To use high-resolution magnetic resonance imaging (MRI) to evaluate the trabecular bone structure in the distal femur and the proximal tibia and its to correlate the findings with different stages of osteoarthritis (OA) of the human knee.
To use high-resolution magnetic resonance imaging (MRI) to evaluate the trabecular bone structure in the distal femur and the proximal tibia and its to correlate the findings with different stages of osteoarthritis (OA) of the human knee.
Axial images of the distal femur and proximal tibia were obtained at 1.5 T in patients without and with mild OA and with severe OA. The spatial resolution was 195 × 195 μm2 with a 1-mm slice thickness. Apparent measures of trabecular bone volume fraction (BV/TV), trabecular number (Tb.N), trabecular separation (Tb.Sp), and trabecular thickness (Tb.Th) were calculated.
Significant differences existed in the trabecular bone structure of the femur and tibia. Differences in trabecular bone structure between the tibia and the femur decreased with the degree of OA. The apparent BV/TV, Tb.N, and Tb.Sp in the femoral condyles could be used to differentiate healthy patients or patients with mild OA from patients with severe OA (P < 0.05). Among individuals, the structural variation of the lateral and medial femoral condyle was indicative of the extent of the disease.
High-resolution MRI of the knee joint can provide a noninvasive assessment of trabecular bone structure. Trabecular bone structure, determined by high-resolution MRI, shows significant variation in patients with varying degrees of OA. The impact of OA on trabecular bone is different in the tibia than in the femur, and this difference depends on the extent of the disease.
Osteoarthritis (OA) is a multifactorial disease characterized by the progressive loss of articular hyaline cartilage and the development of altered joint congruency, subchondral sclerosis, intraosseous cysts, and osteophytes. It affects about 14% of the adult population (1) and is the second most common cause of permanent disability among subjects over the age of 50 years (2). In addition to changes in articular cartilage that occur in OA, it has been suggested that early changes are seen in the adjoining subchondral and trabecular bone (3).
In a guinea pig model, Layton et al (4) observed with microscopic computed tomography (CT) that an initial loss of trabecular bone volume fraction and thinning of trabeculae was followed, in the advanced stages of OA, by an increase of trabecular bone volume fraction via eventual thickening of trabeculae. In a canine OA model induced by anterior cruciate ligament (ACL) transection, Dedrick et al (5) demonstrated an increase in subchondral bone thickness, accompanied by a decrease in trabecular thickness. Using magnification radiographs of humans, Lynch et al (6) showed an increase in the horizontal trabecular thickness of the tibia in early OA, followed by an increase in vertical connectivity of trabeculae in advanced disease. More recently, Buckland-Wright and colleagues (7) showed that the structural changes in the trabecular bone microarchitecture in patients with ACL ruptures are detectable by fractal analysis of radiographs, well before joint space narrowing and other radiologic changes are detectable.
Imaging and assessment of OA are based primarily on plain radiography (8). While radiographic changes reflect the pathologic changes in the bone and joint space, these changes do not generally correlate with the severity of pathologic joint destruction. Furthermore, conventional radiographs only depict gross osseous changes directly, and such changes tend to occur late in the disease. Early changes in the articular cartilage and other articular tissues are not directly visible. Cartilage loss can only be inferred from the development of joint space narrowing. CT provides cross-sectional images of the affected joint, and thus diminishes the problem of overlapping structures. However, CT is limited in terms of soft-tissue contrast and imaging planes.
Magnetic resonance imaging (MRI) is ideal for monitoring OA. It is nonionizing, offers multiplanar capabilities, has high spatial resolution, and provides superior depiction of soft tissue detail. In addition to evaluation of articular cartilage volume, thickness, and degeneration (9–14), recent advances have made it possible to use MRI to assess bony and soft tissue changes. In a guinea pig model, results showed that MR images of trabecular bone accurately reflected the degree of osteopenia and the development of subchondral sclerosis and osteophytes (15). In vitro studies performed on small cubic specimens have shown that some of the parameters derived from high-resolution MR images contributed to the prediction of biomechanical properties of bone (16). In the last decade, numerous studies have also demonstrated the feasibility of MR micro-imaging of trabecular structure in humans in vivo at a resolution adequate for visualizing trabecular bone (17, 18).
The assessment of human trabecular bone structure using MRI techniques showed that it was possible to quantify the trabecular architecture and derive such measures as trabecular width, trabecular bone volume fraction, and mean intercept length, as well as quantitative measures of texture, such as the fractal characteristics of the trabecular bone network (19–21). Majumdar et al (20) derived structural parameters in premenopausal healthy women and in postmenopausal women with osteoporosis and correlated the data with peripheral quantitative CT bone mineral density (BMD) and spinal fracture status. Although MR techniques may result in partial volume effects because the spatial resolution is comparable with trabecular dimensions, Kothari et al (22) have shown that these effects may be minimized when the lower resolution of the slice is selected along the direction of primary trabecular orientation. In this study, dedicated MR techniques were used in vivo to characterize and quantify variations in the trabecular bone structure along the distal femur and proximal tibia of OA patients and normal subjects. Differences of structure between the tibia and femur in joints of patients with different stages of OA were also investigated.
The distal femur and the tibial plateau of 28 subjects divided into 3 groups were investigated. Group I included 10 young healthy subjects (6 men and 4 women; age 129 ± 4.9 years) with no knee impairment. The other 2 groups consisted of patients with OA, classified according to the radiography-based Kellgren/Lawrence (K/L) scale (23). Group II included 8 patients with mild OA (1 man and 7 women; age 68 ± 9.1 years) with K/L scores of 1–2, and group III included 10 patients with severe OA (4 men and 6 women; age 70 ± 6.3 years), with K/L scores of 3–4. Reproducibility was assessed in 4 healthy patients (2 men and 2 women; age 32 ± 5.8 years), with repositioning between 3 measurements.
The examinations were performed in accordance with the rules and regulations of the University of California, San Francisco Human Research Committee. Informed consent was obtained from all patients after the nature of the examinations had been fully explained.
Imaging. The images were acquired on a Signa 1.5T- echo-speed system (General Electric Medical Systems, Milwaukee, WI) equipped with gradients operating at 2.2 G/cm and a rise time of 184 μsec. To ensure accuracy in subject positioning as well as to reduce motion artifacts, the leg of each subject was placed on a dedicated holder (manufactured in-house) and immobilized using Velcro straps. A bilateral dual-phased array coil (USA Instruments, Cleveland, OH) consisting of 4 elements was used to image the knee joint. Coil positioning was critical to maximize the signal-to-noise ratio. A sagittal localizer was used to identify the 2 regions of interest, the proximal tibia and the distal femur, in the axial plane (Figure 1A).
High-resolution MR images were obtained using a 3-dimensional (3-D) fast gradient-echo sequence (17) with a partial echo acquisition (echo time [TE] 4.5 msec, repetition time [TR] 30 msec, 40° flip angle, ±15.6 kHz bandwidth). A total of 60 images (1 mm thick) were obtained in the axial plane with a field of view (FOV) of 10 cm and an imaging matrix size of 512 × 384 pixels, corresponding to a reconstructed spatial resolution of 195 × 195 μm2. Two sets of 60 slices were collected to cover the knee joint, as shown in Figure 1. The scan time was ∼12 minutes per acquisition, amounting to a total of 24 minutes for the entire anatomic coverage.
In addition, routine clinical sequences were used for radiologic grading purposes. Coronal 3-D T1-weighted images (TR 25 msec, TE 3 msec, FOV 10 cm, slice thickness 2 mm, matrix 512 × 256 × 32) were used to detect osteophytes and sclerosis. Sagittal T2-weighted (TR 3 sec, TE 70 msec) and T1-weighted (TR 500 msec, TE 11 msec) images were obtained to detect the presence of edema, subchondral cysts, and joint effusion, and to assess the state of ligaments and menisci. Based on these clinical scans, 2 radiologists (TML and LS) identified by consensus whether the lateral or medial side was most affected by the disease. The total time of a typical imaging session was 60 minutes.
Image analysis. The reconstructed volumetric data were transferred to a Sun workstation (Sun Microsystems, Mountain View, CA). Images were analyzed using processing software developed at our laboratory using IDL (Research Systems, Boulder, CO) programming language.
Since the images were acquired with a set of surface coils, the reception profile was inhomogeneous, leading to an intensity inhomogeneity in the image (Figures 1B and D). Prior to a quantitative analysis of the trabecular bone structure, a 3-D low-pass (LP) filter–based correction algorithm was applied (24). The corrected image was obtained by dividing the initial image by an LP-filtered image obtained from the original using Gaussian k-space filtering. In the ideal case, choice of an appropriate filter bandwidth allowed us to remove variations due to the coil reception profile without loss of anatomic details (Figures 1C and E).
The first (proximal) and last (distal) 5 slices of the volume were eliminated from the analysis to minimize artifacts from slice selection profile imperfection (Figure 1A). Regions of interest (ROIs) fitting the trabecular bone and marrow regions in both the distal femur and the proximal tibia were drawn manually over 35–45 slices of each data set (Figures 2A and B). Individual ROIs were clustered in 4 distinct groups for statistical analysis. Region R1 included ROIs selected in the proximal tibia, starting at the level of the subchondral bone. Region R2 included ROIs in the femur, starting from the point where the medial and lateral condyles merged and going along the shaft. Regions R3 and R4 included ROIs drawn in the medial femoral condyle and lateral femoral condyle, respectively (Figures 2C and D).
In order to use the standard stereologic techniques and quantify the trabecular bone network, images were thresholded and segmented into bone and marrow phases. The problem of segmenting 3-D images, where the image resolution approximates that of the trabeculae, is a critical issue. The threshold depends significantly on the imaging modality and may vary depending on the operator. Based on a dual reference limit and assuming a biphasic model, as previously validated (20), a global threshold was calculated and the images were binarized into a bone phase and a marrow phase. The segmented binary image was used to compute (per slice and ROI) trabecular bone parameters, such as the apparent trabecular bone volume fraction (BV/TV), apparent trabecular number (Tb.N), apparent trabecular separation (Tb.Sp), and apparent trabecular thickness (Tb.Th) analogous to standard bone histomorphometry (25). (All parameters discussed in this study are apparent parameters.) The average time to perform the structure analysis of each volume was 15–20 minutes on an Ultra-Sparc 1 (Sun).
For each region, spatial variations of the structure parameters along the slice direction were obtained to determine localized differences and regional variations, starting at the joint line and extending into the shaft. The average values for regions R1–R4 were calculated for statistical analysis. All statistical computations were performed using JMP software (SAS Institute, Cary, NC). Differences between the individual groups were evaluated using Student's 2-tailed t-tests of significance. P values less than 0.05 were considered significant.
As shown in Figures 1C and E, the MR images clearly depicted the trabecular bone microarchitecture in the femur and the tibia. Coil intensity correction improved the signal homogeneity in the plane of the images. Due to coil sensitivity dropoff on the axial plane, this was a crucial first step to quantitative analysis.
Trabecular structure variations as a function of the distance from the joint line. For each volunteer, the analysis of the trabecular structure varied substantially within each ROI and between the femur and the tibia, especially when evaluated in progression from the metaphysis to the epiphysis. Representative graphs of trabecular spacing, thickness, and number in a young, healthy volunteer are shown in Figure 3, from the shaft to the subchondral region toward the joint line for the tibia and from the subchondral region to the shaft away from the joint line for the femur. In the tibia, the shaft showed a decrease in the Tb.Sp up to the growth plate and an increase in the Tb.N, while the Tb.Th remained approximately constant. A valley in the Tb.Sp and a peak in the Tb.N marked the growth line. In contrast, in the femur, a peak in the Tb.Sp and a valley in the Tb.N were observed at the growth line. Progression from the growth line to the shaft yielded a decrease in the Tb.Sp and an increase in the Tb.N.
Difference in trabecular bone structure between the tibia and the femur. Mean values were calculated per patient group for each morphologic parameter in the tibia (region R1) and the femur (region R2); these are shown in Figure 4. The BV/TV, Tb.N, and Tb.Th were all typically higher in the femur, and the Tb.Sp was lower than in the tibia.
The absolute differences in the mean parameter values between the tibia and the femur are summarized in Table 1. The absolute value was chosen so that the magnitude of the difference could be evaluated. The significance of the differences observed between the tibia and the femur was established using Student's paired t-test. When compared with groups I and II, patients with severe OA (group III) had substantially lower absolute differences in mean values for all parameters, with the exception of Tb.Th. There were significant differences (P < 0.05) between the tibia and the femur for patients from groups I and II for all parameters. For group III, significant difference was obtained only for BV/TV.
Differences in trabecular bone structure between controls and patients with OA. The morphologic parameter variations as a function of group are shown in Figure 4. Based on data reported in Table 2, a significant nonmonotonic variation was observed in the tibia for BV/TV with disease progression, while the other parameters showed no specific trend. Based on a group of healthy patients, no correlation was found between measures of the structure and age.
|I and II||0.040†||0.106||0.012||0.087†|
|II and III||0.032†||0.048||0.018||0.053|
|I and III||0.006||0.058||0.007||0.033|
|I and II||0.041†||0.093||0.015||0.064†|
|II and III||0.007||0.048||0.018||0.023|
|I and III||0.034†||0.140†||0.003||0.087†|
Absolute differences and significance of mean parameter values between the 3 groups are presented in Table 2. The analysis showed that mean values for BV/TV and Tb.Sp were significantly different (P < 0.05) between groups I and II in both the tibia and the femur and between groups I and III in the femur. Additionally, BV/TV differences between groups II and III in the tibia and Tb.N differences between groups I and III in the femur were significant. No significant differences in Tb.Th were seen between any groups.
Differences in trabecular structure between the lateral and medial femoral condyles. In regions R3 and R4, corresponding respectively to the medial and lateral femoral condyles (Figure 2), the mean values of morphologic parameters were also calculated. For each derived trabecular bone structure parameter, the 3 subject groups were compared using the average and absolute difference of the population mean, calculated in the 2 condyles. Graphs illustrating the results are shown in Figure 5.
For patients from group III, the BV/TV in the condyle with the most severe degenerative changes was higher than the average femoral value (region R2), while the BV/TV was lower in the least diseased condyle. However, as seen from Figures 4 and 5, the average value of the 2 femoral condyles and the mean values obtained in region R2 of the femur were nearly the same. As seen from the graphs, the absolute difference for all morphologic parameters was found to be significantly higher (P < 0.05) for group III compared with groups I and II, which had no significant structural differences between the condyles.
In addition to analyses of the mean and difference, the variations in bone structure when progressing from the joint line to the shaft of the femur were also studied in regions R3 and R4. In a normal femoral condyle (as graded by radiologists [TML and LS]), with increasing distance from the joint line, an increase in BV/TV and a decrease in Tb.Sp was observed. A diseased femoral condyle, however, showed a decrease or a very small increase in BV/TV and an increase or a very small decrease in Tb.Sp (Figure 6).
Using linear regression, the slope of the variation in each bone parameter with respect to the distance from the joint line was determined. The results for determined parameters from all 10 patients from group III patients are presented in Figure 6. For BV/TV, all patients with mild OA had a positive slope in both condyles, whereas greater differences (positive and negative slopes) in the condyles of the same patient were measured in the severe OA group. As can be seen in these graphs, there was a correlation between the slope characterizing the BV/TV rate of variation and the side most affected by disease, as shown by the results of the clinical readings of the MR images. Thus, the femoral bone structure variations close to the joint line, as quantified by the slope direction, appeared to be indicative of the disease extent in advanced OA patients.
Reproducibility of the technique, including scan and quantification processes, was assessed using the coefficient of variation (26). The reproducibility of the image acquisition and segmentation with quantification of bone structure parameters in the femur (region R2) was 5.4, 2.9, 2.7, and 5.2% for BV/TV, Tb.N, Tb.Th, and Tb.Sp, respectively, and 4.0, 3.3, 1.4, and 4.6% in the tibia (region R1).
In this study, using high-resolution images coupled with image processing, we evaluated the trabecular bone architecture in the proximal tibia and the distal femur in healthy subjects as well as in patients with mild and severe OA. As discussed earlier, knowing that the predominant trabecular orientation in the femur and tibia is along the shaft as a result of the loading function of these long bones, we acquired images in the axial direction in order to reduce the partial volume effects of the slice dimension being thicker than the in-plane resolution (22).
Due to spatial resolution limitations, parameters related to trabecular bone structure suffer from partial volume effects. The structure parameters assessed differ from those derived using histomorphometry, and therefore are considered “apparent” structure parameters. Majumdar et al (27) have shown in an experimental model that as the resolution decreases, there is an overestimation of the bone volume fraction and trabecular thickness as well as an underestimation of trabecular separation. Although absolute MR-derived measures differ from histomorphometry measures, it has been demonstrated that measures of structure derived from MR images contribute to the assessment of trabecular bone strength with good correlations (22).
All quantified morphologic parameters were found to be significantly different between the tibia and the femur in both groups I and II (Table 1). The BV/TV, Tb.N, and Tb.Th are all typically higher in the femur, and the Tb.Sp is lower than in the tibia. We hypothesize that this is due to differences in loading function. The convexity of the distal femur, and the relative flatness of the proximal tibia have specific implications on both local loading environments. Trabecular bone is more dense in the distal femur, where loading forces are concentrated in the 2 condyles, in contrast with the tibia, where loading forces are evenly distributed on the whole surface of the tibial plateau.
For group III, however, only BV/TV was significantly different between the tibia and the femur (Table 1). Except for the Tb.Th, the differences in bone structure between the tibia and the femur decreased with increased severity of OA. From group I to group III, this was mainly due to a reduction in the BV/TV and an increase in the trabecular spacing in the femur (Figure 4). This finding indicates a change in the loading function with the progression of OA, as a result of which, in late OA, the femoral trabecular bone is lost, and this loss is akin to osteopenic changes compared with the baseline status.
Variations in mean structure parameters between subject groups indicated that the MR-derived structure measures BV/TV and Tb.Sp were the most sensitive parameters (Table 2), with significant differences in the tibia as well as in the femur. These parameters were significantly different in both the tibia and the femur between groups I and II, and in the femur between groups I and III (P < 0.05). Studies correlating biomechanical strength of bone cubes from the spine with MR-derived structure measures (28) yielded similar findings, with the best correlations found for BV/TV and Tb.Sp.
The difference in the mean parameter values between groups I and II was 12.3, 6.6, 6.1, and 13.2% for BV/TV, Tb.N, Tb.Th, and Tb.Sp, respectively, in the femur (region R2) and 13.3, 8.2, 5.3, and 15.6% in the tibia (region R1). Accordingly, differences between groups I and III were 10.2, 10.0, 1.2, and 18.0%, respectively, in the femur and 2.2, 4.4, 3.1, and 6.0% in the tibia. These differences are overall a factor of 2 or more greater than the reproducibility of the measures we have demonstrated.
Comparing the absolute differences of the sample mean between the medial and lateral femoral condyles permits us to differentiate between subjects without OA versus those with severe OA as well as between subjects with mild versus severe OA. These results have been obtained with all of the structure parameters, but the differences were most significant for BV/TV, Tb.Sp, and Tb.Th (Figure 5).
Thus, in summary, we found that as a result of loading differences and biomechanical loads, the structure of trabecular bone is different in the tibia compared with the femur. The bone is denser in the femoral condyles and femur in general; however, with disease progression to severe OA, the trabecular bone structure is lost in the femur compared with the tibial structure. The significance of these results will lie in the management of patients with OA, and once the relationship between bone changes and cartilage changes is established, modification of biomechanical loading and gait could ameliorate the course of the disease.
Another finding of this study was that the axial variation of the bone structure in the femoral condyles was indicative of disease extent. There was an asymmetry of the medial and lateral condyles related to the extent of disease in subjects with severe OA. Indeed, in patients from group III, the slope direction as determined by a linear fit of the morphologic parameters BV/TV and Tb.Sp, with respect to the distance from the joint line, was found to be predictive of the side most affected by the disease (Figure 6). Given that the BV/TV values at the point where the 2 condyles merge (region R2) are the same, a negative slope of BV/TV in 1 condyle corresponds to an increased mean BV/TV on that side (most affected) and vice versa. This indicates that a modification of the trabecular structure leading to bone sclerosis appears in the most affected condyle. This result confirms the findings of a previous study performed by Layton et al (4), in which femoral heads from a guinea pig model of OA showed a highly significant increase in bone volume fraction.
Using macroradiographs, Buckland-Wright et al (29) have recently shown in human tibia that ACL rupture leads to thickening of subchondral horizontal trabeculae in the medial tibial compartment of knees. No significant changes were detected in the lateral compartment. But whether the bony changes detected precede those of early OA is still under investigation. Unlike other imaging modalities, MRI offers multiplanar capabilities, is capable of directly visualizing all parts of the joint simultaneously, and provides information regarding the relationship between bone and cartilage. MRI was previously used to monitor bony changes in OA in a guinea pig model (30). The investigators found that trabecular bone accurately reflected the degree of osteopenia. Trabecular thinning was also noted around the cruciate ligament insertion.
Radin and Rose (31) postulated that increased stiffness in the subchondral bone was responsible for the initiation of cartilage damage in OA. A sharp gradient in stiffness in the bone under articular cartilage increases stress in the cartilage layer. In patients with severe OA, the bone volume fraction values were highest in the most diseased femoral condyle. For non-OA patients, the BV/TV was lower near the subchondral bone, in the epiphysis, and increased toward the shaft, while in OA patients, this gradient was reversed. One explanation for the reversal of the gradient in patients with severe OA is initial trabecular bone hypertrophy, which then increases cartilage stresses, leading to OA.
Seven of the 10 subjects with severe OA constituting group III were predominantly affected on the medial side. This is consistent with the mean loading forces in the knee, which do not act in the middle of the joint but are more centered in the medial compartment (32).
The results presented here provide an insight into the amplitude and trend of trabecular bone structure modifications with disease progression. A potential limitation of this study, however, is that the subjects in group I, which was used as the reference group, were healthy, with no knee impairment, but were not age-matched with groups II and III.
Future studies will extend these MRI and analysis methods to both cross-sectional and longitudinal studies for assessment of the progression of OA. The former is necessary to increase the patient population per group, and thereby increase the statistical relevance of the results. Currently both trabecular bone and cartilage data are being acquired on all subjects for the purpose of investigating the dynamics between changes in the trabecular bone architecture and degeneration of cartilage. Establishing the magnitude of these changes and their interrelationships and chronologic progression with cartilage degeneration may assist in understanding the mechanism for postinjury cartilage and joint degeneration. The objective of this study was to establish that high-resolution MRI could provide an efficient technique for early detection of changes in periarticular bone relative to OA.
MRI is an emerging clinical modality that allows investigation of the clinical relationship between bone and cartilage and the onset and progression of OA. In the future, high-resolution MRI of trabecular bone structure may become a very powerful tool in providing quantitative and objective parameters in the early stages of joint disease. More specifically, derived structure parameters from femoral condyles could be a potential marker for evaluating disease progression. Adjusting biomechanical loads, either by modification of gait or by osteotomies, may potentially modulate the changes induced by the progression of OA. From a clinical perspective, the ability to monitor cartilage and bone changes in 3 dimensions would thus have a tremendous impact on overall patient management. Furthermore, if the intrinsic connection between bone and articular cartilage is established, it may be possible to preserve articular cartilage by moderating bone turnover and quality. These applications, although speculative, do indicate the tremendous potential of MR as a means to monitor not only trabecular and subchondral bone and cartilage, but also the whole joint, in OA.
The authors would like to thank Dr. T. K. Tran, Dr. A. Laib, and Dr. A. Shimakawa for helpful discussions, as well as Dr. Vikas Patel, Andrew Burghardt, and Cynthia Frazier for assistance with data analysis.