瘤内及瘤周MR影像组学联合临床特征预测宫颈癌淋巴脉管间隙浸润  

Intratumoral and peritumoral magnetic resonance imaging radiomics combined with clinical characteristics to predict lymphovascular space invasion in cervical cancer

作　　者：林宝金 吴朝霞[2] 王石[2] 龙先凤 梁莉莉 李弟声 朱超华 LIN Baojin;WU Zhaoxia;WANG Shi;LONG Xianfeng;LIANG Lili;LI Disheng;ZHU Chaohua(Department of Radiotherapy Physical Technology,the People's Hospital of Guangxi Zhuang Autonomous Region,Nanning 530021,China;Department of Engineering Physics,Tsinghua University,Beijing 100084,China)

机构地区：[1]广西壮族自治区人民医院放疗物理技术室,广西南宁530021 [2]清华大学工程物理系,北京100084

出　　处：《中国医学物理学杂志》2024年第7期851-857,共7页Chinese Journal of Medical Physics

基　　金：广西壮族自治区卫生健康委科研课题(Z-A20230042);广西医疗卫生适宜技术开发与推广应用项目(S2022014)。

摘　　要：目的:探讨基于瘤内+瘤周MR影像组学联合临床特征构建的列线图模型,在预测宫颈癌淋巴脉管间隙浸润(LVSI)状态的价值。方法:回顾性分析178例经术后病理证实的宫颈癌患者资料,其中70例LVSI(+)、108例LVSI(-),按照8:2的比例将其划分为训练集[142例,54例LVSI(+)、88例LVSI(-)]和测试集[36例,16例LVSI(+)、20例LVSI(-)]。所有宫颈癌患者术前均行MRI检查,在T2WI序列上手动逐层勾画感兴趣区(ROI),瘤周区域分别等距外扩。对临床因素实行单因素逻辑分析,筛选宫颈癌LVSI(+)的独立预测因子。分别基于瘤内区、瘤周区和瘤内+瘤周区提取影像组学特征,以最佳者构建影像组学模型,比较各瘤周和瘤内+瘤周模型差异。基于最佳瘤内+瘤周模型影像组学评分与临床独立预测因子构建联合模型,绘制列线图。采用受试者工作特征曲线来评价各模型的诊断性能,决策曲线评价模型的临床价值。结果:联合模型在各种模型对比中效果最佳,训练集和测试集的AUC值为0.970和0.803。结论:瘤内+瘤周MR影像组学联合临床特征可以有效预测宫颈癌LVSI。Objective To investigate the value of a nomogram model constructed from intratumoral and peritumoral magnetic resonance imaging radiomics combined with clinical characteristics in predicting the status of lymphovascular space invasion(LVSI)in cervical cancer.Methods A retrospective analysis was conducted on 178 cervical cancer patients confirmed by postoperative pathology,with 70 cases of LVSI(+)and 108 cases of LVSI(-).The patients were divided into a training set[142 cases,including 54 cases of LVSI(+)and 88 cases of LVSI(-)]and a test set[36 cases,including 16 cases of LVSI(+)and 20 cases of LVSI(-)]at a ratio of 8:2.All underwent magnetic resonance imaging before surgery,and regions of interest were manually delineated layer by layer on the T2WI sequence,with the peritumoral region being uniformly expanded outward.Univariate logistic analysis was performed on clinical factors to select independent factors for cervical cancer LVSI(+).Radiomic features were extracted separately from the intratumoral region,the peritumoral region,and the intratumoral-peritumoral region to construct radiomics models,and the differences between the peritumoral and the intratumoral-peritumoral models were compared.A combined model was established based on the radiomics scores of the optimal intratumoral-peritumoral model and clinical independent predictive factors,and a nomogram was plotted.Receiver operating characteristic curves were used to evaluate the diagnostic performance of each model,and decision curve analysis was used to assess the clinical value of the models.Results The combined model demonstrated the best performance among the various models,with AUC of 0.970 in the training set and 0.803 in the test set.Conclusion Intratumoral and peritumoral magnetic resonance imaging radiomics combined with clinical characteristics can effectively predict LVSI in cervical cancer.

关 键 词：宫颈癌 影像组学 磁共振成像 淋巴脉管浸润 

分 类 号：R318[医药卫生—生物医学工程] R737.33[医药卫生—基础医学]