- Email: [email protected]
Use of Axial X-Ray Microcomputed Tomography to Assess Three-Dimensional Trabecular Microarchitecture and Bone Mineral Density in Single Comb White Leghorn Hens
Use of Axial X-Ray Microcomputed Tomography to Assess Three-Dimensional Trabecular Microarchitecture and Bone Mineral Density in Single Comb White Leghorn Hens M. A. Martı´nez-Cummer,* R. Heck,† and S. Leeson*1 *Department of Animal and Poultry Science and †Department of Land Resource Science, University of Guelph, Guelph, Ontario, Canada N1G 2W1
Key words: bone density, laying hen, tomography, computed tomagraphy image 2006 Poultry Science 85:706–711
becular structure) will lead to a more comprehensive and accurate determination of skeletal integrity (Judex et al., 2003). Dual energy x-ray apsorptiometry (DEXA) is a widely accepted technique for the diagnosis of osteopenia and osteoporosis. However, this technique relies on 2-D estimates for bone mineral assessment (Schreiweis et al., 2003; Hester et al., 2004). In addition, the poor resolution of this technique prevents a detailed analysis of bone architecture by excluding key information on structural morphology (i.e., the role of trabecular architecture) and potentially missing key site-specific changes within a given region of interest. Korver et al. (2004) evaluated quantitative tomography and examined bone mineral density (BMD; mg/cc) in different bone compartments based on single slice tomography analysis in live laying hens. Based on single slice computed tomography, these researchers were able to distinguish between cortical bone and cancellous bone. As expected, they found a significantly higher BMD (mg/cc) in cortical bone vs. cancellous bone. This research group was also able to detect changes in cancellous bone in individual hens over time. They did not, however, detect changes in cortical bone over time. Although Korver et al. (2004) were unable to discern between trabecular bone and medullary bone, they detected changes in trabecular bone that might be ex-
INTRODUCTION Osteoporosis is a systemic skeletal disease characterized by low bone mass (Whitehead and Fleming, 2000) and microarchitectural deterioration of bone tissue (Ru¨egsegger et al., 1996). Osteoporosis leads to increased bone resorption, and with time, trabecular plates can perforate and connecting rods can dissolve. Therefore, direct 3-D assessment is necessary to quantitate cancellous bone degeneration. Axial x-ray microcomputed tomography (microCT; model MS, General Electric Medical Systems, London, ON, Canada) is especially suited for applications involving measurements of bone density because of the high signal contrast between bone and soft tissue (Holdsworth and Thornton, 2002). In addition to bone mass, the trabecular structure is an important factor contributing to bone strength. Three-dimensional image reconstructions permit the assessment of trabecular bone architecture. An overall characterization of both bone quantity (bone volume/total volume; BV/TV) and quality (tra-
2006 Poultry Science Association, Inc. Received May 27, 2005. Accepted November 6. 2005. 1 Corresponding author: [email protected]
706
Downloaded from http://ps.oxfordjournals.org/ by guest on March 21, 2015
tions (0.5°) around the object of interest. These groups of views were then used to reconstruct a computed tomography image. A test grid with orthogonal test lines was used to calculate bone volume and bone surface. From these calculations, parallel plate equations were used to derive trabecular architectural parameters such as average trabecular plate thickness and average trabecular plate separation. Three-dimensional microarchitecture was evaluated using specialized stereological analysis software. Significant relationships between apparent bone mineral density (mg/cc) and 3-D structure were observed in femoral specimens from 66-wk-old Single Comb White Leghorn hens.
ABSTRACT Axial x-ray microcomputed tomography is a cost-effective technique with the potential to assess bone mineral density (mg/cc) in both cortical and cancellous bone in Single Comb White Leghorn hens. The technique requires little sample preparation and involves relatively simple data processing. The system described in this research is based on compact fan-beam type tomography, using a tungsten-anode x-ray tube with a relatively small focal spot (∼5 m), coupled with a high-resolution x-ray detector system (∼10 m). To produce a real 3-D data set using microcomputed tomography, x-ray projection views were acquired at 720 equally spaced angular posi-
USE OF AXIAL X-RAY MICROCOMPUTED TOMOGRAPHY
707
plained by medullary bone modeling and remodeling. This paper provides information on the application of microCT to monitor skeletal integrity in laying hens.
MATERIALS AND METHODS Sample Preparation For most applications, the only preparation needed for microCT scans is to ensure that the object fits into the field of view and to immobilize the object during the scanning process. Because the complete scan field for CT is a cylinder, the most effective geometry to scan is also a cylinder (Ketcham and Carlson, 2001). Typically, this is accommodated by packing the object in a cylindrical container with either x-ray transparent filler or with material of similar density. Birds were cared for according to guidelines established by the Canadian Council on Animal Care. Bone mineral density (mg/cc), trabecular BMD (mg/cc), and trabecular architecture parameters of the diaphysis of the right femora harvested at 66 wk of age were measured using microCT. Fresh femoral samples were cleaned of excess flesh, and the total length was recorded using vernier calipers (±0.05 mm). An examination of all anatomical sites was done on harvested and cleaned femurs. The lengths of femoral specimens were measured from the distal (Facies articularis antitrochanterica) to the proximal end (Condylus lateralis). Lines were
then drawn on the bone at 40, 50, and 60% of the distance from the proximal to the distal end of the bone. Bone lying between the 40 and 60% markers was used for subsequent studies. Using a microtome with a diamond-encrusted blade, the excess portions of the proximal and distal regions were removed to obtain anatomically comparable samples and to reduce variability associated with differences in bone length. The resulting femoral diaphyses were used to estimate apparent femoral density by calculating the ratio of total tissue mass to specimen volume; this was achieved by using an electronic top-loader scale (Satorius Laboratory Plus L420 P+, Sartorius GmbH Goettingen, West Germany), and bones were weighed to the nearest 1 mg. Bone fragments were suspended by a thin stainless-steel wire and weighed in air. Once the dry weight was recorded, the specimen’s proximal and distal ends were sealed using a fine layer of petroleum jelly and then fully immersed in a beaker with 15 cc of double deionized water. Caution was taken to ensure that the bone sample did not touch the sides of the beaker. The weight in water was read immediately after immersion. Density (mg/cc) was calculated by dividing the weight in air by the weight in water. After the apparent density was measured, bone fragments were immediately preserved in 10% formalin and rinsed in saline solution prior to scanning. Preliminary tomographic scans revealed that distinction between cortical and trabecular bone was not clear
Downloaded from http://ps.oxfordjournals.org/ by guest on March 21, 2015
Figure 1. Specimen (as shown in Figure 2) with a computer-generated isosurface.
708
MARTI´NEZ-CUMMER ET AL.
Downloaded from http://ps.oxfordjournals.org/ by guest on March 21, 2015
Figure 2. Adequate cancellous bone structure in femoral diaphysis medullary cavity at 66 wk of age. a) Three-dimensional volume; b) cross-sectional view of femoral diaphysis with a 3-D selected volume of interest; c) 3-D volume of interest (trabecular bone).
Figure 3. Depleted trabecular bone structure in femoral diaphysis medullary cavity at 66 wk of age. a) Three-dimensional volume; b) cross-sectional view of femoral diaphysis with a 3-D selected volume of interest; c) 3-D volume of interest (trabecular bone).
USE OF AXIAL X-RAY MICROCOMPUTED TOMOGRAPHY
709
in the medullary cavity. A range of skeletal structures was imposed by feeding birds 2.5, 3.5, or 4.5% calcium from 18 to 24 wk of age together with adequate (0.48%) or inadequate (0.28%) levels of available phosphorus. From 24 to 66 wk of age, all birds were fed 4% calcium and 0.38% available phosphorus.
MicroCT Data Correction
RESULTS
Figure 4. Relationship between femoral diaphysis apparent bone mineral density (mg/cc) and a) femoral medullary bone density (mg/ cc); b) femoral medullary trabecular plate thickness (mm); c) femoral medullary trabecular plate separation (mm); and d) medullary trabecular bone density (%).
in the metaphyseal shell of the proximal femur. Thus, an additional series of scans was conducted to assess the architectural makeup in the diaphyseal region and to obtain specimens with a distinct contrast between cortical and trabecular bone. This situation is imperative to be able to accurately quantitate BMD in both cortical bone located in the diaphyseal shell and trabecular bone
The end result of a complete microCT scan is a 3-D volume array of relative attenuation values (Figure 1). For bone specimen evaluations, 2-D slices along with Coronal (X), Sagital (Z), and Axial (Y) orthogonal planes allow for interpretation of skeletal structures (Figures 2 and 3). Figure 2 depicts adequate medullary bone from 66-wk-old layers; Figure 3 shows comparable orthogonal views for birds with depleted bone reserves. Sample calculations obtained from microCT scans are shown in Table 1. The descriptive statistics observed in these tables are divided into parameters that were either calculated [BMD, BV/TV, bone surface density (BS/BV)] or derived [trabecular plate thickness (TBTh), trabecular plate number (TbN), trabecular plate separation (TbSp)]. Traditional histomorphometric methods were used with bone surface and bone volume ratios and assumed a parallel plate model. The 3-D-calculated (model independent) stereology parameters were calculated using a test grid with orthogonal test lines (oriented along the scan axes) and measured 2 parameters: 1) Pp = PV/TV = number of voxels greater than the threshold divided by the total number of voxels in the volume of interest and 2) P1 = number of intersections of the test lines with the bone-marrow interface divided by the total line length. These test grids were measured for each of the 3 axes described as X, Y, and Z. Equations 1 and 2 were used to calculate BV/TV, which is the ratio between bone volume within a sampled volume of tissue, and BS/BV, which is the relationship between
Downloaded from http://ps.oxfordjournals.org/ by guest on March 21, 2015
It is imperative to establish the characteristics of the x-ray signal as read by the detector under scanning conditions and to reduce geometrical uncertainties of computed tomography images. Therefore, a phantom (Model SB#, Gammex rmi, Middleton, WI) constructed of a clear plastic material designed to hold air, water, and a sample of hydroxyapatite of bone mineral equivalent was utilized. This phantom was used in every scan to calculate a standard calibration curve from which both cortical and cancellous bone density could be interpolated. Once the data collection and imaging correction were completed, values obtained from the phantom were keyed in, and acquisition electronics that digitize the detector signals were sent to a computer for reconstruction. Reconstruction is the mathematical process of converting the single sets of detector readings for all 720 views into 2-D slice images, i.e., 3-D volume array (Ketcham and Carlson, 2001).
MARTI´NEZ-CUMMER ET AL.
710
Table 1. Example of quantitative bone morphometry data obtained from microcomputed tomography scan1,2 BV/TV3
BS/BV4
TbTh5
TbN6
TbSp7
Mean Median Minimum Maximum SD CV
84.85 84.98 74.45 93.84 3.83 4.52
21.44 21.44 9.84 34.56 4.66 21.48
0.0984 0.0933 0.0579 0.2032 0.0253 25.74
9.01 9.12 4.62 12.86 1.57 17.47
0.0166 0.0164 0.0133 0.0210 0.0017 10.09
Results averaged over the 3 directions Mean
84.85
21.30
0.1003
8.93
0.0167
1
3
bone surface present within a volume of tissue examined. Model-dependent parameters such as average TbTh (mm), average TbN, and average TbSp (mm) were derived from BV/TV and BS/BV using Equations 3, 4, and 5, respectively. BV/TV (%) = Pp
[1]
BS/BV (mm2/mm3) = 2 × Pl/Pp
[2]
TbTh (mm) = Pp/Pl = 2/(BS/BV)
[3]
TbN (1/mm) = Pl = (BV/TV)/TbTh
[4]
TbSp (mm) = (1 − Pp)/Pl = (1/TbN) − TbTh
[5]
Figure 4 shows the relationship between femoral BMD as measured by microCT and various other bone parameters. The R2 values are generally high (0.69 to 0.94) with relationships being highly significant (P < 0.01). Trabecular BMD (mg/cc) was found to have similar relationships with average TbTh and average TbSp. As expected, medullary cavity trabecular BMD (mg/cc) and medullary cavity trabecular BV/TV were highly correlated because both of these parameters quantitatively reflect the amount of bone present within an area of interest. These results suggest that femoral diaphysis apparent density (mg/cc) is a function of trabecular BMD (mg/cc), BV/TV (%), average TbTh (mm), and average TbSp (mm).
DISCUSSION Microcomputed tomography was initially developed because of the need for a highly precise and accurate tool for high-resolution reconstruction of the complex architecture of bone tissue (Mu¨ller and Ru¨egsegger,
1997; Hildebrand et al., 1999). Recently, Schreiweis et al. (2003) and Hester et al. (2004) reported DEXA as a tool to monitor changes in skeletal density in commercial Single Comb White Leghorn hens. For several reasons, including low cost, radiation dose, simplicity, and the ability to image skeletal sites, this methodology is used extensively and is considered the standard for the diagnosis of osteopenia and osteoporosis in humans. Hester et al. (2004) reported densitometry assessment in avian tibia using DEXA. Based on the results reported by these scientists, it is evident that DEXA provides integral measurements of trabecular and cortical bone. The precision of DEXA is very high, and results correlate well with traditional destructive tests (i.e., bone breaking force, bone ash); however, its accuracy in assessing bone volumetric density is greatly reduced because of the limitations of 2-D projection. For instance, DEXA is not able to distinguish bone density in different compartments such as compact vs. cancellous bone. Therefore, 3-D imaging techniques, such as axial x-ray microCT as reported in this study, are needed because they offer the ability to separately examine factors that may play independent and important roles in osteoporosis. Microcomputed tomography does allow for measurement of BMD in different bone compartments as well as the determination of 3-D morphometric indices in cancellous bone. The method can provide a more comprehensive analysis based on changes occurring in different bone types in very small regions of interest. Results from our study demonstrate a substantial improvement in overall femoral BMD determination because microCT is able to quantitate bone density in different compartments of the femur and to provide qualitative and quantitative data of the architectural status of bone. Microcomputed tomography used in this research project does have a practical limitation in that all data collected were in postmortem femoral diaphysis, and as
Downloaded from http://ps.oxfordjournals.org/ by guest on March 21, 2015
Total points used: 22,781,250; total volume of examined region: 122.09 mm ; total number of voxels occupied by bone: 1,932,959. 2 Results are averages summed directly over 3 directions. 3 Bone volume to total volume (%; calculated parameter). 4 Bone surface to total volume (mm2/mm3; calculated parameter). 5 Trabecular thickness [mm; model-derived parameter: TbTh (mm) = Pp/Pl = 2/(BS/BV), where Pp = PV/ TV = number of voxels greater than the threshold divided by the total number of voxels in the volume of interest and P1 = number of intersections of the test lines with the bone-marrow interface divided by the total line length]. 6 Trabecular number [1/mm; derived parameter: TbN (1/mm) = P1 = (BV/TV)/TbTh]. 7 Trabecular separation [mm; model-derived parameter: TbSp (mm) = (1 − Pp)/P1 = (1/TbN) − TbTh).
USE OF AXIAL X-RAY MICROCOMPUTED TOMOGRAPHY
REFERENCES Gasser, J. A., and J. Yam. 2002. In vivo non-invasive monitoring of changes in structural cancellous bone parameters with a novel prototype MicroCT in rats. Presented at the 24th Annu. Mtg. Am. Soc. Bone Mineral Res., San Antonio, TX. Hester, P. Y., M. A. Schreiweis, J. L. Orban, H. Mazzuco, M. N. Kopka, M. C. Ledur, and D. E. Moody. 2004. Assessing bone mineral density in vivo: Dual energy x-ray absorptiometry. Poult. Sci. 83:215–221. Hildebrand, T., A. Laib, and R. Muller. 1999. Direct threedimensional morphometric analysis of human cancellous bone: Micro structural data from spine, femur, iliac crest, and calcaneus. J. Bone Miner. Res. 14:1167–1174. Holdsworth, D. W., and M. M. Thornton. 2002. Micro-CT in small animal and specimen imaging. Trends Biotechnol. 20:S34–S39. Judex, S., S. Boyd, Y.-X. Qin, L. Miller, R. Mu¨ller, and C. Rubin. 2003. Combining high-resolution micro-computed tomography with material composition to define the quality of bone tissue. Curr. Osteoporosis Rep. 1:11–19. Ketcham, R. A., and W. D. Carlson. 2001. Acquisition, optimization and interpretation of X-ray computed tomographic imagery: Applications to the geosciences. Comput. Geosci. 27:381–400. Korver, D. R., J. L. Saunders-Blades, and K. L. Nadeau. 2004. Assessing bone mineral density in vivo. Poult. Sci. 83:222–229. Mu¨ller, R., and P. Ru¨egsegger. 1997. Microtomographic imaging for the non-destructive evaluation of trabecular bone architecture. Pages 61–80 in Bone Research in Biomechanics. G. Lowet, P. Ru¨egsegger, H. Weinans, and A. Meunier, ed. IOS Press, Amsterdam, The Netherlands. Ru¨egsegger, P., B. Koller, and R. Mu¨ller. 1996. A microtomographic system for the non-destructive evaluation of bone architecture. Calcif. Tissue Int. 58:24–29. Schreiweis, M. A., J. I. Orban, M. C. Ledur, and P. Y. Hester. 2003. The use of densitometry to detect differences in bone mineral density and content of live White Leghorns fed varying levels of dietary calcium. Poult. Sci. 82:1292–1301. Whitehead, C. C., and R. H. Fleming. 2000. Osteoporosis in cage layers. Poult. Sci. 79:1033–1041.
Downloaded from http://ps.oxfordjournals.org/ by guest on March 21, 2015
a consequence, evaluations over time were not possible. However, a number of scanners are now available that are able to scan small animals at high resolutions and scanning speeds (Gasser and Yam, 2002). For example, 100 sections at a 20- resolution are possible for imaging trabecular bone in live mice in <5 min. The size of current scanning chambers restricts high-resolution scanning to small live animals such as rats and mice. The development of units with scanning chambers large enough to fit a live laying hen is currently underway. These units will allow the assessment of the relative importance of bone architecture and bone mineralization in the characterization of bone quality as a predictor for bone strength in multiple sites. These capabilities are likely to enhance our understanding of the effect of density, morphological, and loading factors in the progression of osteoporosis in Single Comb White Leghorn hens. This enhanced understanding might lead to improved approaches that could significantly reduce the high incidence of fractures currently observed in commercial operations. Nonetheless, future research is likely to encounter inherent complexities in developing an integrative assessment of bone integrity. A complete assessment of bone quality will require the consideration of factors that are seldom considered because of the difficulty or inability to quantify such parameters. For instance, microdamage in bone accumulated by repetitive loading can lead to serious mechanical and biological consequences (Korver et al., 2004). With advances in 3-D bone characterization, the emphasis is now to study functional environments that induce microdamage and the effect of any preventive mechanisms (Judex et al., 2003). These data confirm that microCT is able to quickly and precisely determine bone architecture in the femur of laying hens. Over time, such measurements will be possible on live birds, a situation that will greatly add to our understanding of factors affecting osteoporosis in laying hens.
711
Key words: bone density, laying hen, tomography, computed tomagraphy image 2006 Poultry Science 85:706–711
becular structure) will lead to a more comprehensive and accurate determination of skeletal integrity (Judex et al., 2003). Dual energy x-ray apsorptiometry (DEXA) is a widely accepted technique for the diagnosis of osteopenia and osteoporosis. However, this technique relies on 2-D estimates for bone mineral assessment (Schreiweis et al., 2003; Hester et al., 2004). In addition, the poor resolution of this technique prevents a detailed analysis of bone architecture by excluding key information on structural morphology (i.e., the role of trabecular architecture) and potentially missing key site-specific changes within a given region of interest. Korver et al. (2004) evaluated quantitative tomography and examined bone mineral density (BMD; mg/cc) in different bone compartments based on single slice tomography analysis in live laying hens. Based on single slice computed tomography, these researchers were able to distinguish between cortical bone and cancellous bone. As expected, they found a significantly higher BMD (mg/cc) in cortical bone vs. cancellous bone. This research group was also able to detect changes in cancellous bone in individual hens over time. They did not, however, detect changes in cortical bone over time. Although Korver et al. (2004) were unable to discern between trabecular bone and medullary bone, they detected changes in trabecular bone that might be ex-
INTRODUCTION Osteoporosis is a systemic skeletal disease characterized by low bone mass (Whitehead and Fleming, 2000) and microarchitectural deterioration of bone tissue (Ru¨egsegger et al., 1996). Osteoporosis leads to increased bone resorption, and with time, trabecular plates can perforate and connecting rods can dissolve. Therefore, direct 3-D assessment is necessary to quantitate cancellous bone degeneration. Axial x-ray microcomputed tomography (microCT; model MS, General Electric Medical Systems, London, ON, Canada) is especially suited for applications involving measurements of bone density because of the high signal contrast between bone and soft tissue (Holdsworth and Thornton, 2002). In addition to bone mass, the trabecular structure is an important factor contributing to bone strength. Three-dimensional image reconstructions permit the assessment of trabecular bone architecture. An overall characterization of both bone quantity (bone volume/total volume; BV/TV) and quality (tra-
2006 Poultry Science Association, Inc. Received May 27, 2005. Accepted November 6. 2005. 1 Corresponding author: [email protected]
706
Downloaded from http://ps.oxfordjournals.org/ by guest on March 21, 2015
tions (0.5°) around the object of interest. These groups of views were then used to reconstruct a computed tomography image. A test grid with orthogonal test lines was used to calculate bone volume and bone surface. From these calculations, parallel plate equations were used to derive trabecular architectural parameters such as average trabecular plate thickness and average trabecular plate separation. Three-dimensional microarchitecture was evaluated using specialized stereological analysis software. Significant relationships between apparent bone mineral density (mg/cc) and 3-D structure were observed in femoral specimens from 66-wk-old Single Comb White Leghorn hens.
ABSTRACT Axial x-ray microcomputed tomography is a cost-effective technique with the potential to assess bone mineral density (mg/cc) in both cortical and cancellous bone in Single Comb White Leghorn hens. The technique requires little sample preparation and involves relatively simple data processing. The system described in this research is based on compact fan-beam type tomography, using a tungsten-anode x-ray tube with a relatively small focal spot (∼5 m), coupled with a high-resolution x-ray detector system (∼10 m). To produce a real 3-D data set using microcomputed tomography, x-ray projection views were acquired at 720 equally spaced angular posi-
USE OF AXIAL X-RAY MICROCOMPUTED TOMOGRAPHY
707
plained by medullary bone modeling and remodeling. This paper provides information on the application of microCT to monitor skeletal integrity in laying hens.
MATERIALS AND METHODS Sample Preparation For most applications, the only preparation needed for microCT scans is to ensure that the object fits into the field of view and to immobilize the object during the scanning process. Because the complete scan field for CT is a cylinder, the most effective geometry to scan is also a cylinder (Ketcham and Carlson, 2001). Typically, this is accommodated by packing the object in a cylindrical container with either x-ray transparent filler or with material of similar density. Birds were cared for according to guidelines established by the Canadian Council on Animal Care. Bone mineral density (mg/cc), trabecular BMD (mg/cc), and trabecular architecture parameters of the diaphysis of the right femora harvested at 66 wk of age were measured using microCT. Fresh femoral samples were cleaned of excess flesh, and the total length was recorded using vernier calipers (±0.05 mm). An examination of all anatomical sites was done on harvested and cleaned femurs. The lengths of femoral specimens were measured from the distal (Facies articularis antitrochanterica) to the proximal end (Condylus lateralis). Lines were
then drawn on the bone at 40, 50, and 60% of the distance from the proximal to the distal end of the bone. Bone lying between the 40 and 60% markers was used for subsequent studies. Using a microtome with a diamond-encrusted blade, the excess portions of the proximal and distal regions were removed to obtain anatomically comparable samples and to reduce variability associated with differences in bone length. The resulting femoral diaphyses were used to estimate apparent femoral density by calculating the ratio of total tissue mass to specimen volume; this was achieved by using an electronic top-loader scale (Satorius Laboratory Plus L420 P+, Sartorius GmbH Goettingen, West Germany), and bones were weighed to the nearest 1 mg. Bone fragments were suspended by a thin stainless-steel wire and weighed in air. Once the dry weight was recorded, the specimen’s proximal and distal ends were sealed using a fine layer of petroleum jelly and then fully immersed in a beaker with 15 cc of double deionized water. Caution was taken to ensure that the bone sample did not touch the sides of the beaker. The weight in water was read immediately after immersion. Density (mg/cc) was calculated by dividing the weight in air by the weight in water. After the apparent density was measured, bone fragments were immediately preserved in 10% formalin and rinsed in saline solution prior to scanning. Preliminary tomographic scans revealed that distinction between cortical and trabecular bone was not clear
Downloaded from http://ps.oxfordjournals.org/ by guest on March 21, 2015
Figure 1. Specimen (as shown in Figure 2) with a computer-generated isosurface.
708
MARTI´NEZ-CUMMER ET AL.
Downloaded from http://ps.oxfordjournals.org/ by guest on March 21, 2015
Figure 2. Adequate cancellous bone structure in femoral diaphysis medullary cavity at 66 wk of age. a) Three-dimensional volume; b) cross-sectional view of femoral diaphysis with a 3-D selected volume of interest; c) 3-D volume of interest (trabecular bone).
Figure 3. Depleted trabecular bone structure in femoral diaphysis medullary cavity at 66 wk of age. a) Three-dimensional volume; b) cross-sectional view of femoral diaphysis with a 3-D selected volume of interest; c) 3-D volume of interest (trabecular bone).
USE OF AXIAL X-RAY MICROCOMPUTED TOMOGRAPHY
709
in the medullary cavity. A range of skeletal structures was imposed by feeding birds 2.5, 3.5, or 4.5% calcium from 18 to 24 wk of age together with adequate (0.48%) or inadequate (0.28%) levels of available phosphorus. From 24 to 66 wk of age, all birds were fed 4% calcium and 0.38% available phosphorus.
MicroCT Data Correction
RESULTS
Figure 4. Relationship between femoral diaphysis apparent bone mineral density (mg/cc) and a) femoral medullary bone density (mg/ cc); b) femoral medullary trabecular plate thickness (mm); c) femoral medullary trabecular plate separation (mm); and d) medullary trabecular bone density (%).
in the metaphyseal shell of the proximal femur. Thus, an additional series of scans was conducted to assess the architectural makeup in the diaphyseal region and to obtain specimens with a distinct contrast between cortical and trabecular bone. This situation is imperative to be able to accurately quantitate BMD in both cortical bone located in the diaphyseal shell and trabecular bone
The end result of a complete microCT scan is a 3-D volume array of relative attenuation values (Figure 1). For bone specimen evaluations, 2-D slices along with Coronal (X), Sagital (Z), and Axial (Y) orthogonal planes allow for interpretation of skeletal structures (Figures 2 and 3). Figure 2 depicts adequate medullary bone from 66-wk-old layers; Figure 3 shows comparable orthogonal views for birds with depleted bone reserves. Sample calculations obtained from microCT scans are shown in Table 1. The descriptive statistics observed in these tables are divided into parameters that were either calculated [BMD, BV/TV, bone surface density (BS/BV)] or derived [trabecular plate thickness (TBTh), trabecular plate number (TbN), trabecular plate separation (TbSp)]. Traditional histomorphometric methods were used with bone surface and bone volume ratios and assumed a parallel plate model. The 3-D-calculated (model independent) stereology parameters were calculated using a test grid with orthogonal test lines (oriented along the scan axes) and measured 2 parameters: 1) Pp = PV/TV = number of voxels greater than the threshold divided by the total number of voxels in the volume of interest and 2) P1 = number of intersections of the test lines with the bone-marrow interface divided by the total line length. These test grids were measured for each of the 3 axes described as X, Y, and Z. Equations 1 and 2 were used to calculate BV/TV, which is the ratio between bone volume within a sampled volume of tissue, and BS/BV, which is the relationship between
Downloaded from http://ps.oxfordjournals.org/ by guest on March 21, 2015
It is imperative to establish the characteristics of the x-ray signal as read by the detector under scanning conditions and to reduce geometrical uncertainties of computed tomography images. Therefore, a phantom (Model SB#, Gammex rmi, Middleton, WI) constructed of a clear plastic material designed to hold air, water, and a sample of hydroxyapatite of bone mineral equivalent was utilized. This phantom was used in every scan to calculate a standard calibration curve from which both cortical and cancellous bone density could be interpolated. Once the data collection and imaging correction were completed, values obtained from the phantom were keyed in, and acquisition electronics that digitize the detector signals were sent to a computer for reconstruction. Reconstruction is the mathematical process of converting the single sets of detector readings for all 720 views into 2-D slice images, i.e., 3-D volume array (Ketcham and Carlson, 2001).
MARTI´NEZ-CUMMER ET AL.
710
Table 1. Example of quantitative bone morphometry data obtained from microcomputed tomography scan1,2 BV/TV3
BS/BV4
TbTh5
TbN6
TbSp7
Mean Median Minimum Maximum SD CV
84.85 84.98 74.45 93.84 3.83 4.52
21.44 21.44 9.84 34.56 4.66 21.48
0.0984 0.0933 0.0579 0.2032 0.0253 25.74
9.01 9.12 4.62 12.86 1.57 17.47
0.0166 0.0164 0.0133 0.0210 0.0017 10.09
Results averaged over the 3 directions Mean
84.85
21.30
0.1003
8.93
0.0167
1
3
bone surface present within a volume of tissue examined. Model-dependent parameters such as average TbTh (mm), average TbN, and average TbSp (mm) were derived from BV/TV and BS/BV using Equations 3, 4, and 5, respectively. BV/TV (%) = Pp
[1]
BS/BV (mm2/mm3) = 2 × Pl/Pp
[2]
TbTh (mm) = Pp/Pl = 2/(BS/BV)
[3]
TbN (1/mm) = Pl = (BV/TV)/TbTh
[4]
TbSp (mm) = (1 − Pp)/Pl = (1/TbN) − TbTh
[5]
Figure 4 shows the relationship between femoral BMD as measured by microCT and various other bone parameters. The R2 values are generally high (0.69 to 0.94) with relationships being highly significant (P < 0.01). Trabecular BMD (mg/cc) was found to have similar relationships with average TbTh and average TbSp. As expected, medullary cavity trabecular BMD (mg/cc) and medullary cavity trabecular BV/TV were highly correlated because both of these parameters quantitatively reflect the amount of bone present within an area of interest. These results suggest that femoral diaphysis apparent density (mg/cc) is a function of trabecular BMD (mg/cc), BV/TV (%), average TbTh (mm), and average TbSp (mm).
DISCUSSION Microcomputed tomography was initially developed because of the need for a highly precise and accurate tool for high-resolution reconstruction of the complex architecture of bone tissue (Mu¨ller and Ru¨egsegger,
1997; Hildebrand et al., 1999). Recently, Schreiweis et al. (2003) and Hester et al. (2004) reported DEXA as a tool to monitor changes in skeletal density in commercial Single Comb White Leghorn hens. For several reasons, including low cost, radiation dose, simplicity, and the ability to image skeletal sites, this methodology is used extensively and is considered the standard for the diagnosis of osteopenia and osteoporosis in humans. Hester et al. (2004) reported densitometry assessment in avian tibia using DEXA. Based on the results reported by these scientists, it is evident that DEXA provides integral measurements of trabecular and cortical bone. The precision of DEXA is very high, and results correlate well with traditional destructive tests (i.e., bone breaking force, bone ash); however, its accuracy in assessing bone volumetric density is greatly reduced because of the limitations of 2-D projection. For instance, DEXA is not able to distinguish bone density in different compartments such as compact vs. cancellous bone. Therefore, 3-D imaging techniques, such as axial x-ray microCT as reported in this study, are needed because they offer the ability to separately examine factors that may play independent and important roles in osteoporosis. Microcomputed tomography does allow for measurement of BMD in different bone compartments as well as the determination of 3-D morphometric indices in cancellous bone. The method can provide a more comprehensive analysis based on changes occurring in different bone types in very small regions of interest. Results from our study demonstrate a substantial improvement in overall femoral BMD determination because microCT is able to quantitate bone density in different compartments of the femur and to provide qualitative and quantitative data of the architectural status of bone. Microcomputed tomography used in this research project does have a practical limitation in that all data collected were in postmortem femoral diaphysis, and as
Downloaded from http://ps.oxfordjournals.org/ by guest on March 21, 2015
Total points used: 22,781,250; total volume of examined region: 122.09 mm ; total number of voxels occupied by bone: 1,932,959. 2 Results are averages summed directly over 3 directions. 3 Bone volume to total volume (%; calculated parameter). 4 Bone surface to total volume (mm2/mm3; calculated parameter). 5 Trabecular thickness [mm; model-derived parameter: TbTh (mm) = Pp/Pl = 2/(BS/BV), where Pp = PV/ TV = number of voxels greater than the threshold divided by the total number of voxels in the volume of interest and P1 = number of intersections of the test lines with the bone-marrow interface divided by the total line length]. 6 Trabecular number [1/mm; derived parameter: TbN (1/mm) = P1 = (BV/TV)/TbTh]. 7 Trabecular separation [mm; model-derived parameter: TbSp (mm) = (1 − Pp)/P1 = (1/TbN) − TbTh).
USE OF AXIAL X-RAY MICROCOMPUTED TOMOGRAPHY
REFERENCES Gasser, J. A., and J. Yam. 2002. In vivo non-invasive monitoring of changes in structural cancellous bone parameters with a novel prototype MicroCT in rats. Presented at the 24th Annu. Mtg. Am. Soc. Bone Mineral Res., San Antonio, TX. Hester, P. Y., M. A. Schreiweis, J. L. Orban, H. Mazzuco, M. N. Kopka, M. C. Ledur, and D. E. Moody. 2004. Assessing bone mineral density in vivo: Dual energy x-ray absorptiometry. Poult. Sci. 83:215–221. Hildebrand, T., A. Laib, and R. Muller. 1999. Direct threedimensional morphometric analysis of human cancellous bone: Micro structural data from spine, femur, iliac crest, and calcaneus. J. Bone Miner. Res. 14:1167–1174. Holdsworth, D. W., and M. M. Thornton. 2002. Micro-CT in small animal and specimen imaging. Trends Biotechnol. 20:S34–S39. Judex, S., S. Boyd, Y.-X. Qin, L. Miller, R. Mu¨ller, and C. Rubin. 2003. Combining high-resolution micro-computed tomography with material composition to define the quality of bone tissue. Curr. Osteoporosis Rep. 1:11–19. Ketcham, R. A., and W. D. Carlson. 2001. Acquisition, optimization and interpretation of X-ray computed tomographic imagery: Applications to the geosciences. Comput. Geosci. 27:381–400. Korver, D. R., J. L. Saunders-Blades, and K. L. Nadeau. 2004. Assessing bone mineral density in vivo. Poult. Sci. 83:222–229. Mu¨ller, R., and P. Ru¨egsegger. 1997. Microtomographic imaging for the non-destructive evaluation of trabecular bone architecture. Pages 61–80 in Bone Research in Biomechanics. G. Lowet, P. Ru¨egsegger, H. Weinans, and A. Meunier, ed. IOS Press, Amsterdam, The Netherlands. Ru¨egsegger, P., B. Koller, and R. Mu¨ller. 1996. A microtomographic system for the non-destructive evaluation of bone architecture. Calcif. Tissue Int. 58:24–29. Schreiweis, M. A., J. I. Orban, M. C. Ledur, and P. Y. Hester. 2003. The use of densitometry to detect differences in bone mineral density and content of live White Leghorns fed varying levels of dietary calcium. Poult. Sci. 82:1292–1301. Whitehead, C. C., and R. H. Fleming. 2000. Osteoporosis in cage layers. Poult. Sci. 79:1033–1041.
Downloaded from http://ps.oxfordjournals.org/ by guest on March 21, 2015
a consequence, evaluations over time were not possible. However, a number of scanners are now available that are able to scan small animals at high resolutions and scanning speeds (Gasser and Yam, 2002). For example, 100 sections at a 20- resolution are possible for imaging trabecular bone in live mice in <5 min. The size of current scanning chambers restricts high-resolution scanning to small live animals such as rats and mice. The development of units with scanning chambers large enough to fit a live laying hen is currently underway. These units will allow the assessment of the relative importance of bone architecture and bone mineralization in the characterization of bone quality as a predictor for bone strength in multiple sites. These capabilities are likely to enhance our understanding of the effect of density, morphological, and loading factors in the progression of osteoporosis in Single Comb White Leghorn hens. This enhanced understanding might lead to improved approaches that could significantly reduce the high incidence of fractures currently observed in commercial operations. Nonetheless, future research is likely to encounter inherent complexities in developing an integrative assessment of bone integrity. A complete assessment of bone quality will require the consideration of factors that are seldom considered because of the difficulty or inability to quantify such parameters. For instance, microdamage in bone accumulated by repetitive loading can lead to serious mechanical and biological consequences (Korver et al., 2004). With advances in 3-D bone characterization, the emphasis is now to study functional environments that induce microdamage and the effect of any preventive mechanisms (Judex et al., 2003). These data confirm that microCT is able to quickly and precisely determine bone architecture in the femur of laying hens. Over time, such measurements will be possible on live birds, a situation that will greatly add to our understanding of factors affecting osteoporosis in laying hens.
711