Radiomics and MGMT promoter methylation for prognostication of newly diagnosed glioblastoma

Abstract

We attempted to establish a magnetic resonance imaging (MRI)-based radiomic model for stratifying prognostic subgroups of newly diagnosed glioblastoma (GBM) patients and predicting O (6)-methylguanine-DNA methyltransferase promotor methylation (pMGMT-met) status of the tumor. Preoperative MRI scans from 201 newly diagnosed GBM patients were included in this study. A total of 489 texture features including the first-order feature, second-order features from 162 datasets, and location data from 182 datasets were collected. Supervised principal component analysis was used for prognostication and predictive modeling for pMGMT-met status was performed based on least absolute shrinkage and selection operator regression. 22 radiomic features that were correlated with prognosis were used to successfully stratify patients into high-risk and low-risk groups (p = 0.004, Log-rank test). The radiomic high- and low-risk stratification and pMGMT status were independent prognostic factors. As a matter of fact, predictive accuracy of the pMGMT methylation status was 67% when modeled by two significant radiomic features. A significant survival difference was observed among the combined high-risk group, combined intermediate-risk group (this group consists of radiomic low risk and pMGMT-unmet or radiomic high risk and pMGMT-met), and combined low-risk group (p = 0.0003, Log-rank test). Radiomics can be used to build a prognostic score for stratifying high- and low-risk GBM, which was an independent prognostic factor from pMGMT methylation status. On the other hand, predictive accuracy of the pMGMT methylation status by radiomic analysis was insufficient for practical use.

XGBoost Improves Classification of MGMT Promoter Methylation Status in IDH1 Wildtype Glioblastoma

“…Conventionally, different feature selection methods in the radiomics domain have been extensively used to predict MGMT methylation status, such as LASSO [9,16], two-tailed Mann-Whitney U test [17], chi-square test [19], or principal component analysis (PCA) [20]. However, we evaluated the ability of F-score feature selection in finding an optimal set of radiomics features.…”

“…Table 4 shows the comparative performances between our model and the other state-of-the-art models [9,11,[16][17][18][19][20]. According to this comparison, our model had higher sensitivity and specificity than the works of Jiang et al [16], Crisi et al [17], Ahn et al [19], Korfiatis et al [11], and Sasaki et al [20]. Moreover, we have also used fewer features (nine features) than the work of Xi et al [9] (64 features) and reached a better performance.…”

MRI-Based Deep-Learning Method for Determining Glioma MGMT Promoter Methylation Status

“…4 Although multiple radiomic approaches have also been attempted for MGMT prediction, none, to date, have achieved accuracies sufficient for clinical viability. [5][6][7][8][9] Sasaki et al 10 attempted to establish an MR imaging-based radiomic model for predicting MGMT promoter status of the tumor, but it reached a predictive accuracy of only 67%. Wei et al 11 extracted radiomic features from the tumor and peritumoral edema using multisequence, postcontrast MR imaging but only achieved an accuracy of 51%-74% in predicting MGMT promoter methylation status in astrocytomas.…”

Impact of Inversion Time for FLAIR Acquisition on the T2-FLAIR Mismatch Detectability for IDH-Mutant, Non-CODEL Astrocytomas

“…The first cohort (Cohort 1) comprised of 94 MRI studies from 76 histologically and molecularly confirmed IDH mt, non-CODEL LrGGs from six institutions. This cohort was built utilizing the Kansai Molecular Diagnosis Network for CNS Tumors, a region-based brain tumor tissue collection network that includes Osaka University Hospital ( 8 , 9 , 17 ), and an LrGG cohort provided from the National Cancer Center Hospital. Supplementary Table 1 provides detailed information on this cohort.…”