CT放射组学鉴别诊断周围型小细胞肺癌和非小细胞肺癌的价值研究  

The Study of CT Radiology in Differentiating Peripheral Small Cell Lung Cancer and Non Small Cell Lung Cancer

作　　者：靳强[1] 高俊萍[1] 王欢 江涛 孙红红[1] JIN Qiang;GAO Jun-ping;WANG Huan;JIANG Tao;SUN Hong-hong(Department of Imaging,The Second Affiliated Hospital of Hebei North University,Zhangjiakou 075100,Hebei Province,China)

机构地区：[1]河北北方学院附属第二医院影像科,河北张家口075100

出　　处：《中国CT和MRI杂志》2025年第1期47-49,共3页Chinese Journal of CT and MRI

基　　金：河北省2023年度医学科学研究课题计划(20231618)。

摘　　要：目的 本研究旨在评估C T放射学方法鉴别周围型小细胞肺癌和非小细胞肺癌的研究,以期为临床诊治提供参考。方法回顾性研究原发性肺癌患者的临床资料,利用诊断后、治疗前的肺部CT图像对肺癌进行分割。从基于直方图的统计、肿瘤图像纹理分析及其小波变换中提取放射学特征。特征选择采用最小冗余度和最大相关法。用多层人工神经网络建立预测模型,用受试者工作特征曲线(AUC)下面积评价SCLC/NSCLC腺癌分类器的性能。结果69例小细胞肺癌和34例非小细胞肺癌患者相比,小细胞肺癌组男性患者和吸烟者多于非小细胞肺癌组(P<0.05)。我们的SCLC/NSCLC分类器的总体性能AUC为0.93(95%可信区间=[0.85,0.97],灵敏度=0.85,特异度=0.85)。添加如吸烟史等临床数据,可略微改善性能。排名最靠前的放射学特征主要是纹理特征。结论 CT放射组学能定量反映肿瘤的异质性,可用于肺癌亚型的鉴别诊断,效果满意。用小波变换技术对CT图像进行处理,增强了小细胞肺癌/非小细胞肺癌的放射学特征。Objective This study aims to evaluate the differential diagnosis of peripheral small cell lung cancer and non-small cell lung cancer using CT radiology methods,in order to provide reference for clinical diagnosis and treatment.Methods Retrospective study of clinical data of patients with primary lung cancer,and segmentation of lung cancer using lung CT images after diagnosis and before treatment.Extract radiological features from histogram based statistics,tumor image texture analysis,and wavelet transform.The feature selection adopts the minimum redundancy and maximum correlation methods.Establish a prediction model using a multi-layer artificial neural network,and evaluate the perfo rmance of SCLC/NSCLC adenocarcinoma classifier using the area under the subject ope rating characteristic cu rve(AUC).Results Compared with 69 cases of small cell lung cancer and 34 cases of non-small cell lung cancer,the number of male patients and smokers in the small cell lung cancer group was higher than that in the non-small cell lung cancer group(P<0.05).Our SCLC/N SCLC classifier has an overall performance AUC of 0.93(95% confidence interval=[0.85,0.97],sensitivity=0.85,and specificity=0.85).Adding clinical data such as smoking history can slightly improve performance.The top ra n ked radiological features a re mainly texture features.Conclusion CT radiomics can quantitatively reflect the heterogeneity of tumors and can be used for the differential diagnosis of lung cancer subtypes with satisfactory results.The use of wavelet transform technology to process CT images enhances the radiological featu res of small cell lung cancer/non-small cell lung cancer.

关 键 词：CT 周围型小细胞肺癌 非小细胞肺癌 放射学特征 

分 类 号：R563[医药卫生—呼吸系统]