Loading
Viruses engineered to infect and burst cancer cells while leaving normal cells alone, and to alert the immune system in the process.
A healthy cell that detects a virus mounts an interferon response: it slows its own machinery, calls the immune system and kills itself before the virus spreads. Many cancer cells have already broken parts of that response, because it also restrains growth, and a virus can exploit the break. Stojdl and colleagues showed this directly in 2000 with vesicular stomatitis virus, which killed human tumour lines at interferon doses that fully protected normal cells. Coffey and colleagues showed in 1998 that reovirus needs an activated Ras pathway, which many cancers supply. Martuza and colleagues had already built the first engineered version in 1991, a herpes virus with thymidine kinase deleted so it could only replicate where the host cell supplied the missing function.
Four replicating oncolytic viruses have an approval somewhere. H101 (Oncorine) in China from 2005, on a 160-patient randomised trial reporting response rate and no survival data. Talimogene laherparepvec (Imlygic, HSV-1) from 2015 in melanoma: durable response rate 16.3 per cent against 2.1 per cent for granulocyte-macrophage colony-stimulating factor, with a median overall survival difference of 23.3 against 18.9 months that did not reach significance. Teserpaturev (G47 delta, Delytact) conditionally approved in Japan in 2021 on a single-arm trial of 19 patients with recurrent glioblastoma. Vusolimogene oderparepvec (Tudriqev, RP1), an HSV-1 carrying GM-CSF and a fusogenic protein, given accelerated approval with nivolumab for anti-PD-1-failed melanoma on 6 August 2026, on a 140-patient single-arm cohort with a 32.9 per cent response rate in which uninjected lesions responded too.
Two products often counted in the class are not replicating viruses at all. Nadofaragene firadenovec (Adstiladrin) is a replication-deficient adenovirus that delivers the interferon alfa-2b gene to the bladder lining, and aglatimagene besadenovec (CAN-2409) is a replication-defective adenovirus delivering a prodrug-converting enzyme. They are gene delivery, not oncolysis, and the distinction matters when reading claims about the class. Cretostimogene grenadenorepvec, which does replicate, reported a 75 per cent complete response rate in BCG-unresponsive bladder cancer with carcinoma in situ in a single-arm phase 3 trial.
The randomised record is poor. Adding talimogene laherparepvec to pembrolizumab in melanoma (MASTERKEY-265) did not improve progression-free or overall survival. Toca 5 in recurrent high-grade glioma and PHOCUS in liver cancer were both negative, PHOCUS worse than its control. The unsolved problems are the ones named in 2012: delivery to tumours that cannot be injected, pre-existing and rapidly rising neutralising antibody, the conflict between letting the virus spread and wanting an immune response, and manufacturing yield.
An attenuated or naturally tumour-selective virus replicates in tumour cells whose antiviral signalling is defective, lyses them, and releases tumour antigens, viral pathogen-associated patterns and any encoded transgene into the tumour, converting lysis into an in situ immunisation.
Nothing in the corpus depends on this yet.
Dependencies are what this technology cannot be delivered without: manufacturing steps, instruments, software, upstream methods. See its full chain on the map.
Cretostimogene is a virus engineered to multiply only in bladder cancer cells with a broken RB pathway, instilled into the bladder. It cleared carcinoma in situ in 75% of BCG-unresponsive patients.
Oncorine is Shanghai Sunway Biotech's oncolytic adenovirus, approved in China in 2005 for nasopharyngeal carcinoma with chemotherapy, the first oncolytic virus approved anywhere in the world.
Talimogene laherparepvec (T-VEC, Imlygic) was the first approved oncolytic virus (2015), injected into melanoma skin lesions.
Vusolimogene oderparepvec is an engineered herpes virus injected into melanoma tumours, approved in August 2026 with nivolumab after immunotherapy failure.
The reason this case matters beyond one tumour is that it forced journals to decide a question they had avoided. The answer these authors give is not permission: it is that each case needs assessing against the values ethics committees exist to protect, and that a self-experimenter who expects to publish should seek review beforehand. For a reader the practical point is the one both papers make, that a published case is a record of what happened to one person and not an instruction.
The uninjected-lesion responses are the most important result any oncolytic virus trial has produced, because they are the first strong clinical evidence that the mechanism is systemic immunity rather than local lysis. The caution is the same as always: this is a single-arm cohort in a population with no standard option, and the randomised confirmatory trial has not read out.
A phase 2 signal that looked like a survival benefit, tested in 459 patients, turned out to be nothing, and delaying effective systemic treatment to give the virus first made outcomes worse. The authors' own conclusion is that the arrival of checkpoint inhibitors should direct any further development of oncolytic virus strategies, which is a polite way of saying this design is finished.
One person, one tumour, one report, and it is not evidence that anyone should treat themselves. What it does contribute is unusually well documented: serial imaging through the course, a baseline biopsy and an excised specimen scored by the same pathology department, antibody titres, and a named protocol with doses. The design choices are the interesting part. Two different viruses in sequence to stay ahead of the antiviral antibody response, frequent dosing to keep infectious virus concentrated in the tumour, and the neoadjuvant setting rather than the late metastatic setting in which oncolytic viruses are normally tested. The result also cannot be attributed to the viruses alone: the tumour was surgically removed and a year of trastuzumab followed, and the phenotype change to HER2 3+ is itself a plausible reason the disease behaved differently this time.
The honest summary of the field after its first approval: a tolerable class of agents, a large number of candidates, and one product in routine use in one disease. The distinction the review keeps making, between viruses that replicate in the tumour and viruses used only to deliver a gene, is the distinction most coverage of this field drops.
Rigvir is the field's clearest case of a product sold far ahead of its evidence: a national registration, international marketing to patients, and no randomised trial. This is the laboratory work that should have been done first. It is here because a field that produces striking single cases attracts exactly this, and readers deserve to know that an approval somewhere in the world is not the same as evidence.
This is the evidence behind a national approval, and it is 19 patients in a single arm. The survival figure is genuinely higher than historical expectation in recurrent glioblastoma, and the biopsy findings support the immune mechanism. It is not a randomised comparison, and the imaging pattern means the usual response criteria cannot be used, which makes an uncontrolled result harder rather than easier to interpret.
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.
Query for this technology: (TITLE:"oncolytic virus" OR ABSTRACT:"oncolytic virus" OR TITLE:"oncolytic virotherapy" OR ABSTRACT:"oncolytic virotherapy" OR TITLE:"oncolytic viruses" OR ABSTRACT:"oncolytic viruses"). Results are unfiltered search hits about Oncolytic viruses, not a curated reading list.
Shares A Study of Adjuvant Cretostimogene Grenadenorepvec for Treatment of Intermediate Risk NMIBC Following TURBT, BOND-003, Phase 2 Study to Evaluate Safety and Efficacy of Cretostimogene Grenadenorepvec in High-Risk NMIBC, Cretostimogene grenadenorepvec.
Shares Engineered bacteria as living cancer drugs, In situ vaccination, Make every cold tumour hot: a coordinated programme to reprogramme immune-excluded tumours, Immunotherapy roadmap: Coley's toxins → checkpoint inhibitors → engineered immunity.