To extend the magnetic resonance imaging (MRI) score for assessment of wrist synovitis in juvenile idiopathic arthritis (JIA) by inclusion of the metacarpophalangeal (MCP) joints, and to compare the metric properties of the original and the extended score.
Wrist MRI of 70 patients with JIA were scored by 3 independent readers according to (1) the wrist component of the rheumatoid arthritis MRI synovitis score (comprising distal radioulnar, radiocarpal, and combined midcarpal and carpometacarpal joints); and (2) an extended score including the MCP joints. Thirty-eight patients had a 1-year MRI followup. The concordance between the readers [intraclass correlation coefficient (ICC), 95% limits of agreement (LOA), and weighted Cohen's κ], correlations with clinical variables (Spearman's ρ), and the sensitivity to change [standardized response mean (SRM)] were calculated for both scores.
The interreader agreement was moderate for the original score (ICC 0.77; 95% CI 0.68-0.84) and good for the extended score (ICC 0.86; 95% CI 0.80-0.91). Using 95% LOA, the aggregate score variability was less favorable with relatively wide LOA. Weighted Cohen's κ of the individual joints indicated good agreement for the original score and good to excellent agreement for the extended score. Correlations with clinical variables reflecting disease activity improved for the extended score and its SRM was higher compared to that of the original score.
The extended score showed better reliability, construct validity, and sensitivity to change than the original. Inclusion of the MCP joints should be considered for a more accurate assessment of disease activity and treatment efficacy in JIA.