Using a reference when defining an abnormal MRI reduces false-positive MRI results-a longitudinal study in two cohorts at risk for rheumatoid arthritis.

Objectives: The use of hand and foot MRI in the diagnostic process of RA has been advocated. Recent studies showed that MRI is helpful in predicting progression from clinically suspect arthralgia (CSA) to clinical arthritis, and from undifferentiated arthritis (UA) to RA. Symptom-free persons can also show inflammation on MRI. This study aimed to evaluate if MRI findings in symptom-free volunteers are relevant when defining a positive MRI.
Methods: Two hundred and twenty-five CSA patients and two hundred and one UA patients underwent MRI of MCP, wrist and MTP joints at baseline and were followed for 1 year on progression to arthritis and RA, respectively, as reported previously. MRI was considered positive if ⩾ 1 joint showed inflammation (called uncorrected definition), or if ⩾ 1 joint had inflammation that was present in < 5% of persons of the same age category at the same location (called 5% corrected definition). Test characteristics were compared for both definitions.
Results: By using MRI data of symptom-free volunteers as reference, specificity of MRI-detected inflammation increased from 22 to 56% in CSA patients, and from 10 to 36% in UA patients. The sensitivity was not affected; it was 88 and 85% in CSA patients and 93 and 93% in UA patients. The accuracy also increased, from 32 to 60% in CSA patients and 22 to 44% in UA patients.
Conclusion: The use of a reference population resulted in a substantial reduction of false-positive results, without influencing the sensitivity. Although common for other tests in medicine, this phenomenon is novel for MRI in the early detection of RA.
(© The Author 2017. Published by Oxford University Press on behalf of the British Society for Rheumatology. All rights reserved. For Permissions, please email: journals.permissions@oup.com)