作者：黄维佳1，曹健斌1，李逢昌1，李竞长2，张东伟3，何凌云1，李富骊1，伍义文1

单位：

1.545006广西壮族自治区柳州市人民医院胸外科　2.545006广西壮族自治区柳州市人民医院肿瘤科3.545006广西壮族自治区柳州市人民医院呼吸科

Units：

1.Department of Thoracic Surgery, Liuzhou People's Hospital, Liuzhou 545006, China2.Department of Oncology, Liuzhou People's Hospital, Liuzhou 545006, China3.Department of Respiratory, Liuzhou People's Hospital, Liuzhou 545006, China

关键词：

癌，非小细胞肺；表皮生长因子受体；酪氨酸激酶抑制剂；外周血指标；预后；生存分析；列线图；预测

Keywords：

Carcinoma, non-small-cell lung; Epidermal growth factor receptor; Tyrosine kinase inhibitors; Peripheralindicators; Prognosis; Survival analysis; Nomogram; Forecasting

CLC：

R 730.26

DOI：

10.12114/j.issn.1008-5971.2022.00.211

Funds：

广西壮族自治区中医药管理局自筹经费科研课题（GXZYZ20210120）

摘要：

目的　构建接受EGFR-TKIs治疗的EGFR突变阳性非小细胞肺癌（NSCLC）患者预后预测列线图模型，并验证其预测价值。方法　选取2018—2020年柳州市人民医院胸外科、呼吸科、肿瘤科收治的EGFR突变阳性NSCLC患者332例，按照7∶3的比例随机分为试验组（n=232）和验证组（n=100）。收集患者的临床资料，包括性别、年龄及EGFR-TKIs治疗前2周血常规检查结果。从患者接受EGFR-TKIs一线治疗开始对其进行随访，随访内容为后续治疗效果、是否复发或死亡以及死亡时间。采用ROC曲线分析确定血常规指标预测试验组接受EGFR-TKIs治疗的EGFR突变阳性NSCLC患者死亡的最佳截断值；采用单因素、多因素Cox回归分析探讨试验组接受EGFR-TKIs治疗的EGFR突变阳性NSCLC患者死亡的影响因素；采用R语言（R 4.0.3软件包）构建接受EGFR-TKIs治疗的EGFR突变阳性NSCLC患者预后预测列线图模型，采用一致性指数（CI ）、校准曲线和ROC曲线评价该列线图模型预测接受EGFR-TKIs治疗的EGFR突变阳性NSCLC患者预后的准确性。结果　多因素Cox回归分析结果显示，中性粒细胞/淋巴细胞比值（NLR）、乳酸脱氢酶（LDH）、血小板/淋巴细胞比值（PLR）、纤维蛋白原（FIB）是接受EGFR-TKIs治疗的EGFR突变阳性NSCLC患者死亡的独立影响因素（P ＜0.05）。基于多因素Cox回归分析结果，构建接受EGFR-TKIs治疗的EGFR突变阳性NSCLC患者预后预测列线图模型。该列线图模型预测试验组接受EGFR-TKIs治疗的EGFR突变阳性NSCLC患者1、2、3年生存率的CI 分别为0.86、0.80、0.78，预测验证组接受EGFR-TKIs治疗的EGFR突变阳性NSCLC患者1、2、3年生存率的CI 分别为0.89、0.84、0.80。校准曲线分析结果显示，该列线图模型预测试验组、验证组接受EGFR-TKIs治疗的EGFR突变阳性NSCLC患者1、2、3年生存率与患者实际1、2、3年生存率基本一致。ROC曲线分析结果显示，该列线图模型预测试验组、验证组接受EGFR-TKIs治疗的EGFR突变阳性NSCLC患者生存情况的AUC分别为0.896〔95%CI （0.848，0.945）〕、0.833〔95%CI （0.826，0.940）〕。结论　NLR、LDH、PLR、FIB是接受EGFR-TKIs治疗的EGFR突变阳性NSCLC患者死亡的影响因素，本研究基于上述指标构建的列线图模型能有效预测接受EGFR-TKIs治疗的EGFR突变阳性NSCLC患者的生存率。

Abstract：

　Objective　To construct and verify the nomogram model for predicting the prognosis of patients with EGFRmutation-positive non-small cell lung cancer (NSCLC) treated with EGFR-TKIs. Methods　A total of 332 patients with EGFRmutation-positive NSCLC who were admitted to the Department of Thoracic Surgery, Department of Respiratory Medicine andDepartment of Oncology of Liuzhou People's Hospital from 2018 to 2020 were selected and randomly divided into the experimentalgroup (n=232) and the verification group (n=100) according to the ratio of 7∶3. The clinical data of patients were collected,including gender, age, and routine blood test results 2 weeks before EGFR-TKIs treatment. Patients who received EGFR-TKIsfirst-line treatment were followed up, and the follow-up content included the follow-up treatment effect, whether recurrence ordeath, and the time of death. ROC curve analysis was used to determine the optimal cut-off value of each blood routine indexfor predicting the death in patients with EGFR mutation-positive NSCLC treated with EGFR-TKIs in the experimental group.Univariate and multivariate Cox proportional hazards regression analysis was used to investigate the influencing factors of death inpatients with EGFR mutation-positive NSCLC treated with EGFR-TKIs in the experimental group. R language (R 4.0.3 softwarepackage) was used to construct the nomogram model for predicting the prognosis of patients with EGFR mutation-positive NSCLCtreated with EGFR-TKIs. Concordance index (CI ) , calibration curve and ROC curve were used to evaluate the accuracy of thenomogram model in predicting the prognosis of patients with EGFR mutation-positive NSCLC treated with EGFR-TKIs. Results　The results of multivariate Cox proportional hazards regression model analysis showed that neutrophil-to-lymphocyte ratio (NLR) ,lactate dehydrogenase (LDH) , platelet to lymphocyte ratio (PLR) , and fibrinogen (FIB) were the influencing factors of death inpatients with EGFR mutation-positive NSCLC treated with EGFR-TKIs (P < 0.05) . Based on the results of multivariate Coxproportional hazards regression analysis, a nomogram model for predicting the prognosis of patients with EGFR mutation-positiveNSCLC treated with EGFR-TKIs was constructed. The CI of the nomogram model for predicting the 1, 2, and 3 years survivalrates of patients with EGFR mutation-positive NSCLC treated with EGFR-TKIs in the experimental group was 0.86, 0.80, and 0.78,respectively. The CI for the nomogram model for predicting the 1, 2, and 3 years survival rates of patients with EGFR mutationpositiveNSCLC treated with EGFR-TKIs in the verification group was 0.89, 0.84, and 0.80, respectively. The results of calibrationcurve analysis showed that the 1, 2, and 3 years survival rates of patients with EGFR mutation-positive NSCLC treated withEGFR-TKIs predicted by the nomogram model in the experimental group and the validation group were basically consistent withthe actual 1, 2, and 3 year survival rates of the patients. The results of ROC curve analysis showed that the AUC of the nomogrammodel for predicting the survival of patients with EGFR mutation-positive NSCLC treated with EGFR-TKIs in the experimentalgroup and the validation group was 0.896 [95%CI (0.848, 0.945) ] , 0.833 [95%CI (0.826, 0.940) ] , respectively. Conclusion　NLR, LDH, PLR, and FIB are the influencing factors of death in patients with EGFR mutation-positive NSCLC treated withEGFR-TKIs. The nomogram model construeted based on the above indicators can effectively predict the survival rate of patientswith EGFR mutation-positive NSCLC treated with EGFR-TKIs.

ReferenceList：