Correlation Between Dynamic Contract-Enhanced Magnetic Resonance Imaging Scores Of Inflammation And Ramris Scores Of Synovitis And Bone Marrow Oedema In The Wrist Of Patients With Rheumatoid Arthritis

O. Kubassova1 2, M. Boesen3, R. Bouert4, M.B. Axelsen5, M. Ostergaard5, M.A. Cimmino6, B. Danneskiold Samsoe1, H. Bliddal1 1Department of Rheumatology, The Parker Institute, Copenhagen, Denmark; 2Image Analysis Ltd, Leeds, United Kingdom; 3Dept of Radiology, Frederiksberg Hospital, Copenhagen; 4The Parker Institute; 5Departments of Rheumatology, Copenhagen University Hospitals at Herlev and Hvidovre, Copenhagen, Denmark; 6Department of Rheumatology, Universita di Genova, Genoa, Italy

Objectives: To investigate the relation between semiquantiative scores bone marrow oedema and synovitis, using the RA MRI scoring system RAMRIS and computer aided measures of inflammation derived from dynamic contrast-enhanced (DCE)-MRI data using a novel software program (Dynamika-RA) in patients with rheumatoid arthritis (RA) in the wrist joints.

Methods: 45 patients with established RA had a routine MRI on a dedicated scanner (0.2T Esaote E-scan) of the most symptomatic wrist or, in clinical remission, the dominant wrist. Imaging protocol: Coronal STIR and axial coronal T1 gradient echo pre- and postcontrast images (used for RAMRIS scoring) as well as a series of coronal T1 DCE-MRI images (3 slices every 10 seconds, with 30 repetitions) started at the time of i.v. contrast injection Magnevist (0.1mmol/kg) as well as The Dynamika-RA software (www.image-analysis.org) was used to analyse the DCE-MRI data. The program performs movement correction between temporal slices, which allows for the reduction of artifactual enhancement and removes patient movement artefacts. The data was further analysed in three independent ways: 1) a fully automatic whole joint analysis, 2) a user defined placing of a rough ROI surrounding the wrist joint in the slice with most visual enhancement and 3) placing a smaller ROI in the visually most enhancing synovium. The computed output data used for comparison with RAMRIS scores were: initial rate of enhancement (IRE), maximum enhancement (ME), number of enhancing voxels (N-total) number of voxels with plateau and washout pattern of enhancement (N-plateau and N-washout).Correlation between the static RAMRIS scores and the dynamic results was performed using Spearman's rank correlation. P-values below <0.05 were considered significant.

Results: RAMRIS synovitis scores were in the range 2-9, [possible range 0-9] and bone marrow oedema scores 0-39 [0-42]. The best correlations between RAMRIS scores and the computed dynamic parameters (table 1, p<0.05) were seen with the parameters describing the volume of enhancing pixels (Ntotal, Nplateau, and Nwashout; bold in table 1), especially when a rough ROI was placed around the wrist joint excluding blood vessels. The highest correlation between dynamic parameters (IRE and ME) and scores of synovitis and oedema was seen when a small user-defined ROI was placed in the area of the most visual enhancement.

Conclusion: Computer-aided analysis of DCE-MRI data correlated with RAMRIS scores of synovitis and bone marrow oedema. This technique may be used in conjunction with RAMRIS scoring for automated synovitis and possibly bone marrow oedema and will be further validated.

Disclosure of Interest: None declared

Citation: Ann Rheum Dis 2010;69(Suppl3):454

http://www.abstracts2view.com/eular/view.php?nu=EULAR10L_FRI0430&terms=