THU0616 WHOLE-BODY MRI OF PSORIATIC ARTHRITIS AND RHEUMATOID ARTHRITIS PATIENTS AND HEALTHY CONTROLS – INTERSCAN, INTRAREADER AND INTERREADER AGREEMENT AND DISTRIBUTION OF FINDINGS

Abstract

Background: Whole-body MRI (WBMRI) is a promising tool for monitoring disease activity in inflammatory joint diseases. Earlier studies have shown good correlation with conventional MRI and scoring systems for WBMRI have been developed [1,2]. However, the validation of WBMRI is limited; no studies have evaluated the agreement between repeated scans (interscan agreement) and only few studies have evaluated the intra- and interreader agreement. Objectives: To validate WBMRI by evaluating the interscan agreement in patients with psoriatic arthritis (PsA), rheumatoid arthritis (RA) and healthy controls (HC) and to evaluate the intra- and interreader agreement and determine the distribution of pathologies in the subjects. Methods: WBMRI was performed twice with a one-week interval in 14 patients with PsA, 10 with RA and 16 HC. Coronal images of shoulders, hips, hands and ankles/feet, and sagittal images of knees, ankles, feet and spine were obtained (STIR and pre- and post-contrast T1 weighted spin echo images). Images were anonymized and read in pairs with unknown chronological order by experienced readers (peripheral: IE; spine: SK). WBMRI was scored for 83 peripheral joints and for 33 peripheral entheses according to the OMERACT WBMRI scoring system [1], and according to the CanDen MRI spine scoring system [2]. Ten image sets were re-anonymized for assessment of intra- and interreader agreement (peripheral and spine: MO). Agreement was calculated on lesion level by percentage exact agreement (PEA) and Cohen's kappa with squared weights, and for sum scores by absolute agreement singlemeasure intraclass correlation coefficient (ICC). Results: The age in the PsA/RA/HC was median (range) 48(31-68)/49(26- 58)/35(23-54) years and the symptom duration 10(0-24)/7(3-24)/NA years. WBMRI of the spine and peripheral joints and entheses generally showed moderate to almost perfect interscan agreement with a PEA ranging from 95-100%, kappa ranging from 0.71-1.00 and ICC ranging from 0.95-1.00 (Table 1). Intra- and interreader agreement showed moderate to almost perfect agreement with few exceptions (Table 2). More lesions were found in patients than HC. PsA patients had more lesions in acromioclavicular and sternoclavicular joints, whereas RA patients had more hand lesions (Figure 1). PsA patients had more lesions in the spine (Figure 2). Conclusion: WBMRI of the spine and peripheral joints and entheses showed very good interscan agreement, implying that repositioning between examinations does not markedly affect scoring of lesions. Intraand interreader agreement showed moderate to almost perfect agreement. The distribution of findings in PsA, RA and HC was determined.