MRI-based Radiomics For Prediction Of Molecular Subtypes And Survival In Glioma

Purpose: Gliomas can be classified into five molecular subgroups based on the status of IDH mutation,1p/19 q codeletion,and TERT promoter mutation,whereas they need to be obtained by biopsy or surgery.We aimed to use MRI-based radiomic approach to noninvasively predict the five molecular subgroups and assess its prognostic value in glioma.Methods: We retrospectively identified 357 patients with WHO II-IV gliomas.We extracted radiomic features(n = 12222)from preoperative MRI.Radiomic Features were selected by the least absolute shrinkage and selection operator(LASSO).Single-layered radiomic signatures were generated based on one MRI sequence and one machine learning classifier.Multiparametric MRI radiomic models were built based on the significant radiomic signatures.The combined models were constructed by incorporating preoperative clinical features into the multiparametric MRI radiomic models.We compared the prognostic performance of predictive molecular subgroups with actual molecular subgroups in predicting progression-free survival(PFS)and overall survival(OS).We developed prognostic nomograms based on the predictive molecular subgroups and clinicopathologic data to predict PFS and OS of patients using univariate and multivariate Cox regression analysis.Results: Age,tumor location,and LASSO-based multiparametric MRI radiomic model yielded the highest performance in predicting IDH mutation status(AUC,0.855;95% confidence interval [CI]: 0.792-0.907).Decision tree-based multiparametric MRI radiomic model performed the best in predicting 1p/19 q codeletion(AUC,0.796;95% CI: 0.705-0.871)and TERT promoter mutation(AUC,0.849;95% CI: 0.786-0.907).The predictive molecular subgroups were comparable to actual molecular subgroups in predicting PFS and OS.The prognostic nomogram could individually predict PFS and OS in glioma patients.Conclusion: Multiparametric MRI-based radiomics can be useful for noninvasively predicting molecular subgroups and survival in patients with glioma.