The value of contrast enhanced MRI radiomics in predicting the IDH 1 genotype in high-grade gliomas
10.3760/cma.j.cn112149-20190608-00257
- VernacularTitle:增强MRI影像组学在高级别胶质瘤IDH 1基因型预测方面的价值
- Author:
Na LIU
1
;
Qinglan SUI
;
Xuejun LIU
;
Xiaoming ZHOU
;
Weihua FENG
;
Bao WANG
;
Yingchao LIU
;
Lei NIU
Author Information
1. 青岛大学附属医院放射科 266003
- From:
Chinese Journal of Radiology
2020;54(5):445-449
- CountryChina
- Language:Chinese
-
Abstract:
Objective:To explore the predictive value of a radiomics model based on preoperative contrast enhanced MRI in the assessment of the isocitrate dedydrogenase 1 (IDH 1) genotype in high-grade glioma.Methods:A retrospective analysis was performed on a dataset including 182 patients with high-grade glioma confirmed by surgical pathology between December 2012 and January 2018 in the Affiliated Hospital of Qingdao University. There were 79 patients with IDH1-mutant glioma (45 cases with WHO grade Ⅲ, 34 with WHO grade Ⅳ) and 103 with IDH 1 wild-type glioma (33 cases with WHO grade Ⅲ, 70 cases with WHO grade Ⅳ). All patients had complete preoperative brain contrast enhanced MRI.The cases were divided into a training dataset and a validation dataset at a ratio of 7∶3 using stratified random sampling. Radiomic features were initially extracted using A.K (Analysis Kit, GE healthcare) software, and were selected and excluded using Kruskal-Wallis and Spearman analyses. Using R softwear " GLM" function, the Lasso-logistic model was finally conducted to obtain the optimized subset of the feature to build the radiomics model, and the model was then tested with cross-validation. Receiver operating characteristic (ROC) curve analysis was performed to evaluate the performance of the model in differentiating IDH1-mutant type and wild-type gliomas.Results:The radiomics model showed good performance in IDH genotype differentiation in both the training dataset (AUC 0.870, 95% CI: 0.754 to 0.855, accuracy rate 79.8%, sensitivity 85.5%, specificity 75.4%, positive predictive value 0.734, negative predictive value 0.867) and the validation dataset (AUC 0.860, 95% CI: 0.690 to 0.913, accuracy rate 78.9%, sensitivity 91.3%, specificity 69.0%, positive predictive value 0.700, negative predictive value 0.909).Conclusion:The radiomics model based on the preoperative enhanced MR can provide a way to predict the IDH1 genotype in high-grade gliomas.
