Unilateral Sacroiliitis: Differential Diagnosis Between Infectious Sacroiliitis and Spondyloarthritis Based on MRI Findings

Transcription

1 Musculoskeletal Imaging Original Research Kang et al. Use of MRI to Differentiate Infectious Sacroiliitis From Spondyloarthritis Musculoskeletal Imaging Original Research Yusuhn Kang 1 Sung Hwan Hong Ji Young Kim Hye Jin Yoo Ja-Young Choi Minkyung Yi Heung Sik Kang Kang Y, Hong SH, Kim JY, et al. Keywords: infection, MRI, sacroiliac joint, sacroiliitis, spondyloarthropathies DOI: /AJR Received December 11, 2014; accepted after revision April 3, All authors: Department of Radiology, Seoul National University College of Medicine, 101 Daehak-ro Jongno-gu, Seoul , Korea. Address correspondence to S. H. Hong (drhong@snu.ac.kr). AJR 2015; 205: X/15/ American Roentgen Ray Society Unilateral Sacroiliitis: Differential Diagnosis Between Infectious Sacroiliitis and Spondyloarthritis Based on MRI Findings OBJECTIVE. The purpose of this study was to identify the MRI features that aid in the differentiation between infectious sacroiliitis and unilateral sacroiliitis associated with spondyloarthritis. MATERIALS AND METHODS. The MR images of 54 patients who received a diagnosis unilateral sacroiliitis between August 2001 and August 2013 were reviewed. MR images were evaluated for bone lesions (extent and distribution of bone marrow edema and presence and size of bone erosions), soft-tissue lesions (capsulitis, extracapsular fluid collections, and periarticular muscle edema), and joint space enhancement. The Fisher exact test was used for comparison of categoric data, and multivariate stepwise logistic regression analysis was performed. RESULTS. Thick capsulitis, extracapsular fluid collection, and periarticular muscle edema were all more frequently observed in infectious sacroiliitis (p < 0.001). Iliac-dominant bone marrow edema and joint space enhancement were statistically significantly more common in spondyloarthritis (p < and p = 0.014, respectively). The presence of periarticular muscle edema was the only independently differentiating variable on multivariate stepwise logistic regression analysis. When periarticular muscle edema was the sole predictor, unilateral sacroiliitis in spondyloarthritis was correctly identified in 77.3% of cases, and infectious sacroiliitis was correctly identified in 90.6% of cases. The overall accuracy was 85.2%. CONCLUSION. MRI features of the bone lesions, soft-tissue lesions, and joint space enhancement in unilateral sacroiliitis aid in the differential diagnosis between infection and spondyloarthritis. Among various findings, periarticular muscle edema was the single most important predictor of infectious sacroiliitis. S acroiliitis is commonly the first manifestation of spondyloarthritis [1, 2]. Sacroiliitis associated with spondyloarthritis is almost invariably bilateral and symmetric at later stages of the disease [3]. However, unilateral involvement of the sacroiliac joints has been noted in spondyloarthritis, psoriatic arthritis, reactive arthritis, and the early stages of ankylosing spondylitis [4 6], which poses a challenge in the differential diagnosis with other causes of sacroiliitis. Over the past years, MRI has garnered attention for its ability to detect active inflammation of the sacroiliac joint. With MRI, the earliest signs of spondyloarthritis are identified [2, 7, 8], which may lead to early treatment with such drugs as anti tumor necrosis factor agents, thus slowing progression to structural damage and ankylosis [8, 9]. The new criteria for classification of axial spondyloarthritis defined by the Assessment of SpondyloArthritis International Society in 2009 [10] include MRI findings as one of the major criteria. The four MRI findings of active sacroiliitis are osteitis or bone marrow edema, enthesitis, capsulitis, and synovitis. Among these findings, osteitis or bone marrow edema is considered the single most important criterion in making the diagnosis [10]. However, not all four findings are specific for sacroiliitis in spondyloarthritis, and they may be found in sacroiliac disease of other causes, including infection. Determining the cause of unilateral sacroiliitis may be difficult in the early stage of the disease. Even clinical and laboratory features of infection, including fever, leukocytosis, and elevated C-reactive protein (CRP) levels, can be variable and inconsistent in infectious sacroiliitis [11]. The differentiation between infectious sacroiliitis and spondyloarthritis is of great importance because different therapeutic approaches are required. Although infection is an uncommon cause of sacroiliitis, early recognition is important 1048 AJR:205, November 2015

2 Use of MRI to Differentiate Infectious Sacroiliitis From Spondyloarthritis because the delay in diagnosis results in increased morbidity [12, 13]. Stürzenbecher et al. [14] reported that infectious sacroiliitis shows subperiosteal infiltrations, which are transcapsular infiltrations of juxtaarticular muscle layers, on MRI, and they suggested that these features may be used to differentiate infectious sacroiliitis from sacroiliitis in spondyloarthritis. However, their study included only patients with infectious sacroiliitis, and to our knowledge, no reports are available concerning the differentiation of sacroiliitis in spondyloarthritis and infectious sacroiliitis with MRI. A C D Fig year-old man with unilateral sacroiliitis associated with spondyloarthritis. A and B, Oblique axial T1-weighted (A) and T2-weighted (B) fast spin-echo images show irregularity and small erosions (arrowhead, A) at cartilaginous portion of right sacroiliac joint. C and D, Oblique axial (C) and coronal (D) fat-suppressed contrast-enhanced T1-weighted images of sacroiliac joints show enhancement within sacrum and ilium consistent with bone marrow edema. Distribution of bone marrow edema shows iliac dominance. Subtle hyperintensity is noted in pericapsular region (arrow, D), but lesion does not exceed 5 mm in thickness. In addition, contrast enhancement is noted in sacroiliac joint space (arrowheads, D), finding more commonly noted in spondyloarthritis than in infectious sacroiliitis. Thus, the purpose of this study was to identify the MRI features that aid in the differentiation between infectious sacroiliitis and unilateral sacroiliitis in spondyloarthritis. Materials and Methods This retrospective study was approved by the institutional review board of Seoul National University Hospital, and informed consent was waived. Patient Selection Sixty-nine patients with acute unilateral sacroiliitis diagnosed on MRI between August 2001 and August 2013 were selected from our radiology report database. Inclusion criteria were clinical and radiologic diagnosis of unilateral sacroiliitis and MRI of the sacroiliac joint. Inflammation is well visualized on two different MRI sequences, contrast-enhanced fat-suppressed T1-weighted images and fluid-sensitive sequences, such as fatsuppressed T2-weighted images or STIR [4]. Accordingly, we ensured that all the cases included either a fluid-sensitive sequence or a contrast-enhanced fat-suppressed T1-weighted image in the axial and coronal plane with full coverage of the sacroiliac joint. We excluded 14 patients because of MRI with insufficient coverage of the sacroiliac joints (n = 13) and a history of surgery or trauma B AJR:205, November

3 Kang et al. A C D Fig year-old woman with infectious sacroiliitis due to Streptococcus pneumoniae. A and B, Axial T2-weighted fast spin-echo images acquired at level of sacroiliac joint show small amount of sacroiliac joint effusion (arrowhead, A) and large extracapsular abscess anterior to left sacroiliac joint (arrow, A). Periarticular muscle edema is noted in left psoas (white arrowheads, B) and iliacus (black arrowhead, B) muscles, at level of fifth lumbar vertebra. C and D, Axial T1-weighted fast spin-echo (C) and contrast-enhanced fat-suppressed T1-weighted (D) images show large bone erosion (arrowheads) exceeding 1 cm in its greatest dimension. of the sacroiliac joint (n = 1). As a result, a total of 54 patients were included (mean age, 41 years; range, years), including 23 male patients (mean age, 35.4 years; range, years) and 31 female patients (mean age, 45.7 years; range, years). Diagnoses included infectious sacroiliitis (n = 32) and spondyloarthritis (n = 22). Among the 32 patients with infectious sacroiliitis, the causative organism was proven in 22 patients by culture of tissue obtained from the sacroiliac joint (n = 11), blood (n = 6), or tissue obtained from another primary site of infection (n = 5). The most common causative organism was Staphylococcus aureus (n = 9), followed by Mycobacterium tuberculosis (n = 5), Streptococcus species (n = 4), Pseudomonas aeruginosa (n = 2), Corynebacterium species (n = 1), and Klebsiella pneumoniae (n = 1). For the 10 patients for whom the causative organism could not be proven, the diagnosis of infectious sacroiliitis was based on the clinical course of the disease and the response to antibiotic treatment. Among the 22 patients with spondyloarthritis, 16 (mean age, 26.9 years; age range, years) closely met the Assessment of SpondyloArthritis International Society classification criteria for axial spondyloarthritis [15]; all 16 patients were younger than 45 years, had evidence of active sacroiliitis on MRI, and were positive for human leukocyte antigen B27. In the other six patients with spondyloarthritis (mean age, 46 years; age range, years), the Assessment of SpondyloArthritis International Society classification criteria for axial spondyloarthritis were not met, and there was no evidence suggesting a specific disorder, such as ankylosing spondylitis, psoriasis, or inflammatory bowel disease. However, on the basis of the clinical examinations and MRI findings, the patients were clinically diagnosed as having undifferentiated spondyloarthritis [16]. Image Acquisition Because of the retrospective nature of the study, images were obtained with a variety of MRI scanners, including 1-T units (n = 3; Magnetom Impact Expert, Siemens Healthcare), 1.5-T units (n = 50; Signa HDx or HDxt, GE Healthcare; and Sonata, Siemens Healthcare), and 3-T units (n = 2; Mag- B 1050 AJR:205, November 2015

4 Use of MRI to Differentiate Infectious Sacroiliitis From Spondyloarthritis A B Fig year-old woman with infectious sacroiliitis (pathogen not isolated). A, Oblique axial T2-weighted fast spin-echo image shows small amount of left sacroiliac joint effusion (arrow) associated with mild anterior capsular bulging. Periarticular muscle edema is noted anteriorly along left iliopsoas muscle (arrowhead). B and C, Oblique axial (B) and coronal (C) contrast-enhanced fat-suppressed T1-weighted images show enhancement in pericapsular soft tissue (arrows, B and C) and along iliopsoas (arrowhead, B) and gluteus (arrowhead, C) muscles. However, contrast enhancement is not observed in bone marrow and sacroiliac joint space. netom Verio, Siemens Healthcare). The imaging protocol and parameters also varied from case to case. Images were acquired in the orthogonal axial and coronal planes in 49 patients, whereas in five patients, the images were acquired in oblique axial and oblique coronal planes oriented perpendicular and parallel to the long axis of the sacrum. For the orthogonal axial T1- and T2-weighted fast spin-echo sequences, images were acquired with the following imaging parameters: TR/TE of / and /9 11, respectively; sections; 4.0- to 8.0-mm section thickness; 1.0- to 2.0-mm intersection gap; FOV of mm; matrix of ; two acquired signals; and an echo-train length of and 4, respectively. For the oblique axial T1- and T2-weighted fast spin-echo sequences, the imaging parameters were as follows: TR/TE of / and / , respectively; sections; 4.0-mm section thickness; 0- to 0.2-mm intersection gap; FOV of mm; matrix of ; two acquired signals; and an echotrain length of and 4 5, respectively. Fat-suppressed T1-weighted spin-echo sequences were acquired in the orthogonal or oblique coronal or axial planes after the administration of contrast agent. The image parameters for the orthogonal axial and coronal images were TR/TE of / , sections, 4.0- to 7.0-mm section thickness, 1.0- to 2.0-mm intersection gap, FOV of mm, matrix of , two acquired signals, and echo-train length of 4. The oblique coronal and oblique axial contrast-enhanced images were acquired with a TR/TE of / , sections, 4.0-mm section thickness, 0- to 0.2-mm intersection gap, FOV of mm, matrix of , two acquired signals, and echo-train length of 4 5. Image Analysis The MR images were reviewed by a two observers (radiologists with 3 and 16 years of musculoskeletal MRI experience, respectively) in consensus, on a PACS workstation (Marosis, Marotech). The observers were aware of the diagnosis of unilateral sacroiliitis but were blinded to the cause of sacroiliitis. MR images were evaluated focusing on the presence of the following features: the characteristics of bone lesions, including the extent of subchondral bone marrow edema, the distribution of bone marrow edema, and the presence and size of bone erosions; the presence of soft-tissue abnormalities, including capsulitis, extracapsular fluid collection, and periarticular muscle edema; and the presence A B C Fig year-old man with infectious sacroiliitis due to Pseudomonas aeruginosa. A and B, Axial T2-weighted fast spin-echo images acquired at level of sacroiliac joint show small amount of left sacroiliac joint effusion (arrow, A) associated with minimal anterior capsular bulging. Periarticular muscle edema is noted anteriorly along left iliopsoas muscle (arrowhead, A). Periarticular muscle edema extends cranially along iliacus muscle (arrowhead, B). C, Contrast-enhanced fat-suppressed T1-weighted image shows bone marrow edema equally distributed in sacrum and ilium. Enhancement is noted in pericapsular soft tissue (arrowhead). C AJR:205, November

5 Kang et al. of joint space enhancement. Bone marrow edema was defined as hyperintensity within the sacrum or ilium on STIR or fat-suppressed T2-weighted sequences or contrast-enhanced T1-weighted fat-suppressed sequences [10]. Subchondral bone marrow edema involving more than half of the articular surface of the sacrum and ilium was considered extensive. The distribution of edema was divided into three categories: iliac-dominant, sacral-dominant, or even in distribution. Bone erosion size was evaluated in its greatest dimension and was defined to be large if it exceeded 1 cm. Capsulitis indicated a nonfluid soft-tissue lesion located in the pericapsular area and was classified to be thick if the soft-tissue lesion exceeded 5 mm in thickness. Extracapsular fluid collection was defined as a localized fluid collection located outside the joint capsule, with or without rim enhancement, and was evaluated on STIR or fat-suppressed T2-weighted sequences and contrast-enhanced fat-suppressed T1-weighted sequences. The iliacus, psoas, and piriformis muscles were evaluated for the presence of periarticular muscle edema on STIR or fat-suppressed T2-weighted sequences. MR images were considered positive for joint space enhancement when hyperintensity was noted between the ilium and sacrum on contrast-enhanced T1-weighted fatsuppressed images [10]. Laboratory Findings The medical records of the patients were reviewed for the laboratory test results, including WBC count and CRP level. Only the laboratory test results obtained within 1 week of the MRI were used for analysis. Statistical Analysis Univariate analysis was performed with the Fisher exact test to compare the categoric variables of interest in infectious sacroiliitis and spondyloarthritis. The clinicolaboratory findings, including the age of patient, WBC count, and CRP levels, were compared with the t test. Odds ratios (with 95% CIs) for the differentiation of infectious sacroiliitis from spondyloarthritis were calculated for each MRI criterion. A multivariate stepwise logistic regression model was used to determine the best predictor in the differential diagnosis of infectious sacroiliitis and spondyloarthritis. All statistical analyses were performed with SPSS software (version 18.0 for Windows, SPSS). Results with p < 0.05 were considered statistically significant. Results Univariate Analysis The MRI findings of spondyloarthritis and infectious sacroiliitis are summarized TABLE 1: MRI Findings in Spondyloarthritis and Infectious Sacroiliitis Finding Spondyloarthritis (n = 22) Infectious Sacroiliitis (n = 32) p Bone lesions Extent of bone marrow edema None or minimal 15 (68.2) 13 (40.6) Extensive 7 (31.8) 19 (59.4) Distribution of bone marrow edema a < b Iliac dominance 19 (95.0) 11 (39.3) Even distribution 1 (5.0) 10 (35.7) Sacral dominance 0 (0.0) 7 (25.0) Bone erosion Absent or small 22 (100.0) 26 (81.3) Large (> 1 cm) 0 (0.0) 6 (18.8) Soft-tissue lesions Capsulitis < Absent or thin 14 (63.6) 3 (9.4) Thick (> 5 mm) 8 (36.4) 29 (90.6) Extracapsular fluid collection < Absent 22 (100.0) 19 (59.4) Present 0 (0.0) 13 (40.6) Periarticular muscle edema < Absent 17 (77.3) 3 (9.4) Present 5 (22.7) 29 (90.6) Joint space enhancement c Absent 3 (15.0) 14 (51.9) Present 17 (85.0) 13 (48.1) Mean age at presentation (y) WBC count (cells/μl), mean ± SD d 8365 ± ,904 ± Serum C-reactive protein level (mg/dl), mean ± SD d 2.08 ± ± Note Data are number (%) of patients, unless indicated otherwise. a Six cases in which bone marrow edema was not noted were excluded from the analysis. Twenty cases of spondyloarthritis and 28 cases of infectious sacroiliitis were included. b Iliac dominance versus even distribution and sacral dominance. c Joint space enhancement was assessed in a total of 47 cases (20 spondyloarthritis and 27 infectious sacroiliitis cases) in which contrast-enhanced T1-weighted images were acquired. d Subgroup analysis was done with those with laboratory examination results within 1week of MRI, for a total of 39 patients (11 with spondyloarthritis and 28 with infectious sacroiliitis). in Table 1. Bone marrow edema was noted in the sacrum or ilium or both in 20 of 22 patients with spondyloarthritis (90.9%) and in 28 of 32 patients with infectious sacroiliitis (87.5%). The distribution of bone marrow edema showed a statistically significant difference between infectious sacroiliitis and spondyloarthritis (p < 0.001). An iliac-dominant pattern was more common in spondyloarthritis (19/20 [95.0%]), whereas a noniliac-dominant pattern was more common in infectious sacroiliitis (17/28 [60.7%]) (Fig. 1). However, the extent of bone marrow edema did not statistically significantly differ between infectious sacroiliitis and spondyloarthritis (p = 0.057). Large bone erosions were found only in cases of infectious sacroiliitis (6/32 [18.8%]; p = 0.035) (Fig. 2). Thick capsulitis was observed more frequently in infectious sacroiliitis (29/32 [90.6%]) than in spondyloarthritis (8/22 [36.4%]; p < 0.001), and extracapsular fluid collection was noted exclusively in infectious sacroiliitis (13/32 [40.6%]) (Fig. 2). Another discriminative feature was periarticular muscle edema (Fig. 2B), which was 1052 AJR:205, November 2015

6 Use of MRI to Differentiate Infectious Sacroiliitis From Spondyloarthritis seen more commonly in infectious sacroiliitis (29/32 [90.6%]) than in spondyloarthritis (5/22 [22.7%]; p < 0.001). Among the 29 cases of infectious sacroiliitis showing periarticular muscle edema, 17 cases (58.6%) showed muscle edema in the absence of extracapsular fluid collection. Seven cases of infectious sacroiliitis included in our study showed very subtle or no bone marrow signal alteration. In all seven cases, periarticular muscle edema of the iliacus, psoas, or piriformis muscle was noted (Figs. 3 and 4). Joint space enhancement was more common in spondyloarthritis (17/20 [85.0%]; Fig. 1D) than in infectious sacroiliitis (13/27 [48.1%]; Fig. 3D; p = 0.014). Patients with infectious sacroiliitis (mean age, 47.6 years; range, years) were statistically significantly older than patients with spondyloarthritis (mean age, 32.1 years; range, years; p = 0.005). The mean serum CRP level was statistically significantly higher in patients with infectious sacroiliitis (mean, mg/dl; range, mg/dl) than in those with spondyloarthritis (mean, 2.08 mg/dl; range, mg/dl; p = 0.008). The mean WBC count was higher in patients with infectious sacroiliitis (mean, 11,904 cells/μl; range, ,500 cells/μl) than in patients with spondyloarthritis (mean, 8365 cells/μl; range, ,200 cells/μl), but the difference was not statistically significant (p = 0.061). Multivariate Analysis Multivariate stepwise logistic regression analysis showed that the presence of periarticular muscle edema was the only independently differentiating variable, with an odds ratio of 32.9 (95% CI, ). When periarticular muscle edema was the sole predictor, unilateral sacroiliitis in spondyloarthritis was correctly identified in 77.3% (17/22) of cases, and infectious sacroiliitis was correctly identified in 90.6% (29/32) of cases, with overall accuracy of 85.2% (46/54). Discussion Our study results show that the presence of large bone erosion, thick capsulitis, extracapsular fluid collection, and periarticular muscle edema on MRI suggest infectious sacroiliitis, whereas iliac-dominant bone marrow edema and joint space enhancement favor the diagnosis of unilateral sacroiliitis in spondyloarthritis. In this study, the presence of periarticular muscle edema was the only independent variable that differentiated infectious sacroiliitis from spondyloarthritis, with overall accuracy of 83.6%. Regional muscle swelling has been reported to be indicative of infectious sacroiliitis by previous studies [12, 14, 17, 18]. In a study by Klein et al. [12], all the cases of infectious sacroiliitis showed fluid or inflammation in the iliopsoas muscle that tracked posterior to the iliopsoas muscle. Le Breton et al. [17] reported that swelling of the muscles around the sacroiliac joint, which appeared as a decrease of fat between the iliacus and the psoas muscles, could confirm the diagnosis of infectious sacroiliitis. In our study, for the seven cases of infectious sacroiliitis in which bone marrow edema was either minimal or absent, the presence of periarticular muscle edema strongly suggested the diagnosis of infectious sacroiliitis. Therefore, the surrounding muscle should be thoroughly examined for abnormal signal intensities when evaluating the sacroiliac joints. However, five cases with spondyloarthritis also showed periarticular muscle edema in our study. The Assessment of SpondyloArthritis International Society handbook [10] states that, because the joint capsule gradually continues into the periosteum, capsulitis may extend far medially and laterally along the iliac and sacral periosteum. The periarticular muscle edema in spondyloarthritis may be attributable to the extensive capsulitis that results in reactive edema in the adjacent muscles. Thick capsulitis was also more commonly noted in infectious sacroiliitis than in spondyloarthritis, and extracapsular fluid collections were exclusively found in infectious sacroiliitis. Our results coincide with those of previous studies that have shown the involvement of surrounding soft tissue in infectious sacroiliitis [14, 19]. Stürzenbecher et al. [14] found transcapsular infiltrations of the juxtaarticular muscle groups obscuring the fasciae on MRI in all their patients with infectious sacroiliitis. Abscess formation was also an initial finding in eight of the 11 cases in their study. Inflammation of the sacroiliac joints due to infection crosses anatomic borders [20], whereas in spondyloarthritis, inflammation is limited to the bone and sacroiliac joint space [2]. Prior studies have shown imaging findings suggesting inflammation of the joint capsule in sacroiliitis of spondyloarthritis, but the inflammatory process did not extend to the extracapsular soft tissue [14, 21]. This supports the results of our study showing extracapsular abnormalities more frequently in infectious sacroiliitis. Interestingly, sacroiliitis in spondyloarthritis showed iliac dominance in the distribution of bone marrow edema, whereas infectious sacroiliitis more commonly showed either sacral dominance or an even distribution in our study. The predilection for the ilium of the sacroiliac joint in early spondyloarthritis has been shown in prior studies [4, 22]. The sacroiliac joint cartilage is firmly attached to the adjacent bone by fibrous tissue and may be considered an enthesis [4], which explains the characteristic involvement of the sacroiliac joint in spondyloarthritis. Bowen et al. [23] reported that the iliac cartilage is fibrocartilaginous, whereas the sacral cartilage is hyaline. This feature may explain the iliac dominance of bone marrow edema in early spondyloarthritis. Traditionally, infectious sacroiliitis has also been thought to involve the iliac side more severely. Several hypotheses have been postulated for this predilection, including the protection provided by the thicker sacral cartilage and the slow subchondral circulation of the ilium, which predisposes it to infections [12]. Yet the results of our study showed the contrary, and they are supported by a study by Klein et al. [12]. Among the six cases of infectious sacroiliitis included in their study, four involved the sacrum to a greater extent than the ilium, and two cases showed sacral involvement equal to that of the ilium. Erosions are bony defects seen at the joint margin and reflect structural damage due to inflammation [20]. Subchondral or periarticular erosions are found in both infectious sacroiliitis and sacroiliitis associated with spondyloarthritis, thus complicating the differential diagnosis. In our study, large bone erosions (> 1 cm) were found only in cases of infectious sacroiliitis (n = 6). However, because erosions usually appear later in the course of the disease, their value in the early differential diagnosis of unilateral sacroiliitis is probably limited. In addition, psoriatic sacroiliitis may show considerable erosion and eburnation of subchondral bony surfaces in the sacroiliac joint without joint space narrowing [24]. None of the patients included in our study had psoriatic arthritis, but it is noteworthy that the erosions noted in psoriatic sacroiliitis may simulate the changes seen in infectious sacroiliitis. Joint space enhancement was more frequently found in unilateral sacroiliitis associated with spondyloarthritis than in infectious sacroiliitis in our study. Contrast enhancement in the synovial part of the sac- AJR:205, November

7 Kang et al. roiliac joint is defined as synovitis by the Assessment of SpondyloArthritis International Society [10]. Prior reports have shown that there may be enhancement within the sacroiliac joint in spondyloarthritis from synovitis and pannus, typically involving the inferior aspect of the joint [25, 26]. Histologically, only sparse amounts of synovial tissue are found in the sacroiliac joint near the anterior and posterior joint capsules [27]; therefore, the enhancement noted between the sacral and iliac subchondral bone cannot be explained as the enhancement of the synovium itself. One explanation for the enhancement within the joint is the ingrowth of vessels accompanying the proliferative inflammatory tissue [27]. Another possible explanation may be the contrast enhancement of the articular fibrocartilage itself because the articular fibrocartilage can be considered an enthesis [4, 27]. However, further studies are needed on this matter. Patients with infectious sacroiliitis were statistically significantly older than those with unilateral sacroiliitis associated with spondyloarthritis. The Assessment of SpondyloArthritis International Society classification criteria for axial spondyloarthritis specify the age of onset as younger than 45 years [15], and published data indicate that the mean age of onset of axial spondyloarthritis is around 24 years [28]. Our data showed consistent results; among the 22 patients with spondyloarthritis, 20 patients were younger than 45 years. The two patients older than 50 years may be classified as having late-onset spondyloarthritis, which has a reported prevalence between 3% and 8% [29]. Patients with infectious sacroiliitis encompassed a wide age range, including the age group commonly affected by spondyloarthritis. Even though the mean age of patients was statistically significantly different between the two groups, it was not an independently differentiating variable on multivariate logistic regression analysis. The limitations of our study should be mentioned. First, because of the retrospective nature of our study, the images were obtained with various MRI scanners and protocols, and the imaging planes were not optimized for the evaluation of the sacroiliac joint. The oblique coronal and oblique axial planes, defined with respect to the long axis of the sacrum, are considered the standard imaging planes for evaluating the sacroiliac joint [2, 20]. However, most of the axial and coronal images included in our study were acquired with reference to the long axis of the body. Second, the evaluation of enthesitis was not included in our study. Enthesitis is defined as a hyperintense signal on STIR images or contrastenhanced T1-weighted fat-suppressed images at ligament and tendon attachments, including the retroarticular space, according the Assessment of SpondyloArthritis International Society classification criteria [15]. On retrospective review of the MR images, we found that hyperintensity at the retroarticular space was commonly noted in patients with both infectious sacroiliitis and spondyloarthritis. However, it was difficult to determine whether the hyperintensity is a finding of true enthesitis or an extension of periarticular soft-tissue edema accompanying sacroiliitis, and, in turn, we did not include enthesitis in our analysis. Third, because the MR images included in our study were those of the sacroiliac joint, we could not evaluate the associated changes of the spine. In a clinical setting, the presence of spine lesions, such as the MRI corner sign [30], Andersson lesions, facet joint arthritis, and enthesitis of spinal ligaments [10], may help confirm a diagnosis of spondyloarthritis. We conclude that MRI findings of extensive extracapsular soft-tissue abnormalities, including periarticular muscle edema, thick capsulitis, and extracapsular fluid collections, and large bone erosion may enable reliable differential diagnosis of infectious sacroiliitis from unilateral sacroiliitis associated with spondyloarthritis, whereas the presence of iliac-dominant bone marrow edema and joint space enhancement supports the diagnosis of spondyloarthritis. Among these findings, periarticular muscle edema was the single most important predictor of infectious sacroiliitis. References 1. Braun J, Sieper J. The sacroiliac joint in the spondyloarthropathies. Curr Opin Rheumatol 1996; 8: Navallas M, Ares J, Beltrán B, Lisbona MP, Maymó J, Solano A. Sacroiliitis associated with axial spondyloarthropathy: new concepts and latest trends. RadioGraphics 2013; 33: Resnick D, Kransdorf MJ. Bone and joint imaging, 3rd ed. Philadelphia, PA: Elsevier Saunders, 2005: Muche B, Bollow M, Francois RJ, Sieper J, Hamm B, Braun J. Anatomic structures involved in early- and late-stage sacroiliitis in spondylarthritis: a detailed analysis by contrast-enhanced magnetic resonance imaging. Arthritis Rheum 2003; 48: Canella C, Schau B, Ribeiro E, Sbaffi B, Marchiori E. MRI in seronegative spondyloarthritis: imaging features and differential diagnosis in the spine and sacroiliac joints. AJR 2013; 200: Helliwell PS, Hickling P, Wright V. Do the radiological changes of classic ankylosing spondylitis differ from the changes found in the spondylitis associated with inflammatory bowel disease, psoriasis, and reactive arthritis? Ann Rheum Dis 1998; 57: Bredella MA, Steinbach LS, Morgan S, Ward M, Davis JC. MRI of the sacroiliac joints in patients with moderate to severe ankylosing spondylitis. AJR 2006; 187: Lacout A, Rousselin B, Pelage JP. CT and MRI of spine and sacroiliac involvement in spondyloarthropathy. AJR 2008; 191: Braun J, Sieper J. Ankylosing spondylitis. Lancet 2007; 369: Sieper J, Rudwaleit M, Baraliakos X, et al. The Assessment of SpondyloArthritis International Society (ASAS) handbook: a guide to assess spondyloarthritis. Ann Rheum Dis 2009; 68(suppl 2):ii1 ii Hermet M, Minichiello E, Flipo RM, et al. Infectious sacroiliitis: a retrospective, multicentre study of 39 adults. BMC Infect Dis 2012; 12: Klein MA, Winalski CS, Wax MR, Piwnica- Worms DR. MR imaging of septic sacroiliitis. J Comput Assist Tomogr 1991; 15: Horgan JG, Walker M, Newman JH, Watt I. Scintigraphy in the diagnosis and management of septic sacro-iliitis. Clin Radiol 1983; 34: Stürzenbecher A, Braun J, Paris S, Biedermann T, Hamm B, Bollow M. MR imaging of septic sacroiliitis. Skeletal Radiol 2000; 29: Rudwaleit M, Braun J, Sieper J; Assessment of SpondyloArthritis International Society. ASAS classification criteria for axial spondyloarthritis [in German]. Z Rheumatol 2009; 68: Zochling J, Brandt J, Braun J. The current concept of spondyloarthritis with special emphasis on undifferentiated spondyloarthritis. Rheumatology 2005; 44: Le Breton C, Frey I, Carette MF, et al. Infectious sacroiliitis: value of computed tomography (CT) and magnetic resonance imaging (MRI). Eur Radiol 1992; 2: Rosenberg D, Baskies AM, Deckers PJ, Leiter BE, Ordia JI, Yablon IG. Pyogenic sacroiliitis: an absolute indication for computerized tomographic scanning. Clin Orthop Relat Res 1984; Sandrasegaran K, Saifuddin A, Coral A, Butt WP AJR:205, November 2015

8 Use of MRI to Differentiate Infectious Sacroiliitis From Spondyloarthritis Magnetic resonance imaging of septic sacroiliitis. ankylosing spondylitis is not merely enthesitis. Musculoskelet Radiol 2008; 12:72 82 Skeletal Radiol 1994; 23: Arthritis Rheum 2000; 43: Hermann KG, Bollow M. Magnetic resonance im- 20. Aydingoz U, Yildiz AE, Ozdemir ZM, 23. Bowen V, Cassidy JD. Macroscopic and micro- aging of sacroiliitis in patients with spondyloar- Yildirim SA, Erkus F, Ergen FB. A critical over- scopic anatomy of the sacroiliac joint from em- thritis: correlation with anatomy and histology. view of the imaging arm of the ASAS criteria for bryonic life until the eighth decade. Spine 1981; Rofo 2014; 186: diagnosing axial spondyloarthritis: what the radiologist should know. Diagn Interv Radiol (Ank) 2012; 18: Bollow M, Braun J, Hamm B, et al. Early sacroiliitis in patients with spondyloarthropathy: evaluation with dynamic gadolinium-enhanced MR imaging. Radiology 1995; 194: François RJ, Gardner DL, Degrave EJ, Bywaters EG. Histopathologic evidence that sacroiliitis in 6: Resnick D. Diagnosis of bone and joint disorders, 4th ed. Philadelphia, PA: Elsevier Saunders, 2002: Bollow M, Hermann KG, Biedermann T, Sieper J, Schontube M, Braun J. Very early spondyloarthritis: where the inflammation in the sacroiliac joints starts. Ann Rheum Dis 2005; 64: Tuite MJ. Sacroiliac joint imaging. Semin 28. Ozgocmen S, Khan MA. Current concept of spondyloarthritis: special emphasis on early referral and diagnosis. Curr Rheumatol Rep 2012; 14: Hmamouchi I, Bahiri R, Hajjaj-Hassouni N. Clinical and radiological presentations of late-onset spondyloarthritis. ISRN Rheumatol 2011; 2011: Kim NR, Choi JY, Hong SH, et al. MR corner sign : value for predicting presence of ankylosing spondylitis. AJR 2008; 191: AJR:205, November