OP51METABOLIC ACTIVITY OF THE INVASIVE MICROENVIRONMENT OF GLIOBLASTOMAS DETERMINES TIME PROGRESSION: A MULITOMODAL MRI STUDY

Abstract

INTRODUCTION: Virtually all glioblastoma(GBM)patients will die from progressive disease adjacent to the tumour resection cavity. Our previous work has shown that we can identify the invasive margin with diffusion tensor MR (DTI) and that we can use multimodal MRI to study the microenvironment of this region. This study aims to understand how the microenvironment influences progression and response to therapy. METHOD(S): 50 patients with GBM were recruited prospectively and imaged pre-operatively at 3 Tesla. Imaging involved conventional imaging, DTI, dynamic susceptibility perfusion imaging (to provide rCBV) and multivoxel short echo spectroscopy. Regions of interest from the invasive margin were identified on the DTI and perfusion and metabolite analysis performed. All patients subsequently underwent a 5-ALA guided resection followed by chemoradiotherapy. The progression free survival interval (PFS) was noted for all patients. ROC curves were produced to assess different parameters ability to predict PFS >1 year. Survival was assessed with Kaplan-Meier survival curves and a multivariate Cox regression model used to assess impact of imaging parameters with clinical parameters. RESULT(S): 30% of patients had PFS > 1 year. The best predictor of this was the NAA/Cho ratio (AUC 0.73) - a marker of cellular proliferation. A cut-off of 0.6 differentiated between PFS > 1 year and PFS < 1 year. Where the ratio was <0.6 (i.e. low proliferation) PFS = 230 days, where >0.6 PFS = 497 days (P = 0.01). CONCLUSION(S): Our data suggests that more metabolically active tumour margin is associated with better outcome. One hypothesis is that areas with low Cho/NAA have less proliferative activity and are less likely to respond to cytotoxic therapy.