Glioblastoma phenotypes identified on MRI predict molecular behavior

1. In a retrospective review of brain MRIs of two large cohorts of patients with glioblastoma multiforme (GBM), imaged tumors made up three distinct phenotypic clusters: pre-multifocal, spherical, and rim enhancing, each of which reflect distinct molecular characteristics which corresponded to differential survival and response to therapy.

2. Pre-multifocal GBM was associated with the poorest survival and displayed increased c-Kit expression. Spherical GBM was associated with intermediate survival and displayed decreased c-Kit expression and IL-6 and Ang/Tie2 signaling. Rim-enhancing GBM had the most favorable survival and displayed increased FOXA, PDGFR and VEGFR expression.

Evidence Rating Level: 2 (Good)

Study Rundown: GBM is the most common primary brain malignancy in adults associated with significant mortality and poor response to therapy. Currently, diagnosis and individual prognostication relies on tissue-based molecular biomarkers, which necessitates an invasive brain biopsy. To date, there has been a limited research on the use of non-invasive imaging features as surrogates for molecular biomarkers in characterization of GBM  to predict both outcomes and response to therapy. The current study sought to determine and validate the magnetic resonance imaging (MRI) based GBM phenotypes which represent different molecular backgrounds and their behaviors. The authors performed quantitative imaging analysis of brain MRIs of a large cohort of GBM patients to segment tumor features which were subsequently clustered into three common phenotypic groups by a consensus clustering method. These imaging phenotypes were subsequently validated on a second large GBM cohort from The Cancer Genome Atlas (TCGA). At the conclusion of the study, three distinct imaging phenotypes were determined and validated: pre-multifocal, spherical, and rim enhancing based on quantitative MRI-derived features. Interestingly, each imaging phenotype mapped to unique molecular signaling pathways known in the pathogenesis of GBM based on cross-correlation to gene expression data from TCGA. Furthermore, there was a significant difference in the overall survival of patients between the three imaging phenotype groups, and their response to standard therapeutic regimens. The results of this study support the use of these imaging phenotypes as a potential non-invasive, image-based biomarker for prognostication as well as additional avenues for targeted molecular therapies for GBM. The study is limited by its retrospective and the use of a strict inclusion criterion of unifocal, unilateral tumors, which may reduce the external generalizability of these results. Additional prospective, randomized controlled trials with targeted interventions are needed to confirm these associations.

Click to read the study in Science Translational Medicine

Relevant Reading: Non–Small Cell Lung Cancer: Identifying Prognostic Imaging Biomarkers by Leveraging Public Gene Expression Microarray Data—Methods and Preliminary Results

In-Depth [retrospective study]: This study was a retrospective review of two large GBM patient cohorts in the United States, one recruited from patients with GBM from Stanford University Hospital (n = 121), and another from The Cancer Genome Atlas (TCGA; n = 144), including only patients with de novo GBM, age > 18 years, availability of preoperative MRI data, and presence of solitary, unilateral tumors only. Quantitative imaging analysis was performed on the Stanford cohort (used as the development cohort) for subtype identification via k-means consensus clustering. Identified subgroups were validated with the second patient cohort using in-group proportion (IGP) statistics. Each subgroup was subsequently mapped to known molecular signaling pathways from the Cancer Genome Atlas cohort as well as overall survival and response to standard Stupp protocol therapy. At the conclusion of the study, 3 imaging phenotypes were isolated: 1) pre-multifocal, 2) spherical, and 3) rim-enhancing. There was a significant difference in survival between the three subtypes (p = 0.004) with pre-multifocal group having the lowest survival rate and rim-enhancing group having the highest survival rate. Overall, pre-multifocal GBM was associated with the poorest survival and displayed increased c-Kit stem cell factor receptor pathway expression. Spherical GBM was associated with intermediate survival and displayed decreased c-Kit expression and IL-6 and Ang/Tie2 signaling, among 21 total molecular expression differences. Rim-enhancing GBM had the most favorable survival and displayed increased FOXA, PDGFR and VEGFR expression, among 31 different molecular expression differences.

Image: PD

©2015 2 Minute Medicine, Inc. All rights reserved. No works may be reproduced without expressed written consent from 2 Minute Medicine, Inc. Inquire about licensing here. No article should be construed as medical advice and is not intended as such by the authors or by 2 Minute Medicine, Inc.