Related Experiment Video
Updated: Jan 12, 2026
08:58
Development of a Uterosacral Ligament Suspension Rat Model
Published on: August 17, 2022
5.2K
Predicting Postoperative Stress Urinary Incontinence After Prolapse Surgery via Machine Learning and Regression
Minna Su1, Shuyu Wang1, Xiaochun Liu1
1Gynecology and Obstetrics Department, Third Hospital of Shanxi Medical University, Shanxi Bethune Hospital, Shanxi Academy of Medical Sciences, No. 99 Longcheng Street, Xiaodian District, Taiyuan, 030000, China, 86 13934226668.
JMIR Medical Informatics
|November 3, 2025
Summary
A new machine learning model predicts stress urinary incontinence (SUI) after pelvic organ prolapse (POP) surgery. The support vector machine (SVM) model identifies risk factors to guide decisions on concurrent anti-incontinence procedures.
Area of Science:
- Urogynecology and reconstructive pelvic surgery
- Machine learning applications in healthcare
- Predictive modeling for surgical outcomes
Background:
- Pelvic organ prolapse (POP) and stress urinary incontinence (SUI) frequently coexist.
- SUI resolution post-POP surgery varies; some patients develop de novo SUI.
- Clinical decision-making regarding concomitant anti-incontinence surgery requires better prediction tools.
Purpose of the Study:
- To identify risk factors for SUI following POP surgery.
- To develop and validate machine learning-based prediction models for postoperative SUI.
- To provide a tool for evaluating and predicting SUI risk in patients undergoing POP surgery.
Main Methods:
- Prospective and retrospective data collection from patients undergoing prolapse surgery.
- Analysis of clinical, laboratory, urodynamic, and ultrasound findings.
- Development of prediction models using Lasso regression, random forest, SVM, XGBoost, CART, and logistic regression.
Main Results:
- 286 patients were analyzed; 91 developed postoperative SUI.
- Identified risk factors include preoperative SUI, urge urinary incontinence, urodynamic occult SUI, anti-incontinence surgery, genital hiatus, and anterior colporrhaphy.
- The SVM model showed optimal performance (AUC 0.821 training, 0.846 validation).
Conclusions:
- Five prediction models for postoperative SUI were developed with good internal validation.
- The SVM model is a promising tool for predicting SUI risk after prolapse surgery.
- Further external validation is needed; the model can guide decisions on concurrent surgeries.
Keywords:
de novo stress urinary incontinencemachine learningpelvic organ prolapseprediction modelsupport vector machine
