# Toca 5: a virus plus a prodrug in recurrent brain cancer, tested properly, and it did not work

Source: https://onco.cc/key-papers/paper-cloughesy-toca5-glioma-jamaoncol-2020/  
OnCo record `paper-cloughesy-toca5-glioma-jamaoncol-2020` (Key paper). Data CC BY-NC 4.0, attribute "Data from OnCo (onco.cc)"; commercial use needs a licence.

## TL;DR

Four hundred people with a brain tumour that had come back were randomly given either a virus-delivered enzyme plus a pill it converts into chemotherapy, or standard treatment; survival was the same.

## Summary

Randomised, open-label phase 2/3 trial at 58 centres in the United States, Canada, Israel and South Korea. 403 patients having resection for a first or second recurrence of glioblastoma or anaplastic astrocytoma were randomised 1:1 to vocimagene amiretrorepvec, a retroviral replicating vector injected into the resection cavity wall, followed by cycles of oral flucytosine starting six weeks after surgery, or to investigator's choice of lomustine, temozolomide or bevacizumab.

Median overall survival was 11.10 months with the virus and prodrug and 12.22 months with standard of care, hazard ratio 1.06 (95% CI 0.83 to 1.35), p = 0.62. No secondary endpoint showed a significant difference. Adverse event rates were similar.

The design was the best case for the approach: the virus is injected under direct vision into the cavity where the tumour was, the prodrug converts to fluorouracil only where the virus has spread, and the trial was properly randomised and adequately sized. It still failed.

## Fields

- Kind: Key paper
- Last checked: 2026-09-25
- Journal: JAMA Oncology
- Year: 2020
- DOI: 10.1001/jamaoncol.2020.3161
- Authors: Cloughesy TF, Petrecca K, Walbert T, et al.
- Findings: Median overall survival 11.10 months with vocimagene amiretrorepvec and flucytosine versus 12.22 months with standard of care; hazard ratio 1.06 (95% CI 0.83 to 1.35), p = 0.62.; No secondary endpoint, including durable response rate and 12-month overall survival, showed a significant difference.; Adverse event rates were similar between the arms.; 271 deaths among 403 randomised patients at a median follow-up of 22.8 months.
- What it means: The clearest randomised refutation in the field. Phase 1 data had looked encouraging, the delivery problem was solved by surgical access, and the killing mechanism was a well-understood chemotherapy released in place. None of it translated. Any claim that oncolytic virotherapy works in glioma has to be read against this trial.
- Caveats: Open-label, and the control arm was a choice of three agents, none of them strong in recurrent high-grade glioma.; Vocimagene amiretrorepvec is a replicating retroviral vector delivering a prodrug-converting enzyme rather than a directly lytic virus, so the result does not transfer straightforwardly to lytic platforms.; Phase 1 results that prompted the trial came from a selected population.

## Sources

- JAMA Oncol 2020: https://doi.org/10.1001/jamaoncol.2020.3161
- PubMed: https://pubmed.ncbi.nlm.nih.gov/33119048/
- ClinicalTrials.gov NCT02414165: https://clinicaltrials.gov/study/NCT02414165

## Connected records

- cancers: [Glioma & glioblastoma](https://onco.cc/cancers/glioblastoma/)
- fronts: [Immunotherapy](https://onco.cc/fronts/immunotherapy/)
- technologies: [Oncolytic viruses](https://onco.cc/technologies/oncolytic-virus/)
- people: [Timothy F. Cloughesy](https://onco.cc/people/timothy-cloughesy/)
- journals: [JAMA Oncology](https://onco.cc/journals/jama-oncology/)

---
JSON: https://onco.cc/api/v1/entities/paper-cloughesy-toca5-glioma-jamaoncol-2020.json