Machine Learning Models Improve Survival Predictions for Prostate Cancer Patients

Strategic Insights -

Research from Westlake University has led to the development of machine learning models that significantly improve survival predictions for patients with prostate cancer bone metastases (PCBM). The study, published in Scientific Reports, found that the models achieved a high accuracy in predicting survival outcomes, with area under the receiver operating characteristic curve (AUC) values of 0.76, 0.83, and 0.91 for 1-year, 3-year, and 5-year survival predictions, respectively. The models also identified key prognostic factors, including T stage, grade, age, PSA, and Gleason score.

Key Takeaways:

  • The research aimed to establish machine learning models to improve survival predictions for PCBM patients, which is a highly lethal condition with limited survival.
  • The study extracted data for PCBM patients from the SEER database spanning 2010 to 2019 and conducted univariate and multivariate Cox regression analyses to identify prognostic features.
  • The XGBoost models achieved robust performance in predicting survival for PCBM patients, with AUC values of 0.76, 0.83, and 0.91 for 1-year, 3-year, and 5-year survival predictions, respectively.
  • Key prognostic factors included T stage, grade, age, PSA, and Gleason score.
  • Single patients exhibited a significantly higher mortality risk than their married counterparts, with a hazard ratio (HR) of 1.23 (95% CI 1.19-1.27, p < 0.001).
  • The model's high accuracy and interpretability provide valuable support for developing personalized treatment plans for PCBM patients.

Statistics:

  • AUC values of 0.76, 0.83, and 0.91 for 1-year, 3-year, and 5-year survival predictions, respectively.
  • Hazard ratio (HR) of 1.23 (95% CI 1.19-1.27, p < 0.001) for single patients compared to married patients.
  • 2010-2019 SEER database data used for analysis.

Sources:

  • Interpretable machine learning models for survival prediction in prostate cancer bone metastases. Scientific Reports, 2025;15(1):24150.
  • NewsRx. Westlake University Reports Findings in Personalized Medicine (Interpretable machine learning models for survival prediction in prostate cancer bone metastases). Journal of Engineering. July 21, 2025; p 3808.