Fluid-attenuated inversion-recovery sequence with fat suppression as an alternative to contrast-enhanced MRI in pediatric synovitis

Background Non-contrast magnetic resonance imaging (MRI) fluid-attenuated inversion-recovery sequence (FLAIR) with fat suppression (FS) has not been validated in children. Objective Compare FLAIR to T1-weighted post contrast (T1CE) in the detection of knee synovitis. Methods and materials Institutional review board (IRB) waived consent. Children who underwent T1CE and FLAIR sequences of the knee on a 3-T magnet from April 2021 to December 2021 were included. Two pediatric radiologists assessed axial FLAIR and T1CE images for synovitis and synovial thickness. Reliability and agreement were assessed. Sensitivities, specificities, and accuracy were calculated for FLAIR using T1CE as reference standard. Results In total, 42 knees (39 patients) were assessed (median age 12.9 years (2.3–17.8 years); 62% male, 38% female). Readers judged 20/42 (48%) knees to have synovitis. Sensitivity of FLAIR for reader 1 was 79% (19/24; 95% CI 0.58, 0.93) and 84% (16/19; 95% CI 0.60, 0.97) for reader 2. Specificity of FLAIR for reader 1 was 94% (17/18; 95% CI 0.73, 1) and 83% (19/23; 95% CI 0.61, 0.95) for reader 2. Accuracy for readers 1 and 2 was 86% (36/42; 95% CI 0.71, 0.95) and 83% (35/42; 95% CI 0.69, 0.93), respectively. Inter-reader reliability was good (0.75–0.90) for synovial measurements for FLAIR (ICC = 0.80; 95% CI 0.71, 0.86) and moderate for T1 CE (ICC = 0.62 (95% CI 0.48, 0.73)). Conclusion FLAIR FS depicts synovium in the pediatric knee with similar reliability to T1 CE and may be an acceptable alternative to contrast in the initial diagnosis of synovitis. Graphical abstract