JMIR Publications

ABSTRACT

Background:

Pelvic Organ Prolapse (POP) and Stress Urinary Incontinence (SUI) often concurrently existed. The incontinence of part POP patients will disappear after POP surgery, and that of part of them will persist. And a portion of the patients without SUI before surgery will develop de novo stress incontinence. Whether anti-incontinence surgery should be performed while POP surgery be performed to prevent post-operative incontinence needs a prediction model to guide clinical decision-making.

Objective:

This study aims to analyze the risk factors related to SUI after POP surgery, build prediction models for SUI after POP surgery based on machine learning to provide new tools for evaluating and predicting postoperative SUI.

Methods:

The sample size calculation performed by Riley four-step method. Data of patients undergoing prolapse surgery in Shanxi Bethune Hospital from August 2022 to December 2024 were prospectively collected and from January 2020 to August 2022 were retrospectively collected. General clinical data, relevant laboratory tests, urodynamic examination and pelvic floor ultrasound were collected. Lasso regression, univariate analysis and Logistic analysis were used to screen the predictors of SUI after prolapse surgery. Data was split randomly by 7:3 into training set and the validation set. The training set were used to develop the prediction model for postoperative SUI after POP surgery by Lasso regression, random forest, SVM, XGBoost, CART and Logistic regression, and the validation set was used for internal verification. The final implementation was achieved by developing a Shiny-based application for model deployment.

Results:

A total of 266 patients were enrolled in this study, and 82 patients had postoperative SUI. were used to determine 7 risk factors were determined by univariate analysis, Logistic regression, and Lasso regression, which are preoperative SUI, urge urinary incontinence, urodynamic occult SUI, anti-incontinence surgery, genital hiatus, age, and anterior colporrhaphy. Five prediction models were constructed by using Logistic regression, random forest, XGBoost, SVM and CART. Based on comprehensive evaluation of model discrimination, calibration, and clinical utility, the SVM model demonstrated optimal overall performance, with an AUC of 0.806 in the training set and 0.822 in the validation set.

Conclusions:

This study ultimately developed five prediction models for postoperative stress urinary incontinence (SUI) following prolapse surgery, which demonstrated good performance in internal validation. Among them, the SVM prediction model appears to be the most promising. However, further external validation data are required to assess its generalizability. This model has the potential to become a high-quality clinical risk prediction tool for postoperative SUI in prolapse patients, guiding clinic decisions on whether concurrent prolapse and incontinence surgeries are necessary. Clinical Trial: null