Objectives: To determine the intra- and inter-observer agreement of erosions detected and scored with cone-beam computed tomography (CBCT) of bones in the hands and feet, and to compare CBCT with conventional radiography (CR) for assessment of bone erosions in patients with long-standing rheumatoid arthritis (RA).
Method: Thirty patients with long-standing RA from the Better Anti-Rheumatic PharmacOTherapy (BARFOT) cohort were examined with CBCT and CR of hands and feet at their 15 year follow-up. Intra- and inter-class correlation coefficients (ICCs) were calculated. Erosions were analysed with the total rheumatoid arthritis magnetic resonance imaging erosion score (RAMRIS erosion score) for ICCs with CBCT, and with the modified RAMRIS erosion score (RAMRIS-mod.) for the same locations as used in the Sharp van der Heijde score and Sharp van der Heijde erosion score for CR.
Results: All 30 patients showed erosions on CBCT and 26 on CR. The ICCs for both intra- and inter-observer reliability were 0.92-0.99. CBCT showed numerically more erosions than CR for all regions compared, although a statistically significant difference was found only for the metacarpophalangeal joints [median number of eroded joints 1.0 (range 0-14) with CBCT and 0.5 (0-13) with CR, p = 0.044].
Conclusion: CBCT has high reproducibility and is more sensitive than CR in detecting erosions in this cohort of patients with long-standing RA. CBCT has the potential to become an important tool in the detection and follow-up of erosions in patients with RA.