SAT0392 The diagnostic accuracy of existing grading criteria of sacroiliac joint ct in ankylosing spondylitis

Background Imaging modalities are essential for the diagnosis of ankylosing spondylitis (AS) due to the absence of specific clinical manifestations. Sacroiliac Joint (SIJ) CT has been used to identify sacroiliitis for decades with a higher diagnostic accuracy than radiography in detecting structural changes, and not reducing the specificity like MRI. However, no well-accepted grading system for SIJ CT existed. Objectives We evaluated the diagnostic accuracy of existing grading criteria of sacroiliitis, aiming to provide references for future better reading of SIJ CT in AS. Methods A total of 2714 patients who had received CT scanning for any reasons with complete SIJ structures displaying between June 2012 and December 2015 were included. The CT scans were read by 2 rheumatologists together who had received professional training in radiology. Patients with sacroiliitis were selected and bilateral SIJs of each patient were evaluated separately by the 1984 modified New York (mNY) criteria, the criteria proposed by Lee (Lee criteria)[1] and the criteria from Innsbruck workshop report (Innsbruck criteria)[2], respectively. The grading differences among these criteria were analyzed. Results Amount to 509 patients were detected having sacroiliitis with an average age of 34 years and 64% of male. Among 1018 SIJs of these patients, 45 SIJs graded 1∼3 by mNY criteria were graded 0 by Lee criteria, indicating the better specificity of Lee criteria. Lee criteria was much more convenient and reliable than mNY criteria for its more explicit definitions. The SIJs with definite sacroiliitis estimated by mNY and Lee criteria were 79.37% and 82.91%, respectively, and simply divided into grade 3 or grade 4. Conversely, 85.27% SIJs were identified as definite sacroiliitis and classified into 5 grades, from grade IIB to grade IVB, by Innsbruck criteria, and the percentages of each grade were 8.94%, 26.82%, 20.92%, 10.12% and 18.47%, respectively, which means a higher discrimination capability than the other two criteria. Other than graded by the extent of lesions in mNY criteria or Lee criteria, the grading assessment by Innsbruck criteria was based on the lesion types, which was more consistent with the natural progression of sacroiliitis. Conclusions Lee criteria has a better diagnostic specificity with a lower difficulty in the evaluation process, while Innsbruck criteria is a more detailed grading system, which has a higher consistency with the progression of sacroiliitis in AS. References Lee YH, et al. Rheumatol Int, 2013, 33(4): 1005–1011. Klauser A, et al. J Rheumatol 2004, 31(10): 2041–2047. Disclosure of Interest None declared