Machine learning algorithms to predict heart failure with preserved ejection fraction among patients with premature myocardial infarction

Figures

• 

  Figure 1. Flow chart of the study. AMI, acute myocardial infarction; CA. G, coronary angiography; DCA, decision curve analysis; HF, heart failure; HFpEF, heart failure with preserved ejection fraction; PPCI, primary percutaneous coronary intervention; PMI, premature myocardial infarction; SHAP: shapley additive exPlanations; NRI, net reclassification improvement; IDI, integrated discriminant improvement.

• 

  Figure 2. (A) Lasso regression coefficients vs. λ values; (B) cross-validated λ and classification error plot relationships. Ranking of important variables based on Lasso-Logistic (C), XGBoost (D), RF (E), KNN (F), and SVM (G) models. BMI, body mass index; BNP, brain natriuretic peptide; CLR, C-reactive protein to lymphocyte ratio; cTnT, cardiac troponin T; Cr, creatinine; Fg, fibrinogen; Hb, hemoglobin; HbA1c, glycated hemoglobin; HCT, hematocrit; HDL-C, high-density lipoprotein cholesterol; HR, heart rate; MLR, monocyte to lymphocyte ratio; NEU%, neutrophil percentage; SBP, systolic blood pressure; TC, total cholesterol; UA, uric acid; WBC, white blood cell.

• 

  Figure 3. (A) ROC curve of five algorithms; (B) Radar plot of precision, accuracy, AUC, F1 score, and Brier score of the five models; Calibration curves of the XGBoost model in the training set (C) and testing set (D) DCA of the XGBoost model in the training set (E) and testing set (F).

• 

  Figure 4. (A) Forest plot with 10 important variables based on the XGBoost model. (B) Visual online prediction system based on the XGBoost model. The input interface is on the left, and the output interface is on the right. The upper part of the right represents the prediction probability, and the lower part of the right shows an individualized SHAP diagram. The red feature makes the model recognize the sample as class 1, while the blue feature makes the model identify the sample as class 0. BMI, body mass index; BNP, brain natriuretic peptide; CLR, C-reactive protein to lymphocyte ratio; Fg, fibrinogen; HCT, hematocrit; HF, heart failure; HFpEF, heart failure with preserved ejection fraction; HR, heart rate; MLR, monocyte to lymphocyte ratio.