosis, FNBMD was not a risk factor for a cervical fracture, while FNC and FSC, and NSA were able to discriminate these patients from the controls (p 5 0.002, p 5 0.003, and p ! 0.001, respectively). To conclude, BMD is a risk factor only for trochanteric fractures in non-osteoporotic subjects, while bone geometry is the predisposing factor for cervical fracture within non-osteoporotic individuals. The study supports classification of fracture types for relevant fracture risk assessment.
89 DERIVATION OF STRUCTURE-FUNCTION RELATIONS IN CORTICAL BONE K. Raum,1 Q. Grimal,2 and A. Gerisch3, 1Q-BAM Group, Dept. of Orthopedics, Martin Luther University of Halle-Wittenberg, Halle, Germany, 2UPMC Univ
Journal of Clinical Densitometry: Assessment of Skeletal Health
Volume 12, 2009
126 Paris 06, CNRS UMR 7623, LIP, Paris, France, 3Institut f€ ur Mathematik, Martin Luther University of Halle-Wittenberg, Halle, Germany; The accurate prediction of the mechanical competence of bone requires a detailed understanding of structure-function relations of the tissue. High frequency ultrasound has become one of the most powerful tools for microelastic bone characterization, since it allows not only to image the microstructure, but also to map the heterogeneous anisotropic elasticity of the mineralized tissue. Due to its scalability ultrasound can be applied for large and small animal studies as well as for fundamental studies at the mesoscopic (tissue) and microscopic (lamellar) levels. Even at low frequencies (50 MHz) the mineralized tissue matrix can be separated from the larger pores (Haversian canals) and the elastic coefficient in the probing direction can be measured in two dimensions. Depending on the type of sample surface preparation (flat or cylindrically shaped) the local distribution of a single elastic coefficient or the average transverse isotropic stiffness tensor can be derived. With frequencies in the GHz range the lamellar bone structure can be analyzed. However, at one GHz the acoustic wavelength is still one order of magnitude larger than the individual mineralized collagen fibrils. While the thickness of a lamellar unit and some elastic coefficients can easily be assessed from the acoustic image, the derivation of the anisotropic elastic properties of the resulting tissue compound can be revealed by combination with finite element (FE) deformation analysis. A lamellar bone tissue model with a lamellar unit composed of six sublayers has been developed. The orientation of the symmetry axis between adjacent sublayers is shifted clockwise. A sublayer consists of one to ten layers of parallel oriented mineralized collagen fibrils. The transverse isotropic constants of the fibrils are derived from the GHz measurement. By changing the individual layer thicknesses various degrees of anisotropy could be produced. A good agreement with the lamellar pattern obtained in 1.2-GHz SAM images as well as with the anisotropic elastic coefficients measured at the mesoscopic level (50-MHz ultrasound) was obtained by choosing an asymmetric lamellar unit.
International Workshop Abstracts Tissue-level elastic coefficients can be extracted from acoustic impedance measurements using scanning acoustic microscopy (SAM) [1]. As an extension of previous SAM studies reported by our group, this work aimed at validating acoustic microscopy as a modality to map elastic modulus at the tissue-level. Toward this goal a face-to-face comparison was conducted between SAM and nanoindentation estimates of elastic modulus. Three embedded transverse sections taken from 3 female human femoral mid-diaphysis were explored by SAM (8mm-spatial resolution). The acoustical tissue elastic modulus Ea of cortical bone was obtained by combining acoustic impedance and bone density derived from degree of mineralization of bone (DMB) provided by Synchrotron m-CT (10-mm-spatial resolution), assuming a Poisson’s ratio of 0.3. Nanoindentation measurements (2-mm-depth indents) were done in 2 line scans (30 indents each at 30-mm-interval) across the radial direction extending from the periosteal to the endosteal of each anatomical quadrant. Ea was compared to nanoindentation estimates of elastic modulus En. Indent images, acoustic impedance and DMB maps were digitally matched using a custom developed image fusion and analysis software for subsequent comparison between site-matched estimates of moduli and DMB. Comparison between Ea and En performed on homogeneous calibration materials (aluminium, PMMA and polycarbonate) of known Poisson ratio (n) yielded a difference of less than 1%. Results for the bone samples, indicated in the Table, are in general agreement with published results. Differences between Ea and En may likely be due to the fixed assumed value of the Poisson’s ratio (n 5 0.3), since values comprised between 0.15 and 0.45 have been reported in the literature. Despite these differences, a highly significant linear correlation between Ea and En was found (R2 5 0.64, p ! 0.001, RMSE 5 1.8 GPa) suggesting that SAM can reliably be used as a modality to quantitatively map the local variations of tissue-level Young’s modulus. Tissue type
Z (Mrayl)
Ea (GPa)
En (GPa)
DMB(g.cm-3)
Osteonal old osteonal interstitial Primary interstitial All
8.5 1.8 9.4 1.2 10.2 1.1 9.1 1.7
31 12 36 9 42 8 35 12
18.4 2.9 20.6 2.5 21.4 2.0 19.5 3.0
0.92 0.09 0.98 0.05 0.99 0.06 0.95 0.08
90 CONTINUUM LEVEL ELASTIC PROPERTIES OF CORTICAL BONE DERIVED FROM ACOUSTIC MICROSCOPY IMAGES Q. Grimal,1 K. Raum,2 A. Gerisch,3 and P. Laugier1, 1Universite´ Pierre et Marie Curie-Paris 6, Laboratoire d’Imagerie Parame´trique, Paris, France and CNRS, LIP, Paris, France, 2Q-BAM Group, Dept. of Orthopedics, MartinLuther-Universita¨t Halle-Wittenberg, Halle, Germany, 3Institut f€ ur Mathematik Martin-Luther-Universita¨t Halle-Wittenberg, Halle, Germany; The aim of this project was to elucidate the relationships between the microscopic properties of bone and its elasticity at the continuum scale (mesoscale). Continuum scale is typically the millimetre scale and is especially relevant for macroscopic finite element modelling of the mechanical response of the bone organ. We propose a method to estimate the mesoscale properties of cortical bone based on a spatial distribution of acoustic properties obtained with scanning acoustic microscopy (SAM, resolution in the range 8-25 microns). The procedure used to compute the mesoscopic elasticity coefficients involves i) the segmentation of the pores to obtain a realistic model of the porosity; ii) the construction of a field of anisotropic elastic coefficients at the microscopic scale which reflects the heterogeneity of the bone matrix; iii) finite element computations of mesoscopic homogenized properties. The computed mesoscopic properties compare well with available experimental data. Multilinear regression models are established to predict the elasticity at the continuum level based on porosity and mean microscopic impedance (R2 O 0.9 for all elastic coefficients). Furthermore the model is used to investigate the impact of different micro- and nano- organisation patterns of bone on the continuum level elasticity. In particular, it appears that the tissue anisotropy at the microscopic level has a major impact on the mesoscopic anisotropy.
91 BONE MICROELASTIC PROPERTIES ASSESSED BY SCANNING ACOUSTIC MICROSCOPY: A FACE-TO-FACE COMPARISON WITH NANOINDENTATION F. Rupin,1 A. Sa€ıed,1 D. Dalmas,2 F. Peyrin,3 K. Raum,4 E. Barthel,2 G. Boivin,5 and P. Laugier1, 1UPMC Univ Paris 06, CNRS UMR 7623, LIP, Paris, France, 2Unite´ Mixte CNRS/Saint-Gobain ‘‘Surface du Verre et Interface’’ UMR 125, Saint-Gobain Recherche, Aubervilliers, France, 3INSERM U630, CNRS UMR 5220, Villeurbanne, and ESRF, Grenoble, France, 4Q-BAM Group, Dept. of Orthopedics, Martin Luther University of Halle-Wittenberg, Halle, Germany, 5 INSERM Unite´ 831, Faculte´ de Me´decine Laennec, Universite´ de Lyon, Lyon, France;
Journal of Clinical Densitometry: Assessment of Skeletal Health
92 IN VITRO OSTEOCLASTOGENESIS AND BONE RESORPTION DURING SPACEFLIGHT R. Tamma G. Colaianni C. Camerino A. Di Benedetto G. Greco M. Strippoli R. Vergari A. Grano L. Mancini and A. Zallone, Department of Human Anatomy and Histology, University of Bari Medical School, Bari, Italy Serious effects on human health are experienced after long duration space missions. Prolonged exposure to microgravity seems to affect several physiological systems. Osteoporosis associated with space flight has been widely documented. Bone loss is considerable, with losses of 1-2% of bone mass per month in flight, occurring predominantly in the load bearing regions of the legs and lumbar spine. Microgravity induces an uncoupling of bone remodeling between bone formation and resorption that could lead to bone loss. Both processes are probably involved, but their relative importance and how they are orchestrated remain unclear. In order to fully understand the mechanisms underlying this bone loss we participated to the FOTON M3 mission launched on September 2007 that carried three experiments OSTEO, OCLAST and PITS developed by our team. For the first time the effect of microgravity directly on OC was studied in vitro and our preliminary results indicate that osteoclasts (OC) are directly affected. The OSTEO experiment has been conducted within bioreactors with a perfusion system, where the differentiation of osteoclast precursors in mature osteoclasts has been tested utilizing as a support a synthetic 3D bone-like biomaterial, skelite, that partially reproduces the chemical composition and physical structure of natural bone. The aim was to analyze the gene expression of osteoclast differentiated in microgravity, compared with ground controls. RNA extracted from the cells has been examined by RT-PCR. The preliminary results indicate that gene involved in osteoclast final maturation and activity, as integrin beta3, cathepsin K, MMP9 were several fold higher in microgravity in comparison with the same cultures on ground, while other genes were substantially at the same level. A microarray analysis is under way. In OCLAST and in PITS experiments mature OC were cultured on devitalized bovine bone slices in appropriate temperature conditions. The experiments started in orbit and was stopped after 4 days in microgravity. After landing the amount of collagen telopeptides released from bone by osteoclasts was measured from culture media, while from PITS samples, RNA was obtained for gene screening. After 5 days of space flight an increase in OC activity was found, proved by an increased number of excavated pits VS the ground control and by increased expression of genes related to osteoclast resorptive activity. These results have been further confirmed by simulating microgravity conditions on ground with
Volume 12, 2009
125
known n and provided experimental n values with a precision between 0.1 and 3%. In bone, n was 0.42 0.01 (osteonal tissue n 5 0.41+/-0.04; interstitial tissue n 5 0.43 0.02; p ! 0.001) and corresponding values of Young’s modulus of 17.9 2.5 GPa (osteonal tissue E 5 16.8 2.4 GPa; interstitial tissue E 5 19.2 1.8; p ! 0.001). Our values of n are at the upper limit of the range of values derived from macroscopic mechanical tests or low frequency ultrasound testing. A direct comparison with our tissue-level measurements requires a careful analysis of the impact of cortical porosity on n. Our preliminary results indicate that combination of high-resolution SAM and nanoindentation may be relevant to determine both Poisson’s ratio and Young modulus of bone tissue.
87 PREDICTION OF HIP FRACTURE LOAD FROM RADIOGRAPHS BY COMBINED ANALYSIS OF TRABECULAR BONE STRUCTURE AND BONE GEOMETRY
86 IN VITRO STUDIES OF INFLUENCE OF INTRINSIC PROPERTIES OF THE PORE-FLUID ON THE ULTRASONIC PARAMETERS IN CANCELLOUS BONES
Since conventional radiography is widely available with low imaging cost, it is of considerable interest to discover how well bone mechanical competence can be determined using this technology. We tested the hypothesis that mechanical strength of femur can be estimated by the combined analysis of bone trabecular structure and geometry. The sample consisted of 62 cadaver femurs (34 females, 28 males). After radiography and DXA, femora were mechanically tested in side impact configuration. Fracture patterns were classified as being cervical or trochanteric. Computerized image analysis was applied to obtain structure-related trabecular parameters (trabecular bone area, Euler number, homogeneity index, and trabecular main orientation), and set of geometrical variables (neck-shaft angle, medial calcar and femoral shaft cortex thicknesses, and femoral neck axis length). Multiple linear regression analysis was performed to identify the variables that best explain variation in failure load between subjects. In cervical fracture cases, trabecular bone area and femoral neck axis length explained 64% of the variability in failure loads, while femoral neck BMD also explained 64%. In trochanteric fracture cases, Euler number and femoral cortex thickness explained 66% of the variability in failure load, while trochanteric BMD explained 72%. As a conclusion, structural parameters of trabecular bone and bone geometry predict in vitro failure loads of the proximal femur with similar accuracy as DXA, when using appropriate image analysis technology.
M. Pakula F. Padilla and P. Laugier, Laboratoire d’Imagerie Parametrique, University Paris 6, Paris, France; The paper is focused on comparative in vitro ultrasonic experiments of cancellous bones filled with different fluids, marrow, water and alcohol with the goal of evaluating the contribution of different fluid physical properties to wave attenuation and dispersion. Thirty trabecular plates were prepared from fresh human condyles obtained from 15 individuals (one plate from both the left and right femur of the each individual). Ultrasonic measurements were performed on intact specimens (i.e., with marrow inside) and subsequently on water and alcohol saturated specimens (i.e. the marrow was removed). Whereas the three fluids have similar mass density, they differ by their viscosity (viscosity in marrow is 2 to 3 orders of magnitude higher than in water and alcohol) and elastic properties (bulk modulus of water and marrow is twice higher than for alcohol). The measurements were performed using two pairs of broadband ultrasonic transducers (center frequencies 0.5 and 1 MHz respectively) using the insertion method by scanning the specimens in spatial steps of 2 mm. Then, the wave parameters (frequency-dependent attenuation and velocity) were calculated at each scan point after deconvolution of the pulse transmitted through the specimens by the reference pulse transmitted through water. The spatial variation of the acoustic properties reflects the strong heterogeneity of cancellous bone. Within one specimen the amplitude of the transmitted pulse may vary within a range of 60 dB. Moreover, overlapped wave pulses (fast and slow Biot waves) were often associated with dense areas precluding further signal analysis without their former separation. To overcome the problem of overlapped pulses, we decided to restrict the spectral analysis for the low bone density areas, where only a single pulse was observed. It was found that phase velocity differs between alcohol saturated specimens (1215 7 m/s at 0.8 MHz) on one hand and water (1502 5 m/s) and marrow (1482 16 m/s) saturated specimens on the other hand, reflecting the impact of elastic properties of the fluid filling pores. Given the low density of tested areas, the values are close to the speed of sound in the free fluid. In contrast, the values of the attenuation coefficient remained similar for the three saturating fluid even for intact specimens, where the marrow viscosity in comparison with the other fluids is very high. This result showed that attenuation is poorly affected by the viscous properties of the saturating fluid, suggesting that attenuation is mainly due to absorption by the bone tissue and/or scattering by the microarchitecture. These findings should guide us in the future for the choice of appropriate propagation models, in order to be able to solve inverse problems and monitor these properties from ultrasound measurements. PhaseV elocity[ m/s]
1550 1500 1450 1400 1350 1300 1250 1200
Attenuation[ dB/cm]
0,2
0,4
0,6
0,8
1,0
9,0 8,5 8,0 7,5 7,0 6,5 6,0 5,5 5,0 4,5 4,0 3,5
1,2
1,4
Alcohol Marrow Water
0,2
0,4
0,6
0,8
1,0
1,2
1,4
Frequency [MHz]
P. Pulkkinen,1 T. Ja¨msa¨,1 E.-M. Lochm€ uller,2 V. Kuhn,3 M. Nieminen,4 and F. Eckstein5, 1University of Oulu, Oulu, Finland, 2Ludwig-MaximiliansUniversita¨t, M€ unchen, Germany, 3Medical University, Innsbruck, Austria, 4Oulu University Hospital, Oulu, Finland, 5Paracelsus Medical University, Salzburg, Austria;
88 BMD AND BONE GEOMETRY IN THE DISCRIMINATION OF FRACTURE CASES IN FEMALES WITHOUT OSTEOPOROSIS P. Pulkkinen,1 J. Partanen,2 P. Jalovaara,1 and T. Ja¨msa¨1, 1University of Oulu, Oulu, Finland, 2Oulu Deaconess Institute, Oulu, Finland; Even though majority of fractures occur among people who are not classified as having osteoporosis, there is still limited attention in the fracture risk assessment in individuals with normal BMD. In order to achieve a more detailed figure on the factors associated with increased fracture risk in this population, we investigated BMD and geometrical risk factors in individuals without osteoporosis. The study subjects consisted of 57 postmenopausal females with non-pathologic cervical (n 5 39) or trochanteric (n 5 18) hip fractures (mean age 74.2 years). The control group consisted of 40 age-matched females (mean age 73.7 years). BMD was measured at the femoral neck (FNBMD) and trochanter (TRBMD) with Lunar DPX scanner using standard measurement routines. T-score -2.5 was used to exclude subjects with osteoporosis. Anterioposterior pelvic roentgenograms were taken, and several dimensions were measured from the digitized X-rays, such as hip axis length (HAL), femoral neck axis length (FNAL), femoral neck and shaft cortex thicknesses (FNC and FSC), and femoral neck-shaft angle (NSA). Results showed that 23 cervical (59 %) and 6 trochanteric (33.3 %) fracture cases were classified not to have osteoporosis. In those without osteoporosis, TRBMD was able to discriminate fracture patients from controls in the cases of trochanteric fractures (p ! 0.001). On the contrary, FNBMD predicted a cervical fracture only in individuals with low BMD (osteoporosis). For subjects without osteoporosis, FNBMD was not a risk factor for a cervical fracture, while FNC and FSC, and NSA were able to discriminate these patients from the controls (p 5 0.002, p 5 0.003, and p ! 0.001, respectively). To conclude, BMD is a risk factor only for trochanteric fractures in non-osteoporotic subjects, while bone geometry is the predisposing factor for cervical fracture within non-osteoporotic individuals. The study supports classification of fracture types for relevant fracture risk assessment.
89 DERIVATION OF STRUCTURE-FUNCTION RELATIONS IN CORTICAL BONE K. Raum,1 Q. Grimal,2 and A. Gerisch3, 1Q-BAM Group, Dept. of Orthopedics, Martin Luther University of Halle-Wittenberg, Halle, Germany, 2UPMC Univ
Journal of Clinical Densitometry: Assessment of Skeletal Health
Volume 12, 2009
126 Paris 06, CNRS UMR 7623, LIP, Paris, France, 3Institut f€ ur Mathematik, Martin Luther University of Halle-Wittenberg, Halle, Germany; The accurate prediction of the mechanical competence of bone requires a detailed understanding of structure-function relations of the tissue. High frequency ultrasound has become one of the most powerful tools for microelastic bone characterization, since it allows not only to image the microstructure, but also to map the heterogeneous anisotropic elasticity of the mineralized tissue. Due to its scalability ultrasound can be applied for large and small animal studies as well as for fundamental studies at the mesoscopic (tissue) and microscopic (lamellar) levels. Even at low frequencies (50 MHz) the mineralized tissue matrix can be separated from the larger pores (Haversian canals) and the elastic coefficient in the probing direction can be measured in two dimensions. Depending on the type of sample surface preparation (flat or cylindrically shaped) the local distribution of a single elastic coefficient or the average transverse isotropic stiffness tensor can be derived. With frequencies in the GHz range the lamellar bone structure can be analyzed. However, at one GHz the acoustic wavelength is still one order of magnitude larger than the individual mineralized collagen fibrils. While the thickness of a lamellar unit and some elastic coefficients can easily be assessed from the acoustic image, the derivation of the anisotropic elastic properties of the resulting tissue compound can be revealed by combination with finite element (FE) deformation analysis. A lamellar bone tissue model with a lamellar unit composed of six sublayers has been developed. The orientation of the symmetry axis between adjacent sublayers is shifted clockwise. A sublayer consists of one to ten layers of parallel oriented mineralized collagen fibrils. The transverse isotropic constants of the fibrils are derived from the GHz measurement. By changing the individual layer thicknesses various degrees of anisotropy could be produced. A good agreement with the lamellar pattern obtained in 1.2-GHz SAM images as well as with the anisotropic elastic coefficients measured at the mesoscopic level (50-MHz ultrasound) was obtained by choosing an asymmetric lamellar unit.
International Workshop Abstracts Tissue-level elastic coefficients can be extracted from acoustic impedance measurements using scanning acoustic microscopy (SAM) [1]. As an extension of previous SAM studies reported by our group, this work aimed at validating acoustic microscopy as a modality to map elastic modulus at the tissue-level. Toward this goal a face-to-face comparison was conducted between SAM and nanoindentation estimates of elastic modulus. Three embedded transverse sections taken from 3 female human femoral mid-diaphysis were explored by SAM (8mm-spatial resolution). The acoustical tissue elastic modulus Ea of cortical bone was obtained by combining acoustic impedance and bone density derived from degree of mineralization of bone (DMB) provided by Synchrotron m-CT (10-mm-spatial resolution), assuming a Poisson’s ratio of 0.3. Nanoindentation measurements (2-mm-depth indents) were done in 2 line scans (30 indents each at 30-mm-interval) across the radial direction extending from the periosteal to the endosteal of each anatomical quadrant. Ea was compared to nanoindentation estimates of elastic modulus En. Indent images, acoustic impedance and DMB maps were digitally matched using a custom developed image fusion and analysis software for subsequent comparison between site-matched estimates of moduli and DMB. Comparison between Ea and En performed on homogeneous calibration materials (aluminium, PMMA and polycarbonate) of known Poisson ratio (n) yielded a difference of less than 1%. Results for the bone samples, indicated in the Table, are in general agreement with published results. Differences between Ea and En may likely be due to the fixed assumed value of the Poisson’s ratio (n 5 0.3), since values comprised between 0.15 and 0.45 have been reported in the literature. Despite these differences, a highly significant linear correlation between Ea and En was found (R2 5 0.64, p ! 0.001, RMSE 5 1.8 GPa) suggesting that SAM can reliably be used as a modality to quantitatively map the local variations of tissue-level Young’s modulus. Tissue type
Z (Mrayl)
Ea (GPa)
En (GPa)
DMB(g.cm-3)
Osteonal old osteonal interstitial Primary interstitial All
8.5 1.8 9.4 1.2 10.2 1.1 9.1 1.7
31 12 36 9 42 8 35 12
18.4 2.9 20.6 2.5 21.4 2.0 19.5 3.0
0.92 0.09 0.98 0.05 0.99 0.06 0.95 0.08
90 CONTINUUM LEVEL ELASTIC PROPERTIES OF CORTICAL BONE DERIVED FROM ACOUSTIC MICROSCOPY IMAGES Q. Grimal,1 K. Raum,2 A. Gerisch,3 and P. Laugier1, 1Universite´ Pierre et Marie Curie-Paris 6, Laboratoire d’Imagerie Parame´trique, Paris, France and CNRS, LIP, Paris, France, 2Q-BAM Group, Dept. of Orthopedics, MartinLuther-Universita¨t Halle-Wittenberg, Halle, Germany, 3Institut f€ ur Mathematik Martin-Luther-Universita¨t Halle-Wittenberg, Halle, Germany; The aim of this project was to elucidate the relationships between the microscopic properties of bone and its elasticity at the continuum scale (mesoscale). Continuum scale is typically the millimetre scale and is especially relevant for macroscopic finite element modelling of the mechanical response of the bone organ. We propose a method to estimate the mesoscale properties of cortical bone based on a spatial distribution of acoustic properties obtained with scanning acoustic microscopy (SAM, resolution in the range 8-25 microns). The procedure used to compute the mesoscopic elasticity coefficients involves i) the segmentation of the pores to obtain a realistic model of the porosity; ii) the construction of a field of anisotropic elastic coefficients at the microscopic scale which reflects the heterogeneity of the bone matrix; iii) finite element computations of mesoscopic homogenized properties. The computed mesoscopic properties compare well with available experimental data. Multilinear regression models are established to predict the elasticity at the continuum level based on porosity and mean microscopic impedance (R2 O 0.9 for all elastic coefficients). Furthermore the model is used to investigate the impact of different micro- and nano- organisation patterns of bone on the continuum level elasticity. In particular, it appears that the tissue anisotropy at the microscopic level has a major impact on the mesoscopic anisotropy.
91 BONE MICROELASTIC PROPERTIES ASSESSED BY SCANNING ACOUSTIC MICROSCOPY: A FACE-TO-FACE COMPARISON WITH NANOINDENTATION F. Rupin,1 A. Sa€ıed,1 D. Dalmas,2 F. Peyrin,3 K. Raum,4 E. Barthel,2 G. Boivin,5 and P. Laugier1, 1UPMC Univ Paris 06, CNRS UMR 7623, LIP, Paris, France, 2Unite´ Mixte CNRS/Saint-Gobain ‘‘Surface du Verre et Interface’’ UMR 125, Saint-Gobain Recherche, Aubervilliers, France, 3INSERM U630, CNRS UMR 5220, Villeurbanne, and ESRF, Grenoble, France, 4Q-BAM Group, Dept. of Orthopedics, Martin Luther University of Halle-Wittenberg, Halle, Germany, 5 INSERM Unite´ 831, Faculte´ de Me´decine Laennec, Universite´ de Lyon, Lyon, France;
Journal of Clinical Densitometry: Assessment of Skeletal Health
92 IN VITRO OSTEOCLASTOGENESIS AND BONE RESORPTION DURING SPACEFLIGHT R. Tamma G. Colaianni C. Camerino A. Di Benedetto G. Greco M. Strippoli R. Vergari A. Grano L. Mancini and A. Zallone, Department of Human Anatomy and Histology, University of Bari Medical School, Bari, Italy Serious effects on human health are experienced after long duration space missions. Prolonged exposure to microgravity seems to affect several physiological systems. Osteoporosis associated with space flight has been widely documented. Bone loss is considerable, with losses of 1-2% of bone mass per month in flight, occurring predominantly in the load bearing regions of the legs and lumbar spine. Microgravity induces an uncoupling of bone remodeling between bone formation and resorption that could lead to bone loss. Both processes are probably involved, but their relative importance and how they are orchestrated remain unclear. In order to fully understand the mechanisms underlying this bone loss we participated to the FOTON M3 mission launched on September 2007 that carried three experiments OSTEO, OCLAST and PITS developed by our team. For the first time the effect of microgravity directly on OC was studied in vitro and our preliminary results indicate that osteoclasts (OC) are directly affected. The OSTEO experiment has been conducted within bioreactors with a perfusion system, where the differentiation of osteoclast precursors in mature osteoclasts has been tested utilizing as a support a synthetic 3D bone-like biomaterial, skelite, that partially reproduces the chemical composition and physical structure of natural bone. The aim was to analyze the gene expression of osteoclast differentiated in microgravity, compared with ground controls. RNA extracted from the cells has been examined by RT-PCR. The preliminary results indicate that gene involved in osteoclast final maturation and activity, as integrin beta3, cathepsin K, MMP9 were several fold higher in microgravity in comparison with the same cultures on ground, while other genes were substantially at the same level. A microarray analysis is under way. In OCLAST and in PITS experiments mature OC were cultured on devitalized bovine bone slices in appropriate temperature conditions. The experiments started in orbit and was stopped after 4 days in microgravity. After landing the amount of collagen telopeptides released from bone by osteoclasts was measured from culture media, while from PITS samples, RNA was obtained for gene screening. After 5 days of space flight an increase in OC activity was found, proved by an increased number of excavated pits VS the ground control and by increased expression of genes related to osteoclast resorptive activity. These results have been further confirmed by simulating microgravity conditions on ground with
Volume 12, 2009