肺腺癌临床及CT特征在预测EGFR基因突变中的价值  

The value of clinical and CT features in predicting EGFR mutation of lung adenocarcinoma

作　　者：霍继文 李琦[1] 罗天友[1] 何小群[1] 吕发金[1] HUO Ji-wen;LI Qi;LUO Tian-you(Department of Radiology,the First Affiliated Hospital of Chongqing Medical University,Chongqing 400016,China)

机构地区：[1]重庆医科大学附属第一医院放射科,重庆400016

出　　处：《放射学实践》2021年第11期1375-1381,共7页Radiologic Practice

基　　金：重庆市卫生计生委项目医学科研计划项目面上项目(2017MSXM010);重庆市科学技术局基础科学与前沿技术研究项目(cstc2017jcyjAX0281)。

摘　　要：目的:探讨肺腺癌临床及CT特征在预测表皮生长因子受体(EGFR)基因突变中的价值。方法:回顾性分析2012年2月-2018年10月我院经病理证实且行胸部CT扫描及EGFR基因检测的单发肺腺癌患者803例,根据基因检测结果分为EGFR基因突变阳性组(426例,53.05%)和阴性组(377例,46.95%),比较两组临床及CT特征的差异。联合上述有统计学差异的临床及CT特征,建立Logistic回归模型,利用曲线下面积(AUC)判断该模型在预测EGFR基因突变中的效能。结果:①临床特征与EGFR基因突变的关系:两组患者年龄无统计学差异(62±10 vs 61±11);女性(69.46%vs 39.03%)及不吸烟患者(66.37%vs 36.16%)EGFR基因突变率显著高于男性及吸烟患者,差异有统计学意义(P均<0.001),而Ⅰ~Ⅱ期与Ⅲ~Ⅳ期患者EGFR基因突变率无统计学差异(57.75%vs 50.83%)。②CT特征与EGFR基因突变的关系:在803例患者中,两组肿瘤大小[(3.36±1.83cm)vs(3.79±2.12cm)]差异有统计学意义(P<0.05),且周围型分布(55.28%vs 45.74%)、存在毛刺征(60.45%vs 50.26%)、GGO(66.67%vs 50.87%)、充气支气管征(70.83%vs 49.93%)、血管集束征(76.92%vs 45.39%)、胸膜牵拉征(64.95%vs 39.47%)及双肺多发转移(62.86%vs 51.58%)的肺腺癌EGFR基因突变率较高,而存在含气腔隙(46.67%vs 53.33%)、坏死(33.93%vs 56.15%)、胸腔积液(46.15%vs 55.26%)及胸内淋巴结肿大(48.67%vs 59.87%)的肺腺癌EGFR基因突变率较低,差异均有统计学意义(P均<0.05);在608例增强患者中,EGFR突变组(40.79±23.43)HU与非突变组(35.81±20.72)HU静脉期△CT值差异有统计学意义(P<0.05)。③联合临床及CT特征的Logistic回归分析:女性、不吸烟、GGO、充气支气管征、血管集束征、胸膜牵拉征及双肺多发转移为EGFR基因突变的独立预测因子,ROC曲线分析示该回归模型预测EGFR基因突变的AUC为0.771,其敏感度和特异度分别为74.82%、68.53%。结论:肺腺癌的临床及CT特征对EGFR基因突变状态具有重要的预�Objective:To explore the value of clinical and CT features in predicting epidermal growth factor receptor(EGFR)mutations of lung adenocarcinoma.Methods:A total of 803 patients with solitary lung adenocarcinoma confirmed by pathology from February 2012 to October 2018 were retrospectively analyzed.All the patients have obtained EGFR gene analysis and underwent chest CT scan.According to the reports of genetic tests,all cases were divided into EGFR mutation positive group(426 cases,53.05%)and negative group(377 cases,46.95%),and the differences of clinical and CT characteristics between the two groups were compared.Combining the above-mentioned statistically different clinical and CT features,a logistic regression model was established,and the area under the curve(AUC)was used to present the efficiency of the model in predicting EGFR mutations.Results:Relationship between clinical characteristics and EGFR gene mutation:there was no statistical difference in age(62±10 vs 61±11)between the two groups.The EGFR mutation rate was significantly higher in female(69.46%vs 39.03%)and non-smoking patients(66.37%vs 36.16%)than that in male and non-smoking patients(all P<0.001),while there was no statistically significant difference between stageⅠ-ⅡandⅢ-Ⅳpatients(57.75%vs 50.83%).Relationship between CT characteristics and EGFR mutations:EGFR gene mutation rate was significantly different in tumor size[(3.36±1.83)cm vs(3.79±2.12)cm]between the two groups.EGFR gene mutation rates were more presented in nodules with peripheral(55.28%vs 45.74%),burr sign(60.45%vs 50.26%),GGO(66.67%vs 50.87%),air bronchogram sign(70.83%vs 49.93%),vascular cluster sign(76.92%vs 45.39%),pleural traction sign(64.95%vs 39.47%)and multiple pulmonary metastases(62.86%vs 51.58%),all P<0.05,while the mutation rates of EGFR gene were lower in nodules with cavitation sign(46.67%vs 53.33%),necrosis(33.93%vs 56.15%),pleural effusion(46.15%vs 55.26%)and lymphadenopathy(48.67%vs 59.87%),all P<0.05.In 608 enhanced cases,there was significant differen

关 键 词：肺腺癌 体层摄影术 X线计算机 表皮生长因子受体 基因突变 

分 类 号：R734.2[医药卫生—肿瘤] R814.42[医药卫生—临床医学]