摘要
目的基于术前双序列(T2加权成像及增强T1加权成像)MRI图像进行影像组学分析,并联合临床特征,建立最大径小于4 cm的早期宫颈癌(ⅠB期及ⅡA期)中危因素预测模型,并对该模型进行验证。材料与方法回顾性分析我院2016年6月至2021年6月的宫颈癌患者病例170例,根据术后病理结果将病例分为中危因素组及非中危因素组。将病例按7∶3的比例随机分为训练组(n=119)、验证组(n=51),采用Analysis Kinetics软件对图像特性进行特征提取,并采用多因素分析构建临床模型、影像组学模型及临床-影像组学组合模型,并通过受试者工作特征曲线、校准曲线和决策曲线分析比较并验证三种模型。结果临床-影像组学组合模型术前可以预测宫颈癌的中危因素(曲线下面积为0.853,P<0.01),敏感度为85.5%,特异度为78%,高于临床模型,与影像组学模型相比差异无统计学意义。结论对于小于4 cm的早期宫颈癌(ⅠB、ⅡA期),以双序列MRI图像及临床特征建立的临床-影像组学组合模型有助于预测宫颈癌病理中危因素,有益于临床制订个体化诊疗决策。
Objective:To establish and validate a combined predictive model based on pretreatment dual-sequence MR(T2 weighted imaging and contrast-enhanced T1 weighted imaging)imaging features and clinical features to predict intermediate risk factors in patients with early cervical cancer(ⅠB andⅡA)less than 4 cm.Materials and Methods:A total of 170 patients eligible for inclusion with cervical cancer from our hospital between 2016 and 2021 were retrospectively collected,and were divided into intermediate-risk and non-intermediate-risk groups based on postoperative pathological results.The cases were randomly divided into training group(n=119)and validation group(n=51)according to the ratio of 7:3.Analysis Kinetics software was used to extract radiomics characteristics.Multivariate Logistic regression analysis was used to develop the clinical model,the radiomics signature(Rad-score)and the clinical-radiomics model(the combined model).Performance of the three models were assessed by using receiver operating characteristic curves,calibration curves and decision curve analysis(DCA).Results:The combined pretreatment clinical-radiomics model could predict intermediate-risk cervical cancer(AUC=0.853,P<0.01).Sensitivity of the clinical-radiomics model was 85.5%and specificity was 78%.The combined model showed better performance than clinical model and no significant difference compared with radiomics model.Conclusions:The intermediate risk factors in early cervical cancer(ⅠB andⅡA)less than 4 cm can be predicted with the combined clinical-radiomics model based on dual-sequence MRI and clinical characteristics.Therefore,it could benefit individualized treatment decision-making.
作者
易芹芹
周宙
罗燕
钟淑媛
凌人男
YI Qinqin;ZHOU Zhou;LUO Yan;ZHONG Shuyuan;LING Rennan(Department of Radiology,Shenzhen People's Hospital(the Second Clinical Medical College of Jinan University,the First Affiliated Hospital,Southern University of Science and Technology),Shenzhen 508020,China)
出处
《磁共振成像》
CAS
CSCD
北大核心
2022年第4期124-127,136,共5页
Chinese Journal of Magnetic Resonance Imaging
基金
广东省医学科学技术研究基金项目(编号:B2020004)。
关键词
影像组学
宫颈癌
危险因素
磁共振成像
预测模型
radiomics
cervical cancer
risk factors
magnetic resonance imaging
predicting model
