Related Experiment Videos
Development and temporal validation of an interpretable XGBoost model for predicting cardiovascular death in patients
Xiaoling Wan1, Ting Shi1, Qiao Liu1
1Department of Emergency, Xianning Central Hospital, The First Affiliated Hospital of Hubei University of Science and Technology, Xianning, Hubei, China.
Insights
A new machine learning model accurately predicts cardiovascular death risk in diabetic foot patients. This tool aids early identification and personalized risk assessment for better patient outcomes.
Area of Science:
- Cardiology
- Diabetology
- Medical Informatics
Background:
- Diabetic foot (DF) patients face elevated cardiovascular (CV) mortality risk.
- Existing risk prediction tools do not adequately address this high-risk population.
Purpose of the Study:
- To develop and validate an interpretable machine learning (ML) model for predicting CV death in patients with DF.
- To identify key predictors of CV death in this cohort.
Main Methods:
- A retrospective cohort study of 2,835 DF patients.
- Development and temporal validation of nine supervised ML algorithms, with Extreme Gradient Boosting (XGBoost) selected for optimal performance.
- SHapley Additive exPlanations (SHAP) used for model interpretability.
Main Results:
- The optimal XGBoost model achieved high predictive accuracy, with Area Under the Curve (AUC) values of 0.829 (internal validation), 0.844 (test set), and 0.828 (temporal validation).
- The model demonstrated good calibration and clinical utility via decision curve analysis.
- Key predictors identified include age, serum creatinine, glycated hemoglobin, triglycerides, and BMI.
Conclusions:
- An interpretable XGBoost model effectively predicts CV death in DF patients.
- This model can aid in early risk stratification and personalized management.
- Prospective multicenter validation is recommended prior to widespread clinical adoption.
Abstract:
Patients with diabetic foot (DF) have a high risk of cardiovascular (CV) death, yet dedicated risk-prediction tools for this population are lacking. We developed and temporally validated an interpretable machine learning (ML) model for predicting CV death in patients with DF. This single-center retrospective cohort study included 2,835 patients admitted between February 2017 and May 2025. The development cohort comprised 2,325 patients, including 748 CV deaths, and was divided into training, internal validation, and held-out test sets; an independent temporal validation cohort included 510 patients, including 220 CV deaths. Nine supervised ML algorithms were compared using the area under the receiver operating characteristic curve (AUC). Extreme gradient boosting (XGBoost) showed the best overall performance. The optimal model, incorporating demographic and diabetes-related characteristics, routine laboratory parameters, and DF-specific features, achieved AUCs of 0.829 (95% confidence interval [CI]: 0.752-0.905) in internal validation, 0.844 (95% CI: 0.806-0.881) in the held-out test set, and 0.828 (95% CI: 0.789-0.868) in temporal validation. The model demonstrated good calibration and favorable net benefit on decision curve analysis. SHapley Additive exPlanations (SHAP) identified age, serum creatinine, glycated hemoglobin, triglycerides, and body mass index as the most influential predictors of increased model-predicted risk. This interpretable XGBoost model may support early identification and individualized risk stratification of patients with DF at high risk of CV death; however, prospective multicenter validation is required before clinical implementation.
