Detecting inflammation in the unprepared pediatric colon — how reliable is magnetic resonance enterography?

Background Pediatric inflammatory bowel disease frequently affects the colon. MR enterography is used to assess the small bowel but it also depicts the colon. Objective To compare the accuracy of MR enterography and direct visualization at endoscopy in assessing the colon in pediatric inflammatory bowel disease. Materials and methods We included children with inflammatory bowel disease who had undergone both MR enterography and endoscopy, and we restrospectively assessed the imaging and endoscopic findings. We scored the colonic appearance at MR using a total colon score. We then compared scores for the whole colon and for its individual segments with endoscopy and histology. Results We included 15 children. An elevated MR colonic segmental score predicted the presence of active inflammation on biopsy with a specificity of 90% (95% confidence interval [CI] 79.5–96.2%) and sensitivity of 60% (CI 40.6–77.3%); this compares reasonably with the predictive values for findings at colonoscopy — specificity 85% (CI 73.4—92.9%) and sensitivity 53.3% (CI 34.3%–71.6%). Accuracy did not change significantly with increasing bowel distension. Conclusion MR-derived scores had comparable accuracy to those derived during visualization at colonoscopy for detecting biopsy-proven inflammation in our patient group. MR enterography might prove useful in guiding biopsy or monitoring treatment response. Collapse of a colonic segment did not impair assessment of inflammation.