Development and validation of a clinical cure marker based on negative lymph nodes for gastric cancer after gastrectomy

Abstract

Objective: To explore lymph node (LN)-related derived indicators as clinical cure markers for gastric cancer (GC) after gastrectomy. Methods: Data of resected GC patients were extracted from the SEER database and our own department. Propensity score matching (PSM) was used to balance the baseline differences between the clinical cure and the nonclinical cure groups. The area under the curve (AUC) and decision curve analysis (DCA) were used to choose the optimal marker, and survival analysis was used to validate the clinical value of the most effective marker. Results: After PSM, the differences in age, sex, race, location, surgical type, and histologic type between the two groups were significantly reduced (all P > 0.05), and the AUCs of examined LNs (ELNs), negative LNs (NLNs), ESR (ELNs/tumor size), ETR (ELNs/T-stage), NSR (NLNs/tumor size), NTR (NLNs/T-stage), EPR (ELNs/PLNs) and NPR (NLNs/PLNs) were 0.522, 0.625, 0.622, 0.692, 0.706, 0.751, 7.43, and 7.50, respectively. When NTR was 5.9, the Youden index of 0.378 was the highest. The sensitivity and specificity were 67.5% and 70.3% in the training group and 66.79% and 67.8% in the validation group, respectively. DCA showed that NTR had the largest net clinical benefit, and patients with NTR greater than 5.9 had significantly prolonged overall survival in our own cohort. Conclusion: NLNs, NTR, NSR, ESR, ETR, NPR and EPR can be used as clinical cure markers. However, NTR was the most effective, and the best cutoff value was 5.9.