Machine Learning-Based Prediction of Long-Term Mortality in STEMI Patients Using Clinical, Laboratory, and...

Area of Science:

• Cardiology
• Artificial Intelligence
• Medical Informatics

Background:

• ST-segment elevation myocardial infarction (STEMI) poses a significant long-term mortality risk.
• Primary percutaneous coronary intervention (pPCI) is a standard treatment for STEMI.
• Accurate prediction of mortality risk is crucial for patient management and treatment optimization.

Purpose of the Study:

• To compare the performance of various machine learning (ML) models in predicting long-term mortality in STEMI patients post-pPCI.
• To evaluate the prognostic value of novel inflammatory-metabolic indices in this patient cohort.
• To identify key predictors of mortality for improved risk stratification.

Main Methods:

• Retrospective analysis of 329 STEMI patients who underwent pPCI.
• Development and comparison of five ML algorithms: Logistic Regression (LR), Random Forest (RF), Extreme Gradient Boosting (XGBoost), Support Vector Machines (SVM), and Artificial Neural Networks (ANN).
• Performance evaluation using accuracy, sensitivity, specificity, and ROC-AUC; SHAP analysis for model interpretation.

Main Results:

• The XGBoost model demonstrated superior performance with 98.99% accuracy, 0.999 ROC-AUC, and 100% sensitivity.
• Higher door-to-balloon time (DTBT), Systemic Inflammatory Response Index (SIRI), and pan-immune-inflammation value (PIV) were associated with mortality.
• Lower body mass index (BMI), Prognostic Nutritional Index (PNI), and Advanced Lung Cancer Inflammation Index (ALI) were observed in the mortality group.
• SHAP analysis identified DTBT, albumin, and ALI as the strongest mortality predictors.

Conclusions:

• The XGBoost algorithm is a highly accurate and reliable tool for predicting long-term mortality in STEMI patients.
• Integrating novel inflammatory-metabolic indices, such as ALI and TyG, alongside DTBT into ML models can enhance early risk identification and stratification.
• These findings suggest a potential for improved clinical decision-making and patient outcomes through advanced predictive modeling.