Imaging of sacroiliitis in early seronegative spondylarthropathy: Assessment of abnormalities by MR in comparison with radiography and CT

Transcription

1 Acta Radiologica ISSN: (Print) (Online) Journal homepage: Imaging of sacroiliitis in early seronegative spondylarthropathy: Assessment of abnormalities by MR in comparison with radiography and CT K. BØcker Puhakka, A. G. Jurik, N. Egund, B. Schiottz-Christensen, K. Stengaard-Pedersen, G. vanovereemhansen & J. VallØ Christiansen To cite this article: K. BØcker Puhakka, A. G. Jurik, N. Egund, B. Schiottz-Christensen, K. Stengaard-Pedersen, G. vanovereemhansen & J. VallØ Christiansen (2003) Imaging of sacroiliitis in early seronegative spondylarthropathy: Assessment of abnormalities by MR in comparison with radiography and CT, Acta Radiologica, 44:2, To link to this article: Published online: 16 Sep Submit your article to this journal Article views: 102 View related articles Citing articles: 2 View citing articles Full Terms & Conditions of access and use can be found at Download by: [ ] Date: 04 January 2018, At: 01:07

2 Acta Radiologica 44 (2003) Copyright # Acta Radiologica 2003 Printed in Denmark. All rights reserved ACTA RADIOLOGICA ISSN IMAGING OF SACROILIITIS IN EARLY SERONEGATIVE SPONDYLARTHROPATHY Assessment of abnormalities by MR in comparison with radiography and CT K. BØCKER PUHAKKA 1,A.G.JURIK 1,N.EGUND 1,B.SCHIOTTZ-CHRISTENSEN 2,K.STENGAARD-PEDERSEN 2, G. van OVEREEM HANSEN 3 and J. VALLØ CHRISTIANSEN 4 Departments of 1 Radiology R and 2 Rheumatology, Aarhus University Hospital, Aarhus Kommunehospital, 3 King Christian X Rheumatological Hospital, Graasten, Aarhus and 4 Department of Radiology, Sönderborg Hospital, Sönderborg, Denmark. Abstract Purpose: To analyze the type and frequency of abnormalities of the sacroiliac joint (SIJ) in early seronegative spondylarthropathy (SpA) by MR in comparison with CT and radiography, assess the most appropriate MR sequences to be used, and introduce a new way of grading MR abnormalities of the SIJ. Material and Methods: The SIJs of 41 patients with early SpA (median duration of inflammatory low back pain of 19 months) were evaluated by MR imaging using STIR, T1, T2, and T1 fat saturated (FS) sequences before and after i.v. Gd contrast medium followed by staging of abnormalities. The findings were compared with those obtained by CT and radiography. Results: MR and CT had equal efficacy superior to radiography in staging of erosions and osseous sclerosis. Only MR allowed visualization and grading of active inflammatory changes in the subchondral bone and surrounding ligaments in addition to bone marrow fatty accumulations. T2-weighted sequences did not contribute to assessment of sacroiliitis. Conclusion: MR of the SIJs is reliable in its visualization of joint erosions in early SpA and allows differentiation between active and chronic sacroiliitis. We recommend the following sequences: semicoronal T1 and both semicoronal and semiaxial STIR. If these images are normal, the examination can be finished; otherwise additional semicoronal T1 FS before and after i.v. contrast has to be performed as well as semiaxial post-contrast T1 FS. Key words: Sacroiliitis; MRI; CT. Correspondence: Katriina Bøcker Puhakka, Barthsgade 9. st. mf. th., DK-8200 Aarhus N, Denmark. FAX þ kbpuh@akh.aaa.dk Accepted for publication 20 November The European Spondylarthropathy Study Group (ESSG) defined seronegative spondylarthropathies (SpA) in SpA is divided into five entities (16): ankylosing spondylitis, psoriatic arthritis, reactive arthritis, arthritis associated with inflammatory bowel diseases, and unclassified SpA. The ESSG classification is based on clinical findings. Radiographic evidence of sacroiliitis is included in the definition but is not necessary for diagnosis of SpA. Despite the use of the SpA criteria, the delay in diagnosis may be up to 9 years (13, 36, 41). As SpA occurs mainly in young patients (11) an early diagnosis would be valuable with regard to therapy, prognosis, and evaluation of working capacity. The sacroiliac joints (SIJs) are involved in most cases of SpA and sacroiliitis is commonly the first manifestation of SpA (13). Diagnosis of arthritis in these joints by radiography, CT, or MR therefore 218

3 IMAGING OF SACROILIITIS plays a central role in establishing the diagnosis of early SpA. Since the publication of the diagnostic criteria in 1991, several studies have emphasized the value of CT and MR in the diagnosis of sacroiliitis (4, 7, 8, 18, 38). Both CT and MR have proved their capability to visualize joint erosions earlier than radiography (5, 6, 10, 12, 13, 17, 20, 24, 31, 40). In addition, MR can detect inflammatory changes of the SIJ before any changes are visible by CT. Further, MR is not, as CT, associated with radiation (3, 12, 30, 40). Despite these advantages, MR has not gained wide acceptance in the diagnosis of SpA. This may be explained by difficulties in interpreting MR images of the SIJ partly due to its complex anatomy, being composed of two compartments: a C-shaped cartilaginous portion that lays inferiorly/anteriorly and a ligamentous portion superiorly/posteriorly (9, 23). Besides, analysis of the diagnostic usefulness of different MR sequences seems to be lacking. There is no well-established method for evaluating pathological MR findings of the SIJ, though the need for this is obvious (35). AHLSTRÖM et al. (1) introduced the first method. They divided the findings into two types, mainly based on signal changes on T1, T2, and phase-contrast sequences: Type I lesions were characterized by a low signal intensity on T1-weighted images and a high signal intensity in both phase-contrast and T2-weighted images. Type II lesions consisted of changes presenting with low signal intensity in all sequences. BRAUN et al. (10) described a grading system for disease activity based on quantitative evaluation of contrast enhancement at dynamic MR, and used the New York criteria for joint destruction (22). Other studies have focused on the presence of subchondral bone marrow edema (31) or whether the pathological findings were uni- or bilateral (28). As no studies have assessed the MR findings in detail according to specific localization of signs and grading, there is a need for a new system that can be used to quantify MR signs of both disease activity and joint destruction. The purposes of this study were to analyze articular abnormalities of the SIJs in early SpA by MR in comparison with findings by radiography and CT, assess the most appropriate MR techniques and sequences to be used, and introduce a new way of grading MR abnormalities of the SIJ. Material and Methods In all, 41 patients were included at two different centers, 25 patients at Aarhus University Hospital (Center A) and 16 at the Rheumatological Hospital, Table 1 Clinical features of the 41 patients included in the study No. of patients Median Range Age Males 24 Females 17 Duration of ILBP, months Schober lumbar spine motion test, cm HLA B27 positive 26 Elevated CRP 8 Ongoing NSAID 23 treatment ILBP ¼ inflammatory low back pain. CRP ¼ C-reactive protein. NSAID ¼ non-steroid inflammatory drugs. Graasten (Center B), after referral from the primary ward for either radiography or evaluation at the rheumatological out-patient clinic (Table 1). Patients with radiographic changes stages 0 3 according to the New York criteria (22) were included. All patients fulfilled the ESSG criteria for SpA localized to the SIJ (16) having inflammatory low back pain (ILBP) with at least four of the following manifestations: 1) onset before the age of 45 years; 2) insidious onset; 3) improvement with exercise; 4) associated morning stiffness; and 5) at least 3 months duration. Patients were not included if they showed evidence or had a clinical history of metabolic diseases or malignancies and/or a positive rheumatoid factor. The patients (24 males and 17 females) had a median age of 26 years (20 45 years). The median duration of ILBP was 19 months (3 156 months). Thirty-four of the 41 patients had a disease duration of 36 months or less. The patients gave informed, written consent to participate in a prospective clinical and radiological investigation of SpA. The study was carried out according to the Helsinki Declaration and was approved by the local ethical committees. The present paper is confined to the results of radiography, CT, and MR of the SIJs, performed on all patients at their inclusion in the investigation, with the exception of CT in 2 patients who refused to participate in that examination. A.p. radiography was obtained with a 20 caudal tube tilt. CT was performed in the supine position with a cranial gantry tilt to obtain semicoronal images through both the cartilaginous and ligamentous portions of the SIJs using a Somatom Plus S and a Somatom Plus S4 unit (Siemens) at Centers A and B, respectively. Examination parameters were: contiguous 5-mm slices using 120 kv/ 330 mas at Centers A, and 140 kv/ mas at 219

4 K. PUHAKKA ET AL. Center B. For image reconstruction a standard bone algorithm was used at Center A and a high resolution algorithm at Center B. MR images were obtained with a 1.0 T unit (Signa Highspeed, GE), in 18 patients from Center A and with a 1.5 T unit (Magneton Vision, Siemens) in 7 patients from Center A and in all patients from Center B. The examinations with the 1.0 T unit were performed with a pelvic array coil and consisted of the following sequences: semicoronal short tau inversion recovery (STIR) (TR/TE/TI 4480/66/ 125 ms), semicoronal T1 (TR/TE 400/16 ms), semicoronal T1 with fat saturation (FS) (TR/TE 660/ 16 ms), semiaxial T2 high resolution (TR/TE 5000/ 95 ms), and after administration of i.v. gadolinium (Gd) (Omniscan, 1 mmol/kg, maximum 100 mmol) semicoronal and semiaxial T1 FS (TR/TE 660/ 16 ms). The slice thickness was 4 mm at all sequences with mm spacing. The examinations with the 1.5 T unit were performed with a body array coil. Similar sequences were used, but the parameters were slightly different: TR/TE/TI 4000/30/150 ms for STIR, TR/TE 405/12 ms for T1, TR/TE 660/12 ms for T1 FS, and TR/TE 4100/ 132 ms for T2. The slice thickness was 4 mm at all sequences with spacing of <1 mm. The slice orientations were standardized: semicoronal slices were placed parallel to a line joining the upper dorsal aspect of S1 and S3, and semiaxial slices were perpendicular to the semicoronal plane (Fig. 1). Assessment and grading of images: The images were assessed blindly by two senior radiologists (A.G.J. and N.E.). The MR and CT examinations were analyzed independently, but with the radiographs available at both evaluations to simulate the daily clinical situation. Radiographic abnormalities of the SIJs (Fig. 2a) were graded according to the New York criteria for sacroiliitis in 5 stages: 1) normal; 2) suspicious changes; 3) minimal abnormality in the form of small areas of erosions or sclerosis without alteration in the joint width; 4) unequivocal abnormality moderate or advanced sacroiliitis consisting of erosions, sclerosis, widening, narrowing, and/or partial ankylosis; and 5) severe abnormality in the form of total ankylosis (22). At CT (Fig. 2b) the SIJs were assessed with regard to 1) erosions (destruction of the joint surface); 2) bone marrow sclerosis (increased density); 3) fat accumulation in bone marrow (decreased density); and 4) irregular surface and/or new bone formation at entheses. These findings were graded as 0¼normal, 1¼minimal, 2¼moderate and 3¼severe. A similar grading was used for joint space narrowing, supplemented by a registration of joint space widening (grade 2*). 220 Fig. 1. Scan planes by MR. a) Drawing showing the semicoronal and semiaxial scan planes at the middle of the sacrum together with a lateral view of the whole sacrum with the joint facet. b) Semicoronal and semiaxial scan planes (patient in Fig. 2). Assessment of the MR examinations (Fig. 2c f) included a similar grading of the following findings: 1) erosions; 2) osseous sclerosis (low signal intensity at T1 and/or T1 FS); 3) fat accumulation in the bone marrow (high signal intensity at T1); 4) irregular surface and/or new bone formation at entheses; 5) joint space width; 6) bone marrow edema/reparative processes (high signal intensity at STIR); 7) Gd contrast enhancement in the bone marrow separated in regions with abnormal low and normal signal intensity at precontrast T1 FS, respectively; 8) Gd contrast enhancement in the cartilaginous and ligamentous joint space; and 9) Gd contrast enhancement at entheses outside the joint. All assessments and gradings by CT and MR were performed at four anatomical sites of each SIJ: the sacral and iliac side of the cartilaginous and ligamentous portions of the joint (Fig. 3). The

5 IMAGING OF SACROILIITIS a b c d e f Fig. 2. A 27-year-old HLA B27 positive man with 2 years of ILBP associated with psoriatic skin and nail changes. a) Radiograph of the SIJs showing bilateral sacroiliitis grade 3. b) Semicoronal CT demonstrating bilateral erosions and sclerosis with irregular joint spaces and a slightly decreased density in the sacral bone compatible with fatty marrow changes. c) Semicoronal T1-weighted sequence showing erosions, sclerosis, fatty marrow changes (white arrows) and joint width alteration. d) T1-weighted sequence with FS more clearly visualizes the erosions (white arrows). e) Semicoronal STIR sequence illustrating bone marrow edema (open arrow). f) Semicoronal T1 FS after i.v. Gd demonstrating contrast enhancement in the bone marrow (arrows) and the joint space. 221

6 K. PUHAKKA ET AL. a b c d e Fig. 3. Drawing of the two joint compartments of the sacroiliac joints showing the frequency of the different grades of pathological changes at the various joint positions by CT and MR. Number of patients. a) Erosions. b) Sclerosis. c) Fat accumulation in bone marrow. d) Joint width alteration and contrast enhancement. e) Bone marrow edema and contrast enhancement. severity of erosions, sclerosis, and fat accumulation in bone marrow were calculated as a sum of the score for the abnormality in each position, the total maximal score for each abnormality being 12 for each joint, and 24 for a patient. Similarly, a total score for alteration of joint space width and Gd contrast enhancement in the joint space was obtained, resulting in a total maximal score of 6 222

7 IMAGING OF SACROILIITIS Table 2 Scoring system for MR abnormalities based on assessment at 8 osseous positions (the iliac and sacral sides of the cartilaginous and ligamentous portions of the joints, respectively) and 4 joint space positions (cartilaginous and ligamentous joint spaces), as demonstrated in Fig. 3. An osseous joint position or a joint space One SIJ: 4 osseous positions; 2 joint spaces A patient: 8 osseous positions; 4 joint spaces MR abnormalities Grade range Score range Score range Erosion Sclerosis Joint space width Overall joint destruction score 0 60 Bone marrow edema Gd enhancement in bone marrow Gd enhancement in joint space Overall activity score 0 60 for a joint and 12 for a patient. Widening of the joint space (grade 2*) was given a value of 2. Finally, for MR, an overall score for joint destruction was calculated as a sum of the scores for erosions, sclerosis, and joint width (maximum 60 per patient). Similarly, an overall score for inflammatory activity was calculated as a sum of the scores for bone marrow edema, Gd contrast enhancement in the bone marrow, and in the joint space (maximum score 60 per patient) (Table 2). An overall stage of sacroiliitis with regard to joint destruction and enthesopathy graded 0 4 corresponding to the New York criteria for sacroiliitis (22) was also recorded for each joint at CT and MR to facilitate the comparison with radiography. The MR sequences used were assessed with regard to their capability to visualize the articular and osseous abnormalities of sacroiliitis. Statistics: SPSS software (version 10.0) was used to perform the following tests: McNemar test for differences in frequencies of two variables and nonparametric paired samples t-test to compare means. Kappa test with regard to agreement and McNemar test for disagreement were used to estimate the interand intraobserver variations. A value of p<0.05 was considered significant. Results The distribution and severity of sacroiliitis assessed by radiography is displayed in Table 3, which also shows the frequency of unilateral and bilateral changes as well as symmetric and asymmetric changes. The interobserver agreement for the evaluation of radiographs was 89% with a kappa value of 0.63 and a non-significant McNemar test. Osseous changes in the ligamentous portion of the SIJs were rare by CT and MR (Fig. 3). Therefore, assessment of interobserver agreement at CT and MR was only performed for the evaluation of the Table 3 Frequencies, distribution, and severity of sacroiliitis by radiography, CT, and MR. Radiography and MR data are based on 41 and CT on 39 patients. The MR findings are divided into those related to joint destruction (MRI joint) and activity (MR activ.) and based on the New York criteria and our new MR scoring, respectively No. of patients Grading of changes, maximal stage; no. of patients Sacroiliitis stage Sacroiliitis unilat. bilat. sym. asym Mean New York Radiography CT MR joint New scoring MR joint MR activ unilat., unilateral sacroiliitis. bilat., bilateral sacroiliitis. sym., symmetric sacroiliitis. asym., asymmetric sacroiliitis. 223

8 K. PUHAKKA ET AL. cartilaginous portion of the joint, and the results, based on evaluation of all patients, are shown in Table 4. For both CT and MR there was a poor agreement in the assessment of the joint width. The interobserver agreement for the other CT and MR parameters varied between 71% and 94% with kappa values of MR for sclerosis and inflammatory activity evaluated the lowest agreements. The McNemar test was significant for erosions, sclerosis, edema, contrast enhancement, and overall activity score by MR, and for erosions and alteration of the joint width by CT (Table 4). This was not due to a trend of any of the observers to score more joint abnormalities than the other. There was a tendency for N.E. to assess more erosions by CT and MR and more joint width alterations by CT than A.G.J., while A.G.J. had a tendency to assess more sclerosis, bone marrow edema, and enhancement by MR than N.E. Assessment of the intraobserver agreement of A.G.J. was made blindly with approximately 1 year between the first and the second evaluations. The intraobserver agreements varied between 93% and 100% with kappa values between 0.84 and 1.0 and no significant McNemar tests (Table 4). The intraobserver agreement for the radiographs was 95% (kappa¼0.64, p¼1.0). The following findings are based on the evaluation of A.G.J., instead of an amalgam of the two observers, because of the high inter- and intraobserver agreements. The results of CT and MR appear in Fig. 3. Erosions: were observed in 34 patients by CT and in 33 patients by MR. The distribution of erosions in the different positions and the frequency of the different grades of erosions are shown in Fig. 3a. Except in 1 patient, who had erosions in the ligamentous portion of both joints by CT, erosions were only found in the cartilaginous portion of the joints where they predominated on the iliac side. The mean total joint score for erosions was 1.31 by CT, 1.32 by MR for the 39 patients examined by CT, and 1.29 for all patients (Table 5). Sclerosis was observed in 30 patients by CT and in 29 patients by MR. The distribution of sclerosis in the different positions and the different grades of sclerosis are shown in Fig. 3b. Sclerosis was most often seen in the cartilaginous portion of the joint. The mean total joint score for sclerosis was 1.12 by CT and 1.19 by MR for the 39 patients examined by CT (Table 5). Subchondral osseous sclerosis in the cartilaginous portion of the SIJ graded >2 by CT occurred in 21 positions. In all those positions there was decreased signal intensity on T1 FS before contrast injection. In 19 positions there was additional increased signal intensity by STIR and/or contrast enhancement (18 had both). Increased signal intensity by T1 (sign of fat accumulation in bone marrow) occurred in 15 of the 21 positions, in combination with marrow edema and contrast enhancement in 12. Decreased signal intensity on T1 FS before contrast injection occurred in a further 73 positions, in 37 of which there was osseous sclerosis grade 1 by CT. In 16 of the remaining 36 positions Table 4 Inter- and intraobserver (A.G.J.) agreement for each sign in the cartilaginous portion of the SIJ assessed by MR and CT; McNemar (p) and kappa ( k) tests values Pathological findings MR agreement, % Interobserver p k CT agreement, % p k MR agreement, % Intraobserver p k CT agreement, % Erosion Sclerosis Fat accumulation in bone marrow Alterations of joint width Bone marrow edema Enhancement in bone Enhancement in joint space Sacroiliitis (joint stage) Sacroiliitis (joint score) Sacroiliitis (activity score) p k 224

9 IMAGING OF SACROILIITIS Table 5 Mean and median values for the total joint scores representing the sum of the gradings in each position of the joints. By MR for all joints and for the 78 joints compared with CT MR, 82 joints MR, 78 joints CT, 78 joints MR/CT, 78 joints mean median range mean median range mean median range difference p-value Erosion Sclerosis Fatty marrow < Joint width <0.01 Edema Enhancement, bone Enhancement, joint space there was increased signal intensity by T1 (fat accumulation in bone marrow). Fat accumulation in bone marrow was observed in 2 patients by CT and in 26 patients by MR. The distribution in the different joint positions and the division into grades appear in Fig. 3c. The changes mainly affected the cartilaginous portion of the joint. The mean total joint score was 0.04 by CT and 1.19 by MR (p<0.0001) (Table 5). New bone formation at entheses was observed in 13 patients by CT and in 1 patient by MR (Fig. 4). The new bone formation observed by CT was in all 13 patients located at the insertion of the dorsal ligaments, but 1 patient had additional changes anteriorly which were also visible by MR (Fig. 4). Alterations in joint width: Changes in joint width, either narrowing or widening, were observed in 34 patients by CT and in 26 patients by MR. The distribution in the different joint portions, including a separation between narrowing and widening and a division into grades appears in Fig. 3d. The changes were mainly seen in the cartilaginous portion of the joint. The mean total score for joint alteration was significantly higher by CT than by MR, the values being 1.15 and 0.82, respectively, for the patients evaluated by both modalities (p<0.01) (Table 5). Bone marrow edema was present in 35 patients by MR. The distribution and division into grades appear from Fig. 3e. Marrow edema predominantly occurred in the cartilaginous portion and at the iliac side of the joint. The mean total joint score for edema was 2.11 (Table 5). Enhancement of bone was observed in 37 patients by MR with a joint distribution as shown in Fig. 3e. Enhancement was seen in both the ligamentous and the cartilaginous portions of the joint, but most often in the latter and on the iliac side. The grading of changes appears in Fig. 3e. The mean total joint score for bone enhancement was 1.89 (Table 5). Correlation between bone marrow edema and enhancement: There was a significant association between the occurrence of bone marrow edema and contrast enhancement, both when looking at all joint positions (p<0.001) and when analyzing only the cartilaginous portion of the joints (p<0.001). Bone marrow edema and contrast enhancement occurred simultaneously in 91 and were absent in 55 of the 164 cartilaginous joint positions, corresponding to an agreement of 89%. Enhancement was observed in 8 positions without signs of edema, and lack of enhancement occurred in 10 positions with an edema. Enhancement in the joint space was seen in 26 patients by MR with a joint distribution and different grades of enhancement as shown in Fig. 3d. It occurred in both joint portions, but more frequently in the cartilaginous than in the ligamentous portion of the joint. The mean total joint score for enhancement was Enhancement at entheses outside the joint occurred in 9 patients (Fig. 4). The agreement and discrepancy of CT and MR to detect erosions, sclerosis, and fat accumulation in bone marrow, and alterations of joint width was analyzed statistically corresponding to the cartilaginous portion only, due to the rare occurrence of abnormalities in the ligamentous portion (Fig. 3). The results based on the 39 patients undergoing both examinations, are given in Table 6. Erosions, sclerosis, and sacroiliitis (stage of joint destruction) were recorded equally often by the two examinations, while MR significantly more often than CT demonstrated fat accumulation in bone marrow (p<0.0001). The agreement between CT and MR in displaying joint width alteration approached significance (p¼0.08) in the favor of CT. CT was able to demonstrate new bone formation at entheses, which was only the case in 1 patient by MR (Fig. 4). MR, on the other hand, detected signs 225

10 K. PUHAKKA ET AL. a b c Fig. 4. HLA B27 negative 45-year-old man with ILBP for 5 years and a family history of psoriasis. a) CT images showing new bone formation at the attachment of the posterior ligaments to the sacrum and bridging of new bone formation anteriorly (arrows). b) Semicoronal and (c) semiaxial T1-weighted FS sequences after i.v. Gd demonstrating ligamentous enhancement posteriorly and bridging bone formation anteriorly without accompanying enhancement (arrows). Table 6 Agreement and discrepancy between CT and MR in the detection of pathological changes in the cartilaginous portion of the SIJs of 39 patients No. of locations/ joints Erosion Sclerosis Fat accumulation in bone marrow Changes of joint width Sacroiliitis, joint destruction 226 Agreement CT/MR MR pos. CT neg. CT pos. MR neg. p-value McNemar n % <

11 IMAGING OF SACROILIITIS of inflammation in the form of contrast enhancement at entheses outside the joint in 9 patients (Fig. 4) and also inflammatory changes in the bone and joint spaces, not visible by CT (Fig. 2). Based on the 39 patients undergoing all three examination modalities, MR was found to grade the overall changes of joint destruction significantly higher than either radiography (p<0.001) or CT (p<0.01) (Table 3). Of the 2 patients not evaluated by CT, 1 had bilateral, symmetrical joint destruction stage 2 by both radiography and MR, and a bilateral activity stage of 3 by MR. The other patient had unilateral joint destruction stage 1 by radiography and no signs of activity by MR. By MR the mean overall activity score for all 41 patients was 10.3 (0 37), while the similar mean overall destruction score was 6.5 (0 16). Evaluation of the MR sequences used disclosed that the STIR sequence in all patients was valuable for the detection of inflammation. The postcontrast T1 FS gave further information due to higher contrast resolution and a better delineation of the anatomical structures and visualization of vascularized inflammation. Except in 3 patients, the T1 sequence was considered to give information about fat accumulation in bone marrow that could not be achieved by the other sequences. In all but 2 patients, the transverse slices (T2 and postcontrast T1 FS) gave additional information compared to the semicoronal slices due to a better visualization of the ligamentous portion of the joint, but the T2 sequence never gave additional information compared to the postcontrast T1 FS. Discussion In the course of sacroiliitis both acute and chronic changes can occur, separately or at the same time in different areas of the same joint. Acute inflammatory activity presents with bone marrow edema, which is known to be an early and unspecific sign of various abnormalities in bone marrow (37), and contrast enhancement caused by increased vascularization (25). Erosions, sclerosis, and changes of joint width are regarded as signs of chronicity (13), as well as fat deposition in the bone marrow, which can be regarded as a sign of degeneration and healing (21, 26). Chronic bone changes can be detected and graded by both CT and MR, and our data showed that no essential information is lost by using MR instead of CT. We found that the overall stage of joint destruction using the New York criteria (22) was evaluated higher by MR than by CT, probably because MR was better than CT at detecting fat accumulation in the bone marrow. CT demonstrated osseous proliferation in the ligamentous portion that could not be identified by MR. However, MR was capable of detecting inflammation, so the final diagnosis of sacroiliitis was usually established. Only 1 of the present patients was recorded to have possible sacroiliitis (stage 1) by both conventional radiography and CT, but had no detectable joint destruction or signs of activity by MR. At follow-up a year later there were signs of both joint destruction and activity at MR (33). The advantages of MR compared with CT are that it visualizes active inflammatory changes in the form of bone marrow edema and contrast enhancement, making it possible to diagnose sacroiliitis by MR before definite joint destruction, detectable by CT and radiography, occu (3, 5, 6, 10, 12, 13, 24, 30, 31, 40). To our knowledge, six studies comparing MR and CT of the SIJs with regard to sacroiliitis have been published (1, 3, 30, 34, 39, 40). The number of patients included in these studies varied from 17 to 50 (mean 38). The duration of symptoms was not mentioned in any of these studies, but as most of the patients included had defined SpA, it seems likely that our patients were examined at an earlier stage of the disease as they were referred from the primary ward and most had a relative short disease duration. In one of the six studies, MR and CT were not evaluated independently (1). In the remaining studies different sequences and scan parameters were used, and postcontrast MR images were only obtained in one study (40). The results of the five comparative studies are in agreement with those of the present study, finding that MR is equally good or even better than CT at detecting overall SIJ changes. When looking at different signs of sacroiliitis, CT has been reported superior to MR in detecting sclerosis (40) and bone production (20). The latter finding was consistent with that of the present analysis, but sclerosis, in our study, was detected with nearly equal frequency by CT and MR. The present study included an analysis of the signal intensities by MR corresponding to areas disclosing definite sclerosis by CT. The results of one observer (A.G.J.) showed that the sclerosis detected by CT corresponded to a mixture of inflammation, fatty marrow changes, and/or increased osseous tissue or marrow fibrosis by MR. An explanation for the low interobserver agreement regarding sclerosis may be the consideration of one observer (N.E.) that contrast enhancement in bone marrow with low signal intensity at the precontrast T1 sequence may indicate that there is not any increased amount of trabecular bone. 227

12 K. PUHAKKA ET AL. The great advantage of MR compared with CT is the capability to detect inflammation, which in addition to joint destruction is a sign of sacroiliitis. Besides, MR enables a clear distinction of the two portions of the SIJ based upon a well-delineated differentiation between fatty tissue in the ligamentous portion and cartilage in the cartilaginous portion (32). CT only identifies the two portions of the SIJ by their orientation and position (14). This, combined with the ability of MR to detect inflammatory changes, provides an opportunity to analyze pathological inflammatory changes in the ligamentous portion of the joint that would not be detected by CT. This can be important for the diagnosis of early inflammatory changes in patients where the changes are mainly situated in the ligamentous portion. However, chronic changes in that region in the form of new bone formation at entheses are better visualized by CT than by MR, reflected by the fact that new bone formation seen in 13 of the present patients at CT was only visualized in 1 patient by MR (Fig. 4). This patient had changes both at the attachment of the posterior ligaments and anteriorly. The bridging bone formation anteriorly may, however, be due to bridging osteophytes, as seen in degenerative disease. In concordance with our findings, in a recent study (29) CT was found to be optimal in the detection of erosions and was capable of visualizing calcifications in the ligamentous portion of the SIJ. Taking all aspects of sacroiliitis into consideration, MR has only been found significantly inferior to CT for the diagnosis of sacroiliitis in one study (34). An explanation for the different results could be that study (34) was based on comparison of a high resolution CT technique (slice thickness of 2 mm and 480 mas) and MR performed with a body coil and only T1- and T2-weighted sequences. The capability of MR to distinguish between acute and chronic changes and to estimate the degree of disease activity can be beneficial in monitoring the effect of pharmacological treatment (2). The use of dynamic contrast-enhanced MR to evaluate the level of inflammatory activity has been reported of value (7, 38). This method is, however, difficult to apply in routine clinical practice because the assessment is time-consuming, and reproduction of the findings may be difficult. Most reports on MR examinations of SIJs are obtained with semicoronal sectioning. In this study we added a semiaxial sequence providing obvious advantages, as the anatomy of the ventral and dorsal margins of the cartilaginous portion of the joint only can be assessed by this view (32). Semiaxial views also give a good characterization of the ligamentous portion of the SIJ. A proper examination 228 of the joint should therefore contain both semicoronal and semiaxial views. Based on the present analysis we recommend the following sequences for examination of the SIJs with regard to sacroiliitis: semicoronal T1 and both semicoronal and semiaxial STIR. If the images of these two sequences are normal, the examination is completed; otherwise, additional semicoronal T1 FS before and after i.v. contrast has to be performed as well as semiaxial postcontrast T1 FS. CT of the SIJs is associated with radiation exposure. Although optimal machine settings, the use of only a few slices (15, 19), or a semicoronal scan plane (27) can reduce the radiation dose, it is more desirable to avoid radiation completely. This can be achieved by replacing CT with MR, though still keeping in mind the situations in which CT is superior to MR, i.e., detection of chronic bone changes in the ligamentous portion of the joint (29). The present study described a new way of assessing and grading MR abnormalities of the SIJ in early SpA, which can be used for quantifying both the inflammatory activity and the joint destruction. The method has to be further analyzed before it can be used in daily routine. Conclusion: MR and CT are nearly equally good in detecting joint destruction as part of sacroiliitis in the cartilaginous portion of the SIJ. MR in addition visualizes early active inflammatory changes and is not associated with radiation exposure. When available, MR of the SIJs should therefore be the first choice in suspected early SpA, and we recommend semicoronal T1 and both semicoronal and semiaxial STIR sequences. If the images of these sequences are normal, the examination can be completed; otherwise, additional semicoronal T1 FS before and after i.v. contrast has to be performed as well as semiaxial postcontrast T1 FS. ACKNOWLEDGMENTS We gratefully acknowledge the Danish Rheumatism Association and A.P. Møller og Hustru Chastine Mc-Kinney Møllers Fond til almene Formaal for financial support. We also thank Nycomed Denmark A/S for providing the contrast agent, and engineer, Ph.D. Peter Vestergaard Poulsen for technical assistance in providing the MR sequences. REFERENCES 1. Ahlström H, Feltelius N, Nyman R, Hällgren R. Magnetic resonance imaging of sacroiliac joint inflammation. Arthritis Rheum. 1990; 33: Antoni C, Dechant C, Ogilvie A, Kalden-Nemeth D, Kalden JR, Manger B. Successful treatment of psoriatic arthritis with infliximab in a MRI controlled study. J. Rheumatol. 2000; 27: Battafarano DF, West SG, Rak KM, Fortenbery EJ, Chantelois AE. Comparison of bone scan, computed

13 IMAGING OF SACROILIITIS tomography, and magnetic resonance imaging in the diagnosis of active sacroiliitis. Semin. Arthritis Rheum. 1993; 23: Bigot J, Loeuille D, Chary Valckenaere I, Pourel J, Cao MM, Blum A. Détermination des meilleurs critères diagnostiques des sacro-iliites en IRM. J. Radiol. 1999; 80: Blum U, Buitrago-Tellez C, Mundinger A, et al. Magnetic resonance imaging (MRI) for detection of active sacroiliitis a prospective study comparing conventional radiography, scintigraphy, and contrast enhanced MRI. J. Rheumatol. 1996; 23: Bollow M, Braun J, Biedermann T, et al. Use of contrastenhanced MR imaging to detect sacroiliitis in children. Skeletal Radiol. 1998; 27: Bollow M, Braun J, Hamm B, et al. Early sacroiliitis in patients with spondyloarthropathy: evaluation with dynamic gadolinium-enhanced MR imaging. Radiology 1995; 194: Bollow M, König H, Hoffmann C, Schilling A, Wolf KJ. Erste Erfahrungen mit der dynamischen Magnetresonanztomographie in der Diagnostik entzündlicher Erkrankungen der Sakroiliakagelenke. RÖFO 1993; 159: Bowen V, Cassidy JD. Macroscopic and microscopic anatomy of the sacroiliac joint from embryonic life until the eighth decade. Spine 1981; 6: Braun J, Bollow M, Eggens U, König H, Distler A, Sieper J. Use of dynamic magnetic resonance imaging with fast imaging in the detection of early and advanced sacroiliitis in spondylarthropathy patients. Arthritis Rheum. 1994; 37: Braun J, Bollow M, Remlinger G, et al. Prevalence of spondylarthropathies in HLA-B27 positive and negative blood donors. Arthritis Rheum. 1998; 41: Braun J, Sieper J. The sacroiliac joint in the spondyloarthropathies. Curr. Opin. Rheumatol. 1996; 8: Braun J, Sieper J, Bollow M. Imaging of sacroiliitis. Clin. Rheumatol. 2000; 19: Carrera GF, Foley WD, Kozin F, Ryan L, Lawson TL. CT of sacroiliitis. AJR 1981; 136: Damilakis J, Prassopoulos P, Persinakis K, Faflia C, Gourtsoyiannis N. CT of the sacroiliac joints. Dosimetry and optimal settings for a high-resolution technique. Acta Radiol. 1997; 38: Dougados M, van der Linden S, Juhlin R, et al. The European Spondylarthropathy Study Group preliminary criteria for the classification of spondylarthropathy. Arthritis Rheum. 1991; 34: Fenlon HM, Casserly I, Sant SM, Breatnach E. Plain radiographs and thoracic high-resolution CT in patients with ankylosing spondylitis. AJR 1997; 168: Friedburg H, Meske S, Hennig J, Billmann P, Peter HH, Wenz W. Die Kernspintomographie des Sakroiliakagelenkes. Eine neue Untersuchungstechnik zur Differentialdiagnose rheumatischer Erkrankungen im lumbosacralen Übergang. Radiologe 1987; 27: Friedman L, Silberberg PJ, Rainbow A, Butler R. A limited, low-dose computed tomography protocol to examine the sacroiliac joints. Can. Assoc. Radiol. J. 1993; 44: Geijer M, Sihlbom H, Göthlin JH, Nordborg E. The role of CT in the diagnosis of sacroiliitis. Acta Radiol. 1998; 39: Gillams AR, Chaddha B, Carter AP. MR appearances of the temporal evolution and resolution of infectious spondylitis. AJR 1996; 166: Gofton JP, Bennett PH, Bremner JM, et al. Report from the Subcommittee on Diagnostic Criteria for Ankylosing Spondylosis. 1966: Goss CM. Gray s Anatomy. Lea & Febiger, 1973: Hanly JG, Mitchell MJ, Barnes DC, MacMillan L. Early recognition of sacroiliitis by magnetic resonance imaging and single photon emission computed tomography. J. Rheumatol. 1994; 21: Hurley JV. Inflammation. In: Muir s Textbook of Pathology. London: Edward Arnold, Jevtic V, Rozman B, Kos-Golja M, Watt I. MR-Bildgebung bei seronegativen Spondyloarthritiden. Radiologe 1996; 36: Jurik AG, Hansen J, Puhakka KB. Effective radiation dose from semicoronal CT of the sacroiliac joints in comparison with axial CT and conventional radiography. Eur. Radiol. 2002; 12: Lee SH, Jee WH, Kim WU, et al. Association of sacroiliac MR findings with disease activity in patients with ankylosing spondylitis. Arthritis Rheum. 1999; 42: Mester AR, Makó EK, Karlinger K, et al. Enteropathic arthritis in the sacroiliac joint. Imaging and differential diagnosis. Eur. J. Radiol. 2000; 35: Murphey MD, Wetzel LH, Bramble JM, Levine E, Simpson KM, Lindsley HB. Sacroiliitis. MR imaging findings. Radiology 1991; 180: Oostveen J, Prevo R, den Boer J, van de Laar M. Early detection of sacroiliitis on magnetic resonance imaging and subsequent development of sacroiliitis on plain radiography. A prospective, longitudinal study. J. Rheumatol. 1999; 26: Puhakka KB, Egund N, Melsen F, et al. MR imaging of the normal sacroiliac joint with correlation to histology. Skeletal Radiol. 2003; Puhakka KB, Jurik AG, Egund N, et al. The course of sacroiliitis in early seronegative spondylarthropathy. A radiological and clinical follow-up study. Scand. J. Rheum., in press. 34. Rémy M, Bouillet P, Bertin P, et al. Evaluation of magnetic resonance imaging for the detection of sacroiliitis in patients with early seronegative spondylarthropathy. Rev. Rhum. Engl. Ed. 1996; 63: Stone M, Salonen D, Lax M, Payne U, Lapp V, Inman R. Clinical and imaging correlates of response to treatment with infliximab in patients with ankylosing spondylitis. J. Rheumatol. 2001; 28: Underwood MR, Dawes P. Inflammatory back pain in primary care. Br. J. Rheumatol. 1995; 34: Vahlensieck M, Reiser M. Knochenmarködem in der MRT. Radiologe 1992; 32: Wittram C, Whitehouse GH, Bucknall RC. Fat suppressed contrast enhanced MR imaging in the assessment of sacroiliitis. Clin. Radiol. 1996; 51: Wittram C, Whitehouse GH, Williams JW, Bucknall RC. A comparison of MR and CT in suspected sacroiliitis. J. Comput. Assist. Tomogr. 1996; 20: Yu W, Feng F, Dion E, Yang H, Jiang M, Genant HK. Comparison of radiography, computed tomography and magnetic resonance imaging in the detection of sacroiliitis accompanying ankylosing spondylitis. Skeletal Radiol. 1998; 27: Zink A, Braun J, Listing J, Wollenhaupt J. Disability and handicap in rheumatoid arthritis and ankylosing spondylitis results from the German rheumatological database. German Collaborative Arthritis Centers. J. Rheumatol. 2000; 27: