ORIGINAL RESEARCH n GASTROINTESTINAL IMAGING 98 radiology.rsna.org n Radiology: Volume 280: Number 1—July 2016 1 From the Department of Radiology (A.M.C., L.C., I.B.C., J.H., M.Z.) and Clinical Research Unit (H.B., G.C.), Hôpital Saint-Joseph, 185 rue Raymond Losserand, 75014 Paris, France; Department of Medical Imaging, CHU Lapeyronie, Montpellier, France (I.M., P.T.); Department of Neuroradiol- ogy, AP-HP, Hôpitaux Universitaires Henri Mondor, Créteil, Faculty of Medicine, Université Paris Est Créteil, Créteil, France (J.H.), and Clinical Research Unit, Hôpital Européen Georges Pompidou, Université René Descartes, Paris, France (G.C.). Received May 4, 2015; revision requested June 29; revision received October 9; accepted November 4; final version accepted November 19. Address corre- spondence to M.Z. (e-mail: mzins@hpsj.fr ). q RSNA, 2016 Purpose: To determine whether adding unenhanced computed to- mography (CT) to contrast material–enhanced CT improves the diagnostic performance of decreased bowel wall en- hancement as a sign of ischemia complicating mechanical small bowel obstruction (SBO). Materials and Methods: This retrospective study was approved by the institutional review board, which waived the requirement for informed consent. Two gastrointestinal radiologists independently performed retrospective assessments of 164 unenhanced and contrast-enhanced CT studies from 158 consecutive pa- tients (mean age, 71.2 years) with mechanical SBO. The reference standard was the intraoperative and/or histologic diagnosis (in 80 cases) or results from clinical follow-up in patients who did not undergo surgery (84 cases). Decreased bowel wall enhancement was evaluated with contrast-en- hanced images then and both unenhanced and contrast- enhanced images 1 month later. Diagnostic performance of decreased bowel wall enhancement and confidence in the diagnosis were compared between the two readings by using McNemar and Wilcoxon signed rank tests. Interobserver agreement was assessed by using k statistics and compared with bootstrapping. Results: Ischemia was diagnosed in 41 of 164 (25%) episodes of SBO. For both observers, adding unenhanced images im- proved decreased bowel wall enhancement sensitivity (ob- server 1: 46.3% [19 of 41] vs 65.8% [27 of 41], P = .02; observer 2: 56.1% [23 of 41] vs 63.4% [26 of 41], P = .45), Youden index (from 0.41 to 0.58 for observer 1 and from 0.42 to 0.61 for observer 2), and confidence score (P , .001 for both). Specificity significantly increased for observer 2 (84.5% [104 of 123] vs 94.3% [116 of 123], P = .002), and interobserver agreement significantly increased, from mod- erate (k = 0.48) to excellent (k = 0.89; P , .0001). Conclusion: Adding unenhanced CT to contrast-enhanced CT improved the sensitivity, diagnostic confidence, and interobserver agreement of the diagnosis of ischemia, a complication of mechanical SBO, on the basis of decreased bowel wall enhancement. q RSNA, 2016 Anh Minh Chuong, MD Lucie Corno, MD Hélène Beaussier, PharmD, PhD Isabelle Boulay-Coletta, MD Ingrid Millet, MD, PhD Jérôme Hodel, MD, PhD Patrice Taourel, MD, PhD Gilles Chatellier, MD, PhD Marc Zins, MD Assessment of Bowel Wall Enhancement for the Diagnosis of Intestinal Ischemia in Patients with Small Bowel Obstruction: Value of Adding Unenhanced CT to Contrast-enhanced CT 1 This copy is for personal use only. To order printed copies, contact reprints@rsna.org