Background MRI is higher sensitive for detecting inflammation and damage than clinical examination and X-ray. Since MRI is widely used in clinical trials and clinical practice in rheumatoid arthritis (RA), the EULAR-OMERACT Rheumatoid Arthritis MRI Scoring (RAMRIS) system is recommended for assessing synovitis, bone marrow edema (osteitis) and bone erosion on MRI. However, little is known about which components of histological synovitis did RAMRIS correlate with.
Objectives To investigate the correlation between RAMRIS and histological synovitis change in RA patients.
Methods Consecutive hospitalized patients with active RA were enrolled from October 2013 to January 2015. Bilateral 3.0T MRI of wrist were performed and RAMRIS-synovitis (0–18), osteitis (0–90) and bone erosion (0–300) were scored according to the RAMRIS reference image atlas by a radiologist blinded to clinical and histological data. Synovial tissue samples were obtained by closed-needle biopsy from inflamed knee before MRI examination. Total histological synovitis score and its subscores (lining hyperplasia, inflammatory infiltration and stroma activation) and fiber ratio were scored under H&E staining. Densities of cells with positive staining for CD68 (macrophages), CD3 (T cells), CD20 (B cells) or CD38 (plasma cells) were counted and microvascular count (identified by CD34+ endothelial cells) were determined.
Results (1) Among 34 RA patients enrolled, 79% were female, age (median and IQR, similarly hereinafter) was 54 (46–65) years, disease duration was 12 (10–99) months and DAS28-CRP was 4.8 (3.9–5.7).
(2) RAMRIS-synovitis score was 10 (5–14) and correlated positively with disease duration or 28TJC (r=0.435–0.469, both p<0.05). RAMRIS-osteitis score was 48 (15–118) and correlated positively with disease duration, 28TJC or anti-CCP (r=0.361–0.557, all p<0.05). RAMRIS-bone erosion score was 23 (6–54) and correlated positively with disease duration, 28TJC, DAS28-CRP, CDAI or SDAI (r=0.355–0.653, all p<0.05). RAMRIS-osteitis score was significantly higher in patients with positive anti-CCP (n=29) than in those with negative anti-CCP [n=5, 29 (17–59) vs 6 (2–6), P=0.003] and also significantly higher in patients with positive RF (n=27) than in those with negative RF [n=7, 29 (15–56) vs 4 (2–19), P=0.016].
(3) Synovial tissue with synovial lining and vascularized subintima were evaluated. RAMRIS-synovitis correlated positively with CD34+ microvascular count (r=0.648, p=0.023) or fiber ratio (r=0.468, p=0.005). RAMRIS-osteitis score correlated positively with synovial fiber ratio (r=0.542, p=0.001). RAMRIS-bone erosion score correlated positively with synovial stroma activation (r=0.352, p=0.041) or fiber ratio (r=0.668, p<0.001). Patients with high-grade synovial stroma activation (n=19) had higher RAMRIS-bone erosion score than those with low-grade synovial stroma activation [n=15, 76 (33–140) vs 23 (11–60), p=0.02]. Patients with synovial fiber ratio≥25% (n=16) had higher RAMRIS-synovitis, osteitis or bone erosion score than those with synovial fiber ratio <25% (n=18, all p<0.05).
Conclusions Our preliminary results suggest that EULAR-OMERACT RAMRIS-synovitis, osteitis and bone erosion score correlate with synovial stroma activation or fiber deposit, which may play vital role in RA synovitis.
Disclosure of Interest None declared