Related Experiment Video
Updated: Sep 21, 2025
06:19
Constructing and Visualizing Models using Mime-based Machine-learning Framework
Published on: July 22, 2025
820
Development and Validation of Dynamic Multivariate Prediction Models of Sexual Function Recovery in Patients with
Nnenaya Agochukwu-Mmonu1,2, Adharsh Murali3, Daniela Wittmann4
1Department of Urology, New York University, New York, NY, USA.
European Urology Open Science
|May 31, 2022
Summary
This study developed a predictive model for sexual function recovery after radical prostatectomy (RP), aiding patient decision-making and potentially reducing regret. The model accurately forecasts erectile function outcomes at 12 and 24 months post-surgery.
Area of Science:
- Urology
- Oncology
- Medical Informatics
Background:
- Radical prostatectomy (RP) is a primary treatment for intermediate-risk prostate cancer.
- Erectile dysfunction (ED) is a common complication following RP.
Purpose of the Study:
- Develop and validate predictive models for sexual function recovery at 12 and 24 months post-RP.
- Provide tools for informed decision-making and patient counseling.
Main Methods:
- Utilized Michigan Urological Surgery Improvement Collaborative (MUSIC) registry data (2016-2021).
- Developed dynamic, multivariate random-forest models incorporating baseline characteristics and sexual function assessments.
- Evaluated prediction accuracy using metrics like RMSE, MAE, and AUC.
Main Results:
- The model accurately predicted 12-month sexual domain scores (AUC=0.82) and 24-month scores (AUC=0.81) using baseline data.
- Incorporating post-RP data significantly improved prediction accuracy (AUC=0.91 at 12 months, AUC=0.94 at 24 months).
- A web-based prediction tool is available.
Conclusions:
- The developed model offers a valid method for predicting sexual function recovery post-RP.
- Dynamic, multivariate predictions can aid in surgical decision-making and survivorship care.
- Improved understanding of potential outcomes may reduce patient decision regret.
