能谱CT参数Logistic回归模型预测中晚期非小细胞肺癌EGFR基因突变的价值  

The value of energy spectrum CT parameter Logistic regression model in predicting EGFR gene mutation in advanced non-small cell lung cancer

作　　者：余莹莹[1] 甘晓晶[1] 许晓燕[1] 周诚[1] 周永[1] 文智[1] YU Ying-ying;GAN Xiao-jing;XU Xiao-yan;ZHOU Cheng;ZHOU Yong;WEN Zhi(CT Room,Affiliated Tumor Hospital of Xinjiang Medical University,Urumqi 830011,China)

机构地区：[1]新疆医科大学附属肿瘤医院CT室,新疆乌鲁木齐830011

出　　处：《中国临床医学影像杂志》2022年第4期243-248,共6页Journal of China Clinic Medical Imaging

基　　金：新疆维吾尔自治区自然科学基金(编号:2017D01C405)。

摘　　要：目的:探索能谱CT形态学征象及定量参数与中晚期非小细胞肺癌表皮生长因子受体(EGFR)基因突变的相关性,建立Logistic回归模型预测中晚期非小细胞肺癌EGFR基因突变状态。方法:根据79例中晚期非小细胞肺癌患者EGFR基因突变的情况分为阳性组与阴性组,收集两组能谱CT形态学征象、定量参数的指标,分析能谱CT形态学征象、定量参数与EGFR基因突变的相关性,建立Logistic回归模型预测中晚期非小细胞肺癌EGFR基因突变状态。结果:两组间肿瘤直径、分叶征、肺内转移、纵隔淋巴结转移、远处转移、骨转移、动静脉期的标准化碘浓度(Normalized iodine concentrations,NIC)、能谱曲线斜率(λHU)、70 keV CT值的差异有统计学意义。ROC曲线判断静脉期的NIC诊断效能最大(曲线下面积(AUC)=0.905(P<0.001),灵敏度为73.20%,特异度为93.80%,阈值为0.33 mg/cm^(3))。基于具有统计学意义的指标建立的Logistic回归模型示肺内转移、远处转移、静脉期NIC、静脉期λHU是预测中晚期非小细胞肺癌EGFR基因突变的独立因素,具有较高的诊断效能(AUC=0.953,灵敏度为93.6%,特异度为87.5%)。结论:基于能谱CT形态学征象及定量参数建立的Logistic回归模型在预测中晚期非小细胞肺癌EGFR基因突变中有一定的价值。Objective:To explore the correlation between the morphological features and quantitative parameters of energy spectrum CT and EGFR gene mutation in advanced non-small cell lung cancer,and to establish a Logistic regression model to predict EGFR gene mutation in advanced non-small cell lung cancer.Methods:According to the mutation of EGFR gene,79 patients were divided into positive group and negative group.The differences of morphological features and quantitative parameters of energy spectrum CT between the two groups were collected to explore the correlation between morphological features and quantitative parameters of energy spectrum CT and EGFR gene mutation.Logistic regression model was established to predict EGFR gene mutation.Results:There were significant differences in tumor diameter,lobulation sign,intrapulmonary metastasis,mediastinal lymph node metastasis,distant metastasis and bone metastasis between the two groups in morphological features.Among the quantitative parameters,the values of normalized iodine concentrations(NIC),λHU and 70 keV CT in arterial phase and venous phase were significantly different between the two groups.The diagnostic efficiency of quantitative parameters judged by ROC curve shows that NIC in venous phase is the highest(AUC=0.905(P<0.001),sensitivity is 73.20%,specificity is 93.80%,threshold is 0.33).With the above statistically significant indicators,a Logistic regression model was established,in which intrapulmonary metastasis,distant metastasis,venous phase NIC and venous phaseλHU were independent factors for predicting EGFR gene mutation and had high diagnostic efficiency(AUC=0.953,had a sensitivity of 93.6%and a specificity of 87.5%).Conclusion:The CT regression model based on the morphological features and quantitative parameters of Logistic has a certain value in predicting EGFR gene mutation in advanced non-small cell lung cancer.

关 键 词：癌 非小细胞肺 体层摄影术 X线计算机 

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