1973. Underlying Biological Differences in Glioblastoma Tissue that Enhances With Gadolinium Compared to Ferumoxytol and Flair
Authors * Denotes Presenting Author
  1. Jared Stoller *; Oregon Health and Science University
  2. Cymon Kersch; Oregon Health and Science University
  3. Leslie Muldoon; Oregon Health and Science University
  4. Prakash Ambady; Providence Portland Medical Center
  5. Christina Harrington; Oregon Health and Science University
  6. Edward Neuwelt; Oregon Health and Science University
  7. Ramon Barajas; Oregon Health and Science University
Gadolinium enhanced MRI is standard of care for monitoring glioblastoma tumor response but is of limited utility for defining immune mediated responses which is critical in the era of immunotherapies. The purpose of this study is to assess if noninvasive image phenotypes can capture variance in tumor immune status and other essential biological traits that impact both treatment selection and monitoring of treatment response. We evaluated associations between transcriptional patterns that define immune responses, and MRI T1-enhancement using gadolinium contrast, ferumoxytol contrast, and T2-flair sequences.

Materials and Methods:
Eight patients with glioblastoma IDH-wild type glioblastoma were included in this prospective study. Each patient underwent MRI prior to tumor resection. Tissue sampling was performed using stereotactic image guided technique allowing for qualitative assessment of T1-shortening and T2-flair prolongation. RNA was extracted from the tissue samples and analyzed by gene expression array for transcript identification by the Oregon Health & Science University Gene Profiling Shared Resource. The gene identifications were dichotomized based on sample site; enhancing/FLAIR hyperintense or not. Samples were then processed using Gene Set Enrichment Analysis to find the highest degree of correlation between the imaging methods and Hallmark gene sets. Gene sets with both a normalized enrichment score (NES) >2 and a false discovery rate (FDR) q-value < 0.05 were considered significant results.

Nineteen gene sets registered a high correlation with gadolinium contrast, six of these falling under the process category of immunity. Additionally, the gene sets expressed are consistent with known biological factors associated with general malignant features. Comparable results are seen with ferumoxytol enhancement with 16 highly correlated gene sets, seven of which fall under the process category of immunity. Flair hyperintensity was linked to 13 gene sets, seven of which involve the immune system. Although all three metrics found correlations with the same seven immune system linked gene sets, gadolinium contrast was the most correlated with proliferation, development, and pathway genes, five of which exceeded an NES of 2.5. Ferumoxytol enhancement had the highest correlation with immune gene sets, notably IFN, IFN-a, and inflammatory response with six of the seven gene sets exceeding an NES of 2.5. Flair hyperintensity had two immune associated gene sets with an NES > 2.5 meaning that both ferumoxytol enhancement and Flair hyperintensity more strongly correlated with immune gene sets than gadolinium enhancement.

Each of the imaging metrics viewed are correlated with seven enhancing enriched genes dealing with the immune system, with ferumoxytol enhancement and Flair hyperintensity having higher correlations with the immune response than gadolinium enhancement. Nonenhancing enriched genes were also evaluated for all three metrics however, none were highly correlated (i.e., NES < -2.0 and FDR q-value < 0.05).