Prognostic value of the atherogenic index of plasma combined with lactate-to-albumin ratio for in-hospital mortality in patients with acute myocardial infarction complicated by cardiogenic shock

Figures

• 

  Figure 1. LASSO coefficient profiles of the 18 potential predictors. Each curve represents one candidate variable, and the x-axis shows log(lambda), where larger lambda values impose stronger shrinkage. As lambda increases, coefficients of less informative variables shrink toward zero. Variables with non-zero coefficients at the selected lambda were retained as candidate predictors for subsequent multivariable modeling.

• 

  Figure 2. Selection of the optimal tuning parameter (lambda) in the LASSO model. The plot displays binomial deviance from 10-fold cross-validation versus log(lambda). A dotted vertical line indicates lambda.min, which gives the minimum cross-validated deviance and corresponds to six retained predictors. The selected lambda was used for candidate-variable selection.

• 

  Figure 3. Importance of selected variables. Bar chart showing the standardized coefficients of the variables retained by LASSO regression. Longer bars indicate greater relative contribution to the penalized model, while positive values indicate variables associated with higher mortality risk.

• 

  Figure 4. Receiver operating characteristic (ROC) curves. ROC analysis comparing the discriminatory ability of the key predictors included in the final nomogram (Mechanical Ventilation, AIP, LAR, and LVEF <40%) and the Combined Prediction Model for in-hospital mortality. The Combined Model (AUC = 0.909) showed superior internal discrimination compared with any single predictor, but this finding requires external validation.

• 

  Figure 5. Nomogram for predicting in-hospital mortality in AMICS patients. To use the nomogram, locate the patient's value for each variable on the corresponding axis. Draw a vertical line upward to the “Points” axis to determine the score for that variable. Sum the scores for all variables to get the “Total Points”. Locate the total points on the bottom scale to determine the predicted probability of in-hospital mortality.

• 

  Figure 6. Calibration plot of the nomogram. The x-axis represents the predicted probability of in-hospital mortality, and the y-axis represents the observed probability. The diagonal dotted line represents ideal calibration. The solid line represents apparent calibration across deciles of predicted risk, and the shaded area indicates an approximate 95% interval. Calibration results should be interpreted cautiously because confidence intervals for calibration metrics were relatively wide and no external validation cohort was available.

• 

  Figure 7. Decision curve analysis for the combined prediction model. The x-axis indicates the threshold probability for predicting in-hospital mortality, and the y-axis indicates net benefit. The combined model is compared with the treat-all and treat-none strategies. The analysis suggests apparent clinical net benefit across the displayed threshold range, but external validation is required before clinical implementation.