目的 比较评估3种营养免疫与系统炎症反应指标预后营养指数（PNI），格拉斯哥预后评分（GPS）及控制营养状态（CONUT） 评分在非小细胞肺癌（NSCLC）患者中的预后预测价值。方法 本研究以2014年9月至2018年5月于南通大学附属海安医院接受 根治性手术治疗的186例NSCLC患者为研究对象，根据PNI、GPS和COUNT进行分组。通过Kaplan⁃Meier生存曲线及受试者工作 特征（ROC）曲线分别评估术前PNI、GPS及CONUT评分在NSCLC患者中的预后意义。此外，通过Cox多因素回归分析明确影响 NSCLC患者预后的独立预测因素。结果 基于CONUT评分将NSCLC患者分为两组，结果发现CONUT≥2分组的总生存（OS）期显 著短于CONUT 0~1分组，3年OS率分别为70.6%与94.0%，差异有统计学意义（χ²=29.249；P<0.001)；类似地，以45.0作为PNI的临 界值，将患者分为PNI>45.0组与PNI≤ 45.0组，结果表明PNI≤ 45.0与NSCLC患者不良预后显著相关（3年OS率: 66.5%比 92.6%； χ²=28.686；P<0.001)。另外，Kaplan⁃Meier 生存曲线展示GPS 2分与GPS 0~1分者的3年OS率分别为61.7%和90.7%，差异有统计 学意义（χ²=18.499；P<0.001)。ROC曲线显示术前CONUT评分、PNI值与GPS评分的曲线下面积（AUC）值分别为0.753 (95%CI= 0.659~0.846)、0.734 (95%CI=0.629-0.839）及0.669 (95%CI=0.552~0.786)。多因素的Cox回归分析证实术前CONUT评分(HR=4.068； 95%CI=1.310~12.631; P=0.015), PNI (HR=4.043；95%CI=1.585~10.307；P=0.003）和TNM分期（HR=2.428；95%CI=1.153~5.111, P =0.020）是NSCLC患者预后的独立影响因素。结论 术前PNI与CONUT评分是NSCLC患者的独立预后因素，二者均可作为NSCLC 患者预后评估中一项简便、实用的血液学指标。

Objective To evaluate the prognostic value of three nutritional⁃immune and inflammation⁃related parameters, including prognostic nutritional index (PNI), Glasgow prognostic score (GPS) and controlling nutritional status (CONUT), for non⁃small cell lung cancer (NSCLC) patients. Methods A total of 186 NSCLC patients who were treated with curative resection in our hospital from September 2014 to May 2018 were included in this study. All patients were divided into study groups based on different nutritional⁃and immune⁃related parameters. The prognostic value of PNI, GPS and CONUT for NSCLC patients was determined by Kaplan⁃Meier survival curve and receiver operation characteristics (ROC) curve, respectively. Moreover, the Cox multivariate regression analysis was used to identify the independent prognostic factors for NSCLC patients. Results All NSCLC patients were divided into two groups based on the CONUT score, and the results showed that the overall survival (OS) of patients with CONUT≥2 was significantly shorter than that of patients with CONUT 0~1. The 3⁃year OS of two study groups was 70.6% and 94.0%, respectively (χ²=29.249, P<0.001), with a significantly statistical difference. Similarly, patients were divided into PNI>45.0 and PNI≤ 45.0 group based on the cut⁃off value of PNI. The results indicated that low PNI was significantly correlated to the poor OS of NSCLC patients (3⁃year OS: 66.5% vs 92.6%，χ²=28.686, P<0.001). Additionally, Kaplan⁃Meier curves showed that the 3⁃year OS of GPS 2 and GPS 0-1 group was 61.7% and 90.7%, respectively. There was a significantly statistical difference between the two groups (χ²=18.499, P<0.001). ROC curve showed that the area under curve (AUC) of CONUT, PNI and GPS was 0.753 (95%CI=0.659~0.846), 0.734 (95%CI=0.629~ 0.839) and 0.669 (95%CI=0.552~0.786), respectively. The results of multivariate Cox analysis demonstrated that CONUT (HR=4.068 and 95%CI=1.310~12.631, P=0.015), PNI (HR=4.043 and 95%CI=1.585~10.307, P=0.003) and TNM stage (HR=2.428, 95%CI=1.153~ 5.111, P=0.020) were independently risk factors to the prognosis of NSCLC patients. Conclusion Both PNI and CONUT were independent prognostic factors for NSCLC patients, and they could be used as simple and useful parameter for prognostic prediction of NSCLC patients.