Predicting surgical intervention in pediatric intussusception using machine learning model
Saloua Ammar1,2, Imen Sellami2, Emna Krichen1
1Department of pediatric surgery, Hédi Chaker Hospital, Sfax, Tunisia.
La Tunisie Medicale
|March 5, 2026
Summary
A new model accurately predicts surgical treatment for pediatric intussusception. Key factors include symptom duration, bloody stools, and mass length, aiding pre-operative decisions.
Area of Science:
- Pediatric Surgery
- Medical Informatics
- Clinical Decision Support
Background:
- Intussusception is a common surgical emergency in young children.
- Accurate prediction of surgical intervention is crucial for timely management.
Purpose of the Study:
- To develop and validate a predictive model for surgical treatment of intussusception in children.
- To identify key clinical factors associated with the need for surgery.
Main Methods:
- Retrospective chart review of children under 3 years with ileocolic intussusception.
- Development of a predictive model using logistic regression and machine learning (Knime platform).
- Model validation using a separate dataset.
Main Results:
- The final model identified symptom duration, bloody stools, and intussusception length as independent predictors of surgical treatment.
- The machine learning model achieved 95% sensitivity and specificity.
- No significant differences were observed between training and validation sets.
Conclusions:
- The developed model can assist in pre-operative decision-making for pediatric intussusception.
- Further validation through larger, prospective, multicenter studies is recommended.
