BACKGROUND: High diagnostic accuracy is reported for magnetic resonance enterography (MRE) in Crohn's disease (CD), but few studies have evaluated its role in abdominal fistulae. The primary aim of this study was to assess the reliability of MRE in the identification of internal fistulae in CD. METHODS: One hundred and eighty-six patients with moderate CD (CD Activity Index : 250-400) were prospectively selected from the inflammatory bowel disease clinic of Parma University Hospital. Eligible patients had already undergone nutritional screening, pancolonoscopy, and computed tomography enterography (CTE) in the month before enrollment. MRE was performed according to the study protocol. Additional fluoroscopic contrast-enhanced studies or surgical evaluation were used for discordance between CTE and MRE results. A consensus committee resolved equivocal findings. Surgical findings and/or fluoroscopic contrast-enhanced studies together with the clinical data were considered the composite "reference standard" to which the results of MRE were compared. RESULTS: MRE identified 22 internal fistulae in 21 patients (11%), of whom 4 (19%) also had perianal fistulae and found 7 abscesses (33%). Forty-one (22%) additional patients with perianal fistulae were identified. Thirteen patients (57%) with internal fistulae required enteral nutrition support. No statistically significant differences were found between MRE and CTE in fistula detection. There was also no significant difference between MRE and the composite diagnosis in those who underwent surgery (n = 8) and/or contrast-enhanced studies (n = 7). CONCLUSIONS: CTE and MRE accurately detect internal fistulae in CD. MRE is preferable because it avoids radiation. Reliable identification of internal fistulae by MRE should permit earlier and improved treatment.
BACKGROUND: High diagnostic accuracy is reported for magnetic resonance enterography (MRE) in Crohn's disease (CD), but few studies have evaluated its role in abdominal fistulae. The primary aim of this study was to assess the reliability of MRE in the identification of internal fistulae in CD. METHODS: One hundred and eighty-six patients with moderate CD (CD Activity Index : 250-400) were prospectively selected from the inflammatory bowel disease clinic of Parma University Hospital. Eligible patients had already undergone nutritional screening, pancolonoscopy, and computed tomography enterography (CTE) in the month before enrollment. MRE was performed according to the study protocol. Additional fluoroscopic contrast-enhanced studies or surgical evaluation were used for discordance between CTE and MRE results. A consensus committee resolved equivocal findings. Surgical findings and/or fluoroscopic contrast-enhanced studies together with the clinical data were considered the composite "reference standard" to which the results of MRE were compared. RESULTS: MRE identified 22 internal fistulae in 21 patients (11%), of whom 4 (19%) also had perianal fistulae and found 7 abscesses (33%). Forty-one (22%) additional patients with perianal fistulae were identified. Thirteen patients (57%) with internal fistulae required enteral nutrition support. No statistically significant differences were found between MRE and CTE in fistula detection. There was also no significant difference between MRE and the composite diagnosis in those who underwent surgery (n = 8) and/or contrast-enhanced studies (n = 7). CONCLUSIONS:CTE and MRE accurately detect internal fistulae in CD. MRE is preferable because it avoids radiation. Reliable identification of internal fistulae by MRE should permit earlier and improved treatment.