Bone imaging – the closest thing to art in medicine




James F. GRIFFITH,MB, BCh, BAO, MRCP, FRCR
Department of Imaging and Interventional Radiology
The Chinese University of Hong Kong, HONG KONG


Harry K. GENANT,MD
Thomas M. LINK,MD, PhD
Department of Radiology and Biomedical Imaging
University of California – San Francisco, USA

Bone imaging – the closest thing to art in medicine


by J. . Griffith, T. M. Link, and H. K. Genant , Hong Kong and USA



Advances in bone imaging have had a tremendous impact on our knowledge of skeletal anatomy, physiology, and pathophysiology while at the same time generating images of both aesthetic and scientific interest. Bone imaging for assessing bone quality very much lends itself to multidisciplinary input and collaboration across scientific disciplines, helping to drive technological and analytical advances in the assessment of bone quality. This has allowed a much deeper awareness of the changes that occur in bone quality with increasing age and disease, as well as improved fracture risk prediction and better treatment monitoring. Currently, many high-resolution imaging modalities exist to evaluate bone quality, though all have their particular merits and limitations. The ideal imaging modality, which has yet to fully emerge, would allow an accurate prediction of bone strength, discriminate at-risk individuals, identify which aspects of bone strength are faltering, and precisely monitor the effect of treatment. When this day comes, the occurrence of unheralded debilitating osteoporotic fractures in themiddle-aged and elderly will be seen as an unusual, rather than a usual, event. In the meantime, we can look forward to evenmore aesthetically pleasing images of bone structure, images that help link form to function in the human body and as such administer a helpful dose of science to the art of medicine.

Medicographia. 2012;34:170-177 (see French abstract on page 177)



Over the last 20 years, the digital era has seen an unparalleled expansion in the range and diversity of available imaging modalities and analytical techniques that, for the first time, have allowed detailed, non-invasive 3D assessment of the living human skeleton and soft tissues.1 These sophisticated imagingmodalities are not only successful diagnostic tools but also allow one to appreciate the beauty of human anatomy, especially with isotropic 3D and 4D (3D in real time) multidetector computed tomography (MDCT) and other related imaging techniques.2 The aesthetics of the natural human form is rarely addressed in anatomical and radiological texts, though it was explored enthusiastically by the esteemed Renaissance painters of the 15th century such as Leonardo Da Vinci and Michelangelo, and later by Andreas Vesalius in the 16th century, all of whom vigorously undertook human dissection to discover and paint the secrets and beauty of human anatomy. Radiological imaging, including bone imaging, has led to a new distinctive type of artwork known as radiological art (Figures 1 and 2). High-resolution imaging provides wideranging material for artists to apply their intuition and aesthetic judgment. The pioneers of radiological art have devised means of digitally manipulating and supple- menting clinical radiological images from different body regions to produce visually pleasing images of heightened interest and public appeal. Radiological art draws heavily on the creative qualities and software skills of the producer. Dr Kaihung Fung, an interventional radiologist from Hong Kong, is one of the leading pioneers of radiological art whose images frequently adorn well-known journals such as Radiographics and Leonardo. He invented the rainbow technique, first published in 2006, an image-rendering method in which artifacts are stacked between individual image slices to build a 3D image utilizing a contour line effect, with each contour being rendered in a rainbow of colors.3 In 2009, Dr Fung developed 3D and 4D color Moiré art by enhancing the Moiré interference pattern in 3D computed tomography (CT) and magnetic resonance (MR) datasets.4 Unlike anatomical dissection, which tends to be objective, analytical, and even disturbing to nonmedical observers, radiological imaging provides a palatable means of understanding and appreciating human anatomy, with radiological art often helping to enhance understanding, innuendo, and sentiment. Radiological anatomy, when portrayed as a readily comprehensible image, provides a means of spreading knowledge of the human structure far beyond health-related fields. Further improvements in isotropic 3D imaging and postprocessing digital software will allow increasingly diverse artistic creativity to be applied to baseline imaging data. One can see how the world may well experience an anatomical renaissance with the popularization of digital radiological art, particularly in hospitals, health clinics, and other health-related centers. While a shared inquisitiveness about the natural world drives the pursuit of artists and scientists alike,5 radiological art allows the creative instincts of the artistic mind to sit happily alongside the analytical instincts of the scientific mind.


Figure 1
Figure 1. Radiological art: Looking from the carpal tunnel down
to the fingers.

From 3D computed tomography data. Image courtesy of Dr K. H. Fung.



Figure 2
Figure 2. Radiological art: the femoral shaft.

Longitudinal (A) and axial (B) views of the femoral shaft from computed
tomography data. Image courtesy of Dr K. H. Fung.



Radiological art, in its purest form, deals with portraying human anatomy in an aesthetically pleasing fashion. Fromamore conceptual perspective, bone imaging also allows one to show and appreciate the amalgamation of human anatomy, form, and function. While traditional bone imaging dealt mainly with bone morphology, modern-day bone imaging correlates structure to function at both microscopic and macroscopic levels. Bone imaging as an art form is best exemplified in the field of bone imaging for bone quality assessment with analytical techniques relating morphology and composition features to functional elements such as strength distribution. Modern bone imaging thereby helps showcase the harmonious combination of form and function in the human skeleton. One cannot help but imagine that if Leonardo Da Vinci, the consummate anatomist/painter/scientist, were alive today, he would be taking a keen interest in radiological imaging anatomy and the analytical techniques that link function to formin the human body.

Just as traditional art transcends language and culture, radiological imaging provides a medium not just to make art and medicine interconnect, but also to bridge the gap between basic science, clinical medicine, and other allied scientific fields, and the wider population. Nowhere is this better illustrated than in the field of bone imaging, where scientific input from clinical medicine, anatomy, physiology, chemistry, physics, and computational engineering have contributed to an exponential growth in the knowledge of bone structure and quality over the past 3 decades. It is in part a reflection of this multidisciplinary input that developments in the field of bone imaging have, in many respects, outshone those in many other fields ofmedicine.6 Advances in bone imaging techniques such as dual x-ray absorptiometry (DXA), computed tomography, and magnetic resonance imaging (MRI), have provided hard data to further our understanding of medicine and, in addition to providing images of aesthetic quality, helped bring a large dose of science to the art of medicine. This is best illustrated by highlighting some of the recent advances in bone imaging achieved using these modalities.




Dual-x-ray absorptiometry

The widespread clinical use of dual-x-ray absorptiometry (DXA) to diagnose and gauge the severity of osteoporosis has led to osteoporosis being considered, in some circles, as a disease solely of reduced bone mineral density (BMD). This is not correct since osteoporosis is, by definition, a disease characterized not only by reduced BMD but also by “microarchitectural deterioration of bone.” This “microarchitectural deterioration of bone” is reflected in the term “bone quality,” introduced in 2001 by the Consensus Conference on Osteoporosis of the National Institutes of Health.7 The capability of DXA machines has been expanded in recent years beyond the measurement of BMD (in g/cm2) to provide information on aspects of bone quality such as vertebral fracture assessment and assessment of proximal femoral bone geometry. This additional information regarding vertebral fracture prevalence can be incorporated into the fracture risk assessment (FRAX®) model along with clinical risk factors to improve prediction of the 10-year probability (%) of major osteoporotic fracture (clinical vertebral, distal radius, proximal femur, or proximal humerus) (http://www.shef.ac.uk/FRAX).8 For the hip region, advances in DXA software allow automatic calculation of several proximal femoral structural parameters at the “narrow neck” (ie, the narrowest portion of the femoral neck), the intertrochanteric region, and the subtrochanteric femoral shaft region. Parameters such as hip axis length, outer diameter, endosteal diameter, average cortical thickness, cross-sectional moment of inertia, section modulus, and femoral neck shaft angle can be analyzed. These structural parameters compare favorably with similar measurements obtained by volumetric quantitative computed tomography (vQCT) and can be combined with subject height, weight, and age data to calculate the femoral strength index.9,10 In a study comparing 365 hip fracture patients with more than 2000 control subjects over the age of 50 years, fracture prediction was significantly improved by combining T-score with hip axis length and femoral strength index, compared with T-score alone.11 Geometric data is best achieved from 3D data and, with this in mind, volumetric x-ray absorptiometry (VXA) has evolved.12 In human cadaveric specimen, 3D reconstruction of the proximal femur using frontal and lateral acquisitions from a standard DXA unit can be obtained with good accuracy and precision (Figure 3A).12 A more iterative approach that can be applied in vivo has been developed recently (Figure 3B).13 The steps involved in 3D x-ray absorptiometry (3D-XA) include spatial calibration of a commercially available DXA device, acquisition of DXA images in about four different planes (eg, -21, 0, 20, and 30 degree relative to the coronal plane), identification of the specific contours on both views, and deformation of the 3D generic object until its projected contours match the 2D-identified contours.10 In excised proximal femora, combining areal BMD with 3D geometric parameters (such as femoral head diameter and midfemoral neck cross-sectional area) obtained by 3D-XA improved failure load prediction over density measurements alone.11 VXA shows excellent correlation with vQCT for shape parameters (femoral neck axis length, cross-sectional slice area) and density parameters (volumetric bone mineral density [vBMD]).13 Although VXA cannot currently distinguish cortical from trabecular bone and cannot accurately measure cortical thickness, it does show promise as a lowcost, low-radiation, clinically applicable alternative to vQCT in predicting proximal femoral fracture risk, though its clinical usefulness in this respect still has to be determined.


Figure 3
Figure 3. Volumetric x-ray absorptiometry (VXA) images of the
proximal femur.

(A) Superimposition of a VXA image (darker colors) obtained ex vivo and a 3D
computed tomography image (yellow color). (B) In vivo iterative-approach VXA
reconstruction of the proximal femur. The femoral neck axis (purple line) and
narrow neck region (yellow shading with darker yellow indicating higher density)
are shown.
(A) Image courtesy of Dr S. Kolta. (B) Image courtesy of Dr K. Engelke. Reproduced
from reference 13
: Ahmad et al. J Bone Miner Res. 2010;25:2744-2751.


Computed tomography

he high spatial resolution afforded by MDCT facilitates improved delineation of bone architecture with faster acquisition of near-isotropic vQCT datasets than earlier generations of CT scanners. MDCT scanners with 64 multidetector row spiral technology yield an in-plane resolution of 150-300 μm and a slice thickness of approximately 500 μm. MDCT allows assessment of density, structure, and biomechanical properties separately of trabecular and cortical bone components. It also provides volumetric density measurements (in mg/cm3) as opposed to the areal assessment by standard DXA (in g/cm2). One of the main advantages of whole-body MDCT over smaller, higher-resolution peripheral units is the ability to evaluate bone quality in the biologically relevant central areas of the skeleton that are particularly susceptible to fracture. This is important because changes observed in peripheral bone quality do not necessarily reflect bone quality changes in the central skeleton. MDCT systems correlate highly (R=0.92; P<0.0001) with reference standards for bone volume fraction (BV/TV) and trabecular spacing, though—as expected—far less well with trabecular thickness and number, as the spatial resolution of MDCT is larger than the average trabecular thickness of 50-150 μm and more comparable to the average trabecular spacing of 200-2000 μm. Structural parameters obtained by MDCT provide a better discriminator of clinical change than DXA and may be detected as early as 12 months postbaseline. This benefit was shown in a study of postmenopausal women, where teriparatide increased vertebral apparent BV/ TV by 30.6±4.4%(mean ± SE), and apparent trabecular number (Tb.N) by 19.0±3.2% compared with a 6.4±0.7% increase in DXA-derived areal BMD.14

High-precision software, known as medical image analysis framework, facilitates analysis of vQCT datasets through automatic determination of anatomical coordinates to yield predetermined volumes of interest (VOIs) for analysis (Figure 4, page 174).15,16 This automated anatomical coordinate system facilitates the study of the relative contributions of density, geometry, and trabecular and cortical bone to mechanical failure as well as facilitating longitudinal study. An example of how automated anatomical coordinate systems can be used to facilitate CT image analysis was shown in a study by Engelke et al. By comparing predetermined anatomical areas, one can appreciate how ibandronate treatment for 1 year increased volumetric density in the subcortical and extended trabecular areas of the proximal femur, as well as the extended cortical and superior/inferior trabecular regions of the vertebral body, all of which are mechanically significant areas.17

Although densitometric and morphometric analysis of highresolution imaging data improves assessment of fracture risk and treatment efficacy, a more direct measurement of bone strength would be preferable. Finite element analysis (FEA) modeling is a classic engineering computational technique used in design and failure analysis that provides information on parameters such as stiffness, estimated load failure, and stress distribution (Figure 5, page 174). This technique has been used in bone imaging to improve estimation of bone strength in vivo. Mechanical properties are assigned to each finite element high-resolution CT (or MRI) model following segmentation and decomposition. The finite elements can be hexagonal, tetrahedral, or curved scaled versions of CT voxels and can employ either linear or quadratic nodal displacement formulation. Load vectors typifying habitual ormore spurious overloads, simulating for example a sideways fall, can be used to perform a virtual stress test either to the whole bone or to the cortical or trabecular components separately. Models can be created with or without adjacent soft tissues and bones, and analyses can be run for single or multiple loading conditions.18 FEA analysis of vQCT data has revealed that vertebral body strength decreases with age twice as much in women than in men and that this sex difference is primarily due to a greater decline in cortical bone strength in women while trabecular bone strength declines to a comparable degree in both sexes.19 In other words, relatively greater cortical bone resorption in women may in part account for their increased vertebral fracture prevalence. Compared with nonfracture control subjects, vertebral vBMD, apparent cortical thickness, compressive strength assessment by FEA, and load-to-strength ratio were shown to be less in females with mild vertebral fracture and least in those with moderate to severe vertebral fracture, emphasizing how fracture severity, in addition to the presence of a fracture per se, is indicative of relative vertebral strength.20 Using vQCT data and FEA to study age-related changes in proximal femoral strength, Keaveny et al showed how proximal femoral strength declines with aging, much more than would be predicted on the basis of areal BMD changes alone. This study also showed how low proximal femoral strength is much more prevalent in older subjects than osteoporosis as defined by DXA.21


Figure 4
Figure 4. Volumetric quantitative computed tomography of the lumbar spine.

Automated anatomical coordinates outline the periosteal, endosteal, and juxtaendosteal (“peeled”) contours of the vertebral body. Several different volumes of interest
(VOIs) can be evaluated such as the total, trabecular, peeled, elliptical, and Pacman VOIs in the axial plane as well as the superior, midvertebral, and inferior VOIs
in the sagittal plane.
Images courtesy of Dr K. Engelke; reproduced from reference 6: Griffith and Genant. Curr Rheumatol Rep. 2011;13:241-250.



Figure 5
Figure 5. Volumetric quantitative computed tomography provides
a basis for finite element analysis of the proximal femur.

Note how stress distribution as related to color code is highest along the inferomedial
aspect of the femoral neck and proximal shaft.
Image courtesy of Dr K. Engelke; reproduced from reference 6: Griffith and
Genant. Curr Rheumatol Rep. 2011;13:241-250.


High-resolution peripheral quantitative computed tomography

High-resolution peripheral quantitative computed tomography (HR-pQCT) units (Xtreme CT, Scanco Medical AG, Basserdorf, Switzerland) have been developed that can scan the distal radius or distal tibia in 2.8 minutes, acquiring a stack of 110 images over a 9-mm length with a nominal isotropic voxel size of approximately 90 μm. Scan coverage is standardized to a defined distance from the distal radius or distal tibia (Figure 6). This is the only CT system available capable of acquiring high-resolution structural bone detail in humans in vivo. The structural parameters acquired are Tb.N, thickness, separation, structure model index, connectivity, anisotropy, and cortical thickness, all of which are derived from density measurements assuming a fixed mineralization of 1200 mg HA/cm3. As the analysis programs are density-based, many of the structural parameters will strongly correlate with vBMD, though these structural parameters have been validated against microCT measurements. Limitations of HR-pQCT include movement artifacts, particularly of the radius, and the inability to measure the midshaft of the forearm or leg bones. Structural parameters of cortical and trabecular bone assessed by HR-pQCT at the ultradistal radius can discriminate between women with and without vertebral fractures, partially independently of DXA results. In a 2-year, randomized, double- blind, prospective study comparing strontium ranelate and alendronate in postmenopausal women with osteoporosis, HR-pQCT monitoring revealed a 6.3% increase in cortical thickness and a 2.5% increase in cancellous BV/TV in those treated with strontium ranelate, while these parameters only increased by 0.9% and 0.8%, respectively, in those treated with alendronate.23 Estimated failure load also increased with strontium ranelate (+2.1%; P<0.005) but not with alendronate (–0.6%; P<0.05).23 In this study, values were not adjusted fo strontium content. However, bone strontium content is low after 2 years of treatment (about 1%). Trabecular microarchitectural and biomechanical properties derived from FEA analysis of both the distal radius and tibia are associated (odds ratio, 1.19-2.29) with vertebral and nonvertebral insufficiency fractures in men and women. A similar magnitude of association was seen for these parameters, irrespective of whether they were derived from the distal radius or distal tibia.24,25


Figure 6
Figure 6. HR-pQCT image of distal tibia in normal subject.

The inter trabecular bone, the outer trabecular bone, and the cortical bone
components have been separated allowing individual analysis of each component.
Image courtesy of Ling Qin; reproduced from reference 22: Griffith and Genant.
Best Pract Res Clin Endocrinol Metab. 2008;22:737-764.


Magnetic resonance imaging

MRI has several advantages over CT in the assessment of bone quality: no ionizing radiation is involved, thus making it very acceptable in the clinical or research setting; images can be obtained in orthogonal planes directly; and aspects of bone physiology, particularly those related to the marrow cavity such as marrow fat content, marrow diffusion, marrow perfusion, and water content, are obtainable. The disadvantages are cost, time requirements, and more technically demanding data acquisition and analysis. Due to trade-off issues between spatial resolution, signal to noise ratio (SNR), and radiofrequency signal attenuation, most in vivo MR studies have addressed relatively superficial peripheral sites such as the distal radius, distal tibia, and calcaneus as these trabecular-rich areas are accessible to small high-resolution coils. Nearly all MR-derived structural parameters of the distal radius are better than DXA at differentiating women with and without vertebral fracture.26

High-resolution MRI of the central skeleton is limited by SNR and resolution issues due to the persistence of hematopoietic marrow, which contrasts less well with adjacent trabeculae thanmarrow fat. In an in vitro study of excised proximal femoral specimens, combining MR-derived structural parameters with DXA-derived BMD measures led to improved correlation with bone strength parameters, with R values of up to 0.93 being reached.27 The trabecular structure of the proximal femur has been studied with 3 Tesla MRI, using SNR-efficient sequences with an in-plane resolution of 234 μmx 234 μm and a slice thickness of 1500 μm. Future improvements in resolution and analytical techniques may help advance MRI of biologically relevant sites such as the proximal femur.28

To monitor the effects of treatment, MRI of the trabecular structure of the distal radius and trochanteric region of the proximal femur was performed in postmenopausal women. This revealed preservation of apparent BV/TV, apparent Tb.N, and apparent trabecular spacing in patients treated with calcitonin for 2 years, compared with significant loss in a placebo group.29 Over the same period of time, no significant change in DXA BMD was observed among both groups. This study may help explain the results of an earlier study, which showed substantial reduction in fracture risk with calcitonin treatment despite only a small increase in BMD.30 The longitudinal effects of alendronate on MRI-based trabecular bone structure parameters have been evaluated.31 MR-derivedapparent Tb.N, as well as 4 topographical parameters, showed treatment effects in the distal tibia after 24 months, especially when fuzzy clustering trabecular bone segmentation, rather than dual thresholding trabecular segmentation, was used, emphasizing the importance of carefully choosing the right computational method for analysis.31 Surprisingly, no treatment effect was observed by HR-pQCT.31

MR-based virtual bone biopsy of peripheral bone has also been advocated as a means to monitor treatment. To this effect, reproducibility was assessed for a 13-mm wide axial slab encompassing the distal radial medullary cavity as well as a for a 5-mm cuboid subvolume. Whole-volume-derived aggregate mean coefficient of variation of all structural parameters was 4.4%(range 1.8%-7.7%) and 4.0%for axial stiffness; while mean coefficients of variation for similar parameters in the corresponding data in the subvolume were 6.5% and 5.5%, respectively.32


Figure 7
Figure 7. Color mapping based on amplitude map from pharmacokinetic
modeling of lumbar vertebral body MR perfusion data in
a subject with normal BMD and one with osteoporosis.

Note how perfusion parameters are appreciably reduced in the subject with osteoporosis
compared with the subject with normal BMD.



As well as being used to examine the trabecular bone component, MRI has been applied to the study of the bone cortex. Cortical bone only accounts for about 10% of vertebral body bone, though it accounts for up to 75% of femoral neck bone and 50% of femoral intertrochanteric bone.33 Cortical bone seems to have a relatively greater role in proximal femoral bone strength than vertebral body bone strength. With bone loss, cortical bone becomes thinner and more porous.

Cortical porosity is a challenging parameter to measure in vivo, even with HR-pQCT. The wider capability of MRI has allowed a different approach to the assessment of cortical porosity in that cortical water content may serve as a surrogate measure of cortical porosity. Cortical bone water content assessed by ultrashort echo time MRI correlates well with that measured by isotope exchange.34 Tibial cortical water content in hemodialysis subjects was found to be 135% greater than in premenopausal women and 43% greater than in postmenopausal women, although no difference in cortical BMD was found between the 3 groups. This indicates that cortical water content may prove to be a better indicator of cortical bone loss and cortical porosity than cortical BMD.34

MR also has the capability of assessing marrow fat content, molecular diffusion, and marrow perfusion. MRI studies have shown how perfusion is reduced in nonfractured osteoporotic vertebrae bodies compared with those of normal BMD (Figure 7).35,36 This reduced perfusion is most likely to be due to atherosclerosis, impaired endothelial function, or reduced demand for tissue oxygenation due to a relative decrease in the amount of hemopoietic marrow within osteoporotic vertebral bodies.37,38 MR-based perfusion parameters are reduced in osteoporotic vertebral fractures compared with adjacent nonfractured vertebrae.39 Good perfusion is clearly a prerequisite for normal bone metabolism and fracture healing, including microdamage repair. The smaller the area of enhancing tissue within an acutely fractured vertebral body, the more likely that fractured vertebral body will reduce in height on subsequent follow-up.40

Conclusion

The last two decades have seen an exponential growth in bone imaging with new imaging modalities and analytical techniques helping to improve our perception of bone anatomy, physiology, and pathophysiology as well as providing images of aesthetic quality. Bone imaging serves as a focal point for collaboration between clinical and other allied scientific disciplines, which has led to a much better understanding of bone structure and function as well as appreciation of the changes that may occur with age, disease, and treatment. There is little doubt that further advances in bone imaging will continue to hold center stage in osteoporosis and related research. As it stands, bone imaging is probably the closest thing to art in medicine, whether this is a visual appreciation of the aesthetic qualities of bone imaging; a conceptual appreciation of how bone imaging links structure to form and function; or an appreciation of how advances in bone imaging have succeeded in bringing a large dose of science to the art of medicine. _


References
1. Link TM. The Founder’s Lecture 2009: advances in imaging of osteoporosis and osteoarthritis. Skeletal Radiol. 2010;39:943-955.
2. Marinkovic´ S, Stošic´ -Opinc´al T, Strbac M, Tomic´ I, Tomic´ O, Djordjevic´ D. Neuroradiology and art: a review and personal contribution. Tohoku J Exp Med. 2010;222:297-302.
3. Fung KH. The rainbow technique: an innovative approach to the artistic presentation of 3D computed tomography. Leonardo. 2006;39:101-103.
4. Fung KH. Creating special visual effects with Moiré patterns in stereoscopic 3D and 4D computed tomographic art. Leonardo. 2010;43:306-307.
5. Kemp M. From science in art to the art of science. Nature. 2005;434:308-309.
6. Griffith JF, Genant HK. New imaging modalities in bone. Curr Rheumatol Rep. 2011;13:241-250.
7. NIH Consensus Development Panel on Osteoporosis Prevention, Diagnosis, and Therapy. Osteoporosis prevention, diagnosis, and therapy. JAMA. 2001;285: 785-795.
8. Kanis JA, Johnell O, Oden A, Johansson H,McCloskey E. FRAX and the assessment of fracture probability in men and women from the UK. Osteoporos Int. 2008;19:385-397.
9. Prevrhal S, Shepherd J, Faulkner K, Gaither K, Black D, Lang T. Comparison of DXA hip structural analysis with volumetric QCT. J Clin Densitom. 2008;11: 232-236.
10. Faulkner KG, Wacker WK, Barden HS, et al. Femur strength index predicts hip fracture independent of bone density and hip axis length. Osteoporos Int. 2006;17:593-599.
11. Le Bras A, Kolta S, Soubrane P, Skalli W, Roux C, Mitton D. Assessment of femoral neck strength by 3-dimensional X-ray absorptiometry. J Clin Densitom. 2006;9:425-430.
12. Kolta S, Le Bras A, Mitton D, et al. Three-dimensional X-ray absorptiometry (3DXA): a method for reconstruction of human bones using a dual X-ray absorptiometry device. Osteoporos Int. 2005;16:969-976.
13. Ahmad O, Ramamurthi K, Wilson KE, Engelke K, Prince RL, Taylor RH. Volumetric DXA (VXA): a new method to extract 3D information from multiple in vivo DXA images. J Bone Miner Res. 2010;25:2744-2751.
14. Graeff C, Timm W, Nickelsen TN, et al; EUROFORS High Resolution Computed Tomography Substudy Group. Monitoring teriparatide-associated changes in vertebral microstructure by high-resolution CT in vivo: results from the EUROFORS study. J Bone Miner Res. 2007;22:1426-1433.
15. Kang Y, Engelke K, Fuchs C, Kalender WA. An anatomic coordinate system of the femoral neck for highly reproducible BMD measurements using 3D QCT. Comput Med Imaging Graph. 2005;29:533-541.
16. Engelke K, Mastmeyer A, Bousson V, Fuerst T, Laredo JD, Kalender WA. Reanalysis precision of 3D quantitative computed tomography (QCT) of the spine. Bone. 2009;44:566-572.
17. Engelke K, Fuerst T, Dasic G, Davies RY, Genant HK. Regional distribution of spine and hip QCT BMD responses after one year of once-monthly ibandronate in postmenopausal osteoporosis. Bone. 2010;46:1626-1632.
18. Keaveny TM. Biomechanical computed tomography-noninvasive bone strength analysis using clinical computed tomography scans. Ann N Y Acad Sci. 2010; 1192:57-65.
19. Christiansen BA, Kopperdahl DL, Kiel DP, Keaveny TM, Bouxsein ML. Mechanical contributions of the cortical and trabecular compartments contribute to differences in age-related changes in vertebral body strength in men and women assessed by QCT-based finite element analysis. J Bone Miner Res. 2011;26: 974-983.
20. Melton LJ 3rd, Riggs BL, Keaveny TM, et al. Relation of vertebral deformities to bone density, structure, and strength. J Bone Miner Res. 2010;25:1922-1930.
21. Keaveny TM, Kopperdahl DL, Melton LJ 3rd, et al. Age-dependence of femoral strength in white women and men. J Bone Miner Res. 2010;25:994-1001.
22. Griffith JF, Genant HK. Bone mass and architecture determination: state of the art. Best Pract Res Clin Endocrinol Metab. 2008;22:737-764.
23. Rizzoli R, Chapurlat RD, Laroche JM, et al. Effects of strontium ranelate and alendronate on bone microstructure in women with osteoporosis: results of a 2-year study. Osteoporos Int. 2011 Sept 10. Epub ahead of print.
24. Vilayphiou N, Boutroy S, Szulc P, et al. Finite element analysis performed on radius and tibia HR-pQCT images and fragility fractures at all sites in men. J Bone Miner Res. 2011;26:965-973.
25. Vilayphiou N, Boutroy S, Sornay-Rendu E, et al. Finite element analysis performed on radius and tibia HR-pQCT images and fragility fractures at all sites in postmenopausal women. Bone. 2010;46:1030-1037.
26. Krug R, Carballido-Gamio J, Burghardt AJ, et al. Assessment of trabecular bone structure comparing magnetic resonance imaging at 3 Tesla with high-resolution peripheral quantitative computed tomography ex vivo and in vivo. Osteoporos Int. 2008;19:653-661.
27. Link TM, Vieth V, Langenberg R, et al. Structure analysis of high resolution magnetic resonance imaging of the proximal femur: in vitro correlation with biomechanical strength and BMD. Calcif Tissue Int. 2003;72:156-165.
28. Krug R, Banerjee S, Han E, Newitt D, Link T, Majumdar S. Feasibility of in vivo structural analysis of high-resolution magnetic resonance images of the proximal femur. Osteoporos Int. 2005;16:1307-1314.
29. Chesnut CH 3rd, Majumdar S, Newitt DC, et al. Effects of salmon calcitonin on trabecular microarchitecture as determined by magnetic resonance imaging: results from the QUEST study. J Bone Miner Res. 2005;20:1548-1561.
30. Chesnut CH 3rd, Silverman S, Andriano K, et al. A randomized trial of nasal spray salmon calcitonin in postmenopausal women with established osteoporosis: the prevent recurrence of osteoporotic fractures study. Am J Med. 2000; 109:267-276.
31. Folkesson J, Goldenstein J, Carballido-Gamio J, et al. Longitudinal evaluation of the effects of alendronate on MRI bone microarchitecture in postmenopausal osteopenic women. Bone. 2011;48:611-621.
32. Lam SC, Wald MJ, Rajapakse CS, Liu Y, Saha PK, Wehrli FW. Performance of the MRI-based virtual bone biopsy in the distal radius: Serial reproducibility and reliability of structural and mechanical parameters in women representative of osteoporosis study populations. Bone. 2011;49:895-903.
33. Krug R, Burghardt AJ,Majumdar S, Link TM. High-resolution imaging techniques for the assessment of osteoporosis. Radiol Clin North Am. 2010;48:601-621.
34. Techawiboonwong A, Song HK, Leonard MB, Wehrli FW. Cortical bone water: in vivo quantification with ultrashort echo-time MR imaging. Radiology. 2008; 248:824-833.
35. Griffith JF, Yeung DK, Antonio GE, et al. Vertebral bone mineral density, marrow perfusion, and fat content in healthy men and men with osteoporosis: dynamic contrast-enhanced MR imaging and MR spectroscopy. Radiology. 2005;236: 945-951.
36. Griffith JF, Yeung DK, Antonio GE, et al. Vertebral marrow fat content and diffusion and perfusion indexes in women with varying bone density: MR evaluation. Radiology. 2006;241:831-838.
37. Griffith JF, Wang YX, Zhou H, et al. Reduced bone perfusion in osteoporosis: likely causes in an ovariectomy rat model. Radiology. 2010;254:739-746.
38. Griffith JF, Kumta SM, Huang Y. Hard arteries, weak bones. Skeletal Radiol. 2011;40:517-521.
39. Biffar A, Schmidt GP, Sourbron S, et al. Quantitative analysis of vertebral bone marrow perfusion using dynamic contrast-enhanced MRI: initial results in osteoporotic patients with acute vertebral fracture. J Magn Reson Imaging. 2011;33: 676-683.
40. Kanchiku T, Taguchi T, Toyoda K, Fujii K, Kawai S. Dynamic contrast-enhanced magnetic resonance imaging of osteoporotic vertebral fracture. Spine. 2003;28: 2522-2526.

Keywords: bone imaging; computed tomography; dual x-ray absorptiometry; magnetic resonance imaging; osteoporosis; radiological art