Background: Expression of the enzyme alkylpurine-DNA-N-glycosylase (APNG) has been associated with poor outcome in patients with glioblastoma multiforme (GBM). APNG is part of the base excision repair system, and expression of APNG is related to resistance towards temozolomide (TMZ). This study evaluates the prognostic value of APNG in a population-based cohort of 185 Danish GBM patients diagnosed between 2005 and 2009. Due to GBM’s cellular heterogeneity, APNG contribution from non-tumor cells was excluded. Methods: APNG expression was evaluated using image analysis and a novel quantitative immunohistochemical (IHC) assay (qIHC, Dako), where APNG protein levels were represented through dots. Non-tumor cells, which were found to express APNG, were excluded using an IHC/qIHC double-staining. To verify the qIHC results, APNG was measured in 177 of the patients using a quantitative in-house developed APNG immunofluorescence (IF) assay. MGMT methylation status was obtained in 152 patients by pyrosequencing. Results: Using qIHC, median expression of APNG was 0.31 dots/cell (range 0.099-0.96). An optimal cut-point was identified dichotomizing the patients at an APNG value of 0.24 (25% vs. 75% of the patients). In Cox regression high expression of APNG was associated with better overall survival (OS) (HR 0.53, p = 0.001) adjusting for the effect of age, performance status, tumor crossing midline, treatment and gender. APNG was associated with better OS (HR 0.57, p = 0.012) when adjusting for MGMT status. Retesting the cohort using IF, demonstrated similar results (HR 0.81, p = 0.2). Patients with methylated MGMT promoters and high APNG expression demonstrated a better OS, than with non-methylated MGMT promoters and low APNG expression (HR 0.55, p = 0.04). Conclusions: APNG measured by qIHC was found to be an independent significant prognostic factor for OS in GBMs, This was supported by IF measurements. Removing bias associated with APNG expression in non-tumor cells in both assays has, in this study contributed substantially to analyze the prognostic value of APNG in GBM patients. We expect that APNG qIHC can potentially identify patients who will not benefit from treatment with TMZ.