Using computed tomography enterography to evaluate patients with Crohn's disease: what impact does examiner experience have on the reproducibility of the method?

OBJECTIVE To assess the impact that examiner experience has on the reproducibility and accuracy of computed tomography (CT) enterography in the detection of radiological signs in patients with Crohn's disease. MATERIALS AND METHODS This was a retrospective, cross-sectional observational study involving the analysis of CT enterography scans of 20 patients with Crohn's disease. The exams were analyzed independently by two radiologists in their last year of residence (duo I) and by two abdominal imaging specialists (duo II). The interobserver agreement of each pair of examiners in identifying the main radiological signs was calculated with the kappa test. The accuracy of the examiners with less experience was quantified by using the consensus among three experienced examiners as a reference. RESULTS Duo I and duo II obtained a similar interobserver agreement, with a moderate to good correlation, for mural hyperenhancement, parietal thickening, mural stratification, fat densification, and comb sign (kappa: 0.45-0.64). The less experienced examiners showed an accuracy > 80% for all signs, except for lymph nodes and fistula, for which it ranged from 60% to 75%. CONCLUSION Less experienced examiners have a tendency to present a level of interobserver agreement similar to that of experienced examiners in evaluating Crohn's disease through CT enterography, as well as showing satisfactory accuracy in identifying most radiological signs of the disease.