64-slice multidetector computed tomography evaluation of gastrointestinal tract perforation site: detectability of direct findings in upper and lower GI tract

Abstract

To evaluate wall discontinuity, as observed using 64-slice multidetector-row computed tomography (64-MDCT), as a direct finding (DF) indicating the perforation site in patients with gastrointestinal (GI) tract perforations. We retrospectively studied 41 consecutive patients presenting with acute abdomen and exhibiting extraluminal air (EA) on 64-MDCT. Three readers evaluated the distribution of EA, extraluminal faeces, dirty mass, dirty fat sign, extraluminal fluid collection and bowel wall thickening (i.e. conventional findings, CFs) as well as DFs. Twenty-two cases were surgically or endoscopically confirmed to have upper GI tract perforations, and 19 had lower GI tract perforations. The DFs correctly identified the sites of perforation in 80.5% of patients when 2-mm-thick imaging slices were used. For the detection of upper GI tract perforations, the sensitivity, specificity and accuracy were 95.5%, 94.7% and 95.1% for the DFs and 50.0%, 100% and 73.2% for the CFs, respectively. Significant differences in sensitivity (p < 0.001) and diagnostic accuracy (p < 0.05) were observed between the DFs and CFs for upper GI perforations but not for lower GI tract perforations. DFs of the perforation site by using 64-MDCT were more sensitive and accurate than CFs for the detection of upper GI tract perforations.