Analyzing predictors of in-hospital mortality in patients with acute ST-segment elevation myocardial infarction using

Mengge Gong1, Dongjie Liang1, Diyun Xu1

  • 1Department of Cardiovascular Medicine, The Heart Center, The First Affiliated Hospital of Wenzhou Medical University, Wenzhou, 325000, Zhejiang, China.

PubMed

Insights

A new AGCOSCA-SVM model accurately predicts ST-segment elevation myocardial infarction (STEMI) using key patient data. This advanced framework offers improved diagnostic accuracy for this severe cardiac condition.

Area of Science:

  • Cardiology
  • Computational Intelligence
  • Machine Learning

Background:

  • Acute ST-segment elevation myocardial infarction (STEMI) is a critical condition caused by complete coronary artery blockage.
  • Early and accurate prediction of STEMI is vital for timely intervention and improved patient outcomes.

Purpose of the Study:

  • To develop and validate a novel predictive framework for STEMI using evolutionary computation and machine learning.
  • To enhance the prediction of STEMI by optimizing feature selection and classification.

Main Methods:

  • A hybrid AGCOSCA-SVM model was developed, integrating the AGCOSCA algorithm with Support Vector Machine (SVM).
  • The AGCOSCA algorithm, a variant of the Sine Cosine Algorithm, was enhanced with crossover and observation bee strategies.
  • The framework was trained and tested on a dataset of 3205 STEMI patients, with AGCOSCA refining feature selection for SVM.

Main Results:

  • The AGCOSCA-SVM model achieved high diagnostic performance, with Accuracy (97.83%), Sensitivity (93.75%), and Specificity (96.67%).
  • Key predictors identified include acute kidney injury (AKI) stage, fibrinogen, mean platelet volume (MPV), FT3, diuretics, and Killip class.
  • AGCOSCA-SVM outperformed traditional machine learning methods in predicting STEMI.

Conclusions:

  • The AGCOSCA-SVM framework demonstrates significant potential as a clinical decision support tool for STEMI diagnosis.
  • The model's ability to handle complex, nonlinear data relationships makes it suitable for smaller patient datasets.
  • This approach offers a promising avenue for improving the early detection and management of STEMI.