"But statistics — not anecdotes — rule over drug approvals. In 2005, regulators in China approved an oncolytic adenovirus called H101 to treat head-and-neck cancer, after evidence showed that the treatment could shrink tumours. Those trials stopped short of assessing improvements in patient survival — a measure often required for FDA approval. Since then, a medical-tourism industry has built up in China for people who cannot get the therapy in their home countries.

Then, in May this year, a team supported by biotechnology giant Amgen of Thousand Oaks, California, published promising results from a large clinical trial of T-VEC (R. H. Andtbacka et al. J. Clin. Oncol. 33, 2780–2788; 2015). The virus both shrank tumours in people with advanced melanoma and extended patient survival by a median of 4.4 months. Yet statistically, survival benefits fell just a hair’s breadth of significance. “That raised the question, ‘Well, what is statistical significance? Is this an active agent or not?’” Russell says."

It's possible that a more immunogenic (eliciting immune response) virus would have better results against cancers. Of course, this means that it would also hurt the patient more. I'm sure they know this already.

Though I'm biased, I think that given the current state of viral engineering, immunotherapy (immune cell gene-engineering and transplantation)is a better way of getting at cancer for now-- it's rapidly being proven in the clinic and the lab. I can see a time 10 years from now where we'll be able to engineer certain viruses to be powerfully oncolytic, but for now I think the problems are specificity of targeting (you only want to kill tumor cells) and the inability for prototype oncolytic viruses to infiltrate the tumor microenvironment. The infiltration problem isn't a hard dead end, nor is it a problem unique to viral therapies, but a couple of the bleeding edge immunotherapies and even older radiotherapies and chemotherapies can pierce into the tumor microenvironment with no problem.

Though I expect immunotherapy to beat them to the punch, I can completely see oncolytic viruses joining the combination regimen that is the standard of care, alongside surgery, chemo and radiotherapy, provided that they show a more concrete improvement in survival rate/length.

Is it just me, or does this sound kinda dangerous? That is, engineering a virus specifically to evade the immune system. As long as it stays un-mutated, fine. But what if a mutation occurs that results in it becoming a deadly disease, with the bonus of being "camouflaged from the immune system"?

Very cool, but slightly scarey at the same time.

Kill the big dog by infecting the body with a smaller disease that is lethal to the big dog but (slightly) less lethal to the host.

So this can either go the "World war Z" way or the "I am legend" way - that's a cheery thought