Diagnostic accuracy of magnetic resonance enterography and small bowel ultrasound for the extent and activity of newly diagnosed and relapsed Crohn's disease (METRIC): a multicentre trial

BackgroundMagnetic resonance enterography (MRE) and ultrasound are used to image Crohn's disease, but their comparative accuracy for assessing disease extent and activity is not known with certainty. Therefore, we did a multicentre trial to address this issue.MethodsWe recruited patients from eight UK hospitals. Eligible patients were 16 years or older, with newly diagnosed Crohn's disease or with established disease and suspected relapse. Consecutive patients had MRE and ultrasound in addition to standard investigations. Discrepancy between MRE and ultrasound for the presence of small bowel disease triggered an additional investigation, if not already available. The primary outcome was difference in per-patient sensitivity for small bowel disease extent (correct identification and segmental localisation) against a construct reference standard (panel diagnosis). This trial is registered with the International Standard Randomised Controlled Trial, number ISRCTN03982913, and has been completed.Findings284 patients completed the trial (133 in the newly diagnosed group, 151 in the relapse group). Based on the reference standard, 233 (82%) patients had small bowel Crohn's disease. The sensitivity of MRE for small bowel disease extent (80% [95% CI 72–86]) and presence (97% [91–99]) were significantly greater than that of ultrasound (70% [62–78] for disease extent, 92% [84–96] for disease presence); a 10% (95% CI 1–18; p=0·027) difference for extent, and 5% (1–9; p=0·025) difference for presence. The specificity of MRE for small bowel disease extent (95% [85–98]) was significantly greater than that of ultrasound (81% [64–91]); a difference of 14% (1–27; p=0·039). The specificity for small bowel disease presence was 96% (95% CI 86–99) with MRE and 84% (65–94) with ultrasound (difference 12% [0–25]; p=0·054). There were no serious adverse events.InterpretationBoth MRE and ultrasound have high sensitivity for detecting small bowel disease presence and both are valid first-line investigations, and viable alternatives to ileocolonoscopy. However, in a national health service setting, MRE is generally the preferred radiological investigation when available because its sensitivity and specificity exceed ultrasound significantly.FundingNational Institute of Health and Research Health Technology Assessment. Magnetic resonance enterography (MRE) and ultrasound are used to image Crohn's disease, but their comparative accuracy for assessing disease extent and activity is not known with certainty. Therefore, we did a multicentre trial to address this issue. We recruited patients from eight UK hospitals. Eligible patients were 16 years or older, with newly diagnosed Crohn's disease or with established disease and suspected relapse. Consecutive patients had MRE and ultrasound in addition to standard investigations. Discrepancy between MRE and ultrasound for the presence of small bowel disease triggered an additional investigation, if not already available. The primary outcome was difference in per-patient sensitivity for small bowel disease extent (correct identification and segmental localisation) against a construct reference standard (panel diagnosis). This trial is registered with the International Standard Randomised Controlled Trial, number ISRCTN03982913, and has been completed. 284 patients completed the trial (133 in the newly diagnosed group, 151 in the relapse group). Based on the reference standard, 233 (82%) patients had small bowel Crohn's disease. The sensitivity of MRE for small bowel disease extent (80% [95% CI 72–86]) and presence (97% [91–99]) were significantly greater than that of ultrasound (70% [62–78] for disease extent, 92% [84–96] for disease presence); a 10% (95% CI 1–18; p=0·027) difference for extent, and 5% (1–9; p=0·025) difference for presence. The specificity of MRE for small bowel disease extent (95% [85–98]) was significantly greater than that of ultrasound (81% [64–91]); a difference of 14% (1–27; p=0·039). The specificity for small bowel disease presence was 96% (95% CI 86–99) with MRE and 84% (65–94) with ultrasound (difference 12% [0–25]; p=0·054). There were no serious adverse events. Both MRE and ultrasound have high sensitivity for detecting small bowel disease presence and both are valid first-line investigations, and viable alternatives to ileocolonoscopy. However, in a national health service setting, MRE is generally the preferred radiological investigation when available because its sensitivity and specificity exceed ultrasound significantly.