Drs. Rhodes and Tan contributed equally to this work.
Regional variation and differential response to therapy for knee synovitis adjacent to the cartilage–pannus junction and suprapatellar pouch in inflammatory arthritis: Implications for pathogenesis and treatment
Article first published online: 5 AUG 2004
Copyright © 2004 by the American College of Rheumatology
Arthritis & Rheumatism
Volume 50, Issue 8, pages 2428–2432, August 2004
How to Cite
Rhodes, L. A., Tan, A. L., Tanner, S. F., Radjenovic, A., Hensor, E. M. A., Reece, R., O'Connor, P., Emery, P. and McGonagle, D. (2004), Regional variation and differential response to therapy for knee synovitis adjacent to the cartilage–pannus junction and suprapatellar pouch in inflammatory arthritis: Implications for pathogenesis and treatment. Arthritis & Rheumatism, 50: 2428–2432. doi: 10.1002/art.20444
- Issue published online: 5 AUG 2004
- Article first published online: 5 AUG 2004
- Manuscript Accepted: 22 APR 2004
- Manuscript Received: 2 FEB 2004
- Medical Research Council, UK
- Arthritis Research Campaign, UK
To use magnetic resonance imaging (MRI) to investigate the importance of knee joint synovitis at the cartilage–pannus junction (CPJ) in rheumatoid arthritis (RA) as compared with synovitis at a distant site in the suprapatellar pouch (SPP) and as compared with CPJ synovitis in the spondylarthropathies (SpA), and to assess the relative response of knee joint synovitis to therapy at the CPJ and SPP sites.
Dynamic contrast-enhanced MRI (DEMRI) of actively involved knee joints in 24 patients (13 with RA and 11 with SpA) was undertaken. The area of synovitis was calculated at the CPJ and SPP regions of interest in patients with RA and in patients with SpA. Differences in CPJ and SPP synovitis were determined using calculated DEMRI parameters which included the initial rate of contrast enhancement (IRE) and the maximal enhancement (ME). Changes in the synovial area at the CPJ and SPP were also measured in 10 patients with early RA, following treatment with disease-modifying antirheumatic drugs (DMARDs) (either methotrexate or leflunomide).
In patients with RA or SpA, the area of synovitis was significantly larger immediately adjacent to the CPJ compared with a distant site at the SPP (in RA, mean synovitis area 162 mm2 at the CPJ versus 114 mm2 at the SPP [P = 0.010]; in SpA, mean synovitis area 214 mm2 at the CPJ versus 143 mm2 at the SPP [P = 0.002]), but the differences in the areas of synovitis at these sites were not significant between the RA and SpA patients. The IRE and ME values were also higher at the CPJ compared with the SPP, both in the RA patients (IRE P = 0.054, ME P = 0.018) and in the SpA patients (IRE P = 0.002, ME P = 0.001). A larger reduction in the area of synovitis was seen at the SPP compared with the CPJ following DMARD therapy in the RA patients (mean reduction 35% at the SPP [P = 0.023] and 12% at the CPJ [P not significant]).
The non–disease-specific variations in synovitis and the differential responses to therapy in RA patients have implications for improving our understanding of CPJ synovitis. The results suggest that the pathophysiologic events at the CPJ reflect common anatomic, immune system, or biomechanical factors that play a role in modulating the severity of arthritis, and these events are not specific to RA since the same process was observed in other arthritides.
The disease processes at the cartilage–pannus junction (CPJ) are regarded as pivotal in the pathogenesis of rheumatoid arthritis (RA) (1). Indeed, the demonstration of tumor necrosis factor at the CPJ region, and not just the synovium in general, was considered an essential step toward the development of anti–tumor necrosis factor blockade in RA (2). Inflamed tissue at the CPJ is associated with cartilage and bone destruction, and one study suggested that synovitis adjacent to the CPJ differs from that in the suprapatellar pouch (SPP), with synovitis in the CPJ region being characterized by relatively greater macrophage infiltration (3).
The arthroscopic and histologic studies that are the basis of these observations on the CPJ have limitations. Arthroscopy visualizes only the surface and not the depth of synovitis. Moreover, it does not adequately visualize pathologic processes in the adjacent bone. Likewise, synovial biopsies sample only small regions of superficial synovium, and an estimate of the amount of inflamed synovial tissue is not possible. Furthermore, these techniques are unable to provide functional information about the synovium. Most importantly, the studies that support the central importance of the CPJ in RA were carried out in RA alone without reference to other types of inflammatory arthritis, and therefore the presumed specificity of these observations in RA has not been validated.
The application of magnetic resonance imaging (MRI) in inflammatory arthritis has considerably increased the understanding of the disease; however, MRI has not been used to address the importance of CPJ synovitis in the pathogenesis of RA. MRI parameters of synovitis have been shown to correlate with the extent of synovitis at different sites (4, 5). Unlike histologic studies, dynamic contrast-enhanced MRI (DEMRI) is capable of estimating not only the volume of synovitis, but also the degree and kinetics of enhancement (6), including the initial rate of contrast enhancement (IRE) and maximal enhancement (ME), both of which may reflect synovial tissue vascularity and permeability.
This study used MRI to investigate the importance of synovitis adjacent to the CPJ of the knee joint in RA by comparing it with synovitis at a distant site in the SPP and by comparing it with CPJ synovitis in the spondylarthropathies (SpA). The relative responses of CPJ and SPP synovitis to therapy were also determined. Our findings revealed a predilection for synovitis at the CPJ region that is not disease specific. Furthermore, CPJ synovitis was found to be more difficult to suppress following therapy. These results have implications for improving our understanding of CPJ synovitis, and in particular, for developing the concept that the pathophysiologic events at the CPJ site reflect common anatomic, immune system, or biomechanical factors that play a role in modulating the severity of arthritis, rather than being a feature specific to RA.
PATIENTS AND METHODS
Twenty-four patients with early disease were recruited consecutively for this study. Thirteen of the patients fulfilled the American College of Rheumatology (formerly, the American Rheumatism Association) criteria for RA (7) (mean age 58 years, 5 female, 8 male, 69% rheumatoid factor positive) and 11 fulfilled the European Spondylarthropathy Study Group criteria for SpA (8) (mean age 41 years, 6 female, 5 male, all seronegative for rheumatoid factor). The diagnoses in the SpA group were 6 with psoriatic arthritis, 4 with reactive arthritis, and 1 with undifferentiated SpA. All patients had synovitis of at least 1 knee joint. The mean disease duration was 13.6 months in the patients with RA and 9.7 months in the SpA patients. The mean C-reactive protein level was 39 mg/liter in the patients with RA and 38 mg/liter in the SpA patients.
MRI and image processing.
MRI of the knee was performed using a Philips 1.5T Gyroscan ACS-NT whole-body scanner (Philips Medical Systems, Best, The Netherlands) with a Philips quadranture knee coil. The DEMRI sequence analysis acquired 40 spoiled T1-weighted (T1W) gradient-echo images from each of 5 sagittal slices before, during, and after the bolus injection of the contrast agent gadolinium diethylenetriaminepentaacetic acid (Gd-DTPA). The images were obtained using a repetition time (TR) of 11 msec, an echo time (TE) of 5.3 msec, and a flip angle of 60° that allowed the acquisition of 5-mm–thick slices and a 5-mm gap between slices. The temporal resolution of the DEMRI measurements was 6 seconds for the individual images that constitute the DEMRI data set. Gd-DTPA was administered at a dose of 0.1 ml/kg body weight at 1 second after the acquisition of the first image in the dynamic series.
Commercial software (Analyze; Mayo Clinics, New York, NY) and software developed in-house and described previously (6) were used to calculate values for the ME and IRE on a pixel-by-pixel basis. The calculated values of IRE and ME were displayed as color overlays on conventional images of the knee joint anatomy (Figure 1).
Measurements of MRI parameters within the synovium at the CPJ and proximal SPP.
Measurements were made on the midline sagittal slice at 2 different sites, one at the superior pole of the patella immediately adjacent to the CPJ and a second at a remote site in the proximal SPP. Two rectangular regions of interest, one positioned alongside the CPJ and the second within the SPP, were defined using the image analysis software Analyze. Each region had identical superior–inferior dimensions (15 pixels) but variable anterior–posterior dimensions to accommodate the variable synovitis thickness in the sagittal plane (Figure 1). An investigator who was blinded to the diagnosis (LAR) carried out the data analysis.
MRI-determined bone erosion at the CPJ.
The frequency of bone damage (erosion or bone edema) in early RA and SpA adjacent to the CPJ of the knee joint at the superior patellar pole was investigated to determine whether larger areas of synovitis in larger joints, such as the knee, equated with more erosive disease. Bone erosion at the superior pole of the patella CPJ was assessed using T1W sagittal (TR 580 msec, TE 14 msec), T2 fat-suppressed (FS) sagittal (TR 5,500 msec, TE 100 msec), and T1 FS post–Gd-DTPA axial (TR 400 msec, TE 14 msec) MRI, with erosion being scored as present if evident in 2 adjacent slices.
Response of CPJ and SPP synovitis to therapy.
To determine whether CPJ synovitis was more difficult to suppress as compared with SPP synovitis, the area of synovitis at the CPJ and SPP regions of interest in 10 of the 13 RA patients (mean age 58 years, range 27–72 years, 5 female, 5 male, 70% rheumatoid factor positive), who received either methotrexate (n = 6) or leflunomide (n = 4), was determined pre- and posttreatment. Prior to treatment, the mean disease duration in this group was 9.0 months and the mean C-reactive protein level was 36.7 mg/liter. A comparison was made of the percentage reduction in the area of synovitis at the 2 sites at 4 months posttreatment.
A paired t-test was used to compare the differences between the CPJ and SPP in the RA and SpA patients. Comparisons between synovial volume, IRE, and ME measurements between the CPJ and SPP within each disease group required multiple t-tests to be performed on data that were potentially related. P values less than or equal to 0.02 were therefore considered significant. Although image analysis is fully automated and, as such, is 100% reproducible, the delineation of regions of interest was subjective and was therefore analyzed for intra- and interobserver agreement. Interobserver reliability and intraobserver reproducibility were assessed using intraclass correlation coefficients (ICC[3,1] and ICC[1,1], respectively) (9). Within the framework of a one-way analysis of variance, hypothesis tests were performed to assess whether agreement exceeded predefined minimum values that were not merely zero (10). Sample sizes were calculated with the 2-tailed α level set at 0.05 and with 80% power to show ICCs >0.8 for interobserver reliability and ICCs >0.9 for intraobserver reproducibility. The intraobserver reproducibility was thus calculated from 5 repeated measurements (conducted by LAR) in all 24 patients, and the interobserver reliability was determined from measurements repeated by another observer (ALT) in 19 patients.
The interobserver reliability for the delineation of the regions of interest was significantly better than the desired minimum value (0.8) both at the CPJ (ICC[3,1] 0.95, 95% confidence interval [95% CI] 0.88–0.98, n = 19; F = 4.49, P < 0.001) and at the SPP (ICC[3,1] 0.96, 95% CI 0.90–0.99, n = 18; F = 5.50, P < 0.001). Similarly, the intraobserver reliability was significantly higher than the desired minimum value (0.9) for identifying the regions of interest both at the CPJ (ICC[1,1] 0.98, 95% CI 0.96–0.99, n = 24; F = 12.29, P < 0.001) and at the SPP (ICC[1,1] 0.99, 95% CI 0.97–0.99, n = 22; F = 18.52, P < 0.001).
Areas of synovitis at the CPJ and SPP in RA and SpA.
A larger area of synovitis was evident in sites adjacent to the CPJ compared with distant sites at the SPP, both in the patients with RA and in the patients with SpA (in RA, mean synovitis area 162 mm2 at the CPJ and 114 mm2 at the SPP [P = 0.010]; in SpA, mean synovitis area 214 mm2 at the CPJ and 143 mm2 at the SPP [P = 0.002]). In some cases, an up to 3-fold difference between the 2 sites was observed. The area of synovitis at both the CPJ and the SPP was found to be larger in patients with SpA compared with patients with RA; however, this difference was not found to be significant (P = 0.127 at the CPJ, P = 0.308 at the SPP).
IRE and ME measurements of synovitis at the CPJ and SPP.
With regard to the MRI parameters of synovitis, a higher mean ME and higher mean IRE were observed at the CPJ than at the SPP, both in patients with RA and in patients with SpA (Table 1). The mean ME values (in arbitrary units) were 332.5 at the CPJ and 225.7 at the SPP (P = 0.018) in patients with RA, and 509.4 at the CPJ and 311.7 at the SPP (P = 0.001) in patients with SpA. The differences in the IRE values between the CPJ and the SPP reached statistical significance in the patients with SpA and showed a similar trend in the patients with RA (for RA, mean IRE [in arbitrary units] 3.1 at the CPJ and 2.0 at the SPP [P = 0.054]; for SpA, mean IRE 5.6 at the CPJ and 3.2 at the SPP [P = 0.002]).
|Synovitis area, mm2||162||114||0.010||214||143||0.002|
The IRE values at the CPJ were significantly higher in the SpA patients compared with the RA patients and the ME values showed a trend toward significance (P = 0.009 and P = 0.047, respectively). At the SPP region of interest, both the IRE and the ME values also showed a trend toward significance (P = 0.03 and P = 0.096, respectively, compared with the RA patients). Overall, these parameters correlated with the observation that the synovial area at the CPJ was larger than at the SPP, and also that the synovial area in the knee joints of the SpA cohort was larger than that in the knee joints of the RA patients (Table 1).
Frequency of bone erosion at the CPJ in RA and SpA.
Bone erosion was evident at the CPJ of the superior pole of the patella in 4 of the 13 patients with RA and 2 of the 11 in the SpA group. Three other patients with SpA had diffuse patellar bone edema consistent with an enthesitis-related pathology.
Differential effect of therapy on CPJ and SPP synovitis in RA.
A larger reduction in synovitis was evident in the SPP as compared with the CPJ in the 10 RA patients after treatment (pretreatment mean synovitis area 134 mm2 at the CPJ and 89 mm2 at the SPP; posttreatment mean synovitis area 118 mm2 at the CPJ and 58 mm2 at the SPP) (Figure 1). The mean percentage reduction in synovitis was 35% at the SPP (P = 0.023) and 12% at the CPJ (P = 0.303).
The purpose of this MRI study was to explore key concepts relating to the importance of CPJ synovitis in the pathogenesis of RA. We found a predilection for synovitis at the CPJ in both the RA and SpA knee joints, suggesting that this observation is not disease specific. Common factors, which could be either mechanical or immune system–related, or both, may have an effect on the site-preferential distribution of synovitis in the knee.
There was also a smaller reduction of synovitis at the CPJ compared with the SPP after treatment with disease-modifying antirheumatic drugs (DMARDs) in the RA patients, indicating that CPJ synovitis is more difficult to suppress. This could have been influenced by the greater degree of synovitis at the CPJ before therapy.
Østergaard et al have previously shown that relapse following intraarticular corticosteroid therapy in knee joint synovitis was associated with a higher global synovial volume before treatment (11). A greater response to treatment observed in the SPP compared with the CPJ could therefore be therapeutically important. With the advent of biologic therapy involving blockade of key proinflammatory cytokines, a realistic goal for the treatment of RA should be the complete suppression of synovitis. Failure to do so could lead to persistence of activated lymphocytes in pockets of inflamed tissue that could, analogous to minimal residual disease in malignancy, lead to disease relapse. Although such concepts are far from being tested in RA, the demonstration of the differential response of synovitis to DMARD therapy could have important future implications for RA, since the complete ablation of synovitis appears to be a realistic and feasible goal. The findings of this study reveal that the CPJ region consistently has the greatest amount of inflammation in different diseases, and that this is the case even following therapy, thus suggesting that synovial tissue immediately adjacent to the CPJ should be the preferential site for biopsy.
In the present study, the area of synovitis adjacent to the CPJ was considerably larger than that previously noted at the metacarpophalangeal (MCP) joints, reflecting the larger size of the knee joint (12). Nevertheless, bone damage at the CPJ of the knee joint in early RA was not especially common compared with the MCP and wrist joints, which may reflect factors such as different bone cortex thickness and ligament or tendon pressure on the bone at these different sites.
In addition to investigating synovial areas, this study also assessed other MRI measures of synovitis in the CPJ and SPP in RA and SpA knee joints. Rates of contrast enhancement in RA have been shown to correlate with both inflammatory activity and with the numeric density of blood vessels (5). We noted that the IRE and ME parameters at the CPJ and SPP were generally higher in patients with SpA as compared with patients with RA, which is consistent with the idea of higher density and permeability of synovial blood vessels in SpA patients (13–15). Further studies are needed to assess the use of IRE and ME parameters in the diagnosis of RA and SpA.
In conclusion, this study demonstrates that there is a significant variation in synovitis within the knee, and that this variation is not disease specific but may be related to common factors across the spectrum of inflammatory arthritis. Furthermore, the larger area of synovitis adjacent to the CPJ region is more difficult to control following DMARD therapy. These findings have important implications for future studies that seek a better understanding of the disease mechanisms in the inflamed synovial joints.
- 2Localization of tumor necrosis factor receptors in the synovial tissue and cartilage–pannus junction in patients with rheumatoid arthritis: implications for local actions of tumor necrosis factor α. Arthritis Rheum 1992; 35: 1170–8., , , , , , et al.