Background Synovitis is often present on MRI of OA knees and is an important determinant of pain.
Objectives To evaluate the association between synovitis at contrast enhanced (CE) MRI with microscopic and macroscopic features of synovial tissue inflammation.
Methods A total of 43 patients with symptomatic radiographic knee OA attending the rheumatology or orthopaedic outpatient clinic were included. 22 patients (mean age 61±7yrs, 73% women, mean BMI 30±5kg/mm2) underwent arthroscopy of the knee and macroscopic features (neovascularization, villi, fibrin and hyperplasia) were scored (0-4). Furthermore, 15-20 synovial biopsies per knee were obtained. 21 patients (mean age 60±11yrs, 52% women, mean BMI 30±5kg/mm2) underwent arthroplasty and synovial tissues were collected. After haematoxylin and eosin staining, samples were microscopically scored on features: synovial lining layer hyperplasia (0-3), activation of resident cells/stroma (0-3) and degree of inflammatory infiltrates (0-3). Mean total scores (0-9) and feature score (0-3) by 3 observers were used. Saggital and axial T1-weighted CE MRI images (3T) were used to semi quantitatively score synovitis at 11 sites (0-22) according to Guermazi et al. (Ann Rheum Dis 2011). Self reported pain was assessed by visual analogue scale (VAS, 0-100). For comparison between groups, independent t-test and Mann Witney test were used when appropriate. Pearson adjusted for age or spearman rho correlations were used to calculate correlations for all 43 patients. Statistics were calculated by SPSS 17.0.
Results The mean (SD) synovitis score on MRI was 8.0±3.7, representing a mild synovitis, and was significantly different between patients in arthroscopy group (6.4±2.7) and in arthroplasty group (9.7±3.8). A mean (SD) histology score of 2.5±1.6 was seen, which also differed between groups (2.0±1.4 vs 3.0±1.7, respectively). Median (range) score of macroscopic features (0-4) were 2 (1-4) for neovascularization, 1 (0-3) for hyperplasia, 2 (0-4) for villi and 2 (0-3) for fibrin. Mean VAS pain was significantly lower in the arthroscopy than in the arthroplasty group (44.4±23.0 vs 61.4±14.5, respectively). Synovitis score on MRI correlated with microscopic total synovitis score [r=0.6, P=0.000] and separate features (lining layer [r=0.4, P=0.016], stroma [r=0.3, P=0.033] and infiltrate [r =0.4, P=0.003]). Furthermore MRI synovitis score was significantly correlated with neovascularisation [r=0.6, P=0.001], hyperplasia [r=0.6, P=0.005] and villi [r=0.5, P=0.014], but not with fibrin [r=0.3, P=0.133]. A significant correlation of 0.3 was seen between MRI synovitis score and VAS pain of affected knee.
Conclusions Synovitis severity on T1 weighted CE MRI images is significantly correlated with both macroscopic and microscopic features of synovitis in patients with knee OA. CE MRI evaluation is a reliable, non invasive way to determine synovitis severity in OA patients.
Disclosure of Interest B. de Lange-Brokaar Grant/Research support from: Top Institute Pharma, A. Ioan-Facsinay: None Declared, E. Yusuf: None Declared, W. Visser: None Declared, S. Andersen: None Declared, L. van Toorn: None Declared, G. van Osch: None Declared, A.-M. Zuurmond: None Declared, V. Stojanovic-Susulic: None Declared, M. Reijnierse: None Declared, R. Nelissen: None Declared, T. Huizinga: None Declared, M. Kloppenburg: None Declared