Nomogram for predicting overall survival after curative gastrectomy using inflammatory, nutritional and pathological factors

Purpose To establish a nomogram for predicting the overall survival (OS) in patients with gastric cancer (GC) based on inflammatory, nutritional and pathological factors. Methods GC patients underwent curative gastrectomy from January 2012 to June 2017 in our hospital were included, and were classified into training set and validation set with a ratio of 7:3. Then variables associated with OS were analyzed using univariate and multivariate Cox regression analysis. Nomograms predicting OS were built using variables from multivariable Cox models. Finally, Kaplan–Meier curve and Log-rank test were also conducted to analyze the 1-yr, 3-yr and 5-yr OS to validate the efficiency of risk stratification of the nomogram. Results A total of 366 GC patients were included. After univariate and multivariate Cox regression analysis, age (HR = 1.52, 95% CI = 1.01–2.30, P = 0.044), CA50 (HR = 1.90, 95% CI = 1.12–3.21, P = 0.017), PNI (HR = 1.65, 95% CI = 1.13–2.39, P = 0.009), SII (HR = 1.46, 95% CI = 1.03–2.08, P = 0.036), T stage (HR = 2.26, 95% CI = 1.01–5.05, P = 0.048; HR = 7.24, 95% CI = 3.64–14.40, P < 0.001) were independent influencing factors on the survival time of GC patients. Five factors including CEA, prognostic nutritional index (PNI), systemic immune-inflammation index (SII), ln (tumor size), T stage, and N stage were identified and entered the nomogram, which showed good discrimination and calibration in both sets. On internal validation, 1-yr, 3-yr and 5-yr nomogram demonstrated a good discrimination with an area under the ROC curve (AUC) of 0.77, 0.84 and 0.86, respectively. The AUC for 1-yr, 3-yr and 5-yr nomogram in validation set was 0.77, 0.79 and 0.81, respectively. The OS in low risk group of training cohort and validation cohort was significantly higher than that of intermediate risk group and high risk group, respectively. Conclusions We established a nomogram based on PNI, SII and pathological factors for predicting OS in GC patients. In addition, its efficiency was validated by validation set and stratified analysis.