摘要
目的探讨基于3D磁共振T2WI及动态对比增强DCE序列影像组学模型在鉴别ⅠA期宫颈癌与高级别鳞状上皮内瘤变(HSIL)中的价值。方法回顾性分析经手术病理证实的ⅠA期宫颈癌患者52例与HSIL患者35例,收集患者的临床资料及影像资料,利用汇医慧影平台将患者按10:3随机分成训练集和测试集,在矢状位T2WI和DCE上由三名放射科医生手动勾画宫颈,获得宫颈三维容积感兴趣区(VOI),提取影像组学特征。并进行数据降维。采用KNN算法建立模型并绘制受试者工作特性曲线(ROC曲线),对训练集和测试集ROC曲线下的总面积(AUC)、敏感度、特异度、准确率4个指标展开分析。结果T2WI模型、DCE模型以及T2WI联合DCE模型最终提取出的影像组学特征分别为23、11、28,训练集及测试集的AUC值分别为0.833、0.797,0.877、0.870,0.897,0.813。结论T2WI联合DCE图像影像组学模型比单一模型更具有临床价值,有助于评估病情,并针对病情所处阶段制定个体化治疗方案。
Objective To investigate the value of enhanced imaging model based on magnetic resonance T2WI and Dynamic contrast-enhanced magnetic resonance imaging in differentiating stage IA cervical cancer from high-grade squamous intraepithelial lesions(HSIL).Methods A retrospective analysis was performed on 52 patients with stage IA cervical cancer and 35 patients with HSIL confirmed by surgery and pathology in the First Affiliated Hospital of Bengbu Medical College from January 2019 to January 2023,imaging data and clinical data of patients were collected,and patients were randomly divided into training set and test set by 10:3 using Huiyi Huiying platform,the cervix was manually delineated by three radiologists on sagittal T2WI and DCE,the volume of interest(VOI)of the cervix was obtained,and the image omics features were extracted.Variance Threshold,single variable feature selection and minimum Absolute contraction operator method(Lasso)were used to reduce data dimension.K-nearest neighbor classification(KNN)algorithm is used to establish the model and draw the ROC curve.The area under cuver(AUC),sensitivity,specificity and accuracy of the ROC curve of the training set and test set are analyzed.Results The final image features extracted by T2WI model,DCE model and T2WI combined with DCE model were 23,11 and 28,respectively,and the AUC values of training set and test set of T2WI model and DCE model were 0.833,0.797,0.877 and 0.870,respectively,and the AUC values of the training set and the test set of T2WI combined with DCE image model were 0.897 and 0.813,respectively.Conclusion DCE model alone and T2WI combined with DCE image imaging model have certain value in distinguishing stage IA cervical cancer and HSIL before definite pathological results,and T2WI combined with DCE image imaging model has more clinical value,which is helpful for early clinical assessment of the disease stage and the development of individual treatment plan.
作者
邹文涛
陈艾琪
赵以惠
杜小萌
朱晓云
钱宝鑫
马宜传
ZOU Wen-tao;CHEN Ai-qi;ZHAO Yi-hui;DU Xiao-meng;ZHU Xiao-yun;QIAN Bao-xin;MA Yi-chuan(Department of Radiology,The First Affiliated Hospital of Bengbu Medical College,Bengbu 233099,Anhui Province,China;Department of Radiology,The Second Affiliated Hospital of Bengbu Medical College,Bengbu 233040,Anhui Province,China;School of Medical Imaging,Bengbu Medical College,Bengbu 233000,Anhui Province,China;Huiying Medical Technology Co.,Ltd.,Beijing 100192,China)
出处
《中国CT和MRI杂志》
2023年第12期132-135,共4页
Chinese Journal of CT and MRI
基金
基于3DMRI影像组学模型鉴别IA期宫颈癌与高级别鳞状上皮内瘤变的价值研究(Byycx23092)。
关键词
影像组学
宫颈癌
高级别鳞状上皮内瘤变
磁共振成像
Radiomics
Cervical Cancer
High-grade Squamous Intraepithelial Lesion
Magnetic Resonance Imaging
