Introduction: Despite the widespread use of magnetic resonance imaging (MRI) and Doppler ultrasound for the detection of rheumatoid arthritis (RA) disease activity, little is known regarding the association of imaging-detected activity and synovial pathology. The purpose of this study was to compare site-specific release of inflammatory mediators and evaluate the corresponding anatomical sites by examining colour Doppler ultrasound (CDUS) and MRI scans.
Methods: RA patients were evaluated on the basis of CDUS and 3-T MRI scans and subsequently underwent synovectomy using a needle arthroscopic procedure of the hand joints. The synovial tissue specimens were incubated for 72 hours, and spontaneous release of monocyte chemoattractant protein 1 (MCP-1), interleukin 6 (IL-6), macrophage inflammatory protein 1β (MIP-1β) and IL-8 was measured by performing multiplex immunoassays. Bone marrow oedema (BME), synovitis and erosion scores were estimated on the basis of the rheumatoid arthritis magnetic resonance imaging score (RAMRIS). Mixed models were used for the statistical analyses. Parsimony was achieved by omitting covariates with P > 0.1 from the statistical model.
Results: Tissue samples from 58 synovial sites were obtained from 25 patients. MCP-1 was associated with CDUS activity (P = 0.009, approximate Spearman's ρ = 0.41), RAMRIS BME score (P = 0.01, approximate Spearman's ρ = 0.42) and RAMRIS erosion score (P = 0.03, approximate Spearman's ρ = 0.31). IL-6 was associated with RAMRIS synovitis score (P = 0.04, approximate Spearman's ρ = 0.50), BME score (P = 0.04, approximate Spearman's ρ = 0.31) and RAMRIS erosion score (P = 0.03, approximate Spearman's ρ = 0.35). MIP-1β was associated with CDUS activity (P = 0.02, approximate Spearman's ρ = 0.38) and RAMRIS synovitis scores (P = 0.02, approximate Spearman's ρ = 0.63). IL-8 associations with imaging outcome measures did not reach statistical significance.
Conclusions: The association between imaging activity and synovial inflammatory mediators underscores the high sensitivity of CDUS and MRI in the evaluation of RA disease activity. The associations found in our present study have different implications for synovial mediator releases and corresponding imaging signs. For example, MCP-1 and IL-6 were associated with both general inflammation and bone destruction, in contrast to MIP-1β, which was involved solely in general synovitis. The lack of association of IL-8 with synovitis was likely underestimated because of a large proportion of samples above assay detection limits among the patients with the highest synovitis scores.