摘要
目的分析CT影像组学联合临床特征对晚期非小细胞肺癌(NSCLC)化疗及免疫治疗效果的预测价值。资料与方法回顾性纳入2021年9月-2023年10月内蒙古医科大学附属医院经病理证实的120例晚期NSCLC,患者均接受程序性死亡受体1/程序性死亡配体1抑制剂单免疫治疗或化疗联合免疫治疗,观察患者2个疗程(3~6周)治疗后3个月内增强胸部CT,依据实体瘤疗效评价标准1.1版评估临床疗效,并将患者分为疾病进展组66例和疾病未进展组54例。按7∶3随机分为训练组和验证组,训练组中疾病进展患者46例,疾病未进展38例;验证组中疾病进展患者20例,疾病未进展16例。对开始治疗前1个月内胸部增强CT图像进行感兴趣区分割,提取和筛选影像组学特征,建立支持向量机机器学习模型,采用受试者工作特征曲线分析模型预测疗效的诊断效能,绘制校准曲线和决策曲线,评估模型的预测概率及临床收益。结果临床模型纳入有无肾上腺转移及免疫药物种类2个临床因素,影像组学模型纳入11个最具标签的影像组学特征,两者联合构建联合模型。临床模型、影像组学模型、临床-影像组学联合模型的预测效能分别为:训练组曲线下面积0.65、0.92、0.90,验证组曲线下面积0.64、0.83、0.85;校准曲线显示联合模型预测概率最高;决策曲线显示影像组学模型与联合模型收益相当,均优于单一临床模型。结论CT影像组学很大概率可以无创预测晚期NSCLC化疗及免疫治疗效果,尤其结合临床特征的联合模型预测效能最佳,影像组学可以为晚期NSCLC生存期及改善预后提供有利导向。
Purpose To analyze the predictive value for the efficacy of chemotherapy and immunotherapy in patient with advanced non-small cell lung cancer(NSCLC).Materials and Methods A total of 120 patients with pathologically confirmed advanced NSCLC were retrospectively selected from September 2021 to October 2023 in the Affiliated Hospital of Inner Mongolia Medical University.All patients received programmed death receptor 1/programmed death receptor ligand 1 inhibitor monotherapy or chemotherapy combined with immunotherapy.The enhanced chest CT within three months of two courses(three to six weeks)for the evaluation of the clinical efficacy via the efficacy evaluation criteria of solid tumor version 1.1 was observed.According to the evaluation situation,all patients were divided into disease progression group(n=66)and disease without progression group(n=54).They were randomly divided into the training group and the verification group in a ratio of 7∶3.Among them,there were 46 patients with progression and 38 patients without progression in the training group.In the verification group,there were 20 patients with progression and 16 patients without progression.The regions of interest were segmented on the enhanced CT images of the chest within one month before the start of treatment,and the radiomics features were extracted and screened to establish the support vector machine learning model.The diagnostic efficacy of the model in predicting therapeutic effect was analyzed using the receiver operating characteristic curve,the calibration curve and decision curve were plotted to evaluate the prediction probability and clinical benefits of the model.Results Two clinical factors,namely the presence or absence of adrenal metastasis and the types of immune drugs,were included in the clinical model.Eleven of the most labeled radiomics characteristics were included in the radiomics model.The two were combined to construct a combined model.The area under the curve of the predictive efficiencies of the training group of the clinical model,the radiomics model and the clinical-radiomics combined model were 0.65,0.92 and 0.90,respectively;while the area under the curve of validation group were 0.64,0.83 and 0.85,respectively.Conclusion CT radiomics will have a high probability of non-invasively predicting the effects of chemotherapy and immunotherapy for advanced NSCLC.Especially the model combined with clinical characteristics has the best predictive efficacy.Radiomics can provide favorable guidance for the survival period and prognosis improvement of advanced NSCLC.
作者
刘宇婷
赵磊
刘挨师
LIU Yuting;ZHAO Lei;LIU Aishi(Department of Imaging Diagnosis,the Affiliated Hospital of Inner Mongolia Medical University,Hohhot 010050,China;不详)
出处
《中国医学影像学杂志》
北大核心
2025年第7期758-765,共8页
Chinese Journal of Medical Imaging
基金
内蒙古自治区自然科学基金(2022MS08056)。
关键词
癌
非小细胞肺
影像组学
化学疗法
辅助
免疫疗法
病理学
临床
预测
预后
Carcinoma,non-small-cell lung
Radiomics
Chemotherapy,adjuvant
Immunotherapy
Pathology,clinical
Forecasting
Prognosis Chinese Journal of Medical Imaging,2025,33(7):758-765
