In a preliminary MRI study of healthy volunteers's hands investigators showed that the suppression of the fat signal obtained by the Dixon sequences is greater than that obtained by the "classic" CHESS sequences with a better T1-weighted image quality. The investigators goal now is to compare these sequences in the quantification of the inflammatory activity of rheumatoid arthritis by the RAMRIS score. To investigators knowledge, only one study compared those two types of sequence with many limitations including a limited number of patients included.

Rheumatoid arthritis (RA) is the most common chronic inflammatory arthropathy and it mostly affect young women. The MRI sequences currently validated for the joint inflammation assessment are the fat-suppressed sequences produced either by spectral presaturation (CHESS sequences) or by inversion-recovery (STIR = short tau inversion recovery sequence). The presaturation sequences outweigh the STIR sequence due to higher spatial resolution at the expense of sometimes more random saturation of the fat signal. Recently, a third method of suppressing the fat signal (Dixon method) has reappeared thanks to advances in computing power (post-processing). In a preliminary MRI study of the hands of healthy volunteers, the investigators showed that the suppression of the fat signal obtained by the Dixon sequences is greater than that obtained by the "classic" CHESS sequences in T1 and T2 weighting with better quality. image in T1 weighting. This increased performance, particularly in terms of fat suppression, could lead to better detection of PR lesions on MRI.

Inclusion Criteria:

• - Rheumatoid arthritis diagnosis for less than 3 years

Exclusion Criteria:

• contraindication to Dotarem contrast injection
• contraindication to MRI due to incompatible implanted material (peacemaker, prosthesis,...)

None at the time of intervention. At the end of the study, Blind reading without knowledge of the type: Dixon or "classic" sequences.