Imaging genomic mapping of an invasive MRI phenotype predicts patient outcome and metabolic dysfunction: a TCGA glioma phenotype research group project

BMC Medical Genomics Imaging genomic mapping of an invasive MRI phenotype predicts patient outcome and metabolic dysfunction: a TCGA glioma phenotype research group project Rivka R Colen 1 Mark Vangel 4 Jixin Wang 1 David A Gutman 3 Scott N Hwang 3 Max Wintermark 9 Rajan Jain 8 Manal Jilwan-Nicolas 9 James Y Chen 6 7 Prashant Raghavan 9 Chad A Holder 3 Daniel Rubin 11 Eric Huang 10 Justin Kirby 10 John Freymann 10 Carl C Jaffe 5 Adam Flanders 12 TCGA Glioma Phenotype Research Group Pascal O Zinn 2 0 FCT 16. 5037, Houston, TX 77030 , USA 1 Department of Diagnostic Radiology, M. D. Anderson Cancer Center , 1400 Pressler St; Unit 1482, Rm 2 Department of Neurosurgery, Baylor College of Medicine , 1 Baylor Plaza, Houston, TX 77030 , USA 3 Department of Radiology, Emory University , Atlanta, GA , USA 4 Massachussets General Hospital , Boston, MA , USA 5 NCI/NIH , Rockville, MD , USA 6 Department of Radiology, San Diego Medical Center , San Diego, CA , USA 7 University of California San Diego Health System , San Diego, CA , USA 8 Department of Radiology, New York University Medical Center , New York, NY , USA 9 University of Virginia , Charlottesville, VA , USA 10 Clinical Monitoring Research Program, Leidos Biomedical Research Inc., Frederick National Laboratory for Cancer Research , Frederick, MD 21702 , USA 11 Stanford University , Stanford, CA , USA 12 Department of Radiology, Thomas Jefferson University Hospital , Philadelphia, PA , USA Background: Invasion of tumor cells into adjacent brain parenchyma is a major cause of treatment failure in glioblastoma. Furthermore, invasive tumors are shown to have a different genomic composition and metabolic abnormalities that allow for a more aggressive GBM phenotype and resistance to therapy. We thus seek to identify those genomic abnormalities associated with a highly aggressive and invasive GBM imaging-phenotype. Methods: We retrospectively identified 104 treatment-nave glioblastoma patients from The Cancer Genome Atlas (TCGA) whom had gene expression profiles and corresponding MR imaging available in The Cancer Imaging Archive (TCIA). The standardized VASARI feature-set criteria were used for the qualitative visual assessments of invasion. Patients were assigned to classes based on the presence (Class A) or absence (Class B) of statistically significant invasion parameters to create an invasive imaging signature; imaging genomic analysis was subsequently performed using GenePattern Comparative Marker Selection module (Broad Institute). Results: Our results show that patients with a combination of deep white matter tracts and ependymal invasion (Class A) on imaging had a significant decrease in overall survival as compared to patients with absence of such invasive imaging features (Class B) (8.7 versus 18.6 months, p < 0.001). Mitochondrial dysfunction was the top canonical pathway associated with Class A gene expression signature. The MYC oncogene was predicted to be the top activation regulator in Class A. Conclusion: We demonstrate that MRI biomarker signatures can identify distinct GBM phenotypes associated with highly significant survival differences and specific molecular pathways. This study identifies mitochondrial dysfunction as the top canonical pathway in a very aggressive GBM phenotype. Thus, imaging-genomic analyses may prove invaluable in detecting novel targetable genomic pathways. Radiogenomics; MRI segmentation; Glioblastoma; Imaging genomics; Invasion; Biomarker - Background Recent advances in high throughput whole-genome glioblastoma (GBM) analyses have led to an increase in the understanding of gliomagenesis and elucidation of new molecular pathways [1,2]. However, the dismal prognosis of GBM remains largely unchanged with the median survival less than 2 years [3]. Heterogeneity in cellular composition, differential genomic expression profiles and therapy-resistant cancer stem cells within a single tumor (intra-tumor heterogeneity) and with the same tumor across different individuals (inter-individual heterogeneity) contribute largely to the lack of advancement in therapeutic approaches for these heterogeneous tumors and lie central to the failure of significant progress in the treatment of GBM [4,5]. Cellular invasion remains a significant cause of therapy failure as diffuse dissemination of tumor cells throughout the entire GBM brain including the normal appearing white matter precludes clean surgical margins. Moreover, glioma stem-like cells, thought to be involved in invasion, do not respond to current chemotherapeutic agents which target the active tumor core [6,7]. Magnetic resonance imaging (MRI) has been shown to non-invasively reflect the underlying tumor biological and pathological processes [7,8], tumor microenvironment [9], and genomic cancer composition [10-13]. Multiple preoperative imaging characteristics consistent with invasive tumor growth are documented [13,14]. Qualitatively, these include the 1) presence of either T1 contrast enhancement or increase T2/FLAIR hyperintensity involving the internal capsule, corpus callosum (unilateral, bilateral, or contralateral) or brainstem; 2) the presence of ependymal enhancement; and 3) the presence of pial enhancement [14]. Quantitatively, 3D volumetry of the peritumoral non-enhancing FLAIR hyperintensity has been validated to reflect an increase in gene signatures promoting cellular invasion and angiogenesis, as shown by our group previously [13]. In this current study, we sought to identify the invasive MRI characteristics in GBM and the implicated genes and microRNAs associated with qualitative invasive imaging signatures. Using our large-scale genomic database provided by The Cancer Genome Atlas (TCGA) and the imaging of the corresponding TCGA patients provided by The Cancer Imaging Archive (TCIA), we performed an MRI radiophenotype screen to identify key genes and molecular pathways associated with a very aggressive GBM radiophenotype. Methods The collection of the original material and data of TCGA and TCIA study was conducted in compliance with all applicable laws, regulations and policies for the protection of human subjects, and any necessary approvals, authorizations, human subject assurances, informed consent documents, and IRB approvals were obtained [2]. Patient population and TCGA and TCIA We identified 104 (female 38: male 66; mean age 58 years; age range from 14 to 84 years) treatment-nave GBM patients from TCGA whom had gene expression profiles and corresponding pretreatment MR imaging available in the TCIA. The TCGA is a National Cancer Institute (NCI) sponsored publicly available resource which has produced a multi-dimensional genomic and clinical data set of GBM and other cancers [2]. Image data used in this research were obtained from TCIA (http://cancerimagingarchive.net/) sponsored by the Cancer Imaging Program, DCTD/NCI/NIH [15]. The latter archive repository contains the imaging corresponding to the patients of th (...truncated)