Purpose To investigate whether radiomic features at MRI improve survival prediction in patients with glioblastoma multiforme (GBM) when they are integrated with clinical and genetic profiles. Materials and Methods Data in patients with a diagnosis of GBM between December 2009 and January 2017 (217 patients) were retrospectively reviewed up to May 2017 and allocated to training and test sets (3:1 ratio). Radiomic features (n = 796) were extracted from multiparametric MRI. A random survival forest (RSF) model was trained with the radiomic features along with clinical and genetic profiles (O-6-methylguanine-DNA-methyltransferase promoter methylation and isocitrate dehydrogenase 1 mutation statuses) to predict overall survival (OS) and progression-free survival (PFS). The RSF models were validated on the test set. The incremental values of radiomic features were evaluated by using the integrated area under the receiver operating characteristic curve (iAUC). Results The 217 patients had a mean age of 57.9 years, and there were 87 female patients (age range, 22-81 years) and 130 male patients (age range, 17-85 years). The median OS and PFS of patients were 352 days (range, 20-1809 days) and 264 days (range, 21-1809 days), respectively. The RSF radiomics models were successfully validated on the test set (iAUC, 0.652 [95% confidence interval {CI}, 0.524, 0.769] and 0.590 [95% CI: 0.502, 0.689] for OS and PFS, respectively). The addition of a radiomics model to clinical and genetic profiles improved survival prediction when compared with models containing clinical and genetic profiles alone (P = .04 and .03 for OS and PFS, respectively). Conclusion Radiomic MRI phenotyping can improve survival prediction when integrated with clinical and genetic profiles and thus has potential as a practical imaging biomarker. © RSNA, 2018 Online supplemental material is available for this article. See also the editorial by Jain and Lui in this issue.

OBJECTIVE: Following resection of glioblastoma (GBM), microscopic remnants of the GBM tumor remaining in nearby tissue cause tumor recurrence more often than for other types of tumors, even after gross-total resection (GTR). Although surgical oncologists traditionally resect some of the surrounding normal tissue, whether further removal of nearby tissue may improve survival in GBM patients is unknown. In this single-center retrospective study, the authors assessed whether lobectomy confers a survival benefit over GTR without lobectomy when treating GBMs in the noneloquent area.METHODS: The authors selected 40 patients who had undergone GTR of a histopathologically diagnosed isocitrate dehydrogenase (IDH)–wild type GBM in the right frontal or temporal lobe and divided the patients into 2 groups according to whether GTR of the tumor involved lobectomy, defined as a supratotal resection (SupTR group, n = 20) or did not (GTR group, n = 20). Progression-free survival (PFS), overall survival (OS), and Karnofsky Performance Status (KPS) scores were compared between groups (p ≤ 0.05 for statistically significant differences).RESULTS: The median postoperative PFS times for each group were as follows: GTR group, 11.5 months (95% CI 8.8–14.2) and SupTR group, 30.7 months (95% CI 4.3–57.1; p = 0.007). The median postoperative OS times for each group were as follows: GTR group, 18.7 months (95% CI 14.3–23.1) and SupTR group, 44.1 months (95% CI 25.1–63.1; p = 0.040). The mean postoperative KPS scores (GTR, 76.5; SupTR, 77.5; p = 0.904) were not significantly different. In multivariate analysis, survival for the SupTR group was significantly longer than that for the GTR group in terms of both PFS (HR 0.230; 95% CI 0.090–0.583; p = 0.002) and OS (HR 0.247; 95% CI 0.086–0.704; p = 0.009).CONCLUSIONS: In cases of completely resectable, noneloquent-area GBMs, SupTR provides superior PFS and OS without negatively impacting patient performance.

Object: The methylation status of the methylguanine methyltransferase (MGMT) promoter has been associated with treatment response in glioblastoma. The authors aimed to assess whether MGMT methylation status can be predicted by dynamic contrast-enhanced (DCE) MRI and diffusion tensor imaging (DTI). Methods: This retrospective study included 43 patients with pathologically diagnosed glioblastoma who had undergone preoperative DCE-MRI and DTI and whose MGMT methylation status was available. The imaging features were qualitatively assessed using conventional MR images. Regions of interest analyses for DCE-MRI permeability parameters (transfer constant [Ktrans], rate transfer coefficient [Kep], and volume fraction of extravascular extracellular space [Ve]) and DTI parameters (apparent diffusion coefficient [ADC] and fractional anisotropy [FA]) were performed on the enhancing solid portion of the glioblastoma. Chi-square or Mann-Whitney tests were used to evaluate relationships between MGMT methylation and imaging parameters. The authors performed receiver operating characteristic curve analysis to find the optimal cutoff value for the presence of MGMT methylation. Results: MGMT methylation was not significantly associated with any imaging features on conventional MR images. Ktrans values were significantly higher in the MGMT methylated group (median 0.091 vs 0.053 min⁻¹, p = 0.018). However, Kep, Ve, ADC, and FA were not significantly different between the 2 groups. The optimal cutoff value for the presence of MGMT methylation was Ktrans > 0.086 min⁻¹ with an area under the curve of 0.756, a sensitivity of 56.3%, and a specificity of 85.2%. Conclusions: Ktrans may serve as a potential imaging biomarker to predict MGMT methylation status preoperatively in glioblastoma; however, further investigation with a larger cohort is necessary.