比较MRI和CT检测早期SpA骶髂关节侵蚀病变

原文

译文

SAT0107 DETECTION OF EROSIONS OF SACROILIAC JOINTS BY MAGNETIC RESONANCE IMAGING IN EARLY SPONDYLOARTHRITIS (SPA) OF THE ECHOSPA COHORT: A COMPARISON WITH COMPUTED TOMOGRAPHY W. P. Maksymowych 1,*, D. Loeuille 2, I. Chary-Valckenaere 2, A. Desvignes-Engelbert 2, O. Judet 3, R. Lambert 4, M. Breban 5, M. A. D Agostino 5 1RHEUMATOLOGY DEPARTMENT, Alberta Heritage Foundation for Medical Research, University of Alberta, , Edmonton, Canada, 2RHEUMATOLOGY DEPARTMENT, CHU Brabois, Nancy University, Nancy, 3Radiology, VSQY UNIVERSITY - AMBROISE PARE HOSPITAL, Boulogne-Billancourt, France, 4Radiology, Alberta Heritage Foundation for Medical Research, University of Alberta,, Edmonton, Canada, 5RHEUMATOLOGY DEPARTMENT, VSQY UNIVERSITY - AMBROISE PARE HOSPITAL, Boulogne-Billancourt, France   Background: Evaluation of MRI as a diagnostic tool for SpA has focused primarily on its ability to detect bone marrow oedema as a manifestation of active sacroiliitis. There is little data that has systematically evaluated its utility in assessing structural lesions in the sacroiliac joints (SIJ). One report has suggested that erosions in the SIJ may be detected by MRI in up to 50% of patients with pre-radiographic SpA and that this is the most specific abnormality for sacroiliitis. However, there is little data validating MRI for the detection of erosions in the SIJ in comparison to radiographic imaging. Objectives: To assess the utility of MRI for detection of erosions in the SIJ in comparison to computed tomography (CT) in patients with early SpA. Methods: We developed an atlas of reference MRI images depicting the entire range of abnormalities observed in the SIJ including erosions. Erosions were recorded in an online data entry module that displays schematics of the SIJ allowing the recording of lesions in each quadrant of each coronal slice. Two readers blinded to patient and diagnosis, independently assessed CT and MRI scans of the SIJ (TIW and STIR), and post-gadolinium fat-saturated T1W sequences) from 22 patients followed in the French multi-center ECHOSPA pre-radiographic SpA prospective cohort, in which patients were included for symptoms suggestive of SpA (i.e. IBP, arthritis, enthesitis or dactylitis, B27+ uveitis and family history of SpA). Semi-coronal MRI slices through the synovial and ligamentous portion of the SIJ were read systematically from anterior to posterior as described in a standardized online training module developed by the Spondyloarthritis Research Consortium of Canada (SPARCC). Erosions visible on CT were recorded for each iliac and sacral cortex. Readers answered dichotomously (yes/no):  This scan confirms the presence of SpA? Sensitivity/specificity was calculated according to clinical diagnosis and reliability of detection of erosions between CT and MRI and between readers was assessed using kappa statistics. Results: Sensitivity/specificity of MRI and CT for the clinical diagnosis of SpA was 75%/40% and 50%/30%, respectively, for reader 1 and 75%/10% and 50%/70%, respectively, for reader 2. Reader concordance for diagnosis was 59% for CT and 81.8% for MRI. Percent agreement for detection of erosions was 73.9% for reader 1 who recorded 19 (21.6%) erosions by CT that were not detected by MRI and 4 (4.5%) erosions detected by MRI but not CT. For reader 2, percent agreement was 52.3% with 13 (14.8%) erosions detected by CT that were not detected by MRI and 29 (32.9%) erosions detected by MRI but not CT. Only one reader achieved significant reliability between CT and MRI for detection of erosions (kappa = 0.39 (95% CI 0.19-0.58, P <0.0001) but inter-reader reliability for detection of erosions by MRI was also low (kappa = 0.33 (95% CI 0.15-0.50, P = 0.0003). Conclusions: Reliable detection of erosions in the SIJ using MRI is difficult without substantial reader training and calibration using radiographic imaging.

比较MRI和CT检测早期SpA骶髂关节侵蚀病变   Maksymowych WP, et al. EULAR 2011. Present No: SAT0107.     背景：对于MRI可否作为SpA诊断工具的评估主要集中在它对反映活动性骶髂关节炎的骨髓水肿的检出能力，而少有系统性评估它对骶髂关节（SIJ）结构性损害的检测性能。有一项研究提示MRI可以在多达50%的放射学前(pre-radiographic)SpA病人中发生侵蚀病变，并称侵蚀是骶髂关节炎的最特异病变。然而，目前尚无研究通过比较MRI与放射学影像从而验证MRI对SIJ侵蚀的检测性能。 目的：通过比较MRI与CT对早期SpA病人的检测结果，从而评估MRI在SIJ侵蚀检测中的应用。 方法：我们已发布了一部MRI参考图谱，它包含了包括侵蚀在内的所有SIJ已知病变。侵蚀评分是通过一种在线数据输入模块进行录入的，它可以显示SIJ结构图以方便对每个冠状扫描层面的每个象限进行评估。2名对病人和诊断均不知情的读片师分别独立地判读CT和MRI(T1加权(T1W)、短时反转回复序列(STIR)以及增强抑脂T1W)成像。22例病人均来自一项法国多中心前瞻性研究队列(ECHOSPA 放射学前SpA)，凡是有SpA提示性症状(即IBP、关节炎、附着点炎或指趾炎)的病人就可纳入该队列。半冠状扫描层覆盖SIJ的滑膜和韧带部，读片师自前向后地系统读片，所用方法在SPARCC开发的在线训练模块中有详细描述。CT可见侵蚀分别按髂骨和骶骨进行记录。读片师对诸如“该扫描能否确定SpA？”等二分式提问(是/否)进行回答。依据临床诊断来计算敏感性/特异性，用kappa统计量评估读片师之间以及CT与MRI之间对侵蚀检测的可信度。 结果：经读片师1评判，发现MRI和CT对SPA临床诊断的敏感性/特异性分别为是75％/40％和50％/30％，而依据读片师2评判的结果为75％/10％和50％/70％。两位读片师的CT诊断一致率为59％，MRI诊断一致率为81.8％。关于侵蚀检测，读片师1的检出一致性为73.9%，他发现19(21.6%)个CT阳性/MRI阴性的侵蚀灶，以及4个（4.5%）MRI阳性/CT阴性的侵蚀灶。读片师2用两种影像学技术判读的侵蚀检测一致率为52.3%，13个(14.8%)CT+/MRI-侵蚀，29个(32.9％)MRI+/CT-侵蚀。 只有一个读片师用CT和MRI检测侵蚀之间的一致性达到统计学显著意义（kappa= 0.39（95％CI: 0.19-0.58），P <0.0001），读片师之间MRI侵蚀检测可信度是低的（kappa= 0.33（95 ％CI: 0.15-0.50），P = 0.0003）。 结论：如果读片者没有坚实的培训基础且不用放射学平片作为校准，那么用MRI检测骶髂关节侵蚀的可靠性是差的。