Besides the use of temozolomide and radiotherapy for patients with favourable methylation status, little progress has been made in the treatment of adult glioblastoma. Local control of the disease by complete removal increases time to progression and survival. We assessed the efficacy and safety of a locally applied adenovirus-mediated gene therapy with a prodrug converting enzyme (herpes-simplex-virus thymidine kinase; sitimagene ceradenovec) followed by intravenous ganciclovir in patients with newly diagnosed resectable glioblastoma.

For this international, open-label, randomised, parallel group multicentre phase 3 clinical trial, we recruited patients from 38 sites in Europe. Patients were eligible if they were aged 18–70 years, had newly diagnosed supratentorial glioblastoma multiforme amenable to complete resection, and had a Karnofsky score of 70 or more at screening. We used a computer-generated randomisation sequence to allocate patients in a one-to-one ratio (with block sizes of four) to receive either surgical resection of the tumour and intraoperative perilesional injection of sitimagene ceradenovec (1 × 10¹² viral particles) followed by ganciclovir (postoperatively, 5 mg/kg intravenously twice a day) in addition to standard care or resection and standard care alone. Temozolomide, not being standard in all participating countries at the time of the study, was allowed at the discretion of the treating physician. The primary endpoint was a composite of time to death or re-intervention, adjusted for temozolamide use, assessed by intention-to-treat (ITT) analysis. This trial is registered with EudraCT, number 2004-000464-28.

Between Nov 3, 2005, and April 16, 2007, 250 patients were recruited and randomly allocated: 124 to the experimental group and 126 to the standard care group, of whom 119 and 117 patients, respectively, were included in the ITT analyses. Median time to death or re-intervention was longer in the experimental group (308 days, 95% CI 283–373) than in the control group (268 days, 210–313; hazard ratio [HR] 1·53, 95% CI 1·13–2·07; p=0·006). In a subgroup of patients with non-methylated MGMT, the HR was 1·72 (95% CI 1·15–2·56; p=0·008). However, there was no difference between groups in terms of overall survival (median 497 days, 95% CI 369–574 for the experimental group vs 452 days, 95% CI 437–558 for the control group; HR 1·18, 95% CI 0·86–1·61, p=0·31). More patients in the experimental group had one or more treatment-related adverse events those in the control group (88 [71%] vs 51 [43%]). The most common grade 3–4 adverse events were hemiparesis (eight in the experimental group vs three in the control group) and aphasia (six vs two).

Our findings suggest that use of sitimagene ceradenovec and ganciclovir after resection can increase time to death or re-intervention in patients with newly diagnosed supratentorial glioblastoma multiforme, although the intervention did not improve overall survival. Locally delivered gene therapy for glioblastoma should be further developed, especially for patients who are unlikely to respond to standard chemotherapy.

Ark Therapeutics Ltd.