Change in MRI synovitis correlates with change in pain following intra-articular steroid injection

Abstract

Purpose: Synovitis is a well recognised finding in knee osteoarthritis. It can be identified on Magnetic Resonance Imaging (MRI) as synovial thickening, although injection of contrast material is required to distinguish synovial tissue from synovial fluid. The aim of this analysis was to determinewhether change in synovial tissue volume as assessed using gadolinium (Gd) enhancedMRI imaging correlates with change in knee pain following intra-articular steroid therapy. Methods: Men and women aged 40 years and older who met ACR criteria for the disease, were recruited for participation in an ongoing open label clinical trial of intra-articular steroid therapy. Subjects who took part in the study had significant knee pain and grade 2 or higher knee OA. At baseline visit they completed a questionnaire about their symptoms including KOOS pain scale (100 1⁄4 no pain to 0 1⁄4 extreme pain) and also a VAS score (0 1⁄4 no pain] to 10 1⁄4 extreme pain) for pain on a nominated activity (VASnA). They subsequently had a Gd enhanced MRI immediately prior to having an intra-articular steroid injection with repeat questionnaire and Gd-enhanced MRI scan at follow up visit to assess synovial response usually within 2 weeks after the injection. To assess synovial tissue volume, sagittal (post CE T1W Fat Suppressed images: TR 500ms, TE 17ms; FoV 15.9 x 15.9cm; slice thickness 3mm) scans were obtained. Manual segmentation of the synovial tissue layer was performed on the post contrast knee image by a single observer. Using computer image analysis we excluded cartilage within the segmented space, by thresholding in the associated sagittal (3DWATSc: TR 20ms, TE 7.7ms, FoV 15cm, 288x288) scan. The rest of the segmented space was assumed to be a mixture of fluid and synovial tissue. We calculated the proportion of synovial tissue in every voxel using P 1⁄4 (I mf) / (ms mf) truncated to [0,1], where I is the voxel intensity and mf, ms are the means of the intensity distributions of fluid and synovial tissue volume respectively. We looked at mean synovial volume before and after the steroid injection and looked at the Spearman Rank correlation coefficient between change in synovial volume and change in the level of KOOS-pain and also VASnA. Results: We analysed data from 41 patients. Their mean age was 62.4 years (SD 10.5 years), and 21 were female (51.2%). The median time between baseline and follow up scan was 9 days (IQR 7 to 14 days). The median synovial tissue volume at baseline was 7,556mm (IQR 4,670mm to 11,269 mm), and at follow up was 7,078mm (IQR 3,642mm to 8,541mm); median difference -1,007mm (95% CI -1,909mm to -320mm). Both KOOS pain (24.7pts; 95% CI 18.4 pts to 31.0 pts) and VASnA (-3.2cm; 95% CI -4.1cm to -2.2cm) improved significantly between baseline and follow up. The change in synovial tissue volume correlated with change in VASnA (rs 1⁄4 0.39; p 1⁄4 0.01) though not with KOOS-pain (rs 1⁄4 -0.11 ; p 1⁄4 0.50). Change in VASnA did not appear to correlate with synovial tissue volume at baseline (rs 1⁄4 -0.03; p1⁄40.85); nor did change in KOOS pain (rs 1⁄4 0.17; p1⁄40.30). Conclusions: Synovial tissue volume in knee osteoarthritis, assessed with Gd-enhanced MRI correlates with change in pain assessed as pain on a nominated activity, though not KOOS, in response to intra-articular steroid injection.