Glioblastomas (GBMs), the most common primary malignancies of the central nervous system, are highly aggressive and heterogeneous, and remain a dramatic therapeutic challenge. Markers mirroring the complex molecular profile of GBMs that are predictive of patient outcomes are needed to define novel multi-targeted treatment strategies. Resistance to current GBM therapies is partly due to a subpopulation of stem-like and other self-renewing cells (hereafter called glioma stem-like cancer cells, GSCC), which are therefore of key interest as therapeutic entry points. Wnt and Hedgehog are among the main pathways involved in GSCC renewal. β-catenin and Gli1 are markers of Wnt and Hedgehog activation respectively and both pathways are known to be altered in gliomas. To date, there are no investigations of Gli1 protein expression in GBM tissue, and recently a high expression of β-catenin has been found to have a poor prognostic impact in GBM patients in a study. We have therefore quantified the positivity for β-catenin, Gli1, as well as Ki-67, p53, and EGFR proteins on immunohistochemically-stained GBM sections from 106 patients in an investigation for potential predictive biomarkers. Correlation between these markers and survival was evaluated by pair-wise Pearson correlation coefficient and by bi-dimensional hierarchical clustering, followed by survival estimations using linear regression models and classification trees. We demonstrated that both β-catenin and, for the first time, Gli1 proteins are highly predictive markers of short survival, being found in 75 and 90% of the highly predictive trees, respectively, whereas Ki-67, p53 and EGFR were under 30% and thus, not considered as predictive. Our results indicate a role of β-catenin and Gli1 in GBM malignant behaviour, and suggest that inhibiting members of Wnt and Hedgehog pathways could be a valuable therapeutic strategy for GBM patients.