Finger joint synovitis in rheumatoid arthritis: quantitative assessment by magnetic resonance imaging

Transcription

1 Rheumatology 1999;38:66 72 Finger joint synovitis in rheumatoid arthritis: quantitative assessment by magnetic resonance imaging M. Klarlund1,2, M. Østergaard1,2 and I. Lorenzen1 1Department of Rheumatology and 2Danish Research Centre of Magnetic Resonance, Hvidovre Hospital, University of Copenhagen, Denmark Abstract Objective. To assess quantitatively, by magnetic resonance imaging (MRI), the synovial membrane volume in second to fifth metacarpophalangeal (MCP) joints in patients with rheumatoid arthritis and healthy controls, and to compare the synovial membrane volumes with a more easily obtained semi-quantitative score for hypertrophic synovial membrane. Patients and methods. MCP joints of the dominant hand of 37 patients and five controls were examined clinically and by MRI. Laboratory assessments were performed. Results. Median synovial membrane volumes were considerably larger in clinically active rheumatoid arthritis ( RA) joints (e.g ml in the second MCP joint) than in clinically inactive joints (0.54 ml ) and control joints (0.04 ml ). Nevertheless, group distributions overlapped and marked volume differences were found within clinically uniform groups. The semi-quantitative score was highly correlated with the synovial volumes (Spearman rho = 0.79; P < ). Synovial membrane volumes were poorly related to the presence of rheumatoid factor and to laboratory markers of inflammation. Conclusion. These findings suggest that synovial membrane volumes, as determined by MRI, in finger joints are related to clinical signs of synovitis, but also that the volumes may vary more than what can be accounted for by the clinical appearances. A semi-quantitative score may be sufficient for more routine purposes. KEY WORDS: Arthritis, Gadolinium, Magnetic resonance imaging, Rheumatoid arthritis, Synovitis, Quantitative. Submitted 17 February 1998; revised version accepted 17 September Correspondence to: M. Klarlund, Department of Rheumatology, Hvidovre Hospital, Kettegaard Allé 30, DK-2650 Hvidovre, Denmark. Precise assessment of disease activity in rheumatoid arthritis (RA) is important for monitoring the efficacy of treatment and for predicting the outcome. Conventional radiography shows only the late signs of preceding syno- vitis, i.e. bone erosions. Moreover, histopathological evid- ence of synovitis and radiographic progression of bone erosions have been seen in patients in whom the disease is assumed to be inactive, according to clinical and laboratory evaluation [1 4]. Thus, there may be discrepancies between clinical, laboratory and radiographic assessment of the disease activity and severity. In the RA joint, the pathology of the synovial membrane may determine the extent of subsequent cartilage and bone erosions [5]. Magnetic resonance imaging (MRI) is noninvasive and provides a unique opportunity to visualize the synovial membrane of the affected joints [6, 7]. Quantification of synovitis in the knee and wrist joints, and a relationship between synovial membrane volume and synovial inflammatory activity, have been demonstrated [8 14]. The finger joints are usually among the first joints to be affected in RA [15], and they are considered to be the best markers of overall joint damage in RA [16]. Clinical or radiographic changes in finger joints form part of the American College of Rheumatology s definition of RA and in proposed definitions of early RA [17, 18]. Recently, MRI of the finger joints has become possible at clinical MRI units, but methods to quantify synovial membrane volumes have not yet been employed for the finger joints. The present study was undertaken to determine the synovial membrane volume in the metacarpophalangeal (MCP) joints of patients with RA and healthy controls. As volume quantification is time consuming, volumes were compared with a more easily obtained semi- quantitative score for synovial hypertrophy. In addition, the synovial membrane volumes were compared with clinical and laboratory assessments of disease activity. Patients and methods Patients Thirty-seven patients with RA, all fulfilling the ARA 1987 revised criteria [17], and five healthy controls were British Society for Rheumatology

2 Quantitative MRI of finger joint synovitis 67 entered into the study. The median age of the patients was 55 yr (range yr) and the median duration of disease was 8 yr (range yr). The median age of the healthy controls was 31 yr (range yr). Of the 37 patients, 25 were positive for IgM rheumatoid factor (RF) and 10 were negative (blood sample values were missing in two patients). Before MRI, the second to fifth MCP joints of the patients were examined clinically for the presence or absence of swelling and tenderness. Standard laboratory tests, including erythrocyte sedimentation rate ( ESR) and serum C-reactive protein (s-crp), were performed. ESR was considered to be elevated if >20 mm/h, and s-crp if >95 nmol/l. The study was conducted in accordance with the Declaration of Helsinki, and a signed informed consent was obtained from each patient. MRI MRI was performed on the second to fifth MCP joints of the dominant hand. This was repeated in three patients, within 4 days, to assess reproducibility. The hands were examined with a receive-only, wrap-around flex coil in a 1.0 T Impact MR unit (Siemens, Erlangen, Germany). The subjects were positioned supine with the arm along the side of the body and a specially designed splint was used to eliminate hand movement during the examination. The MCP joints were imaged with slice thicknesses of 3 mm, a field of view of 120 mm and a matrix size of Continuous axial and coronal T1-weighted spin-echo images (repetition time ms; echo time 15 ms) were obtained before and after i.v. injection of gadolinium (Gd)-DTPA ( Magnevist, Schering, Berlin, Germany) 0.1 mmol/kg body weight. The post-contrast axial images were obtained 3 4 min after the injection of Gd-DTPA, followed by repetition of the coronal images. Synovial membrane volume The computer software XPrime, an image-processing program developed in our department for use on UNIX systems, was used for quantitative determination of the FIG. 1. MRI of the MCP joints of an RA patient. Axial T1- weighted 3 mm spin-echo MR images through the second to fifth MCP joints before i.v. injection of Gd-DTPA. FIG. 2. MRI of the MCP joints of the same RA patient as in Fig. 1. Axial T1-weighted 3 mm spin-echo MR images of the MCP joints, after i.v. injection of Gd-DTPA. Numbers 2 5 correspond to the second fifth MCP joints. FIG. 3. The synovial membrane is outlined (white line) on the post-contrast Gd-DTPA enhanced, T1 weighted SE image (same image as Fig. 2). Synovitis in the second flexor tendon sheath is seen (arrow). synovial membrane volume. Tissue at the correct anatomical localization and showing enhancement on the post-gd-dtpa axial images was considered to be the synovial membrane and was delineated manually in each slice of the second to fifth MCP joints (Figs 1 5). The outlining of the synovial membrane was guided by the subtraction image. The computer automatically calculated the areas and the synovial membrane volume was calculated by summation of the areas of enhanced synovial membrane of each slice. This can be summarized by the formula: Vol = S (Ar ST), where Vol is i the synovial membrane volume, Ar is the area of synovium in the slice (i) and ST is the slice thickness. The quantitative assessment of synovial membrane volume was performed without knowledge of the clinical status of the patient or the results of the semi-quantitative assessment. Synovial membrane hypertrophy score The synovial membrane hypertrophy was scored semi- quantitatively by comparing pre-contrast and post-

3 68 M. Klarlund et al. than two groups, the Kruskal Wallis test (one-way analysis of variance by ranks) was employed. In the case of significance of the Kruskal Wallis test (P < 0.05), the Mann Whitney test (two-sample rank sum test) was used for pairwise comparison of groups. The total synovial membrane volume was defined as the sum of the volumes in the four MCP joints. Statistical correlations were assessed by Spearman s test of rank correlation. A probability value (P) of <0.05 was considered to be statistically significant. FIG. 4. Subtraction image (pre-gd-dtpa image subtracted from post-gd-dtpa image, Fig. 2 Fig. 1). The enhancing synovial membrane and the enhancing tendon sheath appear white. contrast axial and coronal images simultaneously while displayed on the Siemens terminal. The images were displayed with identical window settings, taking advantage of the zooming facility on the graphical interface to assess details. Each MCP joint was scored separately on a scale from 0 to 4, as follows: 0, no enhancement; 1, slight enhancement; 2, marked enhancement of small areas; 3, marked enhancement of moderate-sized areas; 4, marked enhancement of large areas. Statistical methods Non-parametric methods were used. Data are presented as median values with the range. In analyses with more Results a FIG. 5. Photomicrograph of a synovial membrane specimen from the same patient as in Figs 1 4 (third MCP joint) obtained at synovectomy. Marked synovial inflammation, with diffuse mononuclear and polymorphonuclear leucocyte infiltration, fibrin exudation, villous hypertrophy and moderate vessel proliferation, is shown. (a) Haematoxylin and eosin stained; original magnification 25. (b) Haematoxylin and eosin stained; original magnification 250. In this patient, the synovial membrane volumes were 0.74, 1.01, 1.08 and 0.53 ml in the second, third, fourth and fifth MCP joints, respectively. A total of 158 RA and 20 control MCP joints were examined. Ninety-seven RA joints were clinically active (with clinical swelling and tenderness), whereas 61 joints were judged to be clinically inactive (no clinical swelling and tenderness). The synovial membrane volumes determined by MRI ranged from 0.0 to 11.2 ml in the RA joints and from 0.0 to 0.3 ml in the control joints. In each of the four MCP joints, the median synovial membrane volume was markedly larger in the clinically active joints than in the clinically inactive joints, which, in turn, showed larger volumes than in the healthy control joints (Table 1). In the second MCP joint, for instance, the median synovial membrane volume in clinically active, clinically inactive and healthy joints was 0.97, 0.54 and 0.04 ml, respectively. Large variations in volumes were found within clinically uniform groups (e.g. from 0.0 to 11.2 ml in clinically active second MCP joints). Furthermore, in the range of volumes for each group, a considerable overlap was found between clinicb

4 Quantitative MRI of finger joint synovitis 69 TABLE 1. Median synovial membrane volume in MCP joints (ml ) Second MCP Third MCP Fourth MCP Fifth MCP Group ( ) 0.94 ( ) 0.64 ( ) 0.54 ( ) MCP joints with swelling n = 29 n = 29 n = 20 n = 19 or tenderness Group ( ) 0.41 ( ) 0.28 ( ) 0.24 ( ) MCP joints without swelling n = 8 n = 8 n = 17 n = 18 or tenderness Group ( ) 0.04 ( ) 0.01 ( ) 0.07 ( ) Healthy control joints n = 5 n = 5 n = 5 n = 5 Groups ( ) 0.62 ( ) 0.33 ( ) 0.36 ( ) n = 37 n = 37 n = 37 n = 37 Mann Whitney tests Groups vs group * 0.002* 0.02* 0.04* Group 1 vs group 2 (0.5) NS (0.07) NS (0.2) NS 0.04* Group 2 vs group 3 (0.05) NS 0.03* 0.002* (0.09) NS Group 1 vs group * 0.002* 0.02* 0.02* Values are medians (range for each group). n is the number of patients in groups 1 and 2, and the number of healthy controls in group 3. For Mann Whitney tests, statistically significant results are presented as P values with an asterisk and insignificant results with NS. ally active and clinically inactive joints. Accordingly, or between RF-positive and RF-negative patients the synovial volumes of these two groups of patients (P > 0.3). No statistically significant correlation was were only significantly different in the fifth MCP joint. found between the disease duration and the total In all MCP joints examined, the synovial membrane synovial membrane volume (Spearman rho = 0.31; volumes were significantly larger in the RA joints than P = 0.06). healthy control joints ( Table 1). Joint effusions were clearly visible in five joints and were not quantified. There was a statistically highly significant correlation Discussion between the semi-quantitative and the quantitative assessment of the synovial membrane volume in The present study introduces quantitative assessment of each of the four MCP joints (Spearman rho > 0.79; the amount of synovial membrane in the small joints of P < ), and between the sum of the MRI scores the hands of patients with RA. Patients with RA had of the four joints and the total synovial membrane statistically significant larger volumes of synovial mem- volume (Spearman rho = 0.79; P < ). brane in the MCP joints than had a group of healthy To test the reproducibility (inter-mri variation) of controls. Moreover, the median volumes of synovial the methods, 12 joints in three patients were scanned membrane were markedly larger in clinically swollen twice within 4 days. With respect to volume measurevolume determinations are time consuming (taking and tender joints than in clinically inactive joints. As ments, a median absolute inter-mri variation of 0.12 ml (range ml ) was found, corresponding to a ~1.5 h per four joints), the synovial membrane volume median relative variation of 22% (range %). The was compared with a more easily obtained (~15 min) largest relative variations expressed in percentages were semi-quantitative score for the hypertrophic synovial found in joints with small volumes. With respect to the membrane. A highly significant statistical correlation semi-quantitative scoring method (five grades), complete was found, which indicates that the semi-quantitative agreement was found in 66.7%, whereas the maximum method may be applicable in longitudinal studies in disagreement between assessments was one grade (in which changes rather than absolute values are consid- 33.3%). ered more important. The semi-quantitative score was also used twice by The MRI assessment revealed large variations in the the same observer on the same set of images within a synovial membrane volumes within clinically uniform time interval of 2 months in 15 patients (intra-observer groups and marked similarities in the synovial mem- variation). A statistically significant positive correlation brane volumes between different clinical groups. Thus, was found between the two readings in all the individual the clinical presentation of individual MCP joints did joints (Spearman rho > 0.87; P < ). Complete not unequivocally reflect the amount of synovium. It is agreement was observed in 56.7%, whereas a deviation likely that the additional information provided by MRI by one grade was found in 31.7% and a deviation by determination of the volumes represents important two grades in 11.6%. differences in the inflammatory process in the joint. Patients with elevated s-crp (P < 0.006) showed Conventional radiographs may show progressive bone statistically significantly larger total volumes of synovial destruction, even though the disease is clinically quies- membrane than did patients with normal s-crp values cent. Inflamed synovial tissue not detected by clinical (Table 2). There were no significant differences in the examination probably accounts for these progressive synovial membrane volume as regards ESR (P = 0.4), destructive changes [1]. In contrast, MRI may allow

5 70 M. Klarlund et al. TABLE 2. Synovial membrane volumes Second MCP Third MCP Fourth MCP Fifth MCP Total Elevated CRP vs normal NS CRP (<95 nmol/l)a Elevated ESR vs normal NS 0.87 NS 0.82 NS 0.41 NS ESR (>20 mm/h)a RF positive vs 0.44 NS 0.59 NS 0.93 NS 0.36 NS 0.70 NS RF negativea Disease durationb 0.29 (0.07) NS 0.36 (0.03) NS 0.37 (0.02) 0.21 (0.20) NS 0.31 (0.06) NS amann Whitney tests, values are P values, significant at P < bvalues are Spearman s rho (P) for the correlation between disease duration and synovial membrane volume in each MCP joint. early identification and assessment of subclinical and coronal images. However, the general impression synovitis. was that it was easier to assess changes in the synovial Earlier publications concerning determination of the membrane on the axial images than on the coronal synovial membrane volume by MRI have been restricted images. This is probably because axial images of the to larger joints [8 13, 19, 20]. In accordance with the MCP joints give fewer partial volume artefacts, as the results of the present study, larger amounts of synovial synovium is predominantly oriented along the axis of tissue have been found in clinically active knee joints the extremity. than in knees in clinical remission induced by intraarticular In contrast to studies of the knee joints [11, 26], only therapy [9, 19]. The validity of determining a few joints in our study showed clearly visible joint disease activity by MRI of the synovial membrane effusions. This may be because the MCP joints are small volumes has previously been tested in the knee joints by with capsules very resistant to expansion [27]. It is, histological verification [14]. It was found that the however, probable that small collections of fluid will synovial membrane volumes mainly reflect the amount not be recognized since these will show homogeneous of inflammatory vascularized and cell-infiltrated subsynovial enhancement rather than rim enhancement with the tissue, but also that the synovial membrane potential risk of an overestimation of the synovial volume might be affected by the cumulative synovial membrane volume [28]. This could be of importance if proliferative activity, e.g. the amount of pannus formed aetiopathogenic implications were to be made. It could [14]. New techniques, such as needle arthroscopy, may be speculated whether T2-weighted images, in which enable similar evaluations of MRI in finger joints [21]. high signal intensity mainly reflects water content, might A previous follow-up study of the effect of intra-articular be helpful. However, it has previously been stated that glucocorticosteroid injections found that the pre- exact differentiation of inflamed synovium and joint treatment synovial membrane volume was inversely correlated effusion is not possible on T2-weighted images, as with the duration of clinical remission [11]. inflamed tissue also shows high signal intensity [6 ]. Longitudinal studies are needed to clarify whether synovial Thus, further efforts are needed to optimize discriminprognostic membrane volumes in the finger joints are of ation of synovial membrane and joint fluid by the value. development of new pulse sequences [29]. Moreover, Although we found that patients with elevated CRP comparison with other techniques, such as highfrequency (<95 nmol/l ) had significantly larger total synovial ultrasound, may be useful. membrane volumes than patients without elevated CRP, In the present study, pre- and post-contrast images no association was detected between CRP and total were visually compared to delineate areas of enhanced synovial membrane volume in a correlation analysis tissue. An alternative to this method could be a computer across the whole range of CRP values (Spearman program that automatically delineates areas of enhanced rho = 0.3, P = 0.08). This may be explained by the fact synovium in the finger joints [9, 30]. However, volume that the synovial membrane volume of the four finger estimates of markedly heterogeneous objects may be less joints represents very little of the total inflammatory accurate by computerized methods than volume estimates load of all the joints. On the contrary, if all patients of homogeneous objects [31]. Moreover, subtle with elevated CRP are considered together, disease changes in the positioning of the patient during MR activity is more likely to occur in the finger joints of scanning will disturb most automated ways of compar- these patients than in patients with normal CRP. ESR ing pre- and post-contrast images. Registration programs was also found to be poorly related to the inflammatory to correct these problems are still not available status of the patients. for use in the finger joints. Thus, human perception is Semi-quantitative assessment has been used to evaluate probably still superior to computerized automated the synovial membrane in the finger joints [22 25], methods. but these findings have not previously been compared MRI may be a more precise method of determining to quantitative measurements of the amount of synovial joint inflammation and hence disease activity. Assessment membrane. In the present study, the semi-quantitative of the amount of inflammatory tissue in finger scoring was carried out by combined evaluation of axial joints, where tissue changes are likely to appear first,

6 Quantitative MRI of finger joint synovitis 71 may be clinically useful. With the arrival of dedicated MR imaging enhanced with Gd-DTPA. Radiology extremity MR scanners, and consequent reduction in 1990;176: costs and patient discomfort, MRI of the finger joints 7. Reiser MF, Bongartz GP, Erlemann R et al. Gadolinium- has the potential of becoming widely used. Nevertheless, DTPA in rheumatoid arthritis and related diseases: first results with dynamic magnetic resonance imaging. Skel to make MRI of the finger joints in arthritis a useful Radiol 1989;18: supplement to clinical and laboratory assessment of the 8. Østergaard M, Hansen M, Stoltenberg M, Lorenzen I. patient with RA, standardization of image interpretation Quantitative assessment of the synovial membrane in the is necessary. rheumatoid wrist: an easily obtained MRI score reflects In summary, quantification of the amount of synovial the synovial volume. Br J Rheumatol 1996;35: membrane in the finger joints is introduced. Synovial 9. Østergaard M, Gideon P, Henriksen O, Lorenzen I. membrane volumes were considerably larger in clinically Synovial volume a marker of disease severity in rheuma- active than in clinically inactive RA joints. However, toid arthritis? Quantification by MRI. Scand J Rheumatol the considerable differences in volumes within clinically 1994;23: uniform groups, as well as similarities between groups, 10. Palmer WE, Rosenthal DI, Schoenberg OI et al. Quantification of inflammation in the wrist with gadolinillustrate that the clinical presentation of a joint does ium-enhanced MR imaging and PET with 2-[F-18]-fluoronot necessarily reflect the status of the synovial mem- 2-deoxy-D-glucose. Radiology 1995;196: brane. This suggests that MRI may provide important 11. Østergaard M, Stoltenberg M, Gideon P, Sorensen K, additional information. An easily obtained semi- Henriksen O, Lorenzen I. Changes in synovial membrane quantitative score for synovial membrane hypertrophy and joint effusion volumes after intraarticular methylprednisolone. was highly correlated with the more time-consuming Quantitative assessment of inflammatory measurement of synovial membrane volumes, which and destructive changes in arthritis by MRI. J Rheumatol indicates that MRI may be acceptable for use in longi- 1996;23: tudinal studies, including evaluation of new drugs in 12. Polisson RP, Schoenberg OI, Fischman A et al. Use of short-term, dose-ranging studies. magnetic resonance imaging and positron emission tomo- graphy in the assessment of synovial volume and glucose Acknowledgements metabolism in patients with rheumatoid arthritis. Arthritis Rheum 1995;38: Waterton JC, Rajanayagam V, Ross BD, Brown D, We thank the Danish Rheumatism Association, the Whittemore A, Johnstone D. Magnetic resonance methods Thomas & Elisabeth Frølund Nielsen Foundation, for measurement of disease progression in rheumatoid Henny og Helge Holgersens Grant, Fonden til arthritis. Magn Reson Imaging 1993;11: Lœgevidenskabens Fremme, Emmy Lange, født Kramps 14. Østergaard M, Stoltenberg M, Løvgreen-Nielsen P, Volck Grant and Gårdejer af Stenløse Peder Laurits Pedersens B, Jensen CH, Lorenzen I. Magnetic resonance imaging- Grant for financial support, and Schering Diagnostika, determined synovial membrane and joint effusion volumes Denmark, for providing the contrast agent. P. Løvgreen in rheumatoid arthritis and osteoarthritis. Arthritis Rheum Nielsen is thanked for providing the photomicrographs 1997;40: Fleming A, Benn RT, Corbett M, Wood PH. Early of the synovial biopsies, and P. Ring and J. Arnth rheumatoid disease. II. Patterns of joint involvement. Ann Jensen for developing the image-processing software Rheum Dis 1976;35: package XPrime. 16. Scott DL, Coulton BL, Popert AJ. Long term progression of joint damage in rheumatoid arthritis. Ann Rheum Dis References 1986;45: Arnett FC, Edworthy SM, Bloch DA et al. The American 1. Scott DL, Grindulis KA, Struthers GR, Coulton BL, Rheumatism Association 1987 revised criteria for the Popert AJ, Bacon PA. Progression of radiological changes classification of rheumatoid arthritis. Arthritis Rheum in rheumatoid arthritis. Ann Rheum Dis 1984;43: ;31: Brower AC. Rheumatoid arthritis. Imaging. In: Klippel 18. Emery P. The Roche Rheumatology Prize Lecture. The JH, Dieppe PA, eds. Rheumatology. London: Mosby- optimal management of early rheumatoid disease: the key Year Book Europe, 1994: to preventing disability. Br J Rheumatol 1994;33: Soden M, Rooney M, Cullen A, Whelan A, Feighery C, 19. Østergaard M, Stoltenberg M, Gideon P et al. Effect of Bresnihan B. Immunohistological features in the synovium intraarticular osmic acid on synovial membrane volume obtained from clinically uninvolved knee joints of patients and inflammation, determined by magnetic resonance with rheumatoid arthritis. Br J Rheumatol 1989;28: imaging. Scand J Rheumatol 1995;24: Creamer P, Keen M, Zananiri F et al. Quantitative 4. de Bois MH, Tak PP, Arndt JW, Kluin PM, Pauwels EK, magnetic resonance imaging of the knee: a method of Breedveld FC. Joint scintigraphy for quantification of measuring response to intra-articular treatments. Ann synovitis with 99mTc-labelled human immunoglobulin Rheum Dis 1997;56: G compared to histological examination. Clin Exp 21. Reece R, Emery P. Needle arthroscopy. [Editorial ] Br J Rheumatol 1995;13: Rheumatol 1995;34: Zvaifler NJ, Firestein GS. Pannus and pannocytes. 22. Corvetta A, Giovagnoni A, Baldelli S et al. MR imaging Alternative models of joint destruction in rheumatoid of rheumatoid hand lesions: comparison with conventional arthritis. Arthritis Rheum 1994;37: radiology in 31 patients. Clin Exp Rheumatol 1992; 6. König H, Sieper J, Wolf KJ. Rheumatoid arthritis: evalu- 10: ation of hypervascular and fibrous pannus with dynamic 23. Foley Nolan D, Stack JP, Ryan M et al. Magnetic

7 72 M. Klarlund et al. resonance imaging in the assessment of rheumatoid arth- joints in patients with rheumatoid arthritis. Clin Radiol ritis a comparison with plain film radiographs. Br J 1993;48: Rheumatol 1991;30: Winalski CS, Palmer WE, Rosenthal DI, Weissman BN. 24. Sharp JT, Wolfe F, Mitchell DM, Bloch DA. The progression Magnetic resonance imaging of rheumatoid arthritis. of erosion and joint space narrowing scores in Radiol Clin North Am 1996;34: rheumatoid arthritis during the first twenty-five years of 29. Peterfy CL, Majumdar S, Lang P, Van Dijke CF, Sack disease. Arthritis Rheum 1991;34: K, Genant HK. MR imaging of the arthritic knee: 25. Rominger MB, Bernreuter WK, Kenney PJ, Morgan SL, improved discrimination of cartilage, synovium, and effu- Blackburn WD, Alarcon GS. MR imaging of the hands sion with pulsed saturation transfer and fat-suppressed in early rheumatoid arthritis: preliminary results. T1-weighted sequences. Radiology 1994;191: Radiographics 1993;13: Østergaard M, Court Payen M, Gideon P et al. 26. Østergaard M, Stoltenberg M, Henriksen O, Lorenzen I. Ultrasonography in arthritis of the knee. A comparison The accuracy of MRI-determined synovial membrane and with MR imaging. Acta Radiol 1995;36: joint effusion volumes in arthritis. A comparison of 31. Disler DG, Marr DS, Rosenthal DI. Accuracy of volume pre- and post-aspiration volumes. Scand J Rheumatol measurements of computed tomography and magnetic 1995;24: resonance imaging phantoms by three-dimensional recon- 27. Jevtic V, Watt I, Rozman B et al. Precontrast and postcon- struction and preliminary clinical application. Invest trast (Gd-DTPA) magnetic resonance imaging of hand Radiol 1994;29: