胸部CT影像组学预测EGFR阳性非小细胞肺癌患者脑转移风险  

Prediction risk of brain metastases in EGFR positive non-small cell lung cancer patients based on thoracic CT radiomics

作　　者：张明珠 孙晓蓉 侯懿宸 郑梅 邢力刚 ZHANG Mingzhu;SUN Xiaorong;HOU Yichen;ZHENG Mei;XING Ligang(Cheeloo College of Medicine,Shandong University,Jinan 250012,China;Shandong Cancer Hospital and Institute,Shandong First Medical University and Shandong Academy of Medical Sciences,Jinan 250117,China;Department of Graduate,Shandong First Medical University and Shandong Academy of Medical Sciences,Jinan 250117,China)

机构地区：[1]山东大学齐鲁医学院医学融合与实践中心,山东济南250012 [2]山东省肿瘤防治研究院(山东省肿瘤医院,核医学科),山东第一医科大学(山东省医学科学院),山东济南250117 [3]山东第一医科大学(山东省医学科学院)研究生部,山东济南250117 [4]山东省肿瘤防治研究院(山东省肿瘤医院,放疗科),山东第一医科大学(山东省医学科学院),山东济南250117

出　　处：《中华肿瘤防治杂志》2023年第10期593-599,共7页Chinese Journal of Cancer Prevention and Treatment

基　　金：国家自然科学基金面上项目(82172866);山东省自然科学基金重点支持项目(ZR2021LZL005);山东省自然科学基金(ZR2019LZL019)。

摘　　要：目的 建立预测Ⅲ/Ⅳ期表皮生长因子受体(EGFR)阳性非小细胞肺癌(NSCLC)患者无脑转移生存期(BM-FS)的影像组学预测模型。方法 收集2014-08-05-2020-11-24山东省肿瘤医院收治的142例首次诊断为EGFR突变型Ⅲ/Ⅳ期的NSCLC患者,构建随访队列进行回顾性队列研究。按照7∶3的比例随机分为训练组(n=99)和验证组(n=43)。收集治疗前胸部CT图像,使用3D-slicer软件提取影像组学特征,采用LASSO算法和5倍交叉验证确定最优特征,根据回归系数计算影像组学评分(Radscores)。构建影像组学模型,绘制受试者工作特征(ROC)曲线,计算曲线下面积(AUC)评价模型效能。结果 共提取出851项影像组学特征,筛选出6个最优影像组学特征用于预测BM-FS。根据Radscores截断值划分的脑转移高危组和低危组患者,2组间BM-FS差异有统计学意义,P<0.001。影像组学模型在训练组中的1、2和3年BM-FS AUC分别为0.774、0.809和0.704,验证组的1、2和3年BM-FS AUC分别为0.815、0.896和0.763,表现出较好的预测效果。结论 胸部CT影像组学模型有助于对Ⅲ/Ⅳ期EGFR阳性NSCLC患者脑转移风险进行分层和预测。Objective To develop a radiomics predictive model to predict brain metastasis free survival(BM-FS) for stage Ⅲ/Ⅳ EGFR positive non-small cell lung cancer(NSCLC) patients.Methods A total of 142 patients with first diagnosed EGFR mutant, stage Ⅲ/Ⅳ NSCLC were collected from 5 August 2014 to 24 November 2020, who were selected to construct a follow-up cohort for a retrospective cohort study.The pretreatment thoracic CT images were collected and the radiomics signatures were extracted using the 3D slicer.Totally 142 patients were randomly divided into training and validation cohorts at a ratio of 7∶3.LASSO regression and 5-fold cross-validation were used for radiomics data to identify optimal features and the radscores were calculated according to the regression coefficients.The area under the curve(AUC) of receiver operating characteristic(ROC) was drawn to evaluate the effectiveness of the radiomics model.Results A total of 851 radiomics features were extracted and 6 optimal features were selected to predict BM-FS.There was significant difference in BM-FS between the high-risk group and the low-risk group divided by the cut-off value of radscores(P<0.001).The 1-year, 2-year, 3-year BM-FS AUC of the radiomics model in training cohort were 0.774,0.809,0.704,respectively and the 1-year,2-year,3-year BM-FS AUC of the radiomics model in validation cohort were 0.815,0.896,0.763,respectively,showing agood predictive value of BM-FS in EGFR positive NSCLC patients.Conclusion The thoracic CT radiomics model helps to the stratification and prediction of the risk of brain metastasis in stage Ⅲ/Ⅳ EGFR positive NSCLC patients.

关 键 词：影像组学 脑转移 表面上皮生长因子受体基因突变 非小细胞肺癌 列线图 

分 类 号：R734.2[医药卫生—肿瘤]