PURPOSE: The purpose of this retrospective study was to determine what gives rise to the periportal free air, and ligamentum teres and falciform ligament signs on CT in patients with gastrointestinal (GI) tract perforation, and whether these specific air distributions can play a clinically meaningful role in the diagnosis of gastroduodenal perforation. MATERIAL AND METHODS: Ninety-three patients who underwent a diagnostic CT scan before laparotomy for a GI tract perforation were included. The readers assessed the presence of specific air distributions on CT (periportal free air, and ligamentum teres and falciform ligament signs). The readers also assessed the presence of strong predictors of gastroduodenal perforation (focal defects in the stomach and duodenal bulb wall, concentrated extraluminal air bubbles in close proximity to the stomach and duodenal bulb, and wall thickening at the stomach and duodenal bulb). The specific air distributions were assessed according to perforation sites, and the elapsed time and amount of free air, and then compared with the strong predictors of gastroduodenal perforation by using statistical analysis. RESULTS: All specific air distributions were more frequently present in patients with gastroduodenal perforation than lower GI tract perforation, but only the falciform ligament sign was statistically significant (p<0.05). The presence of all three specific air distributions was demonstrated in only 13 (20.6%) of 63 patients with gastroduodenal perforation. Regardless of the perforation sites, the falciform ligament sign was present significantly more frequently with an increase in the amount of free air on multiple logistic regression analysis (adjusted odds ratio, 1.29; p<0.001). The sensitivity, specificity, accuracy, and positive predictive and negative predictive values of each strong predictor for the diagnosis of gastroduodenal perforation were higher than those of specific air distributions. The focal wall thickening (accuracy, 95.7%) was the most useful parameter for the diagnosis of gastroduodenal perforation. CONCLUSION: The prediction of the perforation site of the GI tract on CT should be based on the presence of strong predictors of the site of bowel perforation, and the specific free air distribution should be regarded as complementary predictors.
PURPOSE: The purpose of this retrospective study was to determine what gives rise to the periportal free air, and ligamentum teres and falciform ligament signs on CT in patients with gastrointestinal (GI) tract perforation, and whether these specific air distributions can play a clinically meaningful role in the diagnosis of gastroduodenal perforation. MATERIAL AND METHODS: Ninety-three patients who underwent a diagnostic CT scan before laparotomy for a GI tract perforation were included. The readers assessed the presence of specific air distributions on CT (periportal free air, and ligamentum teres and falciform ligament signs). The readers also assessed the presence of strong predictors of gastroduodenal perforation (focal defects in the stomach and duodenal bulb wall, concentrated extraluminal air bubbles in close proximity to the stomach and duodenal bulb, and wall thickening at the stomach and duodenal bulb). The specific air distributions were assessed according to perforation sites, and the elapsed time and amount of free air, and then compared with the strong predictors of gastroduodenal perforation by using statistical analysis. RESULTS: All specific air distributions were more frequently present in patients with gastroduodenal perforation than lower GI tract perforation, but only the falciform ligament sign was statistically significant (p<0.05). The presence of all three specific air distributions was demonstrated in only 13 (20.6%) of 63 patients with gastroduodenal perforation. Regardless of the perforation sites, the falciform ligament sign was present significantly more frequently with an increase in the amount of free air on multiple logistic regression analysis (adjusted odds ratio, 1.29; p<0.001). The sensitivity, specificity, accuracy, and positive predictive and negative predictive values of each strong predictor for the diagnosis of gastroduodenal perforation were higher than those of specific air distributions. The focal wall thickening (accuracy, 95.7%) was the most useful parameter for the diagnosis of gastroduodenal perforation. CONCLUSION: The prediction of the perforation site of the GI tract on CT should be based on the presence of strong predictors of the site of bowel perforation, and the specific free air distribution should be regarded as complementary predictors.