Assessment of anal fistulas with high-resolution subtraction MR-fistulography: comparison with surgical findings

PURPOSE: To investigate a new MR-imaging protocol defined as subtraction MR-fistulography for the detection of fistula-in-ano. MATERIALS AND METHODS: A total of 36 patients (18 men, 18 women) with the clinical diagnosis of anal fistula or abscess who were scheduled for surgical exploration were preoperatively imaged at 1.5 T in the supine position; coronal STIR sequences and axial T1-weighted three-dimensional FLASH sequences before and after intravenous injection of gadobenate dimeglumine (Gd-BOPTA) were obtained. Image subtraction was routinely used. The image findings were compared with the findings at surgical exploration according to a uniform classification system. RESULTS: A total of 46 fistulas were seen on subtraction MR-fistulography, whereas surgery described 41 fistulas. MRI and surgery concordantly depicted a total number of 40 fistulas, six fistulas were only described by MRI, and one fistula was only seen during surgery. A total of 27 abscesses were detected with MRI, compared to 22 during surgery. A total of 22 abscesses were concordantly diagnosed. Complete agreement between subtraction MR-fistulography and surgery occurred in 32 (89%) of the 36 patients. The four patients in whom both procedures disagreed had complex anal sepsis with multiple fistulas and/or abscesses due to Crohn disease. CONCLUSION: High-resolution subtraction MR-fistulography is an important complement to surgical exploration and especially suitable for investigating complex anal sepsis.