Abstract

Aim: Magnetic resonance imaging (MRI) based colonography represents a new imaging tool which has mainly been investigated for polyp screening. To evaluate this approach for patients with inflammatory bowel disease (IBD), we compared MRI based colonography with conventional colonoscopy for assessing the presence and extent of colonic inflammation.

Patients and methods: In 22 consecutive patients with suspected or known IBD, MRI colonography was performed immediately before conventional colonoscopy. After bowel cleansing, a T1 positive contrast agent was applied rectally. In addition to T2 weighted sequences, T1 weighted two dimensional and three dimensional Flash acquisitions as well as volume rendered virtual endoscopy were performed. All images were evaluated with regard to typical MRI features of inflammation. The results were compared with colonoscopy findings.

Results: Distension and image quality was assessed as good to fair in 97.4% of all colonic segments. Only four of 154 segments were considered non-diagnostic. With colonoscopy serving as the gold standard, the sensitivity for correctly identifying inflammation on a per segment analysis of the colon was 31.6% for Crohn’s disease (CD) and 58.8% for ulcerative colitis (UC). In CD, in most cases mild inflammation was not diagnosed by MRI while in UC even severe inflammation was not always depicted by MRI. Virtual endoscopy did not add any relevant information.

Conclusion: MRI based colonography is not suitable for adequately assessing the extent of colonic inflammation in patients with IBD. Only severe colonic inflammation in patients with CD can be sufficiently visualised.