Performance of Postoperative Imaging for Leak Detection Following Repair of Upper Gastrointestinal Perforation

Gastric and duodenal perforations are surgical emergencies with high morbidity and mortality. Postoperative leak is a serious complication, yet there are no guidelines for imaging modality or patient selection. This study compared fluoroscopic upper gastrointestinal series (fUGI) and computed tomography (CT) for postoperative leak detection following surgical repair.

This is a retrospective cohort study of patients who underwent repair of a gastric or duodenal perforation between 2011 and 2023. Surgical interventions, postoperative imaging (UGI vs CT), and clinical outcomes were assessed. Post repair leak was defined by radiographic evidence of contrast extravasation, change in drain character, and/or intraoperative findings. Performance of UGI and CT to identify post repair leak was assessed.

A total of 168 patients were included; 142 underwent open repair and 26 laparoscopic. An omental patch was used in 87.5%, and drains were placed in 81.5%. Postoperative imaging was obtained in 85.7%, most often on postoperative day 5, with 95 fUGI and 49 CT studies. Postoperative leak occurred in 20.2% (34/168). All patients with leaks underwent imaging. For patients with confirmed leaks, fUGI demonstrated a sensitivity of 56.3% and negative predictive value (NPV) of 91.9%, while CT demonstrated a sensitivity of 44.4% and NPV of 75.6%. Discordant imaging results were frequent, and CT contrast protocols were not standardized.

Postoperative imaging after gastric or duodenal perforation repair is common but variable, often performed for surveillance. fUGI demonstrated modestly higher sensitivity and NPV compared with CT, suggesting greater reliability for ruling out early leaks. Both modalities showed limitations, underscoring the need for standardized imaging protocols.