Artificial Intelligence in Emergency Surgery – Predictive models for risk of anastomotic leakage

Aim

To investigate the potential of using artificial intelligence (AI) as a decision-support tool in emergency surgery involving small bowel resections. The aim is to improve intraoperative decision-making regarding anastomosis vs stoma, thereby reducing the risk of anastomotic leakage.

Study design

Retrospective single center cohort study involving all emergency smallbowel resections from 2020-2025.

About the project

Anastomotic leakage (AL) is a serious complication in emergency abdominal surgery, associated with increased morbidity, mortality, and need for reoperation. In emergency patients, this risk is heightened further by compromised physiological states such as hypovolemia or sepsis.

When bowel resection is required due to trauma, ischemia, perforation or obstruction, it falls to the surgeon to decide whether anastomosis or stoma is right for the patient. Anastomosis preserves natural bowel function but carries a risk of leakage, while a stoma eliminates this risk at the cost of additional care needs and its own complications. In the acute setting, this decision often rests on the surgeon's intraoperative judgment alone, without time for thorough preoperative workup or multidisciplinary input.

By collecting data on suspected risk factors for anastomotic leakage, like hypovolemia, anemia, hypoalbuminemia, comorbidity and other factors such as indication for surgery surgical technique and mortality the aim is to train a machine learning model to help analyze and predict the risk of anastomotic leakage.

As AI increasingly supports clinical decision-making in fields like cardiology and infection control, this project explores its potential in emergency surgery. We aim to develop an AI model that can assist the clinical decision making of the surgeon in real time, thus reducing the risk of adverse events.

Research area:

Emergency and Trauma Surgery 

Project Investigator (PI)

Project participants: