Magnetic resonance imaging (MRI) is more sensitive than clinical examination and radiographs to detect early inflammatory changes (synovitis, tenosynovitis, osteitis) and structural damage (bone erosions) in patients with rheumatoid arthritis (RA) [1]. Outcome Measures in Rheumatology (OMERACT), a multi-institution study group, developed a scoring system to assess RA activity at MRI, the Rheumatoid Arthritis MRI score (RAMRIS) [2,3]. To improve accuracy and reproducibility, OMERACT recommends using a standardized MRI acquisition protocol including fat-saturated T2-weighted or STIR images and contrast-enhanced (CE) T1-weighted images, using either two two-dimensional (2D) sequences or less commonly one three-dimensional (3D) sequence [3].

Since gadolinium deposition in the brain has been reported after repetitive administration of gadolinium-based contrast agents (GBCA), growing safety concerns push towards the use of GBCA-free MRI acquisition protocols [4,5]. The use of contrast-free MRI protocols to assess disease activity in hands of patients with RA has already been addressed as it would yield to faster, cheaper, and less invasive examinations [6], [7], [8], [9]. Despite significant agreement between fat-suppressed T2-weighted and CE T1-weighted MRI to perform RAMRIS, the use of GBCA is still recommended by the OMERACT as its absence would lead to lower sensitivity and specificity to detect synovial inflammation in hands [6,7,10].

The Dixon technique, which provides four sets of images with different contrasts in a single MRI acquisition, has been increasingly implemented in routine musculoskeletal MRI protocols partly due to its robust fat signal suppression [11], [12], [13], [14], [15], [16]. MRI protocols only based on CE Dixon imaging have been demonstrated efficient and reproducible to score disease in hands of early RA patients [17,18].

Considering the results of previous studies on non-Dixon imaging [6], [7], [8], we hypothesized that a short MRI acquisition protocol only based on T2-weighted Dixon imaging could be as efficient as short protocols based on CE T1-weighted Dixon sequences to assess disease activity using the RAMRIS scoring system. The aim of our study was to compare 2D T2-weighted, CE 2D T1-weighted and CE 3D T1-weighted Dixon sequences to assess disease activity using the RAMRIS scoring system in hands of patients with early RA.