Increasing efficiency of MRE for diagnosis of Crohn's disease activity through proper sequence selection: a practical approach for clinical trials.

Abstract

To derive the best magnetic resonance enterography (MRE) approach for detecting activity and severe lesions in Crohn's disease (CD) to use for selecting patients and measuring response to treatment in clinical trials. We compared the accuracies of MRE (T2-weighted sequences, DWI (b=800s/mm2) sequences, combined T2-weighted and DWI sequences, combined T2-weighted or DWI sequences, and MaRIA score based on T2-weighted and contrast-enhanced T1-weighted sequences) versus ileocolonoscopy (SES-CD) performed within 1month. Bowel segments were classified as inactive (SES-CD<2), active (SES-CD≥2), or active with severe lesions (ulcers seen at endoscopy). McNemar's test was used to compare the accuracies of the different approaches against endoscopy. 224 segments in 43 patients were analyzed. For detecting active disease, the combination of findings from T2 and DWI sequences results in the highest specific and accurate sequence combination. Combined T2-weighted and DWI sequences had similar sensitivity to those of MaRIA (P=0.25) but lower specificity (P=0.007) and accuracy (P=0.0013) than MaRIA score. For detecting severe lesions, T2-weighted sequences alone had greater accuracy [similar to MaRIA score (P>0.999)] than other noncontrast approaches. T2-weighted sequences should be used as a first screening step, and followed by contrast-enhanced T1-weighted sequences only when abnormal findings are identified; adding DWI does not improve the accuracy of MRE.