Objective: To assess the diagnostic accuracy of MRI for detecting perianal fistulas using intraoperative findings as the reference standard, evaluate interobserver agreement in MRI interpretation, and determine the frequency of associated MRI findings. Methods: This cross-sectional study included 226 patients (January to July 2025) with clinically suspected perianal fistula who underwent pelvic MRI on a 1.5-T scanner. Two experienced radiologists independently interpreted MRI findings, blinded to each other. Fistulas were graded using the St. James University Hospital classification. MRI findings were correlated with operative outcomes. Diagnostic performance indices and interreader agreement (Cohen’s kappa) were calculated. Results: Operative confirmation of perianal fistula was present in 109 patients (48.2%). MRI demonstrated a sensitivity of 85.3%, specificity of 88.9%, positive predictive value of 87.7%, negative predictive value of 86.7%, and overall diagnostic accuracy of 87.2%. False-negative cases were predominantly small intersphincteric fistulas, while false positives were mainly blind-ending sinustracts and inflammatory abscesses. MRI grading showed excellent interobserver agreement (unweighted ê = 0.88; weighted ê = 0.94–0.98). Among operative-confirmed cases, secondary tracts were seen in 33.9%, lymphadenopathy in 11.9%, abscesses in 7.3%, and horseshoe extensions in 4.6%. Conclusion: MRI demonstrates high diagnostic accuracy and excellent interreader reliability for evaluation of perianal fistulas, particularly in complex disease. However, simple intersphincteric fistulas may be missed. In resource-limited settings, selective use of MRI for suspected complex or recurrent fistulas may optimize patient outcomes and healthcare utilization.
Memon et al. (Mon,) studied this question.