PURPOSE: We evaluated the CT criteria useful for the diagnosis of the site of bowel perforation in patients with this or mesenteric injury caused by blunt abdominal trauma. MATERIALS AND METHODS: CT findings of 26patients with blunt abdominal trauma were retrospectively analyzed by two observers who were unaware of operative findings, and the results of their analysis were compared with those findings. Twenty cases of bowel perforation found at the jejunum (8), ileum (9), and colon (3), and six cases of mesenteric injuries were confirmed byoperation. We evaluated CT findings of 1) segmental bowel wall thickening, b) focal mesenteric fat infiltration, c) loculated fluid collection and d) extraluminal air adjacent to the bowel, and in addition analyzed the locations of ascites and free air, and the associated injuries of solid organs. RESULTS: The most common finding at the site of bowel perforation was segmental bowel wall thickening (17 cases), followed by focal mesenteric fatinfiltration (12 cases), loculated fluid (12 cases) and extraluminal air ajacent to the bowel (9 cases). Segmental bowel wall thickening was present at 34 sites, and the perforations were confirmed at 17 of these(50%). Focalmesenteric fat infiltration was present at 19 sites ; the perforations were proven at 12(63%). Loculated fluid collections were confirmed at 12/20 sites(60%), and extraluminal air adjacent to the bowel at 9/12(75%). Thepositive predictive value of criteria a, b, d, c) was 100%, and the positive predictive values of a, b, c) and a,d) were 60 and 67%, respectively. We observed ascites in 16 cases and intraperitoneal free air in 8 cases, the locations of ascites and free air did not, however, significantly correlate with the perforation sites. CONCLUSION: Extraluminal air adjacent to the bowel was the most specific criterion, and segmental bowel wall thickening wasthe most sensitive criterion. of all criteria, the finding 'a, b, c, d' most accurately predicted the site ofbowel perforation.