The preliminary radiogenomics association between MR perfusion imaging parameters and genomic biomarkers, and their predictive performance of overall survival in patients with glioblastoma.

The radiogenomics association of neovascularization is important for overall survival (OS) in glioblastoma patients and remains unclear. The purpose of this study is to assess the association between MR perfusion imaging derived parameters and genomic biomarkers of glioblastoma, and to evaluate their prognostic value. This retrospective study enrolled 41 patients with newly diagnosed glioblastoma. The mean and maximal relative cerebral blood volume (rCBV) ratio (rCBV and rCBV), derived from MR perfusion weighted imaging, of the enhancing tumor, as well as maximal rCBV ratio of peri-enhancing tumor area (rCBV) were measured. The ki-67 labeling index, mammalian target of rapamycin (mTOR) activation, epidermal growth factor receptor (EGFR) amplification, isocitrate dehydrogenase (IDH) mutation and TP53 were assessed. There was a significant correlation between rCBV and mTOR based on Pearson's correlations with Benjamini-Hochberg adjustment for controlling false discovery rate, p = 0.047. The rCBV showed significant correlation with mTOR (p = 0.0183) after adjustment of gender and EGFR status. The mean rCBV value of the patients with OS shorter than 14 months was significantly higher than patients with OS longer than 14 months, p = 0.002. The rCBV and age were the two strongest predictors of OS (hazard ratio = 1.29 and 1.063 respectively) by Cox regression analysis. This study showed that hemodynamic abnormalities of glioblastoma were associated with genomics activation status of mTOR-EGFR pathway, however, the radiogenomics associations are different in enhancing and peri-enhancing area of glioblastoma. The rCBV has better prognostic value than genomic biomarkers alone.