摘要
目的对非小细胞肺癌患者的治疗前CT影像进行影像组学分析,以鉴别肺鳞癌和腺癌。方法回顾性分析185例(93例训练集,92例验证集)NSCLC患者胸部CT增强影像,在人工勾画的肿瘤区域提取620个3D影像组学特征,Pearson相关和LASSO回归选择出6个影像特征建立Logistic预测模型。结果模型鉴别肺鳞癌和腺癌曲线下面积(AUC)在训练集和验证集上分别为0.926(95%CI,0.906~0.945),0.916(95%CI,0.901~0.940)。结论影像组学分析可以在治疗前建立新颖且有效的诊断模型,作为鉴别两种亚型的因子,为非小细胞肺癌患者的个性化诊疗提供决策支持。
Objective To perform radiomics analysis of pre-treatment CT images of patients with non-small cell lung cancer(NSCLC) to identify lung squamous cell carcinoma and adenocarcinoma. Methods A retrospective analysis of 185 patients(93 training cohort,92 validation cohort) with chest CT enhancement images in NSCLC patients,620 3 D radiomics features were extracted from the artificially delineated tumor regions,and 6 image features were selected by Pearson correlation and LASSO regression. A Logistic prediction model was then established. Results The model identified lung squamous cell carcinoma and adenocarcinoma AUC as 0.926(95% CI,0.906-0.945) and 0.916(95% CI,0.901-0.940) on the training and validation cohorts,respectively. Conclusion Radiomics analysis can establish a novel and effective diagnostic model before treatment,as a factor to identify the two subtypes,and provide decision support for personalized diagnosis and treatment of NSCLC patients.
作者
王大鹏
杨晓光
赵磊
杨振兴
刘挨师
WANG Dapeng;YANG Xiaoguang;ZHAO Lei(Department of Imaging Diagnosis,Affiliated Hospital of Inner Morgolia Medical University,Hohhot 010050,P.R.China)
出处
《临床放射学杂志》
CSCD
北大核心
2019年第11期2055-2058,共4页
Journal of Clinical Radiology
基金
内蒙古医科大学附属医院重大科研项目资助(编号:NYFY ZD 2014017)
关键词
影像组学
体层摄影术
X线计算机
肺鳞癌
肺腺癌
Radiomics
Tomography,X-ray computed
Lung squamous cell carcinoma
Lung adenocarcinoma
