PET/CT代谢参数及临床特征用于诊断非小细胞肺癌表皮生长因子受体基因突变

PET/CT metabolic parameters and clinical characteristics for diagnosis of epidermal growth factor receptor gene mutation in non-small cell lung cancer

目的观察PET/CT代谢参数和临床特征用于诊断非小细胞肺癌(NSCLC)患者表皮生长因子受体(EGFR)基因突变的价值。方法纳入127例接受PET/CT检查的NSCLC患者,随机抽取15例为验证集;按7∶3比例将其余112例随机分为训练集(n=77)及测试集(n=35),并根据EGFR基因检测结果分为突变型组(n=55)及野生型组(n=57),比较观察2组PET/CT代谢参数和临床特征。采用单因素和多因素logistic回归分析筛选参数构建模型,评估其诊断EGFR突变的效能。结果突变型组女性及腺癌占比均高于野生型组(P均<0.05),而原发灶最大径、肿瘤代谢体积(MTV)及病灶糖酵解总量(TLG)均小于野生型组(P均<0.05)。以TLG、MTV、最大标准摄取值(SUV_(max))和平均标准摄取值(SUV_(mean))诊断NSCLC患者EGFR突变的曲线下面积(AUC)分别为0.68、0.66、0.63及0.63。经单因素及多因素logistic回归分析,将TLG及性别用于建立模型,诊断训练集、测试集EGFR突变的AUC分别为0.72、0.67,其在训练集、测试集及验证集的准确率分别为67.53%、71.43%及66.67%。结论PET/CT代谢参数如TLG和临床特征如性别等可用于诊断NSCLC EGFR突变。

Objective To observe the value of PET/CT metabolic parameters and clinical characteristics for diagnosis of epidermal growth factor receptor(EGFR)gene mutation in non-small cell lung cancer(NSCLC)patients.Methods A total of 127 NSCLC patients who underwent PET/CT examination were enrolled.Among them,15 patients were randomly selected as validation set,while 112 patients were randomly divided into training set(n=77)and test set(n=35)according to the ratio of 7∶3 and mutant group(n=55)and wild-type group(n=57)according to the results of EGFR gene detection.PET/CT metabolic parameters and clinical characteristics were observed and compared between groups.Univariate and multivariate logistic regression analysis were used to screen indexes and construct a model,and the efficacy of the model for diagnosing EGFR mutation was evaluated.Results The proportions of female and adenocarcinoma in mutant group were both higher than those in wild-type group(both P<0.05),while the maximum diameter of the primary lesions,metabolic tumor volume(MTV)and total lesion glycolysis(TLG)in mutant group were all lower than those in wild-type group(all P<0.05).The area under the curve(AUC)of TLG,MTV,the maximum standard uptake value(SUV_(max))and the mean standard uptake value(SUV_(mean))for diagnosing EGFR mutation in NSCLC was 0.68,0.66,0.63 and 0.63,respectively.After univariate and multivariate logistic regression analysis,TLG and gender were selected to establish the model,and AUC of the model for diagnosing EGFR mutation in training set and test set was 0.72 and 0.67,respectively,while its accuracy in training set,test set and validation set was 67.53%,71.43%and 66.67%,respectively.Conclusion PET/CT metabolic parameters as TLG and clinical characteristics as gender could be used to diagnose EGFR mutation of NSCLC.