Magnetic resonance imaging (MRI) clearly is more sensitive than clinical examination and conventional radiography (x-ray) for detection of inflammation (synovitis, bone marrow oedema (osteitis) and tenosynovitis) and damage (bone erosion and cartilage loss/joint space narrowing) in patients with rheumatoid arthritis (RA). The question is when and how MRI should be used. The present article reviews our knowledge about, and provides suggestions for, the use of MRI in clinical trials, clinical care and clinical registries. In clinical trials, the OMERACT RA MRI scoring system (RAMRIS) is a thoroughly validated method which in less time and with fewer patients than x-ray can discriminate between different therapies regarding structural damage progression, and which on top of this offers detailed assessment of upstream inflammatory drivers of damage. In routine clinical care, MRI can contribute to an earlier diagnosis of RA, can reveal subclinical disease activity, e.g. in the synovium (synovitis) and bone (osteitis), and can provide information of strong prognostic significance for the subsequent disease course, which may be useful when deciding the treatment strategy. Future studies will clarify the benefits of including MRI in treat-to-target strategies. The benefits of incorporating MRI into clinical registries are not yet known, but may include improved knowledge about the real-life advantages of MRI, as well as opportunities to develop better clinical and laboratory composite measures to monitor and predict the disease course in RA. In conclusion, MRI has well-documented relevance in several settings in clinical trials and care, but not yet in clinical registries.