Title:Multivariate Prediction of Small-bowel Ischemia and Necrosis using CT in
Emergent Patients with Small-bowel Obstruction
Volume: 20
Author(s): Bo Li and Zhifeng Wu*
Affiliation:
- Third Hospital of Shanxi Medical University: Shanxi Bethune Hospital, Taiyuan, China
Keywords:
CT, Multivariate prediction, Small-bowel obstruction, Ischemia and Necrosis, ROC.
Abstract:
Background:
It is difficult to accurately determine whether emergent patients with small-bowel obstruction (SBO) have small-bowel ischemia and necrosis
(SBIN). Therefore, in this study, we aimed to assess the ability of abdominal CT scans to predict SBIN and establish a new predictive model.
Methods:
From March 2018 to May 2023, a rigorous posthoc analysis was conducted on whether 177 emergent patients with SBO had SBIN. Four clinical
indexes and 19 CT signs were analyzed, and a multivariate scoring model for predicting SBIN was established using logistic regression analysis. A
receiver operating characteristic (ROC) curve was used to assess the accuracy of this model.
Results:
Multivariate analysis showed that mesenteric edema and effusion (OR=23.450), significant thickening and the target sign on the small-bowel wall
on plain scans (OR=23.652), significant thinning of the small-bowel wall (OR=30.439), significant decrease in small-bowel wall density
(OR=12.885), and significant increase in small-bowel wall density (OR=19.550) were significantly correlated with SBIN (P<0.05). According to
their multivariate ORs, an appropriate “predictive score” was assigned to each sign, and the rates of SBIN among those with a total score of 0-4,
5-6, and 7-8 were 2.2%, 86.4%, and 96.9%, respectively. The AUC of this predictive scoring model for SBIN exceeded 0.980.
Conclusion:
We have developed a predictive scoring model for SBIN, for which the incidence of SBIN increases with increasing predictive scores. This model
can be useful for clinical treatment.
