Development and validation of a nomogram for predicting bleeding risk in patients with pulmonary embolism

Figures

• 

  Figure 1. Flowchart of study cohort and prediction model development.

• 

  Figure 2. Variable selection was performed using LASSO and logistic regression. (A) Coefficient profile plots were generated against the log(lambda) sequence to visualize the variable selection process and identify nonzero coefficient variables based on the optimal lambda value. (B) Dotted vertical lines represent optimal values determined using the 1 standard error of the minimum criteria (lambda.1se). (C) Forest plot displaying final predictors in the bleeding risk model with adjusted odds ratios from multivariable logistic regression. CI, cerebral infarction; RBC, red blood cell count; RI, renal insufficiency; SP, systolic pressure; CRE, creatinine; BH, bleeding history.

• 

  Figure 3. Nomogram for bleeding risk prediction in pulmonary embolism patients. The tool converts six clinical parameters into points: cerebral infarction (Cl), red blood cell count (RBC), renal insufficiency (RI), systolic blood pressure (SP), creatinine (CRE), and bleeding history (BH). Summed points (total points axis) correspond to predicted bleeding probability (bottom axis). Example: A patient with prior bleeding (bleeding history = yes, 37.5 points), cerebral infarction (CI = yes, 16 points), renal insufficiency (RI = yes, 18 points), RBC 2.5 × 1012/L (73 points), systolic pressure 160 mmHg (37 points), and creatinine 500 μmol/L (18 points) would have 199.5 points, corresponding to 44% bleeding risk.

• 

  Figure 4. Model performance metrics in development and validation cohorts. Four-panel evaluation of the bleeding risk prediction model. (A) Development ROC: AUC = 0.756 (95% CI: 0.729–0.784). (B) Validation ROC: AUC = 0.729 (95% CI: 0.685–0.773). (C) Development calibration: Ideal fit (slope = 1.000). (D) Validation calibration: Good agreement (slope = 0.810).

• 

  Figure 5. Clinical utility and impact analysis. (A) Development DCA: Net benefit of model-guided decisions versus “treat-all” and “treat-none” strategies across threshold probabilities. (B) Validation DCA: Replication of net benefit superiority in independent cohort. (C) Development CIC: Proportion classified high-risk versus actual bleeding events captured. (D) Validation CIC: Replication of clinical impact in independent cohort. In the DCAs, the y-axis represents the net benefit. The horizontal lines labeled “None” represent the assumption that no participant experienced bleeding. The lines labeled “All” represent the assumption that all participants had bleeding. The lines labeled “nomogram model” represent the predictive model developed in this study. In CICs, the red curve represents the number of individuals classified as positive (high risk) by the model at each threshold probability, indicating the number of high-risk individuals. The blue curve represents the number of true positives (individuals with the outcome) at each threshold probability.

• 

  Figure 6. ROC curve comparisons. (A) Training cohort. (B) Testing cohort. Nomogram has the maximum AUC. CI, cerebral infarction; RBC, red blood cell count; RI, renal insufficiency; SP, systolic pressure; CRE, creatinine; BH, bleeding history; Nomo, nomogram.