中央型小细胞肺癌与鳞癌CT纹理分析参数特征比较  

Comparison of parameters of CT texture analysis between central small cell lung cancer and squamous cell carcinoma

作　　者：龚恩惠 周永进 李露[1] 曹卓[1] GONG Enhui;ZHOU Yongjin;LI Lu;CAO Zhuo(Department of Respiratory,Lishui People's Hospital in Zhejiang Province,Lishui 323000,China;Department of Radiology,Lishui Central Hospital in Zhejiang Province,Lishui 323000,China)

机构地区：[1]浙江省丽水市人民医院呼吸科,浙江丽水323000 [2]浙江省丽水市中心医院放射科,浙江丽水323000

出　　处：《中国现代医生》2019年第21期13-16,I0002,共5页China Modern Doctor

基　　金：浙江省科技计划公益技术项目(2016C33243)

摘　　要：目的运用基于CT图像的纹理分析方法比较中央型小细胞肺癌和鳞癌量化参数特征。方法回顾性分析64例经病理证实的中央型肺癌(鳞癌49例,小细胞肺癌15例)患者资料,将所有患者分为小细胞肺癌组(SCLC组)、中高分化鳞癌组(SCC-A组)和低分化鳞癌组(SCC-B组)。患者均于治疗前行常规CT平扫和增强扫描,对肿瘤实质部分勾画感兴趣区域(ROI),利用后处理软件分别计算相应的纹理参数平均值、峰度值、偏度值及不均质度值。采用单因素方差分析(one-way ANOVA)或Kruskall-Wallis H检验比较三组肺癌平扫和增强CT纹理参数的差异,并对有统计学差异的参数值绘制受试者操作曲线(ROC),计算相应临界值的敏感度、特异度及曲线下面积(AUC)。结果平扫和增强SCC-A组不均质度值明显高于SCLC组和SCC-B组,峰度值明显低于SCLC组和SCC-B组,差异有统计学意义(P<0.05)。ROC结果显示,平扫图像:鉴别SCC-A组与SCLC组和SCC-A组与SCC-B组,不均质度值和峰度值的AUC分别为0.851(95%CI:0.711~0.940)、0.786(95%CI:0.646~0.890)、0.772(95%CI:0.621~0.885)、0.703(95%CI:0.556~0.825);增强图像:鉴别SCC-A组与SCLC组和SCC-A组与SCC-B组,不均质度值和峰度值的AUC分别0.906(95%CI:0.779~0.973)、0.831(95%CI:0.697~0.923)、0.862(95%CI:0.725~0.947)、0.745(95%CI:0.600~0.859)。结论平扫和增强CT纹理参数峰度值和不均质度值可有效鉴别小细胞肺癌与中高分化鳞癌以及中高分化鳞癌与低分化鳞癌,可以提供比平均CT值更多的量化信息,有助于为临床提供更多鉴别二者的定量信息。Objective To compare the quantitative parameters of central small cell lung cancer and squamous cell carcinoma using CT image-based texture analysis.Methods The data of 64 patients with pathologically confirmed central lung cancer(49 cases of squamous cell carcinoma,15 cases of small cell lung cancer) were retrospectively analyzed.All the patients were divided into small cell lung cancer group(SCLC group),moderately and highly differentiated squamous cell carcinoma group(SCC-A group) and poorly differentiated squamous cell carcinoma(SCC-B group).All patients underwent routine CT scan and enhanced scan before treatment,and the region of interest(ROI) was delineated in the parenchyma of the tumor.The post-processing software was used to calculate the mean,kurtosis,skewness and inhomogeneity of the corresponding texture parameters.One-way ANOVA or Kruskall-Wallis H test was used to compare the differences between the routine CT scan and enhanced CT texture parameters of the three groups of lung cancer.And the receiver operating curve(ROC) was plotted against the statistically significant parameter values.The sensitivity,specificity,and area under the curve(AUC) of the corresponding threshold were calculated.Results The level of heterogeneity in the routine CT scan and enhanced CT SCC-A group was significantly higher than that in the SCLC group and the SCC-B group.The kurtosis value of SCC-A group was significantly lower than that in the SCLC group and the SCC-B group(P<0.05).The ROC results showed that the in the routine CT scan image,the AUC of the heterogeneity value and the kurtosis value of identifying SCC-A group and SCLC group and SCC-A group and SCC-B group was 0.851(95%CI:0.711-0.940),0.786(95%CI:0.646-0.890)、0.772(95%CI:0.621-0.885),0.703(95%CI:0.556-0.825)respectively.In enhanced CT image,the AUC of the heterogeneity value and the kurtosis value of identifying SCC-A group and SCLC group and SCC-A group and SCC-B group,were 0.906(95%CI: 0.779-0.973),0.831(95%CI: 0.697-0.923),0.862(95%CI: 0.725-0.947),0.74

关 键 词：计算机断层影像 纹理分析方法 小细胞肺癌 鳞癌 

分 类 号：R730.44[医药卫生—肿瘤] R734.2[医药卫生—临床医学]