MRI OF POUCH-RELATED FISTULAS IN ULCERATIVE-COLITIS AFTER RESTORATIVEPROCTOCOLECTOMY

Purpose: Our purpose was to determine the value of MRI in diagnosing p ouch-related fistulas in patients with ulcerative colitis and to compa re pulse sequences with and without contrast enhancement in their perf ormance of visualization. Method: Forty-four patients with pelvic symp toms after restorative proctocolectomy underwent MRI. All 26 patients with pouch-related fistulas were treated surgically; 18 patients with pouchitis were treated conservatively. MRI was performed at 1.0 T with T1-weighted FLASH sequences before and after administration of Gd-DTP A, T2-weighted and proton density-weighted turbo SE sequences, and a T 2-weighted fat saturation sequence. Images were analyzed for the prese nce of fistula; pulse sequences were additionally compared for best vi sualization on a four point scale of diagnostic confidence. Results: M RI detected 23 of 26 cases of fistulas; there were no false-positive d iagnoses. Surgery revealed fistulas in three cases in which Ilo pathol ogy was found on MRI. Two patients had a short sinus tract at the pouc h-anal anastomosis, and a third patient had a pouch-vaginal fistula. T he Gd-enhanced FLASH sequence obtained the highest score, and second b est was the T2-weighted fat saturation technique, Conclusion: MRI is a valuable technique for diagnosing pouch-related fistulas. However, th ere are limitations in detection of short sinus tracts and pouch-vagin al fistulas. Highest diagnostic confidence is obtained with a Gd-enhan ced FLASH sequence, which might be helpful after pelvic surgery or if the fat saturation technique is equivocal.