Reproducibility and influence of hand rotation on computer-aided joint space analysis

Abstract Objectives Computer-aided joint space analysis (CAJSA) is a recently developed, semi-automated tool to aid in the measurement of joint space margins based on hand radiographs. The objective of this study is to verify the potential effect of hand rotation during X-ray imaging on the measurement of joint space width using CAJSA and to evaluate the reproducibility of the CAJSA technique in healthy subjects and in patients with rheumatoid arthritis. Methods All joint space distance (JSD) measurements were performed using CAJSA-technology at the metacarpophalangeal articulation based on conventional and digital hand radiographs. Results I. Hand rotation showed an effect on the reproducibility with CV ranging between 0.39% (angulation 1°) and 1.66% (angulation 19°). II. Regarding the overall reproducibility of the CAJSA method, the intra-radiograph reproducibility of JSD was calculated with CV = 0.54% for conventional images and CV = 0.38% for digital images. The inter-radiograph reproducibility error was observed with CV = 0.66% (conventional images) and CV = 0.63% (digital images). III. The study revealed a reproducibility for CAJSA measurements in RA ranging between 0.37% (JSD-MCP ring finger; van der Heijde-modification of the Sharp method Score 1) to 1.37% (JSD-MCP index finger; van der Heijde-modification of the Sharp method Score 3). Conclusion CAJSA measures JSD at the metacarpophalangeal articulation with high reproducibility in healthy subjects and in patients with differing stages of rheumatoid arthritis. Additional findings show that hand rotation during X-ray imaging has an impact on the CAJSA measurements; thus avoiding acquisition of hand radiographs with a rotation error of more than 15 degrees can be recommended.