Related Experiment Videos
Can Machine Learning Reduce Unnecessary Surgeries? A Retrospective Analysis Using Threshold Optimization to Prevent
Ivan Males1,2, Marko Kumric2,3,4, Zvonimir Boban2
1Department of Surgery, Division of Abdominal Surgery University Hospital of Split Split Croatia.
Annals of Gastroenterological Surgery
|July 24, 2026
Summary
Machine learning models can aid in diagnosing acute appendicitis using clinical and lab data, potentially reducing unnecessary surgeries. These models identified patients with low appendicitis risk, suggesting reconsidering immediate surgery.
Area of Science:
- Medical Informatics
- Surgical Oncology
- Machine Learning in Healthcare
Background:
- Acute appendicitis is a common surgical emergency.
- Negative appendectomies contribute to healthcare costs and patient morbidity.
- Accurate diagnostic tools are crucial for timely and appropriate treatment.
Purpose of the Study:
- To develop and validate machine learning (ML) models for diagnosing acute appendicitis.
- To assess the potential of ML models in reducing negative appendectomies.
- To identify predictors of appendicitis using clinical and laboratory data.
Main Methods:
- Retrospective study of adult patients with suspected acute appendicitis (Jan 2020 - June 2024).
- Development and comparison of logistic regression, random forest, balanced random forest, and gradient boosting models.
- Performance evaluation using Area Under the Receiver Operating Characteristic Curve (AUC), sensitivity, specificity, and SHAP analysis for explainability.
Main Results:
- Logistic regression achieved the highest AUC (0.765) for appendicitis detection.
- Random forest showed the best performance (AUC 0.785) for predicting complicated appendicitis.
- Inflammatory laboratory markers were identified as key predictors; models showed stability via bootstrap analysis.
Conclusions:
- ML models using clinical and lab data can support surgical decision-making in suspected acute appendicitis.
- These models may help identify low-risk patients where surgery could be reconsidered.
- Prospective external validation is necessary before clinical implementation.