| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 Department of Diagnostic Radiology, Yale University School of Medicine, 333
Cedar St., PO Box 20804, New Haven, CT 06520-8042.
2 Department of Medical Imaging, Mt. Sinai Hospital and the University Health
Network, Toronto, ON, Canada.
3 Department of Orthopedics and Rehabilitation, Yale University School of
Medicine, New Haven, CT.
Received June 23, 2005;
accepted after revision October 21, 2005.
Address correspondence to A. H. Haims
(Andrew.Haims{at}yale.edu).
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. Three board-certified radiologists blinded to prosthesis type used identical MRI protocols to independently evaluate 21 total knee prostheses: 14 zirconium prostheses and seven cobalt chrome prostheses. The radiologists evaluated the following eight parameters: integrity of the medial and lateral collateral ligaments, the quadriceps and the patellar tendons, presence of a joint effusion, and for evidence of periprosthetic osseous signal abnormality around each of the three components. The reviewers gave their degree of confidence in evaluating each of the findings on a five-point scale, 0 being no confidence and 4 being high confidence in the finding. The degree of confidence was used as the vehicle for comparing the two groups of patients.
RESULTS. The confidence ratings for all MRI variables were significantly higher for the zirconium group than for the cobalt chrome group. The confidence ratings varied less for the zirconium group than for the cobalt chrome group with an SD of 0.45 versus 0.95, respectively. There was greater interobserver agreement in the zirconium group (coefficient of interobserver agreement, 0.82 vs 0.35). The reviewers had the highest degree of confidence when examining for joint effusion in both groups (3.9 for the zirconium group; 3.7 for the cobalt chrome group). The greatest discrepancies between the two groups were in evaluation of periprosthetic osseous signal changes with the greatest difference being between the femoral component of each group with an average confidence rating of 3.3 for the zirconium group and 0.8 for the cobalt chrome group.
CONCLUSION. Reviewers had significantly more confidence, less variability, and greater interobserver agreement in MRI evaluation of periprosthetic structures around zirconium knee prostheses than those around cobalt chrome knee prostheses.
Keywords: arthroplasty • MRI • knee • knee prosthesis
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
After prosthesis placement, clinicians have traditionally been limited in their ability to evaluate patients with knee pain [5]. The most common causes of pain after knee arthroplasty are superficial or deep infection, prosthesis loosening or malalignment, pathologic fractures, and tendon or ligament abnormalities, which can be difficult to discern clinically [6-8]. Many of these complications can be indications for surgical intervention. Therefore imaging is essential for pain evaluation. Imagers traditionally have been limited to conventional radiography, arthrography, and nuclear medicine [5, 9]. Radiography is sufficient for analyzing alignment and prosthesis-bone interfaces but is limited in the identification of soft-tissue anomalies. Moreover, pathologic findings are often nonspecific [9]. Arthrographic findings can be informative, but the procedure is invasive, and the results often are nonspecific. Nuclear medicine can be helpful, but again, the findings are nonspecific [10, 11]. Imaging has moved toward cross-sectional techniques, but beam-hardening artifacts in CT and metal susceptibility artifacts in MRI provide challenges in image interpretation. There have been reports of success in decreasing artifact from metal with specific MRI protocols [12-14]. These results are promising because the current role of MRI is limited in the evaluation of patients who have undergone joint replacement.
There are two main types of knee prostheses. The standard and most commonly used is a cobalt chrome prosthesis consisting of cobalt chrome (cobalt-chromium-molybdenum) femoral components, titanium alloy (titanium-aluminum-vanadium) tibial components, and high-molecular-weight polyethylene patellar components. A second commonly used prosthesis is a zirconium prosthesis consisting of ceramic-surfaced oxidized zirconium femoral components, titanium alloy tibial components, and high-molecular-weight polyethylene patellar components [4]. In our clinical experience, we have found minimal artifact in the evaluation of zirconium prostheses compared with the traditional cobalt chrome types. To our knowledge, decreased metal artifact with zirconium prostheses and improved image quality have not been described in the literature. We evaluated reviewer confidence and interobserver agreement as estimations of image quality in MRI assessment of periprosthetic structures around zirconium and cobalt chrome prostheses in a small patient population.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
In the cobalt chrome group, the ages of the patients ranged from 55 to 82 years (mean, 65.4 years). This group consisted of five women and two men. The indication for surgery was osteoarthritis for all seven patients in this group. The indication for MRI in five patients was knee pain. The other two indications were prosthetic loosening and instability.
All MRI examinations were performed with 1.5-T superconducting magnets (Signa LX Horizon, GE Healthcare) with phased-array extremity coils. The MR sequences and parameters were identical for all of the examinations. The parameters were as follows: sagittal fast spin-echo proton density (TR/TE, 3,000/25; matrix size, 384 x 256; number of excitations, 3; echo-train length, 6; field of view, 16 cm; slice thickness, 3 mm; interslice gap, 0.5 mm; receiver bandwidth, 20.83 kHz); sagittal fast spin-echo T2 (3,300/90; matrix size, 256 x 256; number of excitations, 2; echo-train length, 16; field of view, 16 cm; slice thickness, 4 mm; interslice gap, 1 mm; receiver bandwidth, 31.25 kHz); sagittal fast multiplanar inversion recovery (4,300/40; inversion time, 150 msec; matrix size, 256 x 192; number of excitations, 2; echo-train length, 8; field of view, 16 cm; slice thickness, 4 mm; interslice gap, 1 mm; receiver bandwidth, 15.63 kHz); axial fast spin-echo T2 (2,500/90; matrix size, 256 x 256; number of excitations, 2; echotrain length, 16; field of view, 20 cm; slice thickness, 5 mm; interslice gap, 1 mm; receiver bandwidth, 31.25 kHz); axial fast multiplanar inversion recovery (4,300/40; inversion time, 150 msec; matrix size, 256 x 192; number of excitations, 2; echotrain length, 8; field of view, 20 cm; slice thickness, 4 mm; interslice gap, 1 mm; receiver bandwidth, 15.63 kHz); and coronal fast spin-echo T2 (2,875/90; matrix size, 256 x 192; number of excitations, 2; echo-train length, 16; field of view, 20 cm; slice thickness, 4 mm; interslice gap, 1 mm; receiver bandwidth, 31.25 kHz).
Three board-certified radiologists with training in musculoskeletal imaging and experience ranging from 1 to 20 years were blinded to prosthesis type and independently evaluated 21 total knee prostheses. Fourteen of the prostheses were either standard or high-flex zirconium, and seven were cobalt chrome. The radiologists evaluated for the integrity of the collateral ligaments (normal being thin, low-signal-intensity ligament on T2 or inversion recovery sequences) and of the quadriceps and patellar tendons (normal being low signal intensity on the T2 and inversion recovery sequences without evidence of thickening), for the presence of joint effusion (subjectively more than a physiologic amount of fluid), and for evidence of periprosthetic osseous signal abnormality (defined as high T2 or inversion recovery signal) around each of the three prosthesis components. These structures were evaluated on a binary scale as either normal or abnormal, and the reviewers gave their degree of confidence for each of the findings on a five-point scale, 0 being no confidence and 4 being high confidence. The degree of confidence was used as the vehicle for comparing the two groups.
Means and SDs of the reviewers' confidence ratings for evaluation of the zirconium group and of the cobalt chrome group were obtained. The mean values between the two groups were compared by the Student's t test to determine the statistical significance of the difference between the two populations. In addition, interclass correlation coefficients and coefficients of interobserver agreement for the reviewers' confidence were obtained to evaluate the relative homogeneity among reviewers and assess interrater reliability. In this study, interclass correlation coefficients were calculated by two-way random effects analysis of variance with the consistency definition (as opposed to absolute agreement). Both reviewers and patients were considered random samples, and all reviewers rated each patient [14]. The following classification scheme was used for interclass correlation coefficients: < 0.40 = poor; 0.40-0.59 = fair; 0.60-0.74 = good; > 0.74 = excellent [15]. Coefficient of interobserver agreement, defined as the distance among true values assigned by different reviewers for the same subject, also was calculated [16]. This value indicates how much reviewers vary from one another in evaluation of the same subject [17].
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
|
|
|
|
|
|
The most common findings in the zirconium group were joint effusion (13 patients) and tendon abnormalities (15 tendons). In one patient, a quadriceps tendon tear (Fig. 4) was identified along with periprosthetic low T1- and high T2-weighted signal changes along the tibial component, signs that suggested possible loosening. In addition to the patient with periprosthetic signal changes along the tibial component, another patient had similar periprosthetic signal changes along the patellar component (Figs. 5A, 5B). The most common radiologic finding in the cobalt chrome group was joint effusion, which was seen in all seven patients.
|
|
|
|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Cross-sectional imaging, particularly MRI, may play a pivotal role in the evaluation of patients who have undergone knee replacement. However, the obstacle of metal-related artifacts must be overcome. Techniques and protocols for MR studies that minimize metal artifact from traditional metallic implants have been described [14, 18, 19]. Some reports suggest imaging at lower magnetic field strength and increasing readout gradient field strength (increasing receiver bandwidth) as important variables in decreasing artifacts [13, 18, 20]. Other reports suggest reducing image voxel size, orienting the frequency-encoding direction along the long axis of the device, and using fast spin-echo sequences. Fast spin-echo imaging acquisitions reduce diffusion-related signal loss associated with imaging in the vicinity of metal implants and are immune to the potential artifacts of malrotated signal that occur with multiecho spinecho acquisitions [14, 18, 20]. MRI can also benefit from avoiding the use of spectral fat-saturation techniques and gradient-echo acquisitions because of the severity of metal-related imaging artifacts and signal perturbations associated with these techniques.
With identical imaging parameters for the two groups, we found statistically significant higher confidence in our findings in the use of MRI in the evaluation of eight parameters in knees in a group of patients with zirconium prostheses than we did in a group with cobalt chrome prostheses. The reviewers also had more consistent evaluation of the zirconium group than of the cobalt chrome group as evidenced by a lower variation of confidence (SD, 0.45 vs 0.95) and higher interobserver agreement (coefficient of interobserver agreement, 0.82 for the zirconium group and 0.35 for the cobalt chrome group). The more consistent evaluation with significantly higher confidence suggests improved image quality for the zirconium group. We believe the better image quality was in large part the result of decreased metallic susceptibility artifact.
Our protocol included many techniques for decreasing metallic susceptibility artifact. However, the decrease in metallic artifact is due not only to improvements in imaging protocols but also to the introduction of the ceramic-surfaced oxidized zirconium femoral components with properties that minimize metallic susceptibility. The femoral components of zirconium knee prostheses are made up of an alloy of 97.5% zirconium and 2.5% niobium. The outer layer is oxidized so that the outer 5 µm is made up of oxidized zirconium, which is a ceramic (Oxinium, Smith and Nephew) [4]. The tibial component and the patellar component are composed of a titanium alloy (titanium-aluminum-vanadium) and high-molecular-weight polyethylene, respectively, with a polyethylene insert between the tibial and femoral components. These parts are similar to corresponding components used in cobalt chrome prostheses, and titanium has been shown to produce less artifact than cobalt chrome alloy [21].
We believe that the decrease in metal-related artifacts associated with zirconium prostheses occurs because oxidized zirconium (molar susceptibility, -13.8 x 10-6 cm3/mol-1), which is a ceramic, has molar susceptibilities similar to that of human tissue (bone, -10 x 10-6 cm3/mol-1; muscle, -9x10-6 cm3/mol-1), and the zirconium alloy (zirconium, 120 x 10-6 cm3/mol-1; niobium 208 x 10-6 cm3/mol-1) has molar susceptibilities only slightly greater. The cobalt chrome femoral component is made up of an alloy of four main elements: cobalt, chromium, molybdenum, and nickel. Cobalt, which is the main component of the alloy, is ferromagnetic (as is nickel), but chromium and molybdenum are not, having molar susceptibilities of 167 x 10-6 cm3/mol-1 and 72 x 10-6 cm3/mol-1, respectively [22, 23]. We also believe that the increase in molar susceptibility in the cobalt chrome femoral components not only accounts for an increase in metallic susceptibility and metal-related artifact associated with the femoral component but also causes field inhomogeneity. Field inhomogeneity increases artifact and distortion associated with tibial and patellar components similar to the materials in the zirconium prostheses. This phenomenon would account for our results showing the highest discrepancy in confidence ratings for the periprosthetic osseous signal around the femoral components of the zirconium prostheses compared with the cobalt chrome prostheses and slightly lower discrepancy for the tibial and patellar components. Descriptions of MRI of joint replacements have noted imaging artifacts as being most pronounced surrounding spherical prosthesis components or components of complex 3D geometry, as may be the case with the acetabular component in total hip replacements and the femoral component in total knee arthroplasty [18, 19]. This increase in artifacts is related to complex variations in section thickness and signal loss and image distortions induced by such components at MRI evaluation. In our study, the lowest overall reviewer confidence score for any parameter occurred in evaluation of femoral periprosthetic osseous signal change with standard cobalt chrome prostheses. In contrast, the imaging confidence scores for femoral periprosthetic signal improved significantly in the zirconium prosthesis group, assessment confidence scores approaching those of other parameters evaluated.
Ceramic-surfaced oxidized zirconium femoral components were developed in an attempt to increase component longevity by increasing resistance to roughening, improving biocompatibility, and decreasing friction. Previous biomechanical studies comparing oxidized zirconium with cobalt chrome implants showed that wear of the articular surface was dramatically lower in the oxidized zirconium femoral components, most likely because of higher resistance to roughening. This property led to a decrease in complications related to polyethylene wear and greater prosthetic longevity [24].
There were several limitations to our study. First, the study population was small with only 14 zirconium prostheses and seven cobalt chrome prostheses evaluated. We did not have significant clinical or surgical follow-up on any parameters evaluated. We believe future prospective long-term studies will be performed to evaluate the clinical accuracy of the MRI findings on ceramic-surfaced zirconium prostheses, but our goal was to evaluate the confidence and consistency of the evaluation of these prostheses compared with more conventional knee prostheses. In addition, some of the parameters of our protocol, specifically receiver bandwidth, were not optimized to decrease artifacts from metal implants. However, we chose not to change any parameters during the study to ensure all implants would be imaged with identical parameters.
We believe there are potentially profound implications to our preliminary results not only in evaluating patients with pain after knee replacement surgery but also in the realm of orthopedic oncology. MRI currently is the standard in the evaluation and follow-up of oncologic patients with musculoskeletal tumors. It has been difficult, however, to use this imaging technique for follow-up after insertion of metallic prostheses. Ceramic-surfaced zirconium prosthesis technology with imaging optimization may be an invaluable tool for oncologic patients who need to undergo serial imaging.
In summary, we found increased reviewer confidence and consistency in all variables evaluated for ceramic-surfaced zirconium knee prostheses compared with more traditional cobalt chrome implants. Although we did not have long-term clinical or surgical follow-up results, we believe this increased confidence and consistency will lead to more accurate noninvasive diagnoses and improve clinical decision making in the care of patients with painful zirconium knee prostheses. In addition, we believe improved periprosthetic soft-tissue evaluation will have profound future implications in orthopedic oncology.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  H. Malchau and H. G. Potter How are wear-related problems diagnosed and what forms of surveillance are necessary? J. Am. Acad. Ortho. Surg., July 1, 2008; 16(suppl_1): S14 - S19. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
