影像组学在局限期小细胞肺癌脑预防照射优化中的价值  

The value of radiomics for individualized prophylactic cranial irradiation in limited-stage small cell lung cancer

作　　者：侯庆[1] 魏丽娟 姚宁宁 孙博晨 梁玉[1] 曹欣 谭艳[2] 曹建忠[1] Hou Qing;Wei Lijuan;Yao Ningning;Sun Bochen;Liang Yu;Cao Xin;Tan Yan;Cao Jianzhong(Department of Radiotherapy,Shanxi Province Cancer Hospital/Shanxi Hospital Affiliated to Cancer Hospital,Chinese Academy of Medical Sciences/Cancer Hospital Affiliated to Shanxi Medical University,Taiyuan 030013,China;Department of Radiology,the First Affiliated Hospital of Shanxi Medical University,Taiyuan 030001,China)

机构地区：[1]山西省肿瘤医院/中国医学科学院肿瘤医院山西医院/山西医科大学附属肿瘤医院放射治疗中心,太原030013 [2]山西医科大学第一医院影像科,太原030001

出　　处：《中华放射肿瘤学杂志》2023年第1期8-14,共7页Chinese Journal of Radiation Oncology

基　　金：山西省应用基础研究计划项目(20210302124598);山西省留学回国人才科研活动择优资助项目(晋人社厅[2019]1176号);山西省回国留学人员科研资助项目(晋人社厅[2022]210号);山西省卫健委“四个一批”科技兴医创新计划([2022]37号)。

摘　　要：目的探讨治疗前胸部增强CT影像组学模型对局限期小细胞肺癌(LS-SCLC)脑转移的预测能力以及指导个体化预防性脑照射(PCI)的价值。方法回顾性分析2012年1月至2018年12月在山西省肿瘤医院经病理确诊为小细胞肺癌及影像学检查确定为局限期患者资料97例。基于最小绝对值收缩和选择算子(LASSO)Cox与相关性检验筛选与LS-SCLC脑转移显著相关的影像组学特征构建模型,使用校正曲线、受试者操作特征曲线下面积(AUC)、内部5折交叉验证、决策曲线分析(DCA)与整合布莱尔评分(IBS)评估影像组学模型的预测效能与临床获益,使用Kaplan-Merier曲线和log-rank检验绘制生存曲线和评估组间差异。结果提取出影像组学特征1272个,使用LASSO Cox回归和相关性检验筛选特征,最后通过8个与LS-SCLC患者脑转移发生相关的影像组学特征构建影像组学模型。影像组学模型预测LS-SCLC患者1年与2年脑转移的AUC分别为0.845(95%CI为0.746~0.943)和0.878(95%CI为0.774~0.983)。5折内部交叉验证、校正曲线、DCA以及IBS显示模型有较好的预测效能与临床净获益。基于影像组学模型的风险分层:高危患者PCI组1年、2年脑转移累积发病率分别为0%、18.2%,非PCI组分别为61.8%、75.4%(P<0.001),PCI组1年、2年生存率分别为92.9%、78.6%,非PCI组分别为85.3%、36.8%(P=0.023);低危患者PCI组1年、2年脑转移累积发病率分别为0%、0%,非PCI组分别为10.0%、20.2%(P=0.062),PCI组1年、2年生存率分别为100%、77.0%,非PCI组分别为96.7%、79.3%(P=0.670)。结论增强CT影像组学模型在预测LS-SCLC脑转移和指导个体化PCI方面具有良好的性能。Objective To investigate the predictive value of enhanced CT-based radiomics for brain metastasis (BM) and selective use of prophylactic cranial irradiation (PCI) in limited-stage small cell lung cancer (LS-SCLC). Methods Clinical data of 97 patients diagnosed with LS-SCLC confirmed by pathological and imaging examination in Shanxi Provincial Cancer Hospital from January 2012 to December 2018 were retrospectively analyzed. The least absolute shrinkage and selection operator (LASSO) Cox and Spearman correlation tests were used to select the radiomics features significantly associated with the incidence of BM and calculate the radiomics score. The calibration curve, the area under the receiver operating characteristic (ROC) curve (AUC), 5-fold cross-validation, decision curve analysis (DCA), and integrated Brier score (IBS) were employed to evaluate the predictive power and clinical benefits of the radiomics score. Kaplan-Meier method and log-rank test were adopted to draw survival curves and assess differences between two groups. Results A total of 1272 radiomics features were extracted from enhanced CT. After the LASSO Cox regression and Spearman correlation tests, 8 radiomics features associated with the incidence of BM were used to calculate the radiomics score. The AUCs of radiomics scores to predict 1-year and 2-year BM were 0.845 (95%CI=0.746-0.943) and 0.878 (95%CI=0.774-0.983), respectively. The 5-fold cross validation, calibration curve, DCA and IBS also demonstrated that the radiomics model yielded good predictive performance and net clinical benefit. Patients were divided into the high-risk and low-risk cohorts based on the radiomics score. For patients at high risk, the 1-year and 2-year cumulative incidence rates of BM were 0% and 18.2% in the PCI group, and 61.8% and 75.4% in the non-PCI group, respectively (P<0.001). In the PCI group, the 1-year and 2-year overall survival rates were 92.9% and 78.6%, and 85.3% and 36.8% in the non-PCI group, respectively (P=0.023). For patients at low risk, the 1-ye

关 键 词：癌 小细胞肺 局限期 肿瘤转移 脑 影像组学 预防性脑照射 

分 类 号：R734.2[医药卫生—肿瘤]