Development and validation of a nomogram to predict cardiac death after radiotherapy for esophageal cancer

Figures

• 

  Figure 1. Graphical summary. Esophageal cancer (EC) radiotherapy increased cardiovascular toxicity and mortality in patients. A nomogram was developed and validated to accurately predict cardiac death in EC patients after radiotherapy. Risk factors such as age, year of diagnosis, surgery, sequence of surgery and radiotherapy, chemotherapy, and number of tumors are included in the nomogram.

• 

  Figure 2. Flowchart of patient selection.

• 

  Figure 3. Radiotherapy‐induced cardiac mortality in patients with esophageal cancer: Least absolute shrinkage and selection operator (LASSO) regression analysis. (a) For 16 candidate variables, minimum absolute shrinkage and selection coefficients are plotted, in which six variables with non‐zero coefficients are selected as the best predictors. (b) Ten‐fold cross‐validation via minimum criteria was used in the tuning parameter (λ) selection of the LASSO model.

• 

  Figure 4. Nomograms for predicting 5‐ and 10‐year survival rates. Year: 1, 2000–2005; 2, 2006–2010; 3, 2011–2015; 4, 2016–2018. Sequence: 1, radiation before surgery; 2, intraoperative radiation; 3, radiation after surgery; 4, unknown.

• 

  Figure 5. Prognostic nomogram calibration graph for each group. The concordance index (C‐index) of the development group is 0.708, and that of the validation group is 0.679, suggesting that the model has high accuracy. (a) Five‐ and (b) 10‐year calibration plots for the development group. (c) Five‐ and (d) 10‐year calibration plots for the validation group.

• 

  Figure 6. Cardiac death risk in patients receiving radiation therapy for esophageal cancer: Prognostic Decision Curve Analysis (DCA). DCA of the model's clinical value in prediction of (a) 5‐ and (b) 10‐year survival rate for the development group. DCA of the model's clinical value in prediction of (c) 5‐ and (d) 10‐year survival rate for the validation group.

• 

  Figure 7. Cardiac death risk in esophageal cancer patients receiving radiation therapy depicted by receiver operating characteristic curves (ROC). (a) 5‐ and (b) 10‐year ROC curves for the development group. (c) 5‐ and (d) 10‐year ROC curves for the validation group. FPR, false positive rate; TPR, true positive rate.

• 

  Figure 8. The probability of death estimated by Kaplan–Meier survival analysis. (a) 5‐ and (b) 10‐year survival curves with the cut‐off points of the development group. (c) 5‐ and (d) 10‐year survival curves with the cut‐off points of the validation group.