Related Experiment Video
Updated: Jan 6, 2026
14:56
An Experimental Paradigm for the Prediction of Post-Operative Pain PPOP
Published on: January 27, 2010
21.9K
Predicting postoperative fever in culture-negative patients undergoing mini-PCNL using MAP score-augmented machine
Rong Xu1, Jia-Jia Wang2, Wei-Hong Zhao1
1Department of Urology, Taizhou Hospital of Zhejiang Province Affiliated with Wenzhou Medical University, No.150, Ximen Street, Linhai, Taizhou, 317000, Zhejiang Province, China.
World Journal of Urology
|September 25, 2025
Summary
A new tool combining Mayo Adhesive Probability (MAP) scores with machine learning (ML) accurately predicts postoperative fever after mini-percutaneous nephrolithotomy (mini-PCNL) in patients with sterile urine. This aids in personalized preoperative risk assessment and clinical decision-making.
Area of Science:
- Urology
- Nephrology
- Medical Informatics
- Machine Learning in Healthcare
Background:
- Postoperative fever is a common complication after percutaneous nephrolithotomy (PCNL), even with sterile urine cultures.
- Existing risk assessment tools are inadequate for predicting fever in this specific patient group.
- Accurate prediction of postoperative fever is crucial for patient management and resource allocation.
Purpose of the Study:
- To develop a predictive model for postoperative fever in patients undergoing mini-PCNL with sterile preoperative urine cultures.
- To integrate the Mayo Adhesive Probability (MAP) score with machine learning (ML) techniques for enhanced risk prediction.
- To create an interpretable and clinically applicable tool for preoperative risk assessment.
Main Methods:
- Retrospective cohort study of 730 patients undergoing mini-PCNL with sterile urine cultures.
- Least Absolute Shrinkage and Selection Operator (LASSO) regression to identify key predictive variables.
- Development and comparison of ten ML models, with Shapley Additive exPlanations (SHAP) for interpretability; Logistic Regression (LR) selected as optimal.
- Deployment of an online tool for clinical use.
Main Results:
- Postoperative fever occurred in 17.4% of patients.
- The LR model achieved high predictive performance (AUC=0.914, accuracy=92.1%, specificity=97.3%).
- Key predictors included MAP score ≥3, diabetes mellitus, female sex, positive urine leukocytes, and low lymphocyte-monocyte ratio (LMR).
- SHAP analysis highlighted MAP score and urine leukocytes as most influential factors.
Conclusions:
- The integrated MAP score and ML model significantly improves the prediction of postoperative fever following mini-PCNL in patients with sterile urine.
- The LR model offers strong predictive utility, interpretability, and enables personalized preoperative risk assessment.
- The developed online tool provides a practical and accessible resource for individualized clinical decision-making.
