Machine Learning Tool Predicts Survival in Peripheral Artery Disease Patients

Strategic Insights -

Researchers at the University of Genoa in Italy have developed a machine learning tool that can predict survival in peripheral artery disease (PAD) patients who undergo surgical treatment. The study used data from 1,615 patients who underwent surgery between 2005 and 2020 and found that the tool can accurately predict mortality at one, three, and five years after the first surgery. The tool, based on gradient boosted decision trees (GBDTs), identified disease stage, age, chronic kidney disease status, hospital length-of-stay, and total number of comorbidities as key predictors of mortality.

Key Takeaways:

  • The machine learning tool developed by the University of Genoa uses GBDTs to predict mortality in PAD patients.
  • The tool was trained on data from 1,615 patients who underwent surgery between 2005 and 2020.
  • Disease stage was the most important predictor of mortality, followed by age, chronic kidney disease status, hospital length-of-stay, and total number of comorbidities.
  • The tool was able to predict mortality at one, three, and five years after the first surgery with high accuracy (AUC of 0.86, 0.84, and 0.80, respectively).
  • Presence of dyslipidemia was slightly predictive of one- and three-year mortality.
  • Simple clinical and demographic parameters can be used to train the GBDT model.
  • The study was funded by the University degli Studi di Genova within the CRUI-CARE Agreement.
  • The research has been peer-reviewed and published in the Journal of Cardiovascular Translational Research.

Statistics:

  • 1,615 patients were used to develop and validate the machine learning tool.
  • The tool predicted mortality at one, three, and five years after the first surgery with an AUC of 0.86, 0.84, and 0.80, respectively.
  • Disease stage was the most important predictor of mortality (SHAP values).
  • Age, chronic kidney disease status, hospital length-of-stay, and total number of comorbidities were also important predictors of mortality.
  • Presence of dyslipidemia was slightly predictive of one- and three-year mortality.

Sources:

  • "A Machine Learning Tool To Predict Survival After First Surgery In Peripheral Artery Disease Patients." Journal of Cardiovascular Translational Research, 2025. Journal of Cardiovascular Translational Research can be contacted at: Springer, One New York Plaza, Suite 4600, New York, Ny, United States. (Springer - www.springer.com; Journal of Cardiovascular Translational Research - www.springerlink.com/content/1937-5387/)
  • NewsRx. University of Genoa Reports Findings in Peripheral Artery Disease (A Machine Learning Tool To Predict Survival After First Surgery In Peripheral Artery Disease Patients). Hematology Week. November 3, 2025; p 1757.