I have to say that I found that presentation pretty unconvincing. His one case study only demonstrated that a woman relapsed in the normal time frame that a patient relapses with a brain tumor. The fact that she was ketogenic for a while and then not really doesn't mean anything. Also, the standard of care frequently involves surgery - which he doesn't mention, but i suspect that patient had.
The mouse data strikes me as odd as well. Tumors in the brain are not usually reported by weight. It is impossible to resect out a tumor from the brain and get a precise weight, so I would have rather seen his imaging data. His photos show large extracranial tumors on his mice. This is not a normal occurrence with GBM (it happens because tumor cells that get injected in the brain leak back out of the injection site and grow above the skull) and extracranial tumor growth cannot be used as a surrogate for what is happening in the brain.
I'm not saying the this strategy wouldn't work, although personally I have seen cancer patients go literally weeks without a single bite of food and it had no effect at all on their tumor growth. Perhaps it was simply too late for those folks. Mice and rats will also stop eating when they develop brain tumors.
Studies have shown that current strategies for treating brain tumor barely improve survival. However, they do not shorten life expectancy. So if his statement that the treatments are what are killing people was true then survival should be worse with treatment.
Interesting ideas. I am not impressed with the quality of that data.