Early perfusion MRI predicts survival outcome in patients with recurrent glioblastoma treated with bevacizumab and carboplatin.

Bevacizumab, an anti-angiogenic agent, is FDA-approved for use in patients with recurrent glioblastoma multiforme (rGBM). The radiologic evaluation of tumor response to bevacizumab is complex and there is no validated method of monitoring tumor vascularity during therapy. We evaluated perfusion-weighted MR imaging (PWI) in our cohort of patients enrolled in the CABARET trial, which examined the effectiveness of bevacizumab with or without carboplatin in patients with rGBM. Pre-treatment and early follow-up (4- and 8-week) PWI were used to calculate relative cerebral blood volume (rCBV) histogram statistics of the contrast-enhancing and FLAIR hyperintense tumor volumes. A novel rCBV measurement (load) was developed to estimate the total volume of perfused tumor blood vessels. Changes in all rCBV measures were examined for correlations with progression-free (PFS) and overall survival (OS). All of our 15 patients enrolled in the CABARET trial were included. Median PFS and OS were 23 and 45 weeks respectively. Kaplan-Meier analysis of pre-treatment PWI revealed an 18 week reduction in median OS in patients with high tumor rCBV (p = 0.031). Changes in rCBV measures, especially load, correlated significantly with PFS and OS at both follow-up time-points. Patients with the greatest reduction in rCBVload by 8-weeks of therapy had a significantly increased median OS (30 weeks; p = 0.013). PWI may be of significant clinical utility in managing patients with rGBM, particularly those treated with anti-angiogenic agents such as bevacizumab. These findings need to be confirmed prospectively in larger studies.

Preoperative biomarkers of tumour vascularity are elevated in patients with glioblastoma multiforme.

We investigated the correlation between the circulating and imaging biomarkers of tumour vascularity, and examined whether they are prognostic of outcomes in patients with glioblastoma multiforme (GBM). Despite the increasing use of anti-angiogenic agents within neuro-oncology, there are still no validated biomarkers to monitor for a treatment response or relapse. The pre- and postoperative circulating endothelial cell (CEC) and progenitor cell (CEP) levels were assessed. Preoperative perfusion-weighted MRI (PWI) was also performed, and the relative cerebral blood volume (rCBV) histogram statistics of the contrast-enhancing tumour were analysed. A novel PWI parameter (rCBVload) was developed to estimate the total volume of perfused tumour vessels, and it was hypothesised that this parameter would correlate with CEC and CEP concentrations. In total, 24 GBM patients were included. The mean preoperative CEC concentration was significantly higher in GBM patients than the controls (p=0.019), and it then declined significantly postoperatively (p=0.009). The preoperative CEP levels were significantly correlated with the median tumour rCBV (Spearman rank-order coefficient=0.526; p=0.039). Neither CEC nor CEP was correlated with the total tumour vessel volume, as measured by rCBVload. None of the biomarkers that were investigated showed a significant correlation with progression-free or overall survival. We conclude that CEC are potentially useful biomarkers to monitor GBM patients during treatment. We found that CEC are increased in the presence of GBM, and that CEP levels appear to be proportional to tumour vascularity, as measured on PWI. However, in this study, none of the biomarkers of GBM vascularity were highly prognostic of patient outcomes.