Mortality predicting models for patients with infective endocarditis: a machine learning approach.

Area of Science:

• Cardiovascular Medicine
• Medical Informatics
• Machine Learning in Healthcare

Background:

• Infective endocarditis (IE) is a severe cardiovascular disease with high mortality and varied presentations.
• Existing risk models for IE have limitations in predictive accuracy and clinical applicability.
• There is a need for improved predictive systems for better risk stratification in IE patients.

Purpose of the Study:

• To develop and evaluate machine learning models for predicting prognosis in patients with infective endocarditis.
• To identify key clinical variables that predict in-hospital and 6-month mortality in IE patients.

Main Methods:

• A retrospective observational study involving 1705 IE patients.
• Four machine learning algorithms were employed: LASSO logistic regression, random forest (RF), support vector machine (SVM), and k-nearest neighbors (KNN).
• Model performance was assessed using 10-fold cross-validated area under the receiver operating characteristic curve (AUC-ROC).

Main Results:

• Random forest (RF) demonstrated superior performance in predicting both in-hospital (AUC-ROC: 0.83) and 6-month mortality (AUC-ROC: 0.85).
• Important predictors identified by RF for mortality included total bilirubin, N-terminal pro-B-type natriuretic peptide, albumin, diastolic blood pressure, fasting blood glucose, uric acid, and age.
• The study enrolled 1705 patients, with 119 in-hospital deaths and 178 deaths within 6 months post-discharge.

Conclusions:

• A machine learning-based risk model can effectively predict prognosis in infective endocarditis patients.
• This approach facilitates rapid risk stratification and timely clinical management.
• The identified key predictors offer insights into factors influencing IE patient outcomes.