European validation of the Yonsei Gastric Cancer Prognosis Prediction Model after gastrectomy: Validation with the Netherlands Cancer Registry

Abstract

BACKGROUND: Recently, a novel model predicting 5-year survival after gastrectomy was developed, the Yonsei Gastric Cancer Prognosis Prediction Model, to overcome limitations of the tumor-node-metastases (TNM) staging system. This study aimed to validate this model in a European cohort.

METHODS: All patients who underwent gastrectomy for gastric cancer were selected from the Netherlands Cancer Registry (2005-2015). Patients with 30-day mortality, co-existing cancer, neoadjuvant therapy, or missing data were excluded. The prediction model included gender, age, resection type, pT-stage, pM-stage, number of retrieved lymph nodes, number of metastatic lymph nodes, and tumor histology. The model was validated and compared to the 7th TNM staging system using calibration plots and the concordance index (c-statistic with 95% confidence interval (CI)).

RESULTS: From the 5748 patients who underwent gastrectomy, 2253 were included in this study. Mean age was 72.1 years, most patients had advanced gastric cancer (88%), and in 1102 patients (49%) no proper TNM staging could be performed since <16 lymph nodes were retrieved. Median overall survival was 24.6 months, and the 5-year overall survival was 30%, respectively. Model calibration was accurate in predicting 5-year overall survival, and the degree of discrimination was high (c-statistic = 0.807, 95% CI (0.787-0.826)). The model was superior to the TNM staging system in patients who could be properly staged: c-statistic = 0.861, 95% CI (0.838-0.885) vs. 0.711, 95% CI (0.692-0.729), p < 0.0001.

CONCLUSION: The Yonsei Gastric Cancer Prognosis Prediction Model was superior over the TNM staging system in predicting prognosis after gastrectomy in a European cohort, although it is not applicable to patients treated by neoadjuvant therapy.