Diagnostic accuracies of MR enterography and CT enterography in symptomatic Crohn's disease

Research output: Contribution to journalJournal articleResearchpeer-review


Abstract Objective. In patients, with symptomatic Crohn's disease (CD), valid information about the presence or absence of small bowel disease activity and stenosis is clinically important. Such information supports decisions about medical or surgical therapy and can be obtained with MR enterography (MRE) or CT enterography (CTE). Materials and methods. A total of 50 patients with symptomatic pre-existing CD and a demand for small bowel imaging to support changes in treatment strategy were included in this prospective and blinded study. MRE and CTE were performed on the same day in alternating order and subsequently compared with the gold standard: pre-defined lesions at ileoscopy (n = 30) or surgery with (n = 12) or without (n = 3) intra-operative enteroscopy. Results. A total of 35 patients had active small bowel CD (jejunum 0, ileum 1, (neo)-terminal ileum 34) and 20 had small bowel stenosis. The sensitivity and specificity of MRE for detection of small bowel CD was 74% and 80% compared to 83% and 70% with CTE (p ≥ 0.5). MRE and CTE detected small bowel stenosis with 55% and 70% sensitivities, respectively (p = 0.3) and 92% specificities. Conclusions. MRE and CTE have comparable diagnostic accuracies for detection of small bowel CD and stenosis. In symptomatic patients with CD and high disease prevalence, positive predictive values are favorable but negative predictive values are low. Consequently, MRE and CTE can be relied upon, if a positive result is obtained whereas a negative enterography should be interpreted with caution.
Original languageEnglish
JournalScandinavian Journal of Gastroenterology
Issue number12
Pages (from-to)1449-1457
Number of pages9
Publication statusPublished - 2011


Dive into the research topics of 'Diagnostic accuracies of MR enterography and CT enterography in symptomatic Crohn's disease'. Together they form a unique fingerprint.

Cite this