Correlation between manual 0.2T MRI assessment of synovitis and EULAR-OMERACT scores of the wrist in patients with rheumatoid arthritis

• Published in: Journal of magnetic resonance imaging 2014-05, Vol.39 (5), p.1171-1177
• Main Authors: Wojciechowski, Wadim, Czaplicka, Kamila, Włodarczyk, Justyna, Tabor, Zbisław
• Format: Article
• Language: English
• Subjects: Adult; Aged; Algorithms; Arthritis, Rheumatoid - complications; Arthritis, Rheumatoid - pathology; Female; Humans; Image Interpretation, Computer-Assisted - methods; Magnetic resonance imaging; Magnetic Resonance Imaging - methods; Male; manual segmentation; Middle Aged; Observer Variation; RAMRIS; Reproducibility of Results; rheumatoid arthritis; Sensitivity and Specificity; Severity of Illness Index; synovitis; Synovitis - complications; Synovitis - pathology; Wrist Joint - pathology; wrist MRI
• ISSN: 1053-1807; 1522-2586
• DOI: 10.1002/jmri.24267

Purpose To test the correlation between manually assessed volumes of magnetic resonance (MR) signal‐enhancing regions in postcontrast images and MRI‐based rheumatoid arthritis scoring system (RAMRIS) synovitis scores in 0.2T MR data. Materials and Methods The low‐field (0.2T) extremity MRI scans were acquired in 32 patients (37 study cases) as a part of the rheumatoid arthritis diagnostic procedure in accordance with Outcome Measures in Rheumatology Clinical Trials MRI working group (OMERACT) recommendations. The synovitis in joint regions was assessed based on RAMRIS. Regions of enhanced signals in postcontrast MR images were outlined manually. Results Spearman's rank‐order correlations between RAMRIS scores and both absolute and relative volumes of signal‐enhancing regions were equal to 0.68, 0.78 (moderately positive) and 0.84 (strongly positive) for distal radioulnar joint (DRUJ), intercarpal‐carpometacarpal joints (ICCMJ), and radiocarpal joint (RCJ), respectively. Absolute and relative volumes of signal‐enhancing regions grow significantly with RAMRIS scores. Conclusion Development of algorithms for automated detection of signal‐enhancing regions in 0.2T MR images is a reasonable direction of forthcoming studies on computer‐aided diagnosis (CAD) systems supporting low‐field extremity MR‐based diagnosis of rheumatoid arthritis by identifying its symptoms. J. Magn. Reson. Imaging 2014;39:1171–1177. © 2013 Wiley Periodicals, Inc.