Background: Malignant glioma is the most common and lethal primary brain tumour, with dismal survival rates and no effective treatment. We examined the safety and activity of NSC-CRAd-S-pk7, an engineered oncolytic adenovirus delivered by neural stem cells (NSCs), in patients with newly diagnosed high-grade glioma. Methods: This was a first-in-human, open-label, phase 1, dose-escalation trial done to determine the maximal tolerated dose of NSC-CRAd-S-pk7, following a 3 + 3 design. Patients with newly diagnosed, histologically confirmed, high-grade gliomas (WHO grade III or IV) were recruited. After neurosurgical resection, NSC-CRAd-S-pk7 was injected into the walls of the resection cavity. The first patient cohort received a dose starting at 6·25 × 1010 viral particles administered by 5·00 × 107 NSCs, the second cohort a dose of 1·25 × 1011 viral particles administered by 1·00 × 108 NSCs, and the third cohort a dose of 1·875 × 1011 viral particles administered by 1·50 × 108 NSCs. No further dose escalation was planned. Within 10–14 days, treatment with temozolomide and radiotherapy was initiated. Primary endpoints were safety and toxicity profile and the maximum tolerated dose for a future phase 2 trial. All analyses were done in all patients who were included in the trial and received the study treatment and were not excluded from the study. Recruitment is complete and the trial is finished. The trial is registered with ClinicalTrials.gov, NCT03072134. Findings: Between April 24, 2017, and Nov 13, 2019, 12 patients with newly diagnosed, malignant gliomas were recruited and included in the safety analysis. Histopathological evaluation identified 11 (92%) of 12 patients with glioblastoma and one (8%) of 12 patients with anaplastic astrocytoma. The median follow-up was 18 months (IQR 14–22). One patient receiving 1·50 × 108 NSCs loading 1·875 × 1011 viral particles developed viral meningitis (grade 3) due to the inadvertent injection of NSC-CRAd-S-pk7 into the lateral ventricle. Otherwise, treatment was safe as no formal dose-limiting toxicity was reached, so 1·50 × 108 NSCs loading 1·875 × 1011 viral particles was recommended as a phase 2 trial dose. There were no treatment-related deaths. The median progression-free survival was 9·1 months (95% CI 8·5–not reached) and median overall survival was 18·4 months (15·7–not reached). Interpretation: NSC-CRAd-S-pk7 treatment was feasible and safe. Our immunological and histopathological findings support continued investigation of NSC-CRAd-S-pk7 in a phase 2/3 clinical trial. Funding: US National Institutes of Health.