Background General Synovitis score (GSS) has been developed by Krenn et al. in order to discriminate inflammatory arthritis (IA) and non-inflammatory arthritis (NIA) (1). This score assesses 3 major components of synovitis: lining layer hyperplasia, activation of resident cells (stroma) and inflammatory infiltrate. All components are graded semi-quantitatively from 0 to 3 and the total score is on 9. High-grade synovitis is highly associated with IA and is defined by a score upper than 5 with a sensitivity of 61.7% and a specificity of 96.1%. As immunohistochemistry (IHC) is frequently used to better characterize synovitis, we propose to create a new IMmunologic SYnovitis SCore (IMSYC) adding 5 components to the GSS: CD68, CD3, CD20, CD31 and Ki67 immunostaining.
Objectives Our work aimed to evaluate the diagnostic performance of this new score including IHC, to define the best cut off for inflammatory arthritis recognition, and to compare its diagnostic performance with the GSS.
Methods 53 synovial samples from patients were obtained during surgery (arthroplasty or synovectomy). All patients gave written consent prior surgery. Samples were cut and Hematoxylin and eosin stained. CD68, CD3, CD20, CD31 and KI67 Immunohistochemistry were performed. GSS was assessed for each slide and semi-quantitative 4 scale scores (0–3) were given for each immunostaining, in a blind manner. The score is calculated on 24 (GSS 0–9 points, and 0–3 score for each of the 5 immunostaining). A representative amount of slides was read by 2 observers with a good interobserver variability (spearman correlation coefficient of 0.95, p<0.0005). They then defined a consensual and reproducible scoring atlas.
Results 53 patients were included. 25 were females (47,2%), mean age was 62.1 years [Standard deviation (SD) 13.2 years]. 36 had inflammatory arthritis reparsed as follows: 28 Rheumatoid arthritis (RA), 5 had Psoriatic arthritis, 3 had Undifferentiated arthritis. “Non inflammatory” arthritis group included 10 patients with Osteoarthritis and 7 with ligaments or meniscus injuries. Mean GSS was significantly higher in the IA group 5.70 [SD 0.321] vs.3.51 [SD 0.351]; p<0.001). Mean IMSYC was significantly superior in the IA group 14.94 [SD 0.747] vs. 8.50 [SD 0.639]; p<0.001). In univariate analysis by logistic regression, GSS (Odd Ratio (OR) 2.27; p<0.001), CD3 (OR 4.3; p=0.002), CD68 (OR 4.5; p=0.002), Ki67 (OR 11.8; p<0.001), and CD31 scores (OR 6.5; p=0.001) were significantly associated with IA, however CD20 score was not (OR 0.9; p=0.34). ROC curve analysis determined the score of 10.5 out of 24 as the best cut off for discrimination between IA and non-IA with a sensitivity of 74.3% and specificity of 100%. The area under ROC curves was statistically superior with IMSYC (0.93) compared to GSS (0.81) (p=0.05).
Conclusions We hereby propose a new synovitis score including IHC. This score has a better sensitivity and specificity than the Global synovitis score for discrimination between IA and non-IA. Moreover, this score accurately describes synovial membrane immunophenotype and could therefore give a basis for tissue driven therapies in rheumatic diseases, especially in RA.
Krenn V et al. Synovitis score: discrimination between chronic low-grade and high-grade synovitis.Histopathology. 2006 Oct;49(4):358–64.
Disclosure of Interest None declared