Diagnostic utility of candidate definitions for demonstrating axial spondyloarthritis on magnetic resonance imaging of the spine.

Abstract

A recent consensus statement has suggested ≥3 corner inflammatory lesions (CILs) or several corner fatty lesions (CFLs) as candidate criteria indicative of axial spondyloarthritis (SpA) on magnetic resonance imaging (MRI) of the spine. The aim of this study was to evaluate the diagnostic utility of these cutoffs in nonradiographic axial SpA and ankylosing spondylitis (AS). One hundred thirty consecutive patients with back pain who were ≤50 years of age and newly referred to 2 university clinics (cohorts A and B) were classified according to rheumatologist expert opinion based on results of clinical examination and pelvic radiography as having nonradiographic axial SpA (n = 50), AS (n = 33), or nonspecific back pain (n = 47). Cohort A also included 20 age-matched healthy controls. Four blinded readers assessed MRIs of the spine using the standardized Canada-Denmark module. Readers recorded CILs and CFLs in 23 discovertebral units. We tested the diagnostic utility (mean sensitivity and specificity over 4 readers) of the cutoff for the number of lesions on spinal MRI as proposed in the literature (≥2 or ≥3 CILs and ≥6 CFLs), and we tested for possible thresholds (from ≥1 CIL or CFL to ≥10 CILs or CFLs) for nonradiographic axial SpA and AS patients in both cohorts. None of the spinal thresholds (≥2 or ≥3 CILs and ≥6 CFLs) showed clinically relevant diagnostic utility (positive likelihood ratio [LR] range 1.38-2.36) when comparing patients with nonradiographic axial SpA to patients with nonspecific back pain. A threshold of ≥6 CILs had moderate to substantial diagnostic utility (positive LR 13.26 and 6.74 in cohorts A and B, respectively) in nonradiographic axial SpA, while ≥4 CILs showed small diagnostic utility (positive LR 3.83 and 2.72 in cohorts A and B, respectively) but specificities of >0.90. None of the previously proposed candidate criteria for a positive spinal MRI finding of axial SpA showed clinically relevant diagnostic utility in nonradiographic axial SpA. These results question the value of proposed definitions for a positive finding of SpA based on MRI of the spine alone.