A Novel Inflammatory-Nutritional Prognostic Scoring System for Stage III Gastric Cancer Patients With Radical Gastrectomy Followed by Adjuvant Chemotherapy

PURPOSE: The present study was designed to explore the prognostic value of preoperative inflammatory and nutritional biomarkers in stage III gastric cancer (GC) patients with adjuvant chemotherapy and to develop a novel scoring system called the inflammatory-nutritional prognostic score (INPS). METHODS: A total of 513 patients with pathological stage III GC undergoing radical gastrectomy followed by adjuvant chemotherapy from 2010 to 2017 were enrolled in the study. Clinicopathological characteristics and blood test parameters of individual patients were collected. The least absolute shrinkage and selection operator (LASSO) Cox regression model was used for feature selection to construct INPS. Survival curves were generated using the Kaplan-Meier method with log-rank tests. The nomogram was generated based on the result of the multivariate analysis using Cox’s proportional hazards model. The model was assessed by the concordance index (C-index) and was internally validated by bootstraps. RESULTS: According to the results of Lasso Cox regression and K-M survival curves, INPS was determined as follows: a low body mass index (BMI) (<23 kg/m(2)), a low prealbumin (<180 mg/L), a high neutrophil-lymphocyte ratio (NLR) (≥2.7), a high platelet-lymphocyte ratio (PLR) (≥209.4), a low lymphocyte-monocyte ratio (LMR) (<2.8), and a low prognostic nutritional index (PNI) (<45.1); each were scored as 1, and the remaining values were scored as 0. The individual scores were then summed up to construct the INPS and further divided into 4 groups: Low Risk (INPS 0); Low-medium Risk (INPS 1); High-medium Risk (INPS 2-4); and High Risk (INPS 5-6). In multivariate analysis, INPS was an independent predictor of overall survival (OS) in stage III GC, with the 5-year OS rates of 70.8%, 57.4%, 41.5%, and 30.6%, respectively. The nomogram based on INPS and other independent predictors (gender, pT stage, pN stage, lymphovascular invasion, and CEA level) showed good predicting performance with a C-index of 0.707, which was superior to the TNM stage alone (C-index 0.645, p=0.008) and was internally validated with the corrected C-index of 0.693. CONCLUSION: Preoperative INPS was an independent prognostic factor of stage III GC patients with radical surgery followed by adjuvant chemotherapy. The nomogram based on INPS may serve as a simple and potential model in risk stratification and guiding treatment strategies in clinical practice.