Table 1

Studies investigating prediction of progression of joint erosions in patients with rheumatoid arthritis (RA) using baseline imaging parameters

Association with erosion progression (radiographic or MRI)
ReferenceYearStudy typeDescriptionMRI bone oedemaMRI synovitisPDUS synovitis
Døhn 12010152 patients with biologic-naive RA, disease duration 7 years, followed for 12 months on anti-TNF therapy (adalimumab/methotrexate)Baseline: RR 3.8, p<0.001Baseline: RR 0.68, p=0.79Baseline: RR 7.5, p=0.06
Ever: RR 14.8, p<0.0001Ever: RR 0.24, p=0.30Ever: RR 16.9, p=0.052
AUC score p<0.001AUC score p=0.063AUC score p=0.002
McQueen 62003142 patients with early RA enrolled, disease duration ≤6 months. Full imaging data available for 31. Followed for 6 years on non-biologic DMARDsBaseline bone oedema score was only MRI feature on multivariate analysis to predict 6-year Sharp score: R2=0.20, p =0.01. At each bone OR 6.5 (95% CI 2.78 to 18.1) for MRI erosionBaseline score not predictive of 6-year Sharp score: R2=0.05, p=0.2. At each bone no association with later erosion, p=0.5Not included
Hetland 720092130 patients with early RA, disease duration 3.3 years. Combination non-biologic DMARDs including ciclosporin or placebo. Followed for 2 yearsBaseline bone oedema score was the only independent predictor of 2-year change in Sharp score (multivariate linear regression) coefficient=0.75 (95% CI 0.55 to 0.94), p=0.001Baseline synovitis score did not predict change in Sharp score. Coefficient = 0.20 (95% CI −0.09 to 0.48), p=0.17. No AUC analysisNot included
Mundwiler 82009150 patients with RA recruited; 46 had suitable data, disease duration <5 years. MRI and XR of MTP joints (3–5 bilaterally). Traditional and biologic DMARDs assessed at 12 and 24 monthsBaseline bone oedema predicted MRI erosion: OR (6 months = 34.17; 12 months = 68.0). PPV 0.50, NPV 0.99Synovitis resolved in two-thirds of MTPs when present in isolation. No association with later MRI erosion reportedNot included
Naredo 420081367 patients with RA, complete imaging data in 278. Disease duration 9.6 years. PDUS of 28 joints (shoulders, elbows, wrists, hands, knees). Followed for 12 monthsNot includedNot includedTime-integrated values for PDUS signal and RF predicted XR erosion progression, R = 0.64
Brown 1020081102 patients with RA in clinical remission treated with DMARDs, complete imaging data in 90. Disease duration 7 years. PDUS and MRI of dominant wrist and MCP jointsPrediction of structural deterioration in the MCP joints (OR 2.26, 95% CI 0.98 to 5.22, p=0.057)Prediction of structural deterioration in the MCP joints (OR 2.98, 95% CI 1.49 to 5.97, p=0.002)Prediction of structural deterioration in the MCP joints (OR 12.21, 95% CI 3.34 to 44.73, p<0.001)
Palosaari 92006127 patients with early RA, disease duration ≤12 months, followed up for 1 year and 24 for 2 years with contrast-enhanced MRIBone oedema score only baseline variable to predict erosive progression at 2 years on multivariate regression (OR 4.2, 95% CI 1.3 to 13.8). At each bone, predicted erosion at 1 and 2 years: OR 28 (95% CI 11.7 to 67.1) and 14.9 (95% CI 6.3 to 34.9)Synovitis score (baseline) only predictive of erosion at 2 years on univariate analysis; Spearman correlation coefficient=0.57, p=0.004Not included
  • Study type: 1, observational, longitudinal; 2, randomised clinical trial.

  • AUC, area under the curve; DMARDs, disease-modifying antirheumatic drugs; MCP, metacarpophalangeal joint; MTP, metatarsophalangeal joint; NPV, negative predictive value; PDUS, power Doppler ultrasound; PPV, positive predictive value; RF, rheumatoid factor; RR, relative risk; XR, plain radiography.