局部晚期食管鳞癌放射治疗中CT影像组学特征的变化  

Changes of CT image characteristics in radiotherapy of locally advanced esophageal squamous cell carcinoma

作　　者：刘懿梅 陈美宁 王彬[1] 邱波[1] 张俊 邓小武[1] 彭应林 LIU Yimei;CHEN Meining;WANG Bin;QIU Bo;ZHANG Jun;DENG Xiaowu;PENG Yinglin(Department of Radiation Oncology,Sun Yat-sen University Cancer Center/State Key Laboratory of Oncology in South China/Guangdong Key Laboratory of Nasopharyngeal Carcinoma Diagnosis and Therapy,Guangdong Guangzhou 510060,China)

机构地区：[1]中山大学肿瘤防治中心/华南肿瘤学国家重点实验室/广东省鼻咽癌诊治研究重点实验室,放射治疗科,广东广州510060

出　　处：《现代肿瘤医学》2023年第11期2072-2079,共8页Journal of Modern Oncology

基　　金：国家自然科学基金(编号:12005316);中华国际医学交流基金会肿瘤精准放疗星火计划(编号:2019-N-11-20);广东省广州市科技计划项目(编号:202206010154,202206010180)。

摘　　要：目的:通过提取和筛选食管癌计划CT和每周重复CT图像的影像组学特征,分析其特征变化与放射治疗响应的相关性。方法:共入组行放疗的局部晚期食管鳞癌患者15例,每周治疗结束后对患者重新行模拟CT扫描,由同一名医师在5次重复CT图像上勾画靶区。利用Python编程提取计划CT和重复扫描CT图像中靶区的影像组学变量特征并做统计分析,筛选出与放疗周数、靶区体积变化以及治疗结局(2年生存)因素强相关的组学特征,并用斯皮尔曼相关性和点双列相关性分析法对特征相关性进行统计学分析。结果:影像组学特征分析中,在靶区CTV1和CTV2中各提取了1688个特征,并分别筛选出了10个与放疗周数、靶区体积变化以及治疗结局(2年生存)最强相关的特征。组学特征与治疗次数(放疗周数)呈强负相关(<-0.6),与体积变化呈强正相关(>0.6),与治疗结局(2年生存)呈弱相关,其相关系数分别为(-0.81~-0.67)、(0.72~0.99)和(-0.37~0.52)。结论:食管癌放疗过程中靶区CT影像组学变量特征与治疗时间、靶区体积大小以及治疗结局均存在相关性,基于CT影像组学变量特征变化可潜在地预测食管癌放疗预后。Objective:To extract and screen the radiomic features of esophageal cancer planning CT and weekly repeated CT images,and to analyze the correlation between their feature changes and radiotherapy response.Methods:A total of 15 patients with locally advanced esophageal squamous cell carcinoma who received radiotherapy were re-scanned by simulated CT after weekly treatment,and the target volume were delineated on 5 repeated CT images by the same physician.The radiomic features of the target volume in the planned CT and repeated scan CT images were extracted by Python programming and statistically analyzed,and the radiomic features which were strongly related to the weeks of radiotherapy,the volume change of the target volume and the treatment outcome(2-year survival)were screened,and the correlation of the features was statistically analyzed by Spearman correlation and point biserial correlation analysis.Results:In the radiomic feature analysis,1688 features were extracted from CTV1 and CTV2 respectively,and 10 features were most strongly related to treated week numbers,target volume changes and treatment outcome(2-year survival).The radiomic features were negatively correlated with the weeks of radiotherapy(<-0.6),positively correlated with volume changes(>0.6),and weakly correlated with treatment outcome(2-year survival).The correlation coefficients were(-0.81~-0.67),(0.72~0.99)and(-0.37~0.52),respectively.Conclusion:The CT delta-radiomic features of the target volume during radiotherapy for esophageal cancer are correlated with the treatment times,target volume and treatment outcome.Based on the DRFs of CT,the prognosis of esophageal cancer can be predicted.

关 键 词：局部晚期食管鳞癌 影像组学 放射治疗 计算机断层扫描图像 靶区体积 

分 类 号：R735.[医药卫生—肿瘤]