Emerging Research on Bone Health Using High-Resolution CT and MRI
Today’s most prevalent bone disease in the western hemisphere is osteoporosis. Predominantly postmenopausal women and older men suffer from bone loss caused by an imbalance in the physiological tissue renewal process between bone formation and resorption. As a result, osteoporosis is associated with fragility fractures, disability, impaired bone regeneration and increased mortality. The World Health Organization based the gold standard for diagnosing osteoporosis on bone mineral density (BMD) measurements using dual X-ray absorptiometry. However, BMD measurements are limited in discriminating subjects with and without osteoporotic fractures and have been shown to only partly reflect successful treatment of osteoporotic fractures. Bone microstructure is an integral determinant of bone strength. Today, new high-resolution imaging techniques such as high-resolution peripheral quantitative computed tomography and high-resolution magnetic resonance imaging make it possible to measure three-dimensional bone microarchitecture and volumetric bone mineral density with high accuracy and a relatively low radiation dose.
KeywordsOsteoporosis Multidetector computed tomography (MDCT) Magnetic resonance imaging (MRI) Dark field imaging High-resolution bone imaging
The constant process of bone remodeling is a delicate balance between bone resorption by osteoclasts and bone formation by osteoblasts, which maintains bone mass during adulthood. Osteoporosis is the most prevalent disruption of this complex system in our aging societies, and is characterized by inherent bone loss and an increased risk for fragility fractures, with sites most commonly affected being the spine, wrist and hip . The poor primary stability of the fragile bone in osteoporotic fractures results in associated disability and increased mortality [2, 3, 4]. There is an ongoing discussion on whether the regenerative capacity of osteoporotic bone is likely to be additionally impaired. With unfavorable healing conditions and an increased risk for further bone fractures once a fracture has occurred, consequences are drastic . The rising prevalence of osteoporosis and subsequent healthcare costs are burdens on the individual level and socioeconomically . Approximately 26 % of women aged ≥65 years and over 50 % of women aged ≥85 years are affected with postmenopausal osteoporosis (PMO), and direct and associated costs are estimated to reach $12–18 billion in the US . To counter these challenges, significant efforts are being made to investigate and fully understand the underlying etiopathology, a complex interplay of metabolic factors and local tissue dynamics that exceed the compartmental boundaries of the bone.
Accurate diagnosis and effective monitoring to evaluate treatment response are crucial. The World Health Organization (WHO) based their diagnostic standard on bone mineral density measurements (BMD) using dual X-ray absorptiometry (DXA). Quantitative computed tomography (QCT) is another common alternative to quantify BMD. Both techniques compare individual values of bone density expressed as grams of mineral per area or volume to the peak bone mass of a reference cohort. However, these quantitative measurements may insufficiently assess bone health, as biomechanical strength depends on bone quality parameters that may not be mirrored by density. Trabecular and cortical microarchitecture, turnover, damage (e.g., microfractures) and mineralization play crucial roles in bone stability. Consequently, these additional parameters have been investigated to further investigate fracture probability. The WHO population-based fracture risk assessment tool (FRAX)  integrates clinical risk factors to calculate the 10-year fracture probability, but does not include bone strength and quality parameters . Furthermore, the FRAX algorithm was shown to perform poorly in men with osteoporosis . High-resolution imaging provides visualization of the cortical and trabecular bone microstructure and its three-dimensional rod- and plate-like constituents, which have a size of approximately 50–200 μm, allowing assessment of bone quality.
Various imaging techniques have been introduced to non-invasively evaluate bone architecture, such as micro computed tomography (μCT) systems, high-resolution peripheral quantitative computed tomography (HR-pQCT), high-resolution multidetector computed tomography (MDCT) and high-resolution magnetic resonance imaging (MRI). Furthermore, bone structure may be further analyzed with novel modalities such as quantitative ultrasound (QUS) and dark field imaging, which are currently under investigation.
High-Resolution CT Imaging Modalities
Microarchitecture has been described to be a significant contributor to bone strength. Substance loss, trabecular thinning, but also modified trabecular topology, such as the change of plate-like trabeculae into rod-shaped trabeculae and the loss of trabecular connectivity, result in the loss of bone stability. The deterioration of cortical structures also seems to contribute to increased fracture risk. Different modalities based on CT technology are currently in use to acquire high-resolution bone images evaluating micro-architecture: μCT, HR-pQCT and MDCT. Bone trabeculae have a diameter between about 50–200 μm, and the cortical bone presents with a thickness between 0.2 and 5 mm. Therefore, spatial resolution is the limiting factor of any system to perform microstructure analysis.
Micro computed tomography (μCT) is a high-resolution technique limited to in vitro studies. Small invasively harvested bone biopsy cores from the iliac crest can be analyzed with an isotropic voxel size below 8 μm to depict trabecular microstructure . However, there are significant disadvantages due to the invasiveness and small sample size, which are usually harvested from locations not affected by fractures. Eckstein et al.  demonstrated that morphological and mechanical properties vary among anatomical locations, indicating site-specific patterns of trabecular microarchitecture. Site-specific measurements at adequate anatomical locations may therefore be less heterogeneous and more precise in predicting fracture risk. In vitro and in vivo studies have shown that fracture risk is assessed better by a combination of structural parameters and BMD derived from site-specific high-resolution imaging modalities [20, 21, 22].
High-Resolution Peripheral Quantitative CT
Since the year 2000, HR-pQCT devices have become commercially available, specifically designed for the imaging of bone microstructure in the peripheral skeleton. HR-pQCT enables a non-invasive, low-radiation assessment of bone providing information on the microarchitecture and volumetric BMD in cortical and trabecular compartments of the distal radius and distal tibia . Currently, one manufacturer dominates the market for commercially available scanners performing at a resolution sufficient to analyze human bone microarchitecture in vivo (SCANCO Medical AG, Brüttisellen, Switzerland). HR-pQCT allows for a significantly higher SNR and spatial resolution compared to MDCT and MRI with an isotropic spatial resolution of 82 μm3 (the actual spatial resolution is approximately 130 μm near the center of the field of view) in vivo with relatively low effective radiation doses of approximately 4 μSv per scan, a dose that is several orders of magnitude lower compared to whole body CT . A great benefit of HR-pQCT is the possibility for morphometric analysis similar to classical histomorphometry, calculated from the binary trabecular bone images . To permit data interpretation, normative databases providing reference for bone microarchitecture data had to be established enabling population-based comparisons of individual measurements: the Calgary cohort of the (CaMOS) cohort, the Rochester, Minnesota, cohort and the Cambridge, UK, cohort [26, 27]. Unlike MRI and MDCT, which are limited by a large slice thickness compared to their in-plane resolution, HR-pQCT enables direct measurements of microarchitecture parameters in a fairly large bone volume and can be paired with computer-based finite element analysis modeling (FEM) for non-invasive assessment of fracture risk. HR-pQCT has been shown to be able to differentiate between women and men with and without fractures and has increased our understanding of bone architecture and its structural changes related to age, gender, various metabolic disorders and in response to drug therapies [27, 28, 29, 30].
Due to high acquisition and maintenance costs and the need for regular phantom calibration, only few systems are clinically used, and currently most scanners are found at research institutions. This is in part because the relatively long scan times (approximately 3 min) frequently result in motion artifacts. Furthermore, the technique is limited to the evaluation of extremities, providing no information about the more central sites commonly affected by osteoporotic fragility fractures.
Multidetector CT is routinely applied in clinical practice, but to achieve adequate spatial resolution at central regions of the skeleton, considerable radiation doses are required, limiting the technique’s applicability in vivo. Studies examining vertebral microstructure using high-resolution MDCT reported estimated effective doses in the range of 3 mSv, compared to an exposure of approximately 0.1–0.3 mS delivered through a standard QCT of the lumbar spine . However, the radiation doses can be significantly reduced using novel iterative reconstruction algorithms [32, 33]. Baum et al.  have demonstrated in a study conducted on 187 proximal femur specimens that models combining DXA and MDCT-derived trabecular bone structure parameters performed better at predicting failure load than DXA alone. Furthermore, FEMs from in vivo MDCT spine images have been shown to reliably assess bone strength and to differentiate subjects with and without fragility fractures better than BMD measurements alone [22, 34]. In studies measuring the vertebral bone strength assessing the therapeutic effects of teriparatide, alendronate and risedronate, MDCT-based FEMs (as shown below) also provided more information than BMD measurements alone [35, 36].
Magnetic Resonance Imaging
MR-based structural parameters of the radius have been shown to improve the prediction of radial bone strength, outperforming DXA-derived BMD . With the more widespread use of high-field MRI and progress in sequence and coil development, previous limitations of deeper body locations such as low SNR and radio frequency signal attenuation by surrounding tissue can now be overcome . Studies have shown promising results, but evaluation of common fracture sites such as the proximal femur remains challenging because of the presence of hematopoietic bone marrow . Annihilating the positive background contrast of fatty bone marrow with its dark signal, hematopoietic marrow limits the visualization of the trabeculae. Its content in vertebral bodies is even higher, resulting in insufficient contrast to depict microstructures. Additional bone marrow quantification using MR spectroscopy to evaluate bone marrow fat content may complement osteoporosis imaging [45, 46]. A study by Wehrli et al.  demonstrated that MRI-based trabecular bone structure parameters provide a promising tool to evaluate structural treatment responses. The parameters revealed drug effects partly not captured by BMD measurements, indicating feasibility to monitor osteoporosis therapy. Chestnut et al.  showed similar results of successful osteoporotic treatment monitoring at the radius using MRI-based trabecular bone structure parameters.
Parameters to quantify bone structure
To identify the cortical and trabecular bone compartments, regions of interest (ROIs) usually have to be defined first, and images for longitudinal studies need to be registered to minimize reproducibility errors. Various structural parameters have been investigated to assess bone microstructure: scale parameters, represented by bone volume, thickness of the trabeculae and the spaces in between; topological parameters differentiating plate- and rod-like trabeculae, and orientation parameters characterizing the amount of anisotropy within the depicted volume. Standard structural parameters can be computed from both MDCT and MR images similar to classic bone histomorphometry [49, 50]: bone volume divided by total volume (BV/TV; bone volume fraction), trabecular number (Tb.N), trabecular separation (Tb.Sp) and trabecular thickness (Tb.Th) .
Finite Element Modeling (FEM)
An alternative to the structure measurements described above to evaluate bone strength is to calculate elastic and shear moduli using finite element models (FEM). FEMs can be computed within defined ROIs of trabecular or cortical bone, or for both compartments combined, based on 3D data from HR-pQCT, MDCT and MRI [69, 70, 71]. FEMs are based on the volumetric distribution of density parameters and bone geometry. Loading conditions are then implemented into the model, simulating either static loading conditions or a localized impact caused by a fall to the side, e.g., the greater trochanter of the femur. Various types of FEMs have been used. FEMs of the spine and proximal femur have been studied based on MDCT images . Furthermore, micro-FEMs (μFEMs) have been computed from in vivo MR images . Extensive computational power and post-processing resources are needed to perform the complex processing, and in some cases interpolation to higher apparent resolution is required. So far, FEM and μFEM analyses have been mainly limited to research institutions, but hardware and software improvements allow for future clinical application. Recent advancements, such as the new p-version and the finite cell method, have improved the h-version simulation algorithm, providing more accurate prediction of bone stability.
Current Trends and Future Developments
MRI-derived structural measures also demonstrated high accordance with μCT. Krug et al.  used ex vivo and in vivo peripheral trabecular bone structure parameters derived from 3-T MRI compared to HR-pQCT as standard of reference. Eight human specimens and 11 volunteers were imaged with both modalities at a voxel size of 156 × 156 × 500 μm3 at 3 T MRI and 82 μm3 at HR-pQCT. MRI- and HR-pQCT-derived bone structure parameters showed high compliance (R 2 > 0.8). Another study conducted by Phan et al.  investigated trabecular bone structure parameters at the calcaneus derived from in vitro μCT compared to 1.5- and 3-T MRI. The correlation of bone microstructure parameters derived from gradient echo sequences at 3-T MRI with μCT measures was higher than the association between bone structure parameters obtained at 1.5-T MRI and μCT measurements.
SNR increase and spatial resolution at high field strengths are a trade-off, and larger susceptibility effects alter the structural measurements because of artificially thickened trabeculae. Krug et al. investigated the impact of SNR on bone imaging using high-field MRI and found a significant SNR increase at 7 T . Baum et al.  found similar reproducibility errors at 1.5- and 3-T MRI, but absolute parameter values were significantly different, so it remains questionable whether microarchitecture parameters derived from 3- and 1.5-T MRI are comparable.
Various novel imaging techniques and continuous refinement of conventional hardware as well as software and post-processing algorithms enrich our understanding of bone health, pushing the boundaries of imaging resolution and functional quantification. A new generation of flat panel scanners has been introduced, combining standard MDCT gantries with two-dimensional flat panel detectors allowing for fast continuous acquisitions at high spatial resolution [80, 81]. Iterative reconstruction algorithms may allow for radiation dose reduction of clinical MDCT without compromising the resolution of trabecular and cortical structures.
Quantitative ultrasound has emerged as a promising technique without ionizing radiation, and QUS-based measurements have been shown to be highly correlated with BMD: Phalangeal bone structure has been shown to influence the velocity (SoS), shape (number of peaks) and amplitude of the ultrasound signal [82, 83, 84]. QUS parameters measured are bone transmission time (BTT) and pure speed of sound (pSoS). QUS was shown to perform well in fracture risk prediction, and based on QUS measurements, subjects with and without spine and hip fractures could be differentiated [85, 86]. Ingle et al.  have demonstrated good precision over time for the follow-up monitoring of osteoporosis therapies with different drugs such as alendronate and oestradiol. Unfortunately, QUS is limited to peripheral sites (the radius, tibia, calcaneus and the phalanx are currently being studied), and the clinically most important fracture sites (spine and hip) cannot be evaluated directly .
Phase-Contrast and X-Ray Dark-Field Imaging
Recently, it has been shown that complementary conventional X-ray contrast modalities such as phase-contrast and dark-field contrast may be applied in a clinical environment [89•]. Phase-contrast imaging has been shown to be a sensitive soft tissue imaging modality. Dark-field contrast however may have great potential in bone imaging, providing additional information about the micro-morphology. Dark-field imaging is based on the physical process of scattering at features in the micrometer range. Without resolving the individual features directly, it allows drawing conclusions about the number of structures, their size and their anisotropy . In addition to plain acquisitions, tomography can also be performed .
In the recent years, progress in high-resolution bone imaging has been tremendous. CT and MRI systems are widespread in clinical practice and are therefore potentially available for bone microstructure analyses. Bone microarchitecture parameters and finite element model data derived from high-resolution CT images, MDCT data and MR images have improved the assessment of trabecular and cortical bone architecture beyond BMD measurements, allowing for more accurate diagnosis with limited radiation exposure. In addition, microstructure parameters have been shown to be more sensitive to structural changes associated with pharmacotherapy effects compared to BMD measurements alone and thus may also be more effective in monitoring treatment.
Newly developed hardware, advancements in image post-processing, MRI scanners with high field strengths of up to 7 T in combination and newly developed software such as UTE sequences create new promising possibilities in evaluating bone microstructure of both trabecular and cortical bone compartments also at more central body sites. HR-pQCT provides in vivo images with extremely high resolution from peripheral sites at short acquisition times. Novel reconstruction algorithms using iterative reconstruction may allow for reliable assessment of trabecular bone at decreased radiation dose levels with MDCT imaging, and new flat panel scanners are available.
In conclusion, high-resolution bone imaging is crucial for the investigation of bone disease and the underlying etiopathologies.
We acknowledge financial support from the Deutsche Forschungsgemeinschaft (DFG) Cluster of Excellence Munich-Centre for Advanced Photonics (MAP), the DFG Gottfried Wilhelm Leibniz program and the European Research Council (ERC, FP7, StG 240142). This work was supported in parts by grants of the Deutsche Forschungsgemeinschaft (DFG BA 4085/1-2).
Compliance with Ethics Guidelines
Conflict of Interest
Hans Liebl, Thomas Baum, Dimitrios C. Karampinos, Janina Patsch, Andreas Malecki, Florian Schaff, Elena Eggl, Ernst J. Rummeny, Franz Pfeiffer and Jan S. Bauer declare no conflicts of interest.
Human and Animal Rights and Informed Consent
This article does not contain any studies with human or animal subjects performed by any of the authors.
Papers of particular interest, published recently, have been highlighted as: • Of importance
- 8.WHO Study Group. Assessment of fracture risk and its application to screening for postmenopausal osteoporosis. Report of a WHO Study Group. World Health Organ Tech Rep Ser. 1994;843:1–129. Google Scholar
- 12.• Graeff C, Marin F, Petto H, et al. High resolution quantitative computed tomography-based assessment of trabecular microstructure and strength estimates by finite-element analysis of the spine, but not DXA, reflects vertebral fracture status in men with glucocorticoid-induced osteoporosis. Bone. 2013;52(2):568–77. Graeff et al. conducted the first study in which MDCT was used to study not only bone forming but also antiresorptive treatment. CT-based FE analysis estimating vertebral strength showed highly consistent results for vertebral strength improvement under therapy demonstrating effective monitoring of therapeutic drug effects, which were more sensitive than BMD measurements derived from DXA. Google Scholar
- 14.NIH Consensus Development Panel on Osteoporosis. NIH Consensus Development Panel on osteoporosis prevention, diagnosis, and therapy, March 7–29, 2000: highlights of the conference. South Med J. 2001;94(6):569–73. Google Scholar
- 15.• Saito M, Marumo K. Bone quality in diabetes. Front Endocrinol (Lausanne). 2013;4:72. Saito et al. provide evidence to suggest that advanced glycation end products (AGEs) could explain the molecular link between primary osteoporosis and diabetes. This may help to explain why type 2 diabetes is characterized by normal or high BMD, but is associated with an increased risk of fracture. Google Scholar
- 29.Hansen S, Hauge EM, Beck Jensen JE, et al. Differing effects of PTH 1-34, PTH 1-84, and zoledronic acid on bone microarchitecture and estimated strength in postmenopausal women with osteoporosis: an 18-month open-labeled observational study using HR-pQCT. J Bone Miner Res. 2013;28(4):736–45.PubMedCrossRefGoogle Scholar
- 33.Mueck FG, Michael L, Deak Z, et al. Upgrade to iterative image reconstruction (IR) in MDCT imaging: a clinical study for detailed parameter optimization beyond vendor recommendations using the adaptive statistical iterative reconstruction environment (ASIR) Part2: the chest. Rofo. 2013;185(7):644–54.PubMedCrossRefGoogle Scholar
- 35.Chevalier Y, Quek E, Borah B, et al. Biomechanical effects of teriparatide in women with osteoporosis treated previously with alendronate and risedronate: results from quantitative computed tomography-based finite element analysis of the vertebral body. Bone. 2010;46(1):41–8.PubMedCrossRefGoogle Scholar
- 42.Bauer JS, Monetti R, Krug R, et al. Advances of 3T MR imaging in visualizing trabecular bone structure of the calcaneus are partially SNR-independent: analysis using simulated noise in relation to micro-CT, 1.5 T MRI, and biomechanical strength. J Magn Reson Imaging. 2009;29(1):132–40.PubMedCrossRefGoogle Scholar
- 48.Chestnut 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(9):1548–61.Google Scholar
- 57.Boehm HF, Raeth C, Monetti RA, et al. Local 3D scaling properties for the analysis of trabecular bone extracted from high-resolution magnetic resonance imaging of human trabecular bone: comparison with bone mineral density in the prediction of biomechanical strength in vitro. Invest Radiol. 2003;38(5):269–80.PubMedGoogle Scholar
- 67.• Patsch JM, Burghardt AJ, Yap SP, et al. Increased cortical porosity in type 2 diabetic postmenopausal women with fragility fractures. J Bone Miner Res. 2013;28(2):313–24. Patsch et al. conducted a study on postmenopausal diabetic (type 2) women with and without fragility fractures using DXA and micro-FEM analysis derived from HR-pQCT. Diabetic women with fractures exhibited significantly higher cortical porosity and pore-related deficits in stiffness, failure load and cortical load fraction at the ultradistal and distal tibia, and the distal radius than diabetic women without fractures. Google Scholar
- 85.• Grimal Q, Grondin J, Guerard S, et al. Quantitative ultrasound of cortical bone in the femoral neck predicts femur strength: results of a pilot study. J Bone Miner Res. 2013;28(2):302–12. Grimal et al. used quantitative ultrasound (QUS) of the femoral neck in correlation with DXA and QCT measurements to investigate the relation of QUS parameters to the geometric and material properties of the cortical shell (cortical thickness, tissue elasticity and porosity). QUS was significantly correlated with cortical QCT variables, indicating new possibilities to assess bone strength with QUS even at deep locations such as the femoral neck. Google Scholar
- 89.• Pfeiffer F, Bech M, Bunk O, et al. Hard-X-ray dark-field imaging using a grating interferometer. Nat Mater. 2008;7(2):134–7. Pfeiffer et al. describe a novel technique termed dark-field imaging using conventional X-ray tubes in combination with a grating interferometer, which may potentially allow new insights into bone micro-structure, providing complementary and otherwise inaccessible structural information at the micrometer and submicrometer length scale. Google Scholar