School of Medicine

In magnetic resonance imaging (MRI) gadolinium-based contrast agents are administered to identify areas of blood brain barrier compromise that result in brain tumor enhancement. The most common primary brain tumor, glioblastoma, is characterized on MRI by its hterogeneous uptake of gadolinium contrast. Both the heterogeneity and magnitude of contrast enhancement helps clinicians to distinguish blioblastoma from other primary brain tumors as well as intracranial metastases. Unfortunately, the precise molecular mechanisms which govern unique patterns of contrast enhancement remain poorly understood. Furthermore, it has been observed that blioblastoma tumors demonstrate widely different imaging characteristics anddegrees of contrast enhancement. Previous studies have demonstrated relationships between contrast enhancement and overall survival as well as tumor gene expression of pathways related to cell division and hypoxia. However, these studies were limited by small sample sizes and coarse qualitative metrics. Using The Cancer Imaging Archive (TCIA), a method of automated segmentation developed by our laboratory, and matched tumor genomic data from The Cancer Genome Atlas (TCGA) this study examined the genomic correlates of quantitative measures of contrast enhancement (CE). Metrics for contrast enhancement included relative CE (rCE), CE heterogeneity, and ratio of total CE volume to the volume of the filled CE ring (CE:CEfilled). Analyses revealed relationships between rCE and inflammation, CE heterogeneity and angiogenesis, and CE:CEfilled with cell division. Age-adjusted cox regression found a significant overall survival benefit to patients with higher CE heterogeneity (HR = 0.36, p < 0.01). Uncovering relationships between contrast enhancement and genomics may better characterize glioblastomas and improve the understanding of its pathophysiology, which may lead to future therapies and better patient care.