Non-contrast MRI and surgical concordance in fistula-in-ano

Abstract

Fistula‑in‑ano is an abnormal connection between the anal canal or rectum and the perianal skin, often resulting from infection in the anal glands. While clinical examination provides some insights, MR fistulogram is essential for detailed assessment and reducing recurrence rates after surgery.

Objective – to compare and correlate the pre‑operative non‑contrast MR fistulogram findings with surgical findings, focusing on concordance rates for fistula type, craniocaudal extent of tracts, number and clock position of internal and external openings, and presence of complicating features like secondary tracts, supralevator extension, presence and location of abscesses.

Materials and methods. We retrospectively analysed 236 patients with fistula‑in‑ano who underwent both MR fistulogram and subsequent surgery within a span of 1 month over one year. MRI scans were reviewed by an experienced radiologist blinded to surgical findings. Parameters assessed included fistula type (Parks, St. James, simple vs. complex), number and clock position of internal and external openings, craniocaudal level of internal openings, puborectalis involvement, secondary tracts, presence of secondary tracts, and location of abscess, if any. Concordance between MRI and surgical findings was evaluated using percentage agreement and weighted kappa coefficients.

Results. Our study cohort had a mean age of 41.7 years, with the majority being men (89%) and cryptoglandular etiology (93.6%). Transsphincteric fistula was the most common type (64%). Complex fistulas were seen in 63.6%. Secondary tracts, abscesses, or multiple tracts were seen in 45%, 30.5%, and 11%, respectively. There was almost perfect agreement between MRI and surgical findings in identifying fistula type, clock position of internal and external openings, secondary tracts, and location of abscesses (k=0.98, 0.93, 0.94, 0.88 and 0.98, respectively), substantial agreement for the craniocaudal level of internal opening (k=0.72), and only moderate agreement for the number of internal and external openings (k=0.56 and 0.51, respectively).

Conclusions. Non‑contrast MR fistulogram, with its excellent soft tissue resolution, accurately depicts the type of fistula‑in‑ano, localises the internal and external openings, and identifies the presence of any complicating features with almost perfect agreement between MRI and surgical findings.