Background: Drug delivery is a major problem in the treatment of glioblastoma (GBM). Tumour pharmacokinetics (PK) of small molecule targeted agents in GBM are not well understood, and poor activity may result from lack of biological efficacy or adverse PK. Olaparib, a small molecule inhibitor of the DNA repair enzyme poly(ADP-ribose) polymerase (PARP), has potential to overcome treatment resistance of GBM. Despite radiological responses in brain metastases, GBM penetration by olaparib has not been studied. Methods: Preclinically, blood-brain barrier penetration was assessed by directional transport of [¹⁴C]-olaparib across MDCKII cells expressing MDR1 and autoradiography of rats and mice treated with [¹⁴C]-olaparib. Clinically, 8 patients with recurrent GBM underwent dynamic contrast enhanced (DCE) MRI at baseline followed by tumour resection after 4 days of oral olaparib (tablet: 100 mg QD, n=5; 200 mg BID, n=3). Olaparib levels were measured in tumour and plasma by LC-MS. Results: Olaparib was a substrate for MDR1 and efflux was blocked by the MDR1 inhibitor ketoconazole. Radioactivity was not detected in the central nervous systems (CNS) of rats or mice after single dose [¹⁴C]-olaparib, but significant levels were measured in subcutaneous HCT-116 tumour xenografts up to 96 hrs. Olaparib was detected in 24/24 resected GBM specimens from 8 patients (Table) at concentrations similar to those in previous breast cancer studies in which PARP inhibition and tumour responses were observed. Pre-treatment DCE-MRI showed increased vascular permeability in tumours, and tumour cellularity parameters correlated with olaparib levels. Conclusions: Olaparib is excluded from the CNS under normal conditions but reliably penetrates recurrent GBM at therapeutic levels. Small molecule PK in GBM are poorly predicted by standard pre-clinical models. Clinical trial information: NCT01390571.