Computed tomography detection and localization of gastrointestinal tract injury in blunt abdominal trauma: a retrospective single-center series.

PURPOSE: To determine how often contrast-enhanced computed tomography (CT) localizes a definite gastrointestinal tract injury after blunt trauma, how accurately the named segment matches operative findings, and whether indirect CT signs are present when no segment can be localized. METHODS: Single-center retrospective series of 23 patients with blunt abdominal trauma who underwent portal venous phase contrast-enhanced abdominopelvic CT during the index admission and exploratory laparotomy between October 2022 and October 2024. Two radiologists with 10 years of trauma center experience independently re-read all 23 examinations, irrespective of the original report or operative diagnosis, and were blinded to operative findings; disagreements were resolved by consensus. CT was considered localization-positive only when consensus named at least one definite injured gastrointestinal segment; indirect signs were recorded separately. Proportions are reported with Wilson 95% confidence intervals (CIs). RESULTS: Gastrointestinal tract injury was confirmed at laparotomy in 22 of 23 patients (95.7%), with multi-segment injury in six. CT localized at least one injured segment in 16 of 22 patients (localization-based sensitivity 72.7%; 95% CI 51.8-86.8). All six patients without segment localization nevertheless had extraluminal air, indicating indirect evidence of injury. The only patient without operative gastrointestinal tract injury had no CT-localized bowel segment; because only one such patient was included, specificity and predictive values were not estimated. Among the 16 localization-positive examinations, the segment named on CT matched the operative finding exactly in 15 (93.8%; 95% CI 71.7-98.9) and was incomplete in one. Localization was lowest for ileal (6/9) and rectal (0/1) injury. Free fluid was present in 18 patients (78.3%) and extraluminal air in 15 (65.2%). Diagnosis was delayed in three patients (13.0%). CONCLUSION: In this surgically selected cohort, CT localized an injured segment in approximately three-quarters of affected patients and was usually exact when localization was possible. Failure to localize did not mean absence of CT evidence: all six unlocalized injuries had extraluminal air. No false-positive segment localization occurred, but only one patient lacked operative gastrointestinal tract injury. In this series, readers reviewed both soft-tissue and lung windows, and every unlocalized injury nonetheless showed extraluminal air on this dual-window review, supporting routine use of lung windows to aid recognition of subtle extraluminal gas; failure to localize ileal injury should not override clinical concern; the apparent poor localization of rectal injury is based on a single case and should be interpreted with caution.