Objective To confirm feasibility and assess intravoxel incoherent motion (IVIM) to differentiate active sacroiliitis and ankylosing spondylitis.. Methods Forty-one patients were divided into two groups, an active group (n=20) and a chronic group (n=21), according to the Bath Ankylosing Spondylitis (AS) Disease Activity Index (BASDAI) and laboratory parameters. In addition, 21 healthy volunteers were chosen as the control group. Tissue diffusivity (D slow ), perfusion fraction (f), and pseudo-diffusion coefficient (D fast ) values were obtained for all three groups. Oneway analysis of variance and receiver operating characteristic analysis were performed for all parameters. Results There was good interobserver agreement on the measurements between the two observers. The optimal cutoff values (with respective AUC, sensitivity, specificity, positive likelihood ratio, and negative likelihood ratio) between active and chronic groups were D slow =0.53×10 −3 mm 2 /s (0.976, 90 %, 95.2 %, 18.9, 0.10) and f= 0. 09 (0.545, 20 %, 95.5 %, 4.2, 0.84), and between chronic and control groups were D slow = 0.22 × 10 −3 mm 2 /s (0.517, 9.52 %, 100 %, no number, 0.9) and f = 0.09 (0.935, 95.24 %, 80.95 %, 5, 0.059).Conclusion D slow and f of IVIM diffusion-weighted (DW)-MRI in AS show a significant difference in the values of diffusion of water molecules and fractional perfusion-related volume among the three groups.
Key Points• D slow can be used to differentiate the activity of AS.• With perfusion fraction, the sensitivity of differentiating the AS activity is improved. • IVIM DWI plays an important role in detecting the activity in patients with AS.