Preoperative and Noninvasive Prediction of Gliomas Histopathological Grades and IDH Molecular Types Using Multiple MRI Characteristics

Background and PurposeGliomas are one of the most common tumors in the central nervous system. This study aimed to explore the correlation between MRI morphological characteristics, apparent diffusion coefficient (ADC) parameters and pathological grades, as well as IDH gene phenotypes of gliomas. MethodsPreoperative MRI data from 166 glioma patients with pathological confirmation were retrospectively analyzed to compare the differences of MRI characteristics and ADC parameters between the low-grade and high-grade gliomas (LGGs vs. HGGs), IDH mutant and wild-type gliomas (IDHmut vs. IDHwt). Multivariate models were constructed to predict the pathological grades and IDH gene phenotypes of gliomas and the performance was assessed by the receiver operating characteristic (ROC) analysis. ResultsTwo multivariable logistic regression models were developed by incorporating age, ADC parameters, and MRI morphological characteristics to predict pathological grades, and IDH gene phenotypes of gliomas, respectively. The Noninvasive Grading Model classified tumor grades with areas under the ROC curve (AUROC) of 0.934 (95% CI=0.895-0.973), sensitivity of 91.2%, and specificity of 78.6%. The Noninvasive IDH Genotyping Model differentiated IDH types with an AUROC of 0.857 (95% CI=0.787-0.926), sensitivity of 88.2%, and specificity of 63.8%. ConclusionMRI features were correlated with glioma grades and IDH mutation status. Multivariable logistic regression models combined with MRI morphological characteristics and ADC parameters may provide a noninvasive and preoperative approach to predict glioma grades and IDH mutation status.

Automated summary

This study explored the correlation between MRI morphological characteristics, apparent diffusion coefficient (ADC) parameters, and pathological grades as well as IDH gene phenotypes of gliomas. Multivariable logistic regression models combining MRI features and ADC parameters were developed to predict glioma grades and IDH mutation status, providing a noninvasive and preoperative approach.