Background/Purpose: This cross-sectional study aimed to assess the agreement between Magnetic Resonance Imaging (MRI) and ultrasound (US) findings of bone erosions in the metatarsalphalangeal (MTP) joints of patients with early rheumatoid arthritis (RA).

Methods: Men and women with early RA (DMARD naïve) according to ACR criteria were recruited. Each patient underwent an MRI scan (1.0 Tesla peripheral MRI) and an US scan of MTPs 2-5 of the most symptomatic foot. MRI scans were scored by a musculoskeletal trained radiologist using OMERACT RAMRIS criteria for erosions (0-10). Using OMERACT criteria, US images were assessed for erosions (present/absent) on each joint by an experienced rheumatologist. To assess agreement using kappa statistics, MRI erosions which scored 0 were considered to have no erosions while those ≥1 were considered to have an erosion.

Results:  The study included 39 patients (n=33 women), mean (standard deviation(sd)) age = 51.6 (10.3) yrs. MRI detected an erosion with a score ≥1 on 30 MTP2 joints, 28 MTP3 joints, 25 MTP4 joints and 15 MTP5 joints. US detected an erosion on 2 MTP2 joints, 1 MTP3 joints, 0 MTP4 joints and 6 MTP5 joints. The proportion of erosions detected by MRI that were also detected by US in MTPs 2, 3, 4, and 5 were 6.5%, 3.5%, 0% and 37.5%, respectively. In MTPs 2, 3 and 4, there were no erosions detected by US that were not also detected by MRI. However, in MTP5, there were 3 erosions detected by US that were not detected by MRI. Kappa statistics of agreement are presented in Table 1. Table 1: Agreement in Erosions detected by MRI and Ultrasound

*denotes kappa is significant (p<0.05)

Conclusion:  There was a high prevalence of bone erosions in MTPs 2-5 detected by MRI in this cohort of early RA patients. The proportion of erosions detected on MRI that were detected by US was generally very low. MRI appeared to show a greater sensitivity, detecting many more erosions than MRI which may help to explain the lack of agreement between modalities. Lack of agreement may also be explained by the physical size of the lesions that can be detected by MRI versus US due to resolution or because MRI shows more signal variation by tissue composition than US. The higher agreement for the 5^th MTP may be because the US probe can almost encircle the bone allowing better visualization of the bone. In general, however, using MRI as the gold standard, the agreement between these two imaging modalities is poor, suggesting that US detected erosions likely underestimate the true prevalence of erosions in MTPs 2-5.

« Back to 2016 ACR/ARHP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/assessing-bone-erosions-in-the-metatarsophalangeal-joints-of-patients-with-early-rheumatoid-arthritis-a-comparison-between-mri-and-ultrasound/