Three-Dimensional Isotropic Versus Conventional Multisequence 2-Dimensional Magnetic Resonance Imaging of Sacroiliac Joints in Suspected Axial Spondyloarthritis

Objectives The aims of the study were to systematically evaluate the sacroiliac joints of a consecutive series of patients with clinically suspected axial spondyloarthritis and to compare the diagnostic efficacy of 3-dimensional (3D) imaging versus conventional multisequence 2D techniques for sacroiliitis. Methods In this cross-sectional multireader validation study, imaging studies of 104 consecutive patients with clinically suspected axial spondyloarthritis obtained with a rheumatology lumbosacral magnetic resonance imaging protocol were evaluated. Three-dimensional coronal oblique reconstructions parallel to the long axis of the sacrum were created. Two blinded readers evaluated 2D and 3D images separately for each study, and findings were compared with the reference standard for the final diagnosis and disease activity. Cohen kappa and diagnostic performance measures on 2D and 3D imaging were evaluated and compared. Results With 2D versus 3D imaging, respectively, the sensitivity for detecting inflammatory cases was 57% versus 62% (P = 0.1007), the sensitivity for detecting mechanical cases was 89% versus 76% (P = 0.0312), the sensitivity for detecting active sacroiliitis was 49% versus 64% (P = 0.0014), and the specificity for detecting active sacroiliitis was 87% versus 76% (P = 0.0016). Interreader agreement of 2D imaging showed an overall range of kappa = 0.27-0.85, and it was best for evaluation of bony ankylosis (kappa = 0.85). Interreader agreement of 3D imaging was in an overall range of kappa = 0.15-0.64. Conclusions Single 3D isotropic magnetic resonance imaging provided adequate assessment of sacroiliac joints in this study and demonstrated significantly higher sensitivity and lower specificity for detecting active sacroiliitis. The initial results indicate that 3D imaging could be used for future works for screening sacroiliitis and abdominopelvic enthesopathy.