LAUSR

Background: In the light of new therapeutic strategies, it has become essential to detect spondyloarthritis (SpA) in its earliest stages in order to initiate treatment as early as possible. Magnetic resonance imaging (MRI), with its ability to detect active lesions, is usually considered as a key tool for early recognition of sacroiliitis. However, a growing number of studies reported the lack of specificity of bone marrow edema (BME) leading sometimes to a false “positive MRI” of sacroiliac joints (SIJ). In addition, many recent studies have pointed out the utility of detecting structural lesions, especially erosions, reported as the earliest structural change on SIJ in the course of sacroiliitis. Objectives: We aimed to assess the reliability of early recognition of erosions in patients with suspected SpA. Methods: Consecutive patients, aged 16 and over, consulting from February 2014 to February 2017 for symptoms suggestive of SpA (inflammatory back pain, enthesitis or dactylitis…) were enrolled in this cohort. They were referred for computed tomography (CT) and MRI of the SIJ. The CT and MR images were reviewed for the presence of erosions by 2 musculoskeletal radiologists blinded to clinical findings. After a follow-up of 2 years, 2 experienced rheumatologists confirmed or excluded the diagnosis of SpA. Diagnostic utility of erosions for diagnosis of SpA was determined by calculating sensitivity, specificity, positive and negative likelihood ratio with final clinical diagnosis made by rheumatologists as golden standard. Results: Fifty-four patients were included, 13 men and 41 women. The mean age was 39 years [17-71]. The mean duration of symptoms was 75 months (6 years). Cervical, thoracic, lumbar and buttock pain were noted respectively in 46.3%, 37%, 87%, and 57.4% of the studied patients. Morning stiffness was noted in 55.5% of patients. The prevalence of HLA–B27 was 23.4%. After a follow-up of 2 years, the referring rheumatologists made a diagnosis of SpA in 77.8% of patients, whereas SpA was excluded in 22.2%. Among the 42 patients classified as having a confirmed SpA, erosions were detected on SIJ by CT in 64.3% (n=27) and by MRI in (n=18) 42.85% of patients. Among the 12 patients in whom Spa was excluded, erosions were detected on SIJ by CT in 25% (n=3) of patients and were not detected by MRI in any patient. Sensitivity, specificity, positive and negative likelihood ratio of erosions detected by CT were respectively estimated at 64.3%, 75%, 90% and 37.5%. Those of erosions detected by MRI were estimated at 42.9%, 100%, 100%, 33.33%. A statistically significant association was observed between erosions and rheumatologists’ diagnosis of SpA (p= 0.05 for CT and p= 0.012 for MRI). The likehood ratio was estimated respectively for CT and MRI at 3.8 and 7.6. Conclusion: In our study, detection of erosions showed a high specificity for recognition of sacroiliitis, especially when detected by MRI. This study highlights the utility of structural SIJ lesions, particularly erosions, to reduce the number of false positives. Disclosure of Interests: None declared