Nomogram analysis and external validation to predict the risk of lymph node metastasis in gastric cancer

Abstract

Aim: To identify risk factors for lymph node metastasis using a nomogram for gastric cancer patients to predict lymph node metastasis.

Results: The Chi-square test and the logistic regression showed that the Boarrmann type, preoperative CA199 level, T stage and N stage by CT scan were independent risk factors. The concordance index (C-index) was 0.786 in the internal validation of the Nomogram model. In the external validation, the C-index was 0.809, and the AUC was 0.894. The total accuracy of the prediction was 82.2%, and the false-negative rate was 5.4% with a cut-off value set at 0.109.

Materials and Methods: The study consisted of 451 patients with a histological diagnosis of gastric cancer with 0 or 1 lymph node metastasis from the Sun Yat-sen University Cancer Center as the development set, and the validation set consisted of 186 gastric cancer patients from the Sixth Affiliated Hospital of Sun Yat-Sen University. A Chi-square test and a logistic regression analysis were used to compare the clinicopathological variables and lymph node metastasis. The C-index and ROC curve were computed for comparisons of the nomogram’s predictive ability.

Conclusions: We developed and validated a nomogram to predict lymph node metastasis in gastric cancer before surgery. This nomogram can be broadly applied, even in general hospitals, and is useful for decisions regarding treatment programs for patients.