目的:探讨白蛋白与球蛋白比值（AGR）联合预后营养指数（PNI）对可切除胃癌（rGC）患者预后的预测价值。方法:选取2015年1月至2017年12月在南通大学附属如皋医院行胃癌根治术治疗的rGC患者158例，随访截至2020年5月或患者死亡。术前计算AGR、PNI，绘制受试者工作特征（ROC）曲线确定AGR、PNI最佳预测界值，并进行分组。分析AGR联合PNI（AGR-PNI）与临床病理特征的关系，Kaplan-Meier法计算累积总生存率，并进行Log-rank检验和单因素预后分析，Cox比例风险回归模型分析预后影响因素。结果:ROC曲线分析表明AGR、PNI最佳预测界值分别为1.19、43.70，AGR-PNI判断患者预后的ROC曲线下面积（AUC）大于AGR、PNI（n Z = 2.596，n P = 0.009；n Z = 2.403，n P = 0.016）。基于AGR、PNI的最佳预测界值，对患者进行AGR-PNI评分并分组：0分组（AGR ≥ 1.19、PNI ≥ 43.70，75例），1分组（AGR ≥ 1.19、PNI < 43.70，或AGR < 1.19、PNI ≥ 43.70，55例），2分组（AGR < 1.19、PNI < 43.70，28例）。三组年龄、性别构成、肿瘤直径、TNM分期、组织分化程度比较差异有统计学意义（ n P < 0.05）。多因素分析结果显示，年龄（≥ 60岁比< 60岁， n HR = 1.878，95% n CI 1.011~3.491，n P = 0.046）、TNM分期（Ⅲ期比Ⅰ期，n HR = 2.148，95% n CI 1.074~4.296，n P = 0.031）、组织分化程度（中等或好比差，n HR = 0.399，95% n CI 0.211~0.753，n P = 0.005）、AGR-PNI（2分组比0分组，n HR = 2.729，95% n CI 1.303~5.715，n P = 0.008）是影响rGC患者生存的独立危险因素。n 结论:AGR-PNI可作为rGC患者预后的有效预测指标，AGR-PNI高分组提示患者预后不良。“,”Objective:To investigate the prognostic value of albumin to globulin ratio (AGR) combined with prognostic nutrition index (PNI) in patients with resectable gastric cancer (rGC).Methods:From January 2015 to December 2017, 158 patients with rGC who underwent radical gastrectomy in Rugao Hospital were included, and followed up until May 2020 or died. The preoperative AGR and PNI were calculated. The cut-off value of AGR and PNI were determined by receiver operating characteristic (ROC) curve. The relationship between AGR-PNI and clinicopathological indicators were analyzed. The Kaplan-Meier method was used to calculate the cumulative overall survival rate. Log-rank test and univariate prognostic analysis were used. The Cox proportional hazards regression model was used for multivariate prognostic analysis.Results:The cut-off value of ROC curve of AGR and PNI were 1.19 and 43.70 respectively. The area under ROC curve (AUC) of AGR-PNI was larger than that of AGR and PNI (n Z = 2.596, n P = 0.009; n Z = 2.403, n P = 0.016). Patients were divided into three groups: group 0 (AGR ≥ 1.19, PNI ≥ 43.70; 75 cases), group 1 (AGR ≥ 1.19, PNI < 43.70, or AGR < 1.19, PNI ≥ 43.70; 55 cases), group 2 (AGR <1.19, PNI < 43.70; 28 cases). The age, sex, tumor diameter, TNM stage and degree of tissue differentiation were significantly different among three groups ( n P < 0.05). The results of multivariate analysis showed that age (≥ 60 years old vs. < 60 years old, n HR = 1.878, 95% n CI1.011-3.491, n P = 0.046), TNM stage (Ⅲ stage vs. Ⅰ stage, n HR = 2.148, 95% n CI 1.074-4.296, n P = 0.031), degree of tissue differentiation (moderate or good vs. poor, n HR = 0.399, 95% n CI 0.211-0.753, n P = 0.005), AGR-PNI (group 2 vs. group 0, n HR = 2.729, 95% n CI 1.303-5.715, n P = 0.008) were independent risk factors for survival of patients with rGC.n Conclusions:AGR-PNI can be used as an effective predictor of the prognosis of rGC patients, and high grouping indicates poor prognosis.