摘要
目的探讨基于MRI T1增强影像的影像组学预测较低级别胶质瘤(LerGG;包括WHO分级Ⅱ级弥漫性胶质瘤和间变型胶质瘤)异柠檬酸酸脱氢酶(IDH)基因型的价值。方法回顾性分析2017年1月至2020年6月手术切除并经术后病理证实的170例LerGG胶质瘤的临床资料和影像学资料,根据术前MRI T1增强影像,应用影像组学方法构建影像组学预测模型。结果170例中,WHO分级Ⅱ级例60例,Ⅲ级110例。从MRI T1增强影像的感兴趣区中共提取了1595个影像组学特征,降维分析后筛选5个影像组学标签,并构建预测模型,验证集和训练集分析显示模型预测IDH突变的曲线下面积分别为0.84、0.82,准确度分别为79%、80%,敏感度分别为81%、88%,特异度分别为75%、66%。结论基于MRI T1增强影像的影像组学模型对WHO分级Ⅱ~Ⅲ级胶质瘤的IDH型具有较好的预测能力。
Objective To investigate the value of radiomics model based on T1-weighted contrast-enhanced MRI for predicting iso-citrate dehydrogenase(IDH)gene genotype in lower grade gliomas(LerGG;including WHO gradeⅡdiffuse gliomas and intermediate-grade gliomas).Methods The clinical and imaging data of 170 patients with LerGG who were surgically resected from January 2017 to June 2020 and confirmed by postoperative pathological examination were retrospectively analyzed.The prediction model was established using radiomics methods based on the preoperative T1-weighted contrast-enhanced MRI.Results Of 170 patients with gliomas,60 patients had gradeⅡgliomas and 110 patients had gradeⅢgliomas.A total of 1595 radiomics features were extracted from the regions of interest of preoperative T1-weighted contrast-enhanced MRI.After dimension reduction analysis,5 radiomics tags were screened out,and the prediction model of IDH mutation was constructed.The verification and training sets analyses showed that the area under the curve of the model to predict IDH mutation was 0.84 and 0.82,respectively;the accuracy was 79%and 80%,respectively;the sensitivity was 81%and 88%,respectively;the specificity was 75%and 66%,respectively.Conclusions The radiomics model based on T1-weighted contrast-enhanced MRI has a certain value to predict IDH genotype of patients with LerGG.
作者
赵志勇
张婧
曹云太
张国晋
孙建清
李昇霖
周俊林
ZHAO Zhi-yong;ZHANG Jing;CAO Yun-tai;ZHANG Guo-jin;SUN Jian-qing;LI Sheng-lin;ZHOU Jun-lin(Department of Neurosurgery,Lanzhou University Second Hospital,Lanzhou 730030,China;Medical Imaging Center,The Fifth Affiliated(Zhuhai)Hospital of Zunyi Medical University,Zhuhai 276000,China;Medical Imaging Center,Qinghai University Affiliated Hospital,Xining 810001,China;Medical Imaging Center,Sichuan Academy of Medical Sciences,Sichuan Provincial People's Hospital,Chengdu 610020,China;United Imaging Medical Technology Co.,Ltd.,Shanghai 201800,China;Department of Radiology,Lanzhou University Second Hospital,Lanzhou 730030,China)
出处
《中国临床神经外科杂志》
2023年第3期145-149,共5页
Chinese Journal of Clinical Neurosurgery
基金
甘肃省卫生健康行业科研计划项目(GSWSKY2021-006)。
作者简介
通讯作者:周俊林,E-mail:ery_zhoujl@lzu.edu.cn。