基于视觉能力伦勃朗图像MRI特征模型在高级别胶质瘤IDH-1突变中的预测价值分析  

Analysis of the predictive value of Visually Accesable Rembrandt Images MRI feature model in high-grade glioma with IDH-1 mutations

作　　者：徐蕊 赵伟 罕迦尔别克·库锟 帕哈提·吐逊江 谢超 常一凡 王云玲 XU Rui;ZHAO Wei;Hanjiaerbieke·Kukun;Pahati·Tuxunjiang;XIE Chao;CHANG Yifan;WANG Yunling(Imaging Center,the First Affiliated Hospital of Xinjiang Medical University,Urumqi 830054,China)

机构地区：[1]新疆医科大学第一附属医院影像中心,新疆乌鲁木齐830054

出　　处：《实用放射学杂志》2025年第1期5-8,共4页Journal of Practical Radiology

基　　金：科技创新领军人才(2023TSYCLJ0027);新疆维吾尔自治区自然科学重点项目(2024D01D20)。

摘　　要：目的探讨基于构建视觉能力伦勃朗图像(VASARI)MRI特征模型预测高级别胶质瘤(HGG)中异柠檬酸脱氢酶-1(IDH-1)突变的应用价值。方法将242例HGG患者按照7︰3的比例随机分为训练集和验证集,提取患者的VASARI特征,评价HGG中IDH-1突变型及野生型VASARI特征的统计学差异,采用最小绝对收缩和选择算子(LASSO)回归方法对VASARI特征进行降维,采用逻辑回归(LR)机器学习模型对单个VASARI特征及联合特征构建预测模型,利用受试者工作特征(ROC)曲线评估模型的预测性能。结果在VASARI特征中有11个特征与HGG的IDH-1突变相关,差异均有统计学意义(P<0.05)。最终筛选出7个密切相关特征,对7个VASARI特征联合后建立的预测模型曲线下面积(AUC)较高,训练集为0.908,验证集AUC为0.872。结论VASARI MRI特征模型能够较好地预测HGG的IDH-1突变,并且预测效能较高,具有较大的应用价值。Objective To explore the application value of constructing Visually Accesable Rembrandt Images(VASARI)MRI feature model for the prediction of isocitrate dehydrogenase-1(IDH-1)mutations in high-grade glioma(HGG).Methods A total of 242 patients with HGG were randomly divided into training set and validation set according to the ratio of 7︰3,the VASARI features of the patients were extracted to evaluate the statistical difference between IDH-1 mutant and wild-type VASARI features of the HGG.The least absolute shrinkage and selection operator(LASSO)regression method was used to reduce the dimension of VASARI features and the logistic regression(LR)machine learning model was used to construct the prediction model for single VASARI feature and combined features.The receiver operating characteristic(ROC)curve was used to evaluate the predictive performance of the model.Results A total of 11 VASARI features were associated with IDH-1 mutations in HGG,with statistically significant differences(P<0.05).Seven closely related features were finally screened out,and the area under the curve(AUC)of the prediction model established after the combination of the seven VASARI features was higher,and the AUC of training set was 0.908,and the AUC of validation set was 0.872.Conclusion The VASARI MRI feature model can better predict IDH-1 mutations of HGG with high predictive efficacy and has greater application value.

关 键 词：视觉能力伦勃朗图像 胶质瘤 异柠檬酸脱氢酶 磁共振成像 

分 类 号：R339.146[医药卫生—人体生理学] R739.41[医药卫生—基础医学] R445.2