Background/Purpose: There is limited data outlining the anatomic characteristics of bone erosions of the small joints in rheumatoid arthritis (RA). In particular, there is minimal utilization of 3-dimensional imaging modalities, such as magnetic resonance imaging (MRI) for the purpose of describing erosion patterns in the metacarpophalangeal (MCP) joints. This is of particular interest for understanding the role of biomechanical forces exerted by the adjacent anatomy in contributing to erosion initiation and progression. Our objective was to describe the anatomical features of erosions in the MCP joints to explore the importance of biomechanical stress in erosion pathogenesis.

Methods: MR images of the 2^nd through 5^th MCP joints were acquired bilaterally from 22 patients who met early referral criterial for RA. Maximum depth and width of erosions were determined using the semi-automated segmentation software Early Erosions in Rheumatoid Arthritis (EERA). Anatomical features were recorded, including: affected hand and MCP joint, metacarpal head vs. phalangeal base involvement, radial- vs. ulnar-sided cortical breakage, and palmar vs. dorsal cortical breakage. One-way ANOVA was used to compare means across groups. Bivariate correlation between maximum erosion depth and width was calculated as an estimate of erosion geometric sphericity.

Results: Patients were 91% female with mean (SD) age of 55.9 (9.3) years, symptom duration of 4.8 (5.3) years, and disease activity score of 4.8 (1.5). Thirty right-handed and 19 left-handed erosions were identified; all involved the metacarpal head, with none involving the phalangeal base. Erosion frequencies across MCP joints 2 through 5 were: 15, 22, 7 and 5 erosions, respectively. Radial-sided cortical breakage was observed in 42 (86%) erosions, and ulnar-sided breakage was observed in 4 (8%) erosions, 3 of which were on the 5^th metacarpal. Palmar cortical breakage was observed in 3 (6%) erosions, and dorsal breakage was observed in 1 (2%) erosion. Mean (SD) maximum erosion depth and width were 4.3 (1.5) mm and 4.6 (2.1) mm, respectively. Maximum depth and width correlated moderately, Pearson’s r = 0.51 (p < 0.05), and with the exclusion of one large erosion outlier, only correlated weakly, r = 0.36 (p < 0.05). Neither erosion depth nor width were significantly associated with MCP joint number, or location of cortical break (p > 0.05 for all comparisons).

Conclusion: Predominance of erosions on the right (mostly dominant) hand, 2^nd and 3^rd MCP joints (higher stresses compared to 4^th and 5^th MCP), and the radial side of metacarpal heads (higher stabilizing stress from the adjacent radial collateral ligament) except on the 5^th metacarpal, all support the hypothesis that anatomical areas subjected to greater biomechanical stress are more susceptible to erosive damage. The weak to moderate correlation between maximum depth and width reflects the geometric asymmetries and irregularities of erosions of the MCP joints. Application of MRI to characterizing other small joints in RA may provide insight to the unique ways in which biomechanical stresses influence bone erosion pathogenesis and progression.