Background

To assess the diagnostic accuracy of double inversion recovery (DIR) magnetic resonance imaging (MRI) sequences for synovitis of the wrist joints in patients with rheumatoid arthritis (RA).

Material and method

Participants with newly diagnosed RA were enrolled between November 2019 and November 2020. MRI examinations of the wrist joints were performed using a contrast-enhanced T1-weighted imaging sequence (CE-T1WI) and DIR sequence. We measured synovitis score, number of synovial areas, synovial volume, mean synovium-to-bone signal ratio (SBR), and synovial contrast-to-noise ratio (SNR). The inter-reviewer agreement rated on a four-point scale was evaluated by calculating the weighted k statistics. Two MRI sequences were assessed using Bland–Altman analyses, and the diagnostic performance of DIR images was calculated using the chi-square test.

Results

A total of 47 participants were evaluated, and 282 joint regions in 5076 images were reviewed by two readers. There was no significant difference in synovitis scores (P = 0.67), number of synovial areas (P = 0.89), and synovial volume (P = 0.086) between the two MRI sequences. DIR images showed better SBR and SNR (all P < 0.01). There was good agreement between the two reviewers in terms of synovitis distribution (κ = 0.79). The synovitis was well agreed upon by the two readers according to Bland–Altman analyses. Using CE-T1WI as the reference standard, DIR imaging demonstrated a sensitivity of 94.1% and a specificity of 84.6% at the patient level.

Conclusion

The non-contrast DIR sequence showed good consistency with CE-T1WI and potential for evaluating synovitis in patients with RA.

Keywords: Magnetic resonance, DIR, rheumatoid arthritis