Perianal infections: Reader agreement and diagnostic performance of fat‐suppressed contrast‐enhanced MRI

Abstract

We aimed to determine the diagnostic performance of the contrast-enhanced fat-suppression technique in the detectability of perianal infections and to compare this technique with different MR sequences used for this purpose. Thirty consecutive patients with clinically suspected anorectal infections were examined with fast spin-echo (FSE) T2-weighted, short-inversion-time inversion recovery (STIR), and fat-suppressed and non-fat-suppressed T1-weighted spoiled gradient-echo (SGE) (in-phase) dynamic contrast-enhanced sequences. The results of MRI were correlated with the findings of surgery, which was considered as the standard of reference. Receiver-operating-characteristic curves were reconstructed to describe and compare the diagnostic value of each MR technique. The values of kappa were used as a measure of observer reliability. Diagnostic performances of STIR, FSE T2-weighted and fat-suppressed and non-fat-suppressed, contrast-enhanced SGE T1-weighted techniques showed statistically insignificant differences in detection of perianal infections. There was almost perfect interobserver agreement regarding the presence of lesions on each MR technique. Fat-suppressed, contrast-enhanced SGE T1-weighted MRI showed adequate diagnostic performance in the detection of perianal infections. However, its significance is not different from the other MR sequences used for the present study.