Prediction of cervical anastomosis failure after McKeown esophagectomy

Objective. To identify the risk factors of cervical anastomosis failure after McKeown esophagectomy and to create a prognostic model predicting the risk of anastomosis failure. Material and methods. There were 69 esophagectomies between January 1, 2019 and December 31, 2023 for cancer of the middle thoracic esophagus. Significant predictors of cervical anastomosis failure were identified. Results. The incidence of anastomosis failure after McKeown surgery was 18.8%. Logistic regression analysis showed that platelet-lymphocyte ratio (OR=1.141; 95% CI: 1.003—1.297), prognostic nutritional index (OR=0.576; 95% CI: 0.339—0.979), and systemic immune-inflammation index (OR=1.010; 95% CI: 1.004—1.016) were independent risk factors of anastomosis failure after McKeown esophagectomy. AUC was 0.953 ± 0.042 (95% CI 0.872—1.000). The model was significant (p<0.001). Sensitivity and specificity of the model were 84.6% and 87.5%, respectively. Furthermore, we developed a nomogram for predicting cervical anastomosis failure after McKeown surgery using artificial intelligence and machine learning. A comprehensive analysis and validation of prognostic model demonstrated its significance. Conclusions. PNI, NLR and SII are significant predictive factors of cervical anastomotic leakage after McKeown esophagectomy. A nomogram was developed to predict the probability of anastomotic leakage.