Inflammatory and structural evaluation in spondyloarthritis: magnetic resonance imaging analysis of axial and...

To determine the magnetic resonance imaging (MRI) criteria of most value in the assessment of patients with spondyloarthropathy (SpA) with axial or peripheral involvement.

Fat suppressed (FS)-T2 and pre- and postinjection FS-T1 images were obtained in the most symptomatic region (axial or peripheral) of patients requiring tumor necrosis factor-a blockers. Thirty-eight MRI (21 axial and 17 peripheral) were blindly scored at synovial (S) and entheseal (E) sites by 2 experienced observers screening for 7 inflammatory and 7 structural predefined criteria, which were evaluated for frequency (N) and intra- and interobserver reproducibility.

In peripheral regions, synovitis (S; N = 69.4%), ligament inflammation (E; N = 39.7%), bone marrow edema (S; N = 22.1%; E; N = 15%), and tenosynovitis (S; N = 21%) were recorded with good to excellent intraobserver reproducibility [intraclass correlation coefficient (ICC) 0.49-0.93] and moderate to good interobserver reproducibility (ICC 0.49-0.66). With regard to structural criteria, erosions (S; N = 17.1%) and enthesophytes (E; N = 13.9%) exhibited good to excellent intraobserver (ICC 0.71-0.85) and moderate interobserver reproducibility (ICC 0.54-0.49); the reproducibility of fat inflation (N = 1.4%) was good (ICC 0.76-0.78). In axial regions, no inflammatory criteria achieved good interobserver reproducibility. However, fat inflation (S; N = 86%), chondral lesions (S; N = 85.8%), enthesophytes (E; N = 76.7%), fusion (S; N = 41.2%), and erosions (S; N = 25.1%) showed excellent intraobserver reproducibility (ICC 0.81-0.98), and moderate to excellent interobserver reproducibility (ICC 0.50-0.96).

In terms of intra- and interobserver reproducibility, MRI is a reliable tool with which to assess synovitis, bone edema, ligament inflammation, tenosynovitis, erosion, enthesophytes, and fat inflation in patients with peripheral involvement. In those with axial involvement, inflammatory criteria lack interobserver reproducibility, but chondral lesions, erosion, fat inflation, fusion, and enthesophytes are relevant.