In vivo high resolution MRI of the calcaneus: differences in trabecular structure in osteoporosis patients

Noninvasive assessment of bone structure

Bone fragility is determined by bone mass and trabecular structure. While bone mass can be readily measured as bone density, bone trabecular structure cannot be easily assessed by currently available methods. The realization of the importance of bone structure in determining fracture risk has led to the development of several imaging modalities aimed at evaluating the contribution of bone quality to its biomechanical strength and fragility. High-resolution magnetic resonance imaging and computed tomography have limited spatial resolution and high cost but have a potential to generate true three-dimensional images of trabecular structure in vivo. Bone radiographs subjected to various forms of texture analysis have higher resolution and lower cost but provide only a two-dimensional representation of bone structure. Both two- and three-dimensional methods have been shown to predict biomechanical strength in vitro and to differentiate between subjects with and without fractures in vivo. Therefore, all of these methods deserve closer evaluation and also need further technical improvements before they can be considered for use in clinical practice.

Trabecular bone structure of the calcaneus: preliminary in vivo MR imaging assessment in men with osteoporosis

PURPOSE

To use magnetic resonance (MR) imaging to evaluate potential differences in bone structure between men with and men without osteoporosis.

MATERIALS AND METHODS

Sagittal MR images of the calcaneus were obtained in 50 men (26 patients with osteoporosis and 24 age-matched healthy control subjects). Osteoporosis was defined as a low bone mineral density (at least 2.5 SDs below the normal value for young adults at either the lumbar spine or proximal femur) as measured with dual-energy x-ray absorptiometry. Seventeen patients had a history of osteoporotic fractures. For each participant, 10 consecutive sagittal three-dimensional gradient-echo MR sections were analyzed by using a rectangular region of interest. Twenty structural measurements were obtained from these images. Additionally, density measurements at the calcaneus were obtained in 46 participants. The significance of differences between the two groups was calculated by using the unpaired Student t test. The odds ratios for fracture per 1 SD decrease in the control group were calculated with logistic regression analysis. Adjustment for participant weight and height was performed if necessary.

RESULTS

Thirteen of 20 structural parameters, especially connectivity parameters, showed significant differences between control subjects and patients (P <.05). Differences between the two groups were more significant (P <.001) for apparent bone marrow skeleton length, apparent node count, apparent node-to-node strut count, and apparent terminus-to-terminus strut count. Odds ratios for 11 of 13 structural parameters but not for calcaneus density were significant (P <.05). After adjustment for calcaneus density, these parameters were still significant predictors of osteoporotic fracture.

CONCLUSION

Structural measurements derived from MR images of the calcaneus may be used in vivo to characterize trabecular bone architecture in men with osteoporosis.

Trabecular rod thickness by direct measurement from 3D SEM anaglyphs

This study presents a methodology for measuring the thickness of trabecular rods directly from anaglyphs. Macerated sagittal slices of T12 vertebral bodies from 15 subjects were examined by scanning electron microscopy (SEM). Two digital images (the second image tilted 5 degrees ) were recorded, and a 3D anaglyph was created. The thickness of the trabecular rods (Tb.Th((rods))), and the anatomical orientation of the trabecular rods were measured using an image analyser. Conventional 2D histomorphometry was performed on adjacent bone slices. A total of 1559 rod measurements were made from the 15 vertebral bone slices, with a mean Tb.Th((rods)) of 123 +/- 36 microm. The rod thickness in males (128 +/- 34 microm) was significantly greater than that in females (119 +/- 37 microm, P < 0.001). Tb.Th((rods)) changed significantly with age in the males: the thicker rods in the younger men reduced with age to a thickness similar to that in women. 3D measurements were significantly larger than the 2D estimates, and there was no correlation between the two methods of measurement. An inverse correlation was found between the number of rods and the bone volume fraction (BV/TV), indicating that decreased BV/TV is associated with an increased number of rods. The vertical rods (132 +/- 39 microm) were significantly thicker than the horizontal rods (116 +/- 33 microm, P < 0.001). The determination of rod numbers, and their orientation and individual thicknesses enables a greater understanding of cancellous bone architecture in both individuals and populations, and will allow more reliable finite element modelling. Direct measurements from 3D anaglyphs of intact specimens provide new data that show previously unrecognised age- and sex-related changes.

Prediction of frequency-dependent ultrasonic backscatter in cancellous bone using statistical weak scattering model

The goal of this study was to propose a model for the ultrasonic frequency-dependent backscatter coefficient in cancellous bone. This model allows us to address the inverse problem and to predict the mean trabecular thickness. A weak scattering model is used and the backscatter coefficient is expressed in terms of an autocorrelation function of the medium. Different autocorrelation functions (Gaussian, exponential and densely populated media) were used to compute the backscatter coefficient and comparison is made with experimental data for 19 specimens and for frequency ranging from 0.4 to 1.2 MHz. For each specimen, a nonlinear regression was performed and the mean trabecular thickness is estimated. Experimental data and theoretical predictions were averaged over the 19 specimens. A good agreement between experimental data and predictions was found for both the magnitude and the frequency-dependence of the backscatter coefficient. We also found a good agreement between the experimental mean trabecular thickness (Tb. Th = 130 +/- 6.5 micro m) derived from the analysis of bone 3-D microarchitecture using high-resolution microtomography and theoretical predictions (d(Gauss) = 140 +/- 10 micro m, d(exponential) = 153 +/- 12.5 micro m and d(dense) = 138 +/- 6.5 micro m). These results open interesting prospects for the estimation of the mean trabecular thickness from in vivo measurements.

Quantitative magnetic resonance imaging in the calcaneus and femur of women with varying degrees of osteopenia and vertebral deformity status

Quantitative magnetic resonance imaging (QMRI) allows measurement of two parameters that are related to the integrity of the trabecular bone: R2*, the rate constant of the free induction signal, and trabecular bone volume fraction (BVF), the counterpart of apparent density. In this work, R2* and BVF were measured in 68 women (mean age, 58.2 +/- 9.5 years) of varying spinal bone mineral density (BMD) T scores (mean, -1.37 +/- 1.54) and vertebral fracture status on a commercial 1.5 T whole-body imager using customized image acquisition and processing techniques. Twenty-five of the patients had vertebral fractures, characterized by the total cumulative deformity burden exceeding 200%. R2* was measured in the calcaneus and proximal femur and BVF could be measured in the calcaneus only. On a pixel-by-pixel basis, calcaneal R2* and BVF within each subject were highly positively correlated (r2 = 0.61 +/- 0.11) but the correlation of region-of-interest (ROI) means for different calcaneal sites among patients was weaker (r2 = 0.34; p < 0.0001). The strongest discriminator of vertebral deformity was R2* of the calcaneus, which was lower in the fracture group, consistent with lower trabecular density. Among the calcaneal sites examined, the subtalar region, a location characterized by dense nearly horizontal trabeculae that transmit the stresses imparted by body weight from the tibia to the heel, best discriminated the two groups (p = 0.0001), with 77% diagnostic accuracy as determined from the area under the receiver operating characteristic (ROC) curve (compared with 66% for vertebral BMD). The cavum calcanei, an anterior site of low trabecular density, and the tuber calcanei (the location ordinarily used for ultrasound measurements) also had significantly reduced R2* in the fracture group (p < 0.005 and p = 0.01, respectively). The R2av*, computed as the average of all pixels in the calcaneus, was a strong discriminator as well (p < 0.005). On the other hand, calcaneal BVF was only marginally discriminating (p = 0.05). Among the BMD sites examined, the lumbar spine (average L1-L4) was significant (p = 0.005, 66% diagnostic accuracy), as was the femoral neck (p = 0.01). The data suggest the calcaneus to be suited as a surrogate site to assess vertebral osteoporosis and that R2* is sensitive to alterations in bone quality not captured by density.

Local differences in the trabecular bone structure of the proximal femur depicted with high-spatial-resolution MR imaging and multisection CT

RATIONALE AND OBJECTIVES

The authors performed this study to investigate structural variations in the trabecular bone of the proximal femur at high-resolution magnetic resonance (MR) imaging and high-resolution multisection computed tomography (CT).

MATERIALS AND METHODS

Bone mineral density (BMD) was measured in 36 proximal human femur specimens by using dual x-ray absorptiometry. High-resolution MR imaging was performed at 1.5 T with an in-plane spatial resolution of 0.195 x 0.195 mm and a section thickness of 0.3 and 0.9 mm. Multisection CT was performed with an ultra-high-resolution protocol; images were obtained with an in-plane spatial resolution of 0.25 mm and a section thickness of 1 mm. In a subset of these specimens, micro CT was performed with an isotropic spatial resolution of 30 microm. Identical regions of interest (ROIs) were used to analyze images obtained with MR imaging, multisection CT, and micro CT. Trabecular bone structural parameters were obtained, and the parameters from the individual imaging modalities and BMD were correlated.

RESULTS

Significant differences concerning the trabecular microarchitecture between the individual ROIs were demonstrated with multisection CT and MR imaging. A number of the correlations between structural parameters derived with multisection CT, MR imaging, micro CT, and BMD measurements were significant. For MR imaging, threshold technique and section thickness had an effect on structural parameters.

CONCLUSION

Structural parameters obtained in the proximal femur with multisection CT and high-resolution MR imaging show regional differences. These techniques may be useful for depicting the trabecular architecture in the diagnosis of osteoporosis.

High-resolution magnetic resonance imaging to assess trabecular bone structure in patients after transplantation: a review

After organ transplantation patients have a higher incidence of osteoporotic fractures. Bone mineral density (BMD) measurements to assess fracture risk are of limited value in these patients. On the other hand, structure-based techniques have shown promise. In this review, the use of high-resolution magnetic resonance imaging in the analysis of the trabecular bone structure in patients before and after renal and cardiac transplantation cross-sectionally is presented. The analyses of calcaneal trabecular structure were compared with BMD with regard to the prediction of therapy-induced bone loss and osteoporotic fracture status. Sagittal and axial T1-weighted spin-echo sequences with a voxel size of 0.2 x 0.2 x 1 mm were performed at 1.5 T and structure measures analogous to bone histomorphometry were calculated. In addition, fracture status of the spine and of the peripheral skeleton was assessed. Structure measures showed significant differences between healthy controls and patients before and after renal and cardiac transplantation (p < 0.01). Osteoporotic fractures were found in approximately 35% of the transplant patients; the percentage was higher in the cardiac transplants. Structure measures and BMD were lower in patients with fractures; differences were more significant in the cardiac transplant patients. Using receiver operating characteristic analyses the diagnostic performance in differentiating patients with and without fractures was highest when BMD and structure measures were combined. Structure measures performed better than BMD in the cardiac transplant patients, whereas results were comparable in the renal transplant patients. In conclusion, structure measures determined in high-resolution magnetic resonance images may be useful in assessing changes of trabecular bone after organ transplantation and may improve the prediction of fracture risk.

Three-dimensional-line skeleton graph analysis of high-resolution magnetic resonance images: a validation study from 34-microm-resolution microcomputed tomography

The resolution achievable in vivo by magnetic resonance imaging (MRI) techniques is not sufficient to depict precisely individual trabeculae and, thus, does not permit the quantification of the "true" trabecular bone morphology and topology. Nevertheless, the characterization of the "apparent" trabecular bone network derived from high-resolution MR images (MRIs) and their potential to provide information in addition to bone mineral density (BMD) alone has been established in studies of osteoporosis. The aim of this work was to show the ability of the three-dimensional-line skeleton graph analysis (3D-LSGA) to characterize high-resolution MRIs of trabecular bone structure. Fifteen trabecular bone samples of the distal radius were imaged using the high-resolution MRI (156 x 156 x 300 microm3) and microcomputed tomography (microCT; 34 x 34 x 34 microm3). After thresholding, the 3D skeleton graph of each binary image was obtained. To remove the assimilated-noise branches of the skeleton graph and smooth this skeleton graph before it was analyzed, we defined a smoothing length criterion (l(c)), such that all "termini" branches having a length lower than l(c) were removed. Local topological and morphological LSGA measurements were performed from MRIs and microCT images of the same samples. The correlations between these two sets of measurements were dependent on the smoothing criterion l(c), reaching R2 = 0.85 for topological measurements and R2 = 0.57-0.64 for morphological measurements. 3D-LSGA technique could be applied to in vivo high-resolution MRIs of trabecular bone structure, giving an indirect characterization of the microtrabecular bone network.

Subvoxel processing: a method for reducing partial volume blurring with application to in vivo MR images of trabecular bone

Partial volume blurring precludes accurate measurement of structural dimensions in the limited-resolution regime in which image voxel size is larger than the typical structural element to be resolved. Since acquiring images at increased resolution often exacts an unacceptable signal-to-noise ratio (SNR) penalty, methods to alleviate the adverse effects of partial volume blurring are instrumental for the accurate measurement of architectural parameters in applications such as predicting the mechanical competence of trabecular bone networks. In the current work, a novel post-processing method, referred to as "subvoxel processing," is described for increasing apparent image resolution. The method is applicable to volumes of interest containing material phases of two discrete signal intensities. The principal strategy consists of subdividing voxels and assigning voxel intensities to each subvoxel on the basis of local neighborhood criteria and strict mass conservation. In the current work, the method's accuracy has been evaluated using microcomputed tomography images (22 x 22 x 22 microm(3) voxel size) of human trabecular bone. The results demonstrate that subvoxel processing is significantly more accurate than trilinear interpolation in decreasing apparent voxel size, especially in the presence of noise. In addition, the method's effectiveness is illustrated with MR images of human trabecular bone acquired in vivo at 137 x 137 x 350 microm(3) voxel size. The subvoxel-processed images are shown to have architectural features characteristic of images acquired at higher spatial resolution.

Ultrasonic characterization of human cancellous bone using transmission and backscatter measurements: relationships to density and microstructure

The present study was designed to evaluate the relationships between ultrasonic backscatter, density, and microarchitecture of cancellous bone. The slopes of the frequency-dependent attenuation coefficient (nBUA), ultrasound bone velocity (UBV), the frequency-averaged backscatter coefficient (BUB) were measured in 25 cylindrical cancellous bone cores. Bone mineral density (BMD) was determined using X-ray quantitative computed tomography. Microarchitecture was investigated with synchrotron radiation microtomography with an isotropic spatial resolution of 10 microm. Several microstructural parameters reflecting morphology, connectivity, and anisotropy of the specimens were derived from the reconstructed three-dimensional (3D) microarchitecture. The association of the ultrasonic variables with density and microarchitecture was assessed using simple and multivariate linear regression techniques. For all ultrasonic variables, a strong association was found with density (r = 0.84-0.90). We also found that, with the exception of connectivity, all microstructural parameters correlated significantly with density, with r values of 0.54-0.92. For most microstructural parameters there was a highly significant correlation with ultrasonic parameters (r = 0.33-0.91). However, the additional variance explained by microstructural parameters compared with the variance explained by BMD alone was small (Delta r(2) = 6% at best). In particular, no significant independent association was found between microstructure and backscatter coefficient (a microstructure-related ultrasonic parameter) after adjustment for density. The source for the unaccounted variance of quantitative ultrasound (QUS) parameters remains unknown.

Reconsidering the effects of antiresorptive therapies in reducing osteoporotic fracture

Concepts of what constitutes osteoporosis have evolved from the single criterion of low bone mass to a more inclusive consideration of bone strength, based on both quantity and quality. The evidence driving this shift is drawn from many sources. For example, recent studies of bone geometry have shown what engineers have always known: material properties and structural strength are inseparable. Genetic factors also argue against a one-dimensional (ID) view of osteoporosis. Large-scale family studies present a strong case for genetic influences on bone mass and predisposition to fracture. The contribution of aging to fracture risk has long been known, but we are only now beginning to understand what happens to bone remodeling and microstructure in an aging skeleton. The recognition that osteoporosis is far more complex than previously thought suggests that factors in addition to bone mineral density (BMD) may be useful for evaluating bone fragility and therapeutic effectiveness. Although assessment of BMD is noninvasive and widely available, the degree of increase in BMD alone fails to account for the broader effectiveness of antiresorptive agents in reducing the risk of fractures related to osteoporosis. Indeed, the very multiplicity of factors that determine fracture risk implies that response to therapy may be equally complex. Studies of response to antiresorptive agents and the cellular processes they induce are at best preliminary at this time. Although new technologies have been applied to studying bone microarchitecture, their invasive nature limits wide use. New methods are needed to provide insight into the causes and effects of bone fragility. The definition of osteoporosis, meanwhile, must still be considered a work in progress.

Digital topological analysis of in vivo magnetic resonance microimages of trabecular bone reveals structural implications of osteoporosis

Osteoporosis is a disease characterized by bone volume loss and architectural deterioration. The majority of work aimed at evaluating the structural implications of the disease has been performed based on stereologic analysis of histomorphometric sections. Only recently noninvasive imaging methods have emerged that provide sufficient resolution to resolve individual trabeculae. In this article, we apply digital topological analysis (DTA) to magnetic resonance microimages (mu-MRI) of the radius obtained at 137 x 137 x 350 microm3 voxel size in a cohort of 79 women of widely varying bone mineral density (BMD) and vertebral deformity status. DTA is a new method that allows unambiguous determination of the three-dimensional (3D) topology of each voxel in a trabecular bone network. The analysis involves generation of a bone volume fraction map, which is subjected to subvoxel processing to alleviate partial volume blurring, followed by thresholding and skeletonization. The skeletonized images contain only surfaces, profiles, curves, and their mutual junctions as the remnants of trabecular plates and rods after skeletonization. DTA parameters were compared with integral BMD in the lumbar spine and femur as well as MR-derived bone volume fraction (BV/TV). Vertebral deformities were determined based on sagittal MRIs of the spine with a semiautomatic method and the number of deformities counted after threshold setting. DTA structural indices were found the strongest discriminators of subjects with deformities from those without deformities. Subjects with deformities (n = 29) had lower topological surface (SURF) density (p < 0.0005) and surface-to-curve ratio (SCR; a measure of the ratio of platelike to rodlike trabeculae; p < 0.0005) than those without. Profile interior (PI) density, a measure of intact trabecular rods, was also lower in the deformity group (p < 0.0001). These data provide the first in vivo evidence for the structural implications inherent in postmenopausal osteoporosis accompanying bone loss, that is, the conversion of trabecular plates to rods and disruption of rods due to repeated osteoclastic resorption.

Osteodensitometry of human heel bones by MR spin-echo imaging: comparison with MR gradient-echo imaging and quantitative computed tomography

The aim of the study was to investigate whether quantitative magnetic resonance (MR) fast spin-echo (FSE) imaging with moderate spatial resolution enables osteodensitometry in peripheral yellow bone marrow. Signal intensities in T1-weighted FSE images from yellow bone marrow indicate the amount of adipose tissue per volume. The signal intensity in marrow regions with spongy bone was assessed and compared to signal intensity of pure fatty marrow (100%). Heel bones of 30 patients with suspected osteoporosis were analyzed and the FSE images were compared with results from parallel MR gradient-echo (GE) imaging and quantitative computed tomography (QCT) examinations. High correlation was found between FSE imaging and QCT [r = 0.91 in the dorsal region of interest (ROI); r = 0.86 in ventral ROI]. Linear correlation coefficients between GE imaging and QCT were slightly lower in the dorsal part (r = -0.86) and considerably lower in the ventral part (r = -0.68). Correlation between the two MR techniques amounted to r = -0.72/-0.61 (dorsal/ventral). The high correlation between FSE imaging and bone mineral density (BMD) allows possible clinical applications of FSE imaging for diagnosis of osteoporosis. Further improvements of the accuracy using reference phantoms might be possible.

Trabecular structure assessment in lumbar vertebrae specimens using quantitative magnetic resonance imaging and relationship with mechanical competence

The purpose of this study was to use quantitative magnetic resonance imaging (MRI; high-resolution [HR] and relaxometry) to assess trabecular bone structure in lumbar vertebrae specimens and to compare these techniques with bone mineral density (BMD) in predicting stress values obtained from mechanical tests. Fourteen vertebral midsagittal sections from lumbar vertebrae L3 were obtained from cadavers (aged 22-76 years). HR images with a slice thickness of 300 microm and an in-plane spatial resolution of 117 microm2 x 117 microm2 were obtained. Transverse relaxation time T2' distribution was measured by using an asymmetric spin-echo (ASE) sequence. Traditional morphometric measures of bone structure such as apparent trabecular bone fraction (app. BV/TV), apparent trabecular bone number (app. Tb.N), apparent trabecular bone separation (app. Tb.Sp), and apparent trabecular bone thickness (app. Tb.Th) as well as the directional mean intercept length (MIL) were calculated. Additionally, BMD measurements of these sections were obtained by dual-energy X-ray absorptiometry (DXA) and biomechanical properties such as directional stress values (to fracture) were determined on adjacent specimens. With the exception of T2', all morphological parameters correlated very well with age, BMD, and stress values (R between 0.79 and 0.92). However, in the direction perpendicular to the magnetic field, T2' values enhanced the adjusted R2 correlation value with horizontal (M/L) stress values in addition to BMD from 0.70 to 0.91 (p < 0.05).

Autocorrelation analysis of bone structure

We propose a method called spatial autocorrelation analysis (SACA) to determine the spatial anisotropy of the trabecular bone in order to investigate osteoporosis. For demonstrating the potential of SACA we first evaluate the method on rectangular, simulated test patterns as a simple model for the anisotropic pore structure of the bone. As a next step towards biomedical application, photographic reference images of human vertebral bone were investigated by SACA. Osteoporotic bone structure could be clearly differentiated from non-osteoporotic sample images. Moreover, for demonstration of the applicability and potential of the method for in vivo characterization of osteoporosis, the microstructure of the human calcaneus was investigated by MR-microimaging on a young healthy male subject and an osteoporotic female. The measurements were performed using a high-field (3T) whole-body MR tomograph equipped with a special, strong head gradient system. The signal was acquired with a surface coil mounted on an in-house-built device for convenient immobilization of the subject's foot. Using a 3D gradient echo sequence a resolution of 0.254 x 0.254 x 2.188 mm3 was achieved in vivo. Selected images were inverted, gradient corrected for the inhomogeneous but sensitive detection by the surface coil, and subsequently analyzed by SACA. The anisotropy of bone structure detected by SACA is a possible candidate for noninvasive determination of the osteoporotic status, potentially complementing standard bone mineral density measurements.

Trabecular bone volume fraction mapping by low-resolution MRI

Trabecular bone volume fraction (TBVF) is highly associated with the mechanical competence of trabecular bone. TBVF is ordinarily measured by histomorphometry from bone biopsies or, noninvasively, by means of high-resolution microcomputed tomography and, more recently, by micro-MRI. The latter methods require spatial resolution sufficient to resolve trabeculae, along with segmentation techniques that allow unambiguous assignment of the signal to bone or bone marrow. In this article it is shown that TBVF can be measured under low-resolution conditions by exploiting the attenuation of the MR signal resulting from fractional occupancy of the imaging voxel by bone and bone marrow, provided that a reference signal is available from a marrow volume devoid of trabeculation. The method requires accurate measurement of apparent proton density, which entails correction for various sources of error. Key among these are the spatial nonuniformity in the RF field amplitude and effects of the slice profile, which are determined by B(1) field mapping and numerical integration of the Bloch equations, respectively. By contrast, errors from variations in bone marrow composition (hematopoietic vs. fatty) between trabecular and reference site are predicted to be small and usually negligible. The method was evaluated in phantoms and in vivo in the distal radius and found to be accurate to 1% in marrow volume fraction. Finally, in a group of 12 patients of varying skeletal status, TBVF in the calcaneus was found to strongly correlate with integral bone mineral density of the lumbar vertebrae (r(2) = 0.83, p < 0.0001). The method may fail in large imaging objects such as the human trunk at high magnetic field where standing wave and RF penetration effects cause intensity variations that cannot be corrected. Magn Reson Med 46:103-113, 2001.

Future methods in the assessment of bone mass and structure

There have been major advances in the diagnosis of osteoporosis over the last few decades not only in the definitions that are now used but also in the technology that is available. The future will see further development of the techniques currently in common clinical use, such us dual energy X-ray absorptiometry and quantitative ultrasound. In addition new techniques for assessing bone structure, including MRI and fractal analysis of X-rays, may add significantly to our understanding of the pathophysiology of osteoporosis and to the prediction of fracture risk.

Does the trabecular bone structure depicted by high-resolution MRI of the calcaneus reflect the true bone structure?

RATIONALE AND OBJECTIVES

The purpose of this study was to compare trabecular bone structure parameters assessed with high-resolution magnetic resonance imaging (HR-MRI) with those determined in specimen sections.

METHODS

High-resolution MR images were obtained for 30 calcaneus specimens with a three-dimensional, T1-weighted spin-echo sequence (spatial in-plane resolution 0.195 mm, slice thicknesses of 0.3 and 0.9 mm). Thirty-eight sections were obtained from the specimens, and contact radiography was performed. In the corresponding sections, structural parameters analogous to bone histomorphometry were determined.

RESULTS

Significant correlations between MRI-derived structural parameters and those derived from macro pathological sections were found: r values of up to 0.75 were obtained (P < 0.01). The highest correlations were found for apparent bone volume/total volume and trabecular thickness. Image thresholding techniques showed a significant impact on these correlations (P < 0.01). The thinner MR sections were less susceptible to the different thresholding algorithms.

CONCLUSIONS

Trabecular bone structure depicted by HR-MR images is significantly correlated with that shown in macro sections (P < 0.01); however, a number of limitations have to be considered, including the substantial impact of thresholding techniques and slice thickness.

Changes in calcaneal trabecular bone structure after heart transplantation: an MR imaging study

PURPOSE

To use high-spatial-resolution magnetic resonance (MR) imaging to analyze the trabecular bone structure of the calcaneus in patients before and after heart transplantation and to compare this technique with bone mineral density (BMD) measurement in predicting therapy-induced bone loss and vertebral fracture status.

MATERIALS AND METHODS

High-spatial-resolution 1.5-T MR imaging of the calcaneus was performed in 40 men 11-120 months after heart transplantation, in 11 men before heart transplantation, and in 10 age-matched male volunteers. Sagittal and transverse T1-weighted spin-echo images with a voxel size of 0.195 x 0.195 x 1.000 mm were obtained, and structure measurements analogous to bone histomorphometric values were calculated. In addition, the BMD of the lumbar spine was determined in the transplant recipients pre- and postoperatively by using quantitative computed tomography, and vertebral fracture status was assessed.

RESULTS

Significant differences in structure and BMD measurements were found between patients before and after heart transplantation (P <. 05). In 17 (42%) of 40 transplant recipients, vertebral fractures were found. Although structure measurements were significantly different between patients with and those without fractures (P <.05), BMDs were not. Correlations between time after transplantation and some structure measurements were moderately significant (P <. 05), but such correlations with BMD measurements were not.

CONCLUSION

MR imaging-derived structure measurements in the calcaneus are useful for monitoring bone changes after heart transplantation and assessing vertebral fracture status.

In vivo assessment of trabecular bone structure using fractal analysis of distal radius radiographs

Our purpose in this study was (i) to measure trabecular bone structure using fractal analysis of distal radius radiographs in subjects with and without osteoporotic hip fractures, and (ii) to compare these measures with bone mineral density (BMD) as well as with measures of trabecular bone structure derived from high resolution magnetic resonance (MR) images. Distal radius radiographs were obtained using semi-industrial films (55 kVp, 400 mAs) in 30 postmenopausal patients, who had suffered osteoporotic hip fractures (74.8+/-8.2 years) in the last 24 months and 27 postmenopausal age-matched (74.6+/-6.6 yr) normal volunteers. Radiographs were digitized at 50 microm. A Fourier power spectrum-based fractal dimension (FD) characterizing the trabecular pattern was measured in a region of interest proximal to the joint line. The fractal dimension was calculated over two spatial frequency (f) ranges: FD1 was calculated over 0.5<log(f)<l.0, FD2 over the higher range 1.0<log(f)<1.5. Trabecular BMD in the radius was obtained using peripheral quantitative computed tomography (pQCT) (Stratec GmbH, Germany). In addition BMD of the proximal femur was determined using dual x-ray absorptiometry (DXA) (QDR 2000, Hologic, MA). In a subset of patients (16 controls and 18 with hip fractures), high resolution MR imaging of the distal radius (spatial resolution of 156 x 156 x 500 microm) was used to obtain measures analogous to bone histomorphometry. There were significant differences (p<0.05) between the fracture and nonfracture groups in the total femur BMD (13%), trabecular BMD in the distal radius (4%), and the fractal dimension in the radiographs (FD2) (3%). The correlations between FD2 and the total femur BMD as well as trabecular bone BMD in the distal radius were -0.48 (p<0.006) and -0.22 (p<0.33); respectively; FD1 increased with BMD and showed lower correlations. FD2 showed good correlations with App. Tb.N (-0.71) and App. Tb.Sp (0.69) (p<0.01), moderate correlation with App BV/TV (-0.53) (p<0.05), and no significant correlation with App. Tb.Th. The correlations between structural measures and FD1 showed the inverse trend and were typically lower. The odds ratios for a hip fracture were 2.44 for total femur BMD, 1.5 for trabecular BMD (radius), and 1.5 for FD2, respectively. In summary, the fractal measures derived from radiographs of the radius show differences between subjects with and without hip fractures, the predictive power of measures in the distal radius are comparable to radial trabecular BMD but lower than that of total hip BMD.

Does high-resolution computed tomography image analysis of the distal radius provide information independent of bone mass?

This study aimed to investigate the usefulness of computed tomography (CT) image analysis of the distal radius for comparing two groups of postmenopausal women matched for age and bone mineral density at both the lumbar spine and femoral neck. The first one consisted of 16 women with at least one vertebral fracture and the second consisted also of 16 women without disease affecting bone mass or bone metabolism. Eight slices were selected in each patient: four consecutive coronal slices and four consecutive axial slices. Bone texture analysis was performed using structural methods leading to the measurement of 24 features. Most of the structural variables derived from histomorphometric parameters and were measured after segmentation from a binary or a skeletonized image. Nine variables were significantly different between the two groups on axial slices: valley number, valley surface area, apparent bone volume/tissue volume (BV/TV), apparent trabecular separation, apparent trabecular number, trabecular bone pattern factor, trabecular skeletal length, node count, and node-to-node strut count. Also four variables were significantly different between osteoporotic women and controls on coronal slices: apparent BV/TV, trabecular partition, node-to-node strut count, and terminus-to-terminus strut count. In conclusion this study suggests that bone texture analysis could yield additional data on bone mass for explaining bone strength and therefore could be used for improving the prediction of fracture risk.

In vivo comparison between computed tomography and magnetic resonance image analysis of the distal radius in the assessment of osteoporosis

In a prospective case-control cross-sectional study, we investigated the usefulness of both computed tomography (CT) and magnetic resonance (MR) image analysis of the distal radius for distinguishing a small sample of fractured osteoporotic women from control women regardless of bone mineral density. The study population included 12 subjects who were divided into two groups according to their bone status. The first group consisted of six women with at least one vertebral fracture occurring in the absence of high-energy trauma, and the second group consisted of six women without disease affecting bone mass or bone metabolism. Cross-sectional and coronal slices were obtained from both CT and MR systems. For CT images, the slice thickness was 1000 jim and the plane resolution was approx 200 jim x 200 jim. MR images were obtained from a 1.5-T imager with a two-dimensional spin-echo Ti-weighted sequence leading to a slice thickness of 2000 jim and a plane resolution of 195 jim x 195 jim. Bone texture analysis was performed using fractal and structural methods leading to the measurement of 23 features. Most of the structural variables derived from histomorphometric parameters and were measured after segmentation from a binary or a skeletonized image. Bone densitometry was measured by dual-energy X-ray absorptiometry both at the lumbar spine and the nondominant femoral neck. On cross-sectional slices, 12 variables, mainly obtained from structural analysis, were significantly different between the two groups for CT images (p < 0.05) against two variables only for MR images (p < 0.05). The number of variables statistically different between the two groups was significantly higher for CT images compared with MR images (p = 0.003). In the same way, odds ratios for fracture per 1 control group standard deviation decrease were significant for 10 variables on CT images, whereas, in contrast, none of the variables measured on MRI images led to significant odds ratios. The results obtained for the two methods on coronal slices were poorer without a difference between either CT or MR images in terms of discrimination between fracture and nonfracture subjects. In conclusion, this study suggests that bone texture analysis obtained from CT compared with MRI offers a best discrimination between controls and osteoporotic patients, probably the result to the spatial resolution. which is better for CT than for MR images.

Comparison of ultrasound and bone mineral density assessment of the calcaneus with different regions of interest in healthy early menopausal women

This study investigated the effect of different sized regions of interest (ROIs) on quantitative ultrasound (QUS) variables of the calcaneus. The effect on QUS of using a fixed ROI as opposed to an ROI adjusted for foot length was also assessed. Eighty Caucasian women, aged 50-57 yr (mean 53 +/- 2) who were healthy and within 0. 5-5 yr of the onset of menopause participated in this study. Using the QUS-1(trade mark) Ultrasonometer (Metra Biosystems, Mountain View, CA), we assessed broadband ultrasound attenuation ([BUA] and UBI-4, dB/MHz), the average transit time through the heel ([TTH], mus) and a multiple-factor index (UBI-4T = UBI-4/TTH, dB/[MHz. mus]). The QUS measurement results were calculated from three different sizes of ROI as well as one in a fixed location and one adjusted for foot size. Bone thickness, bone width, bone mineral content ([BMC], g/cm), bone mineral density area ([BMD(a)], g/cm(2)), and bone mineral density volume ([BMD(v)], g/cm(3)) were measured by single-energy photon absorptiometry. Lateral radiography of the foot was used to ensure the QUS scanning location in a subgroup. The results showed that there was a 1.4-5.9% difference in QUS parameters among different ROIs (p = 0.076-0.001). No significant differences between fixed and adjusted location were found regarding the mean values of QUS. The correlation between the fixed and adjusted locations was very strong, although there was a 12-42% unexplained variation. On the other hand, QUS in the size-adjusted ROI increased the correlation with BMC/BMD compared to the fixed QUS assessments. After controlling for body weight and height, a significant correlation between QUS and bone mass variables remained, and in some cases correlations became stronger. Lateral radiography showed that when using a fixed location to scan a large foot, the scanning area might be close to the bone edge, an area of higher BMD and potential acoustic artifacts. When scanning a small foot, the scanning area was confined to the middle of the calcaneus. Our results indicate that bone size has a modest effect on BUA. There is a better correlation with BMD when the measurement region is appropriately located in the calcaneus. This suggests that measurement location based on foot size may improve the accuracy of the measurements, resulting in good diagnostic sensitivity.