Analysis of three-dimensional microarchitecture and degree of mineralization in bone metastases from prostate cancer using synchrotron microcomputed tomography

Bone 35 (2004) 432 – 438 www.elsevier.com/locate/bone

Analysis of three-dimensional microarchitecture and degree of mineralization in bone metastases from prostate cancer using synchrotron microcomputed tomography Teruki Sone, a,* Tsutomu Tamada, b Yoshimasa Jo, c Hidenao Miyoshi, a and Masao Fukunaga a b

a Department of Nuclear Medicine, Kawasaki Medical School, Kurashiki, Okayama 701-0192, Japan Department of Diagnostic Radiology, Kawasaki Medical School, Kurashiki, Okayama 701-0192, Japan c Department of Urology, Kawasaki Medical School, Kurashiki, Okayama 701-0192, Japan

Received 6 October 2003; revised 2 March 2004; accepted 13 May 2004

Abstract Bone architecture and mineralization are generally considered to be important components of bone quality, and determine bone strength in conjunction with bone mineral density. Although the features of bone quality have recently been studied under conditions in which bone density decreases, such as osteoporosis, little is known in osteosclerotic diseases. In this study, we compared the trabecular bone microarchitecture and degree of mineralization between osteoblastic bone metastasis and degenerative osteosclerosis using synchrotron radiation microcomputed tomography (SR-ACT). Small cubes of lumbar vertebrae were excised postmortem from the sites of osteoblastic metastasis, degenerative osteosclerosis, and comparative sites of normal subjects without skeletal lesions. The samples were imaged at high spatial resolution (voxel size = 6 Am) using the SR-ACT system developed at the synchrotron facility (SPring-8), Hyogo, Japan. The threedimensional (3D) image data were then analyzed for the morphological parameters and the degree of mineralization of bone (DMB). Trabecular bone in metastatic lesions showed a highly connected and isotropic network pattern compared with the normal samples. Although the trabecular surface was markedly irregular in osteoblastic metastases, no significant difference was found in the mean trabecular thickness (Tb.Th) between osteoblastic metastases and normal tissue. The DMB of trabeculae in metastatic lesions had a broader range and lower mean than that of the normal tissue. In contrast, trabecular bone in degenerative osteosclerotic lesions showed a similar degree of anisotropy (DA) and connectivity to the normal tissue, whereas the trabecular thickness was greater in the degenerative osteosclerotic lesions. No significant difference in DBM between degenerative osteosclerosis and normal tissue was detected. These results characterize the difference in bone quality between osteoblastic bone metastasis and degenerative osteosclerosis. Further study on the relationship between bone quality and bone strength in these osteosclerotic lesions would improve our understanding of the pathogenesis of bone fragility. D 2004 Elsevier Inc. All rights reserved. Keywords: Trabecular microarchitecture; Degree of mineralization; Osteoblastic metastases; Prostate cancer; Synchrotron radiation microtomography

Introduction It has been well established that prostate cancer frequently metastasizes to bone, inducing osteosclerotic lesions. However, recent studies have shown that synchronous osteolysis is present in bone metastases of prostate cancer, even when the overall character appears to be osteoblastic [1]. For example, histomorphometric studies have shown that some of the sclerotic lesions are actually * Corresponding author. Department of Nuclear Medicine, Kawasaki Medical School, 577 Matsushima, Kurashiki, Okayama 701-0192, Japan. Fax: +81-86-462-1199. E-mail address: [email protected] (T. Sone). 8756-3282/$ - see front matter D 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.bone.2004.05.011

mixed in nature, with increased activities of both osteoblasts and osteoclasts [2,3]. A number of studies have shown that patients with advanced prostate cancer exhibit elevated levels of bone resorption markers in urine and blood [4,5]. Urinary excretion of pyridinoline and hydroxypyridinoline is augmented in patients with active prostate carcinoma but not in those with localized or controlled disease [4]. Furthermore, the excretion of these bone resorption markers is somewhat better correlated with the extent of bone metastases, as determined by bone scintigraphy, than is plasma alkaline phosphatase, an index of bone formation. The evidence for new bone formation in association with prostate cancer comes from several sources, such as X-rays

T. Sone et al. / Bone 35 (2004) 432–438

showing extensive osteoblastic activity and the elevation of bone formation markers [6], while bone histomorphometry has also confirmed this quantitatively. Thus, increases in osteoid surface and volume have been demonstrated in bone biopsies displaying multiple prostate cancer lesions but not in biopsies from prostate cancer patients that contain minimal or no evidence of tumor [7]. Although bone resorption is of great magnitude in the region of bone metastases, the relative speed of bone formation results in increased overall bone volume [3]. The histomorphometric procedure for examining the structural properties of cancellous bone is based on twodimensional (2D) sections of bone specimens. Threedimensional (3D) morphometric indices are derived from 2D images using stereological methods. Recent advances in microcomputed tomography (ACT) for imaging of cancellous bone have made possible true 3D quantification of trabecular bone microarchitecture [8]. Synchrotron radiation ACT (SR-ACT) may provide 3D images of bone samples with a spatial resolution as high as 1 Am, and allows a geometric reconstruction of the trabecular bone microarchitecture of bone specimens by reconstructing the images in 3D to yield morphological and topological quantitative parameters [9 –11]. SR-ACT may also provide quantitative information regarding the degree of mineralization of bone (DMB) [12]. The technique relies on the physical properties of SR, which allows reconstruction of quantitative maps of the 3D distribution of the linear absorption coefficient within volumetric samples. Because absorption depends on the amount of mineral in bone, suitable calibration is able to relate the reconstructed gray levels in the SR-ACT images to the local DMB. To date, SR-ACT has been used to study the relationships between 3D architecture and the elastic properties of trabecular bone in humans [13,14] and rats [15,16]. In the present study, we used SR ACT, for the first time, to analyze the 3D microarchitecture and DMB in osteoblastic metastases from prostate cancer and compared the results with benign degenerative osteosclerotic lesions.

433

Immediately after removal, soft X-ray radiographs of these lumbar vertebrae were taken, and the regions of osteoblastic metastases and benign degenerative osteosclerosis were identified. A total of 19 cancellous bone cubes, 4 mm per side, were dissected from the vertebral bodies using a precision diamond-bladed band saw along the anatomical axes in the craniocaudal, mediolateral, and anteroposterior planes; 6 from metastatic prostate cancer, 6 from benign osteosclerosis, and the rest from normal sites. Bone samples were prepared by confirming the osteosclerotic changes on soft X-ray radiographs in the anterior region of the vertebral body or from the corresponding sites of normal subjects without skeletal lesions, and excluding the cortical bone and endplate. The bone cubes were then fixed with 10% formaldehyde neutral buffered solution. After scanning, hematoxylin- and eosin-stained sections were prepared on the axial plane from the central portion of each metastatic cube, and the existence of prostate cancer cells identified in all specimens. SR-lCT ACT measurements were carried out on the BL20B2 beam-line at the facility of the Japan Synchrotron Radiation Research Institute (SPring-8), Hyogo, Japan. The BL20B2 beam-line is equipped with a Si(311) double-crystal monochromator and provides monochromatic X-rays of energy from 12 to 72 KeV. A detailed description of the ACT system is reported elsewhere [17]. The system is based on 3D parallel tomographic acquisition. Three-dimensional imaging was performed by taking two-dimensional radiographs as the specimens were rotated through 180j in 0.5j increments. The energy was set to 20 KeV. The radiographs were reconstructed one slice at a time using a convolution back projection algorithm with the Shepp – Logan filter. Each reconstructed slice was then stacked, providing data for 750
750
750 3D images with an isometric voxel size of 6 Am, which is the highest resolution available in our ACT system. Determination of 3D structural indices and DMB

Materials and methods Bone samples The lumbar vertebrae were harvested from 11 autopsies, comprising 2 subjects (aged 67 and 77 years) with multiple osteoblastic metastases from prostate cancer, 4 male subjects (aged 37– 64 years) with benign degenerative osteosclerosis, and 5 male subjects (aged 38– 75 years) with no macroscopic pathological changes in the lumbar spine. All patients, except for those with prostate cancer, had died in accidents or of acute diseases without long-term immobilization. Specimens from subjects who were on medication or had chronic diseases affecting calcium metabolism were not used.

After scanning, image data were transferred to a workstation, and structural indices and DMB calculated using a 3D image analysis system (TRI/3D-BON; Ratoc System Engineering Co. Ltd., Tokyo, Japan). TRI/3D-BON builds 3D models from serial tomographic datasets for visualization and morphometric analysis [18]. It calculates 3D morphometric parameters of cancellous bone based on ACT scan datasets. CT density analysis is also provided within any volume of interest (VOI). The rectangular VOI was set as large as possible within the internal part of cancellous bone to avoid cracks due to the cutting procedures. The number of voxels of the VOI ranged from 250 to 500 in each direction. The gray-scale images were segmented using a median filter to remove noise and a fixed

434

T. Sone et al. / Bone 35 (2004) 432–438

threshold to extract the mineralized bone phase. Subsequently, the isolated small particles in marrow space and the isolated small holes in bone were removed using a cluster-labeling algorithm. The trabecular bone was then analyzed for structural indices. Bone surface area (BS) was calculated using surface triangulation of the binary data based on the marching cubes method [19]. Bone volume (BV) was calculated using tetrahedrons corresponding to the enclosed volume of the triangulated surface. Total tissue volume (TV) was the entire volume of the analysis. The normalized indices, bone surface to volume ratio (BS/BV), and trabecular bone volume fraction (BV/TV) were then calculated from these values. Trabecular thickness (Tb.Th) was determined according to the method described by Hildebrand and Ru¨egsegger [20]. Trabecular number (Tb.N) was estimated based on the plate model [21]. In addition to the computation of metric parameters, values of nonmetric parameters were calculated to describe the 3D nature of the trabecular bone sample. The fractal dimensions (FD) of the trabecular bone were measured as a representative of complexity using the box-counting method [22] that was developed threedimensionally. The 3D connectivity density (ConnD) was determined for the original ACT data. The ConnD expresses the number of connections per cubic millimeter (1/mm3) and was evaluated using the Euler – Poincare´ formula based on the Euler number (v) as follows [23]: ConnD = (1  v)/TV. In cancellous bone, the Euler number is defined as v = b0  b1 + b2 where b0 is the number of bone particles (traditionally assumed to be 1), b1 is the connectivity, that is, the maximum number of connections that must be broken to split the structure into two parts, and b2 is the number of marrow cavities fully surrounded by bone. The degree of anisotropy (DA) was determined from the ratio between the maximal and minimal radii of the mean intercept length (MIL) ellipsoid [24]. To evaluate the local unevenness on the trabecular bone surface, we introduced a parameter of surface irregularity: SuIr = (BS  BSV)/BS where BSV was the bone surface area after the surface smoothing procedure, which employs a mean filter of 5
5. A high SuIr value indicates an irregular trabecular bone surface, and SuIr is zero in the case of a flat trabecular plate. In addition to trabecular bone microarchitecture, we also quantified the DMB, that is, the mineral density at the level of calcified bone tissue. Due to the properties of synchrotron radiation, the reconstructed images may be interpreted as maps of the 3D distribution of linear absorption coefficient. Cylindrical phantoms containing K2HPO4 solution at various concentrations were imaged by `ıCT as described by Nuzzo et al. [12]. The linear relationship between the K2HPO4 concentrations (0.1 – 0.8 g/cm3) and absorption (1.4 – 5.6 cm1) was verified using experimental measurements. The DBM was calibrated using this relationship. To perform quantitative comparisons and statistical analysis from different DMBs, we

produced gray level histograms from the CT images to show the frequency (given as a percentage of BV) of occurrence of voxels of a certain gray level, and calculated DBM parameters characterizing the total distribution: the weighted mean DBM value (DBMMean), the most frequent DBM value (DBMPeak), and the width of the distribution (DBMWidth). The artifact due to a partial volume effect caused by surface elements on the trabecular bone was corrected by striping a layer of voxels off the bone surfaces when producing histograms. Statistical analysis The differences among three groups were compared using one-way analysis of variance followed by Scheffe´’s post hoc test for intergroup comparisons. Statistical significance was defined as P < 0.05. Statistical analyses were performed the SPSS software, version 11.5 for Windows (SPSS Inc., Chicago, IL, USA).

Results 3D trabecular bone architecture The 3D parameters of trabecular bone architecture are summarized in Table 1. BV/TV, BS, Tb.N, ConnD, and FD in bone metastases were significantly higher than in the normal tissue ( P < 0.001). DA was lower in bone metastases than normal tissue ( P < 0.001). No significant differences in Tb.Th and BS/BV were observed between the two groups. In contrast, Tb.Th in degenerative osteosclerosis was significantly higher than in normal tissue ( P < 0.001), whereas Tb.N was similar to the normal samples. Typical features of Table 1 3D parameters of trabecular bone architecture Normal tissue

Osteoblastic metastasis

Degenerative osteosclerosis

BV/TV 10.30 F 2.01 38.10 F 15.75*** 30.20 F 1.58**,y (%) BS 9.49 F 1.60 34.19 F 8.20*** 17.59 F 2.85 *,y 2 (mm ) 25.92 F 1.74 28.41 F 4.74 16.47 F 3.73***,y BS/BV 1 (mm ) Tb.Th 103.21 F 10.30 89.48 F 15.00 160.58 F 32.18***,y (Am) Tb.N 1.47 F 0.37 4.77 F 1.15*** 2.46 F 0.40y 1 (mm ) ConnD 3.29 F 2.10 119.72 F 46.73*** 7.69 F 5.13y 3 (mm ) FD 2.06 F 0.01 2.13 F 0.01*** 2.05 F 0.01y DA 1.64 F 0.07 1.38 F 0.07*** 1.58 F 0.03y SuIr (%) 0.53 F 0.25 1.41 F 0.30*** 0.60 F 0.19y All data represent mean F SD. * Significantly different from the normal tissue samples at P < 0.05. ** Significantly different from the normal tissue samples at P < 0.01. *** Significantly different from the normal tissue samples at P < 0.001. y Significantly different from the osteoblastic metastases at P < 0.001.

T. Sone et al. / Bone 35 (2004) 432–438

435

Fig. 1. Normal control subject. The soft X ray (A) and 3D volume-rendered SR-ACT (B) images of normal cancellous bone from the L2 vertebral body taken from a 37-year-old man. The 3D image clearly shows both rod-like and plate-like structures of the trabecular bone.

3D trabecular microstructure in each cancellous bone sample are illustrated in Figs. 1– 3. The image from the sample of bone metastases demonstrates a marked increase in variously shaped trabecular bones with irregular surfaces. The image of degenerative osteosclerosis shows thickened trabecular bones with a smooth surface. The irregularity of the trabecular bone surface in osteoblastic metastases was supported by the SuIr result, which was significantly greater in bone metastases than normal tissue or degenerative osteosclerosis (Table 1).

the other two groups (Table 2). No significant difference in DBM between degenerative osteosclerosis and normal tissue was detected. Fig. 5 shows the transaxial slices through the reconstructed 3D images illustrated in Figs. 1 –3. Differences in gray levels are evident on the slices of all three samples. The distribution of less mineralized bone in the metastatic sample as well as the small holes suggesting bone packets in the normal sample are clearly visible.

3D mineral density measurements

Discussion

We compared the DMB of cancellous bone samples among the three groups. The difference among the DMBs of osteoblastic metastases, degenerative osteosclerosis, and normal tissue is shown in Fig. 4. DMBMean and DMBPeak were significantly lower ( P < 0.01) and DMBWidth was significantly higher ( P < 0.001) in bone metastases than

We three-dimensionally analyzed the microarchitecture and degree of mineralization in osteosclerotic lesions. The salient findings of this study are as follows. First, 3D trabecular microarchitecture is quite different between osteosclerotic lesions of bone metastases and benign degenerative changes. Second, the degree of mineralization of

Fig. 2. Benign degenerative osteosclerosis. The soft X ray (A) and 3D volume-rendered SR-ACT (B) images of degenerative osteosclerotic cancellous bone from the L5 vertebral body taken from a 37-year-old man. In the 3D image, regularly thickened trabecular bones are demonstrated.

436

T. Sone et al. / Bone 35 (2004) 432–438

Fig. 3. Osteoblastic metastasis. The soft X ray (A) and 3D volume-rendered SR-ACT (B) images of metastatic cancellous bone from the L3 vertebral body, taken from a 67-year-old man with bone metastases from prostate cancer. The Tb.N was greater than that in normal tissue, whereas no difference in Tb.Th was observed. The 3D image shows an increased number of trabecular bones with an irregular surface.

osteoblastic metastases is lower than that of degenerative osteosclerotic lesions or normal tissue. The osteosclerotic changes detected by conventional radiographs in osteoblastic metastases tend to assume the thickening of trabecular bones [25,26], but to our knowledge, there has been no study on the 3D evaluation of trabecular microarchitecture in osteoblastic metastases. The 3D analysis in this study indicated that the bone volume increased in both benign osteosclerotic lesions and osteoblastic metastases, observed as strong osteosclerotic changes on X-rays, more than in normal regions, but there were differences in the trabecular microarchitecture between

Fig. 4. The frequency distributions of the degree of mineralization in partial volume-corrected 3D image of cancellous bone samples illustrated in Figs. 1 – 3: (A) normal control subject, (B) benign degenerative osteosclerosis; (C) osteoblastic metastasis.

benign and metastatic osteosclerosis. With increases in the bone volume, Tb.N increased without changes in Tb.Th in the osteoblastic metastases, while in the benign osteosclerotic lesions, Tb.Th increased without changes in Tb.N. The samples from the osteoblastic metastases showed higher values of FD than the normal tissue. Trabecular bone is not an entirely fractal structure [27,28], but FD may be used as a measure of the ‘‘complexity’’ of the trabecular network for a given range of spatial resolutions [22]. The increase in FD in osteoblastic metastases implies their complex trabecular network. The surface irregularity of trabecular bones assessed by image processing methods was higher in the osteoblastic metastases than the normal tissue or benign osteosclerotic lesions. The local irregularity of the trabecular bone surface can be increased with accelerated bone resorption. The BS/BV, another common parameter in characterizing the complexity of trabecular structure, did not show any significant difference between osteoblastic metastases and normal tissue, whereas BS/BV was lower in benign osteosclerotic lesions. Since BS/BV is also dependent on the trabecular structure, such as plates and rods, the characteristics identified by BS/BV differ from fractal complexity and surface irregularity. A parameter of trabecular connections, ConnD, also increased in the osteoblastic metastases. This increase in ConnD can be caused by fenestration of the trabecular plate through bone resorption as well as by trabecular bridging through de novo bone formation [29]. Taken together, these findings indicate that the numerous trabecular bones were newly formed in the osteoblastic metastases, whereas the layer of new bone was formed on the preexisting trabeculae in the benign osteosclerotic lesions. Furthermore, the increase in surface irregularity and connectivity of trabeculae suggests that the strong effects of bone resorption occurred in the osteoblastic metastases. Another parameter of trabecular microarchitecture, structural anisotropy, was evaluated by the mean intercept length

T. Sone et al. / Bone 35 (2004) 432–438 Table 2 Comparison of DMB among cancellous bone samples

DMBPeak DMBMean DMBWidth

Normal tissue

Osteoblastic metastasis

Degenerative osteosclerosis

0.944 F 0.044 0.909 F 0.046 0.192 F 0.028

0.840 F 0.052* 0.799 F 0.055* 0.291 F 0.041**

0.959 F 0.045y 0.926 F 0.043y 0.183 F 0.024yy

DMB is calibrated to the concentrations (g/cm3) of K2HPO4 solution. All data represent mean F SD. * Significantly different from the normal tissue samples at P < 0.01. ** Significantly different from the normal tissue samples at P < 0.001. y Significantly different from the osteoblastic metastases at P < 0.01. yy Significantly different from the osteoblastic metastases at P < 0.001.

method. The ratio between the maximum-to-minimum radii of the MIL ellipsoid was significantly lower in the osteoblastic metastases than in the normal tissue or benign degenerative osteosclerosis, indicating that the trabecular bone is more isotropically arranged in osteoblastic metastases. This reduced anisotropy implies that the bone formation in osteoblastic metastases is less affected by mechanical stress, such as weight bearing, than in normal tissue. Increases in osteoid surface and osteoid volume have been demonstrated in bone biopsies displaying osteoblastic metastases from prostate cancer [7]. These increases in osteoid deposition are associated with an elevation in the mineral apposition rate, demonstrating the production of mineralized new bone. In the accelerated state of bone formation, the osteoid is not fully mineralized, and woven bone is formed, which has a lower level of mineral density than the more organized form of lamellar bone. In the present study, the degree of mineralization in cancellous bone was compared among osteoblastic metastases, benign degenerative osteosclerosis, and normal tissue. The mean and peak values of DMB were decreased and the width of the distribution was increased in osteoblastic metastases compared with the normal tissue. The broadening of the distribution of DBM indicates a higher percentage of newly formed, but not fully mineralized bone, suggesting accelerated bone formation. In contrast, benign osteosclerotic lesions exhibited no significant difference in the distribution of DBM compared with the normal tissue. The DBM depends on how recently

437

the bone tissue was formed and on the extent of remodeling activity [30,31]. The change in bone turnover due to therapeutic intervention can result in a change in DBM. For instance, an increase or decrease in DBM has been demonstrated in osteoporotic patients treated with bisphosphonate [32] or parathyroid hormone [33], respectively. The results of the present study suggest that the bone turnover and time course of the mineralization process are similar between benign osteosclerotic lesions and normal tissue. A previous study by Grynpas et al. [34] revealed that osteoarthritis is associated with a thickening of the subchondral bone with abnormally low mineralization. This discrepancy may be attributed to the difference in skeletal site for the determination of DMB, that is, the subchondral bone in femoral head versus cancellous bone in the vertebral body. Bone architecture and mineralization are generally considered to be important components of bone quality, and determine bone strength in conjunction with bone mineral density [35]. Although the relationship between bone quality and bone strength has recently been studied under conditions in which bone density decreases [36], such as osteoporosis, little is known about this relationship in osteosclerotic diseases. The present findings characterize bone architecture and mineralization in osteosclerotic lesions. Further study on the relationship between bone quality and bone strength in osteosclerotic lesions would improve our understanding of the pathogenesis of bone fragility. In conclusion, we have analyzed and compared the 3D microarchitecture and DMB of osteoblastic metastases from prostate cancer and benign degenerative osteosclerotic lesions. The trabecular thickness in degenerative osteosclerosis was greater than in normal tissue; however, no significant difference was found between osteoblastic metastases and normal tissue. Trabecular bone in metastatic lesions showed an increase in connectivity and surface irregularity, and a decrease in DMB. These results suggest that the strong effects of bone resorption, as well as bone formation, occur in osteoblastic bone metastases. Further study is required to elucidate whether these characteristics in trabecular microarchitecture and mineral den-

Fig. 5. 2D tomograms of vertebral cancellous bone samples illustrated in Figs. 1 – 3: (A) normal control subject, (B) benign degenerative osteosclerosis; (C) osteoblastic metastasis.

438

T. Sone et al. / Bone 35 (2004) 432–438

sity change according to therapeutic intervention such as bisphosphonates.

Acknowledgments The synchrotron radiation experiments were performed at the Spring-8 with the approval of the Japan Synchrotron Radiation Research Institute (JASRI) (Proposal No. 2001B0240-NL-np). We thank Dr. H. Nagatsuka and Dr. N. Nagai for their assistance in preparing samples, and Dr. K. Uesugi for his help in microtomography.

[15]

[16]

[17]

[18]

[19]

References [1] Goltzman D. Mechanisms of the development of osteoblastic metastases. Cancer 1997;80:1581 – 7. [2] Urwin GH, Percival RC, Harris S, Beneton MN, Williams JL, Kanis JA. Generalised increase in bone resorption in carcinoma of the prostate. Br J Urol 1985;57:721 – 3. [3] Clarke NW, McClure J, George NJ. Morphometric evidence for bone resorption and replacement in prostate cancer. Br J Urol 1991;68: 74 – 80. [4] Ikeda I, Miura T, Kondo I. Pyridinium cross-links as urinary markers of bone metastases in patients with prostate cancer. Br J Urol 1996;77: 102 – 6. [5] Takeuchi S, Arai K, Saitoh H, Yoshida K, Miura M. Urinary pyridinoline and deoxypyridinoline as potential markers of bone metastasis in patients with prostate cancer. J Urol 1996;156:1691 – 5. [6] Nakashima J, Sumitomo M, Miyajima A, Jitsukawa S, Saito S, Tachibana M, et al. The value of serum carboxyterminal propeptide of type 1 procollagen in predicting bone metastases in prostate cancer. J Urol 1997;157:1736 – 9. [7] Clarke NW, McClure J, George NJ. Osteoblast function and osteomalacia in metastatic prostate cancer. Eur Urol 1993;24:286 – 90. [8] Odgaard A. Three-dimensional methods for quantification of cancellous bone architecture. Bone 1997;20:315 – 28. [9] Engelke K, Graeff W, Meiss L, Hahn M, Delling G. High spatial resolution imaging of bone mineral using computed microtomography. Comparison with microradiography and undecalcified histologic sections. Invest Radiol 1993;28:341 – 9. [10] Bonse U, Busch F, Gunnewig O, Beckmann F, Pahl R, Delling G, et al. 3D computed X-ray tomography of human cancellous bone at 8 microns spatial and 10(4) energy resolution. Bone Miner 1994; 25:25 – 38. [11] Peyrin F, Salome M, Nuzzo S, Cloetens P, Laval-Jeantet AM, Baruchel J. Perspectives in three-dimensional analysis of bone samples using synchrotron radiation microtomography. Cell Mol Biol (Noisyle-grand) 2000;46:1089 – 102. [12] Nuzzo S, Peyrin F, Cloetens P, Baruchel J, Boivin G. Quantification of the degree of mineralization of bone in three dimensions using synchrotron radiation microtomography. Med Phys 2002;29: 2672 – 81. [13] Luo G, Kinney JH, Kaufman JJ, Haupt D, Chiabrera A, Siffert RS. Relationship between plain radiographic patterns and three-dimensional trabecular architecture in the human calcaneus. Osteoporos Int 1999;9:339 – 45. [14] Nuzzo S, Lafage-Proust MH, Martin-Badosa E, Boivin G, Thomas T, Alexandre C, et al. Synchrotron radiation microtomography allows the analysis of three-dimensional microarchitecture and degree of mineralization of human iliac crest biopsy specimens:

[20]

[21]

[22] [23]

[24]

[25]

[26] [27]

[28] [29]

[30]

[31]

[32]

[33]

[34] [35] [36]

effects of etidronate treatment. J Bone Miner Res 2002;17: 1372 – 82. Kinney JH, Ryaby JT, Haupt DL, Lane NE. Three-dimensional in vivo morphometry of trabecular bone in the OVX rat model of osteoporosis. Technol Health Care 1998;6:339 – 50. Kinney JH, Haupt DL, Balooch M, Ladd AJ, Ryaby JT, Lane NE. Three-dimensional morphometry of the L6 vertebra in the ovariectomized rat model of osteoporosis: biomechanical implications. J Bone Miner Res 2000;15:1981 – 91. Uesugi K, Tsuchiyama A, Nakano T, Suzuki Y, Yagi N, Umetani K, et al. Development of microtomography imaging system for rock and mineral samples. Proc SPIE 1999;3772:214 – 21. Joo YI, Sone T, Fukunaga M, Lim SG, Onodera S. Effects of endurance exercise on three-dimensional trabecular bone microarchitecture in young growing rats. Bone 2003;33:485 – 93. Lorensen WE, Cline HE. Marching cubes: a high resolution 3D surface construction algorithm. Comput Graph 1987;21:163 – 9. Hildebrand T, Ru¨gsegger P. A new method for the model-independent assessment of thickness in three-dimensional images. J Microsc 1997;185:67 – 75. Parfitt AM, Drezner MK, Glorieux FH, Kanis JA, Malluche H, Meunier PJ, et al. Bone histomorphometry: standardization of nomenclature, symbols, and units. Report of the ASBMR Histomorphometry Nomenclature Committee. J Bone Miner Res 1987;2:595 – 610. Fazzalari NL, Parkinson IH. Fractal dimension and architecture of trabecular bone. J Pathol 1996;178:100 – 5. Odgaard A, Gundersen HJ. Quantification of connectivity in cancellous bone, with special emphasis on 3-D reconstructions. Bone 1993;14:173 – 82. Harrigan TP, Mann RW. Characterization of microstructural anisotropy in orthotropic materials using a second rank tensor. J Mater Sci 1984;19:761 – 7. Aoki J, Yamamoto I, Hino M, Shigeno C, Kitamura N, Itoh H, et al. Sclerotic bone metastasis: radiologic – pathologic correlation. Radiology 1986;159:127 – 32. Hudson TM. Radiologic – pathologic correlation of musculoskeletal lesions. Baltimore: Williams and Wilkins; 1987. Cross SS, Rogers S, Silcocks PB, Cotton DW. Trabecular bone does not have a fractal structure on light microscopic examination. J Pathol 1993;170:311 – 3. Chung HW, Chu CC, Underweiser M, Wehrli FW. On the fractal nature of trabecular structure. Med Phys 1994;21:1535 – 40. Kinney JH, Ladd AJ. The relationship between three-dimensional connectivity and the elastic properties of trabecular bone. J Bone Miner Res 1998;13:839 – 45. Boivin G, Meunier PJ. Changes in bone remodeling rate influence the degree of mineralization of bone. Connect Tissue Res 2002;43: 535 – 7. Boivin G, Meunier PJ. The degree of mineralization of bone tissue measured by computerized quantitative contact microradiography. Calcif Tissue Int 2002;70:503 – 11. Boivin GY, Chavassieux PM, Santora AC, Yates J, Meunier PJ. Alendronate increases bone strength by increasing the mean degree of mineralization of bone tissue in osteoporotic women. Bone 2000;27:687 – 94. Misof BM, Roschger P, Cosman F, Kurland ES, Tesch W, Messmer P, et al. Effects of intermittent parathyroid hormone administration on bone mineralization density in iliac crest biopsies from patients with osteoporosis: a paired study before and after treatment. J Clin Endocrinol Metab 2003;88:1150 – 6. Grynpas MD, Alpert B, Katz I, Lieberman I, Pritzker KP. Subchondral bone in osteoarthritis. Calcif Tissue Int 1991;49:20 – 6. NIH Consensus Development Panel. Osteoporosis prevention, diagnosis, and therapy. JAMA 2001;285:785 – 95. Seeman E. Bone quality. Osteoporos Int 2003;14(Suppl. 5):3 – 7.