Prediction of anastomotic leakage after esophagectomy for esophageal cancer: a nomogram study integrating systemic inflammation indices and clinical factors

Figures

• 

  Figure 1. Flowchart of patient selection. A total of 685 patients undergoing esophagectomy were screened. After exclusions for prior malignancy, incomplete data, and emergency surgery, 650 eligible patients were included and randomly allocated to the training (n=455) and validation (n=195) cohorts.

• 

  Figure 2. Restricted cubic spline analyses. RCS analyses illustrating the associations between age, CALLY, hypertension, neoadjuvant radiotherapy, NLR, NMR, PLR, and the risk of AL. Solid lines represent adjusted estimates, and shaded areas indicate 95% confidence intervals. Dashed horizontal lines denote the reference level (zero effect). Red dots indicate reference values. Nonlinear relationships were observed for age, inflammatory indices, and nutritional markers, while binary variables showed discrete risk differences.

• 

  Figure 3. Nomogram for predicting the risk of postoperative AL. Nomogram for predicting the probability of AL based on age, CALLY, NLR, NMR, PLR, hypertension, and neoadjuvant radiotherapy. Each variable corresponds to a weighted score, and the total score is used to estimate individual risk. Asterisks indicate levels of statistical significance.

• 

  Figure 4. Discrimination of the nomogram model. ROC curves evaluating the discriminative performance of the predictive model in the training (A) and validation (B) cohorts. The model achieved an AUC of 0.813 (95% CI, 0.746–0.880) in the training group and 0.763 (95% CI, 0.683–0.842) in the validation group. The diagonal dashed line represents the line of no discrimination (AUC = 0.5).

• 

  Figure 5. Validation of the nomogram model in the validation cohort. Calibration plots assessing agreement between predicted and observed probabilities in the training (A) and validation (B) cohorts. The apparent and bias-corrected curves are shown alongside the ideal reference line.

• 

  Figure 6. DCA of the nomogram model. DCA plot illustrates the clinical utility of the nomogram in predicting postoperative AL for the training (A) and validation cohorts (B). The x-axis denotes the threshold probability, while the y-axis indicates the standardized net benefit. The red curve corresponds to the nomogram model, the gray line represents the scenario in which all patients are assumed to experience the event, and the black line signifies the scenario in which none of the patients are assumed to experience the event.