Risk Assessment for Upper Urinary Tract Deterioration in Patients with Neurogenic Bladder: A Retrospective Study...

Area of Science:

• Urology
• Medical Informatics
• Machine Learning

Background:

• Upper urinary tract deterioration (UUTD) is a significant complication for patients with neurogenic bladder (NB).
• Accurate risk assessment is crucial for timely intervention and management of UUTD in NB patients.

Purpose of the Study:

• To develop a practical, machine learning-based tool for assessing the risk of UUTD in patients with NB.
• To identify key clinical and urodynamic predictors of UUTD in NB patients.

Main Methods:

• Utilized clinical and urodynamic data from 2393 NB patients for model development and validation.
• Employed Boruta, Lasso regression, and decision tree algorithms to identify consistent risk factors.
• Developed and validated a Gaussian Naive Bayes (GNB) machine learning model.

Main Results:

• Identified six key risk factors for UUTD, including bladder management method and detrusor sphincter dyssynergia (DSD).
• The GNB model demonstrated strong performance with an AUC of 0.878 in the external test set.
• The model showed consistent performance across various subgroups and was presented as a web-based tool.

Conclusions:

• A machine learning model effectively assesses UUTD risk in NB patients.
• Bladder management method and urodynamic evaluation are critical for UUTD risk assessment.
• The developed tool offers practical, individualized risk assessment for NB patients, though external validation is recommended.