Great article! A related thought occured to me after I read the 'Death Of Cancer' by Vince DeVita. One reason the NCI program was successful is because the NCI operated a research hospital; the hospital was free for the patients in exchange for their participation in clinical research programs. Physician-scientists developed protocols, including combination chemotherapy, while basically experimenting on patients in ways that would not be allowed today. With our current regulations, its inconceivable that combination chemotherapy could've been developed.

My takeaway from that was that perhaps we could re-create a similar institution, but for dogs, as a circumvention measure [I continue to think that the best animal disease model for humans are humans]. The failure mode of such a research program would of course be that treatments are developed in a very targeted manner for dogs that doesn't translate to human cancers.

A pan cancer research program that experiments on dogs on the other could be more promising in that the learnings would translate to humans much more readily.

Yes. I argued that the causes for the malaise are over-determined, and medical liability and close-mindedness are both reasons. There are integrative and functional doctors who are more willing to prescribe off-label medications. But they are rare and I haven't researched them in depth yet. 

The causes for the malaise are over-determined but include among other reasons:  institutional inertia & red tape, lack of funding for repurposing generic drugs, regulations that often force companies to test only one new drug at a time against the standard of care (SoC), oncologists who refuse to deviate from SoC due to some combination of fear of medical liability and close-mindedness and coordination problems amongst pharmaceutical companies.

But if there were 10x more clinical trial results, there'd be a greater universe from which one could prescribe off-label medications, and the quality of off-label recommendations would be higher. And the FDA certainly is a blocker there.

Only 1 way to figure out: test it on (consenting) new patients. The drugs are generic (so cheap), and don't have intolerable side effects. It has anecdotally worked. So it's positive EV for patients with terminal prognosis. And yet its explained away as a curious  'spontaneous' remission, using your very valid objection.

This is after all my bigger claim / criticism of the status quo. My objective is to make a meta-level criticism more so than to advocate for one particular protocol as the potential cure. My claim is that a medical system that acted with urgency and that wanted to maximize patient survival odds, would have at the very least attempted to replicate the protocol in a Phase II clinical trial. Instead, it was ignored due to status quo bias. Not because it lacked a scientific justification, but because the drugs are generic. 

In other words, I use the lack of a replication attempt to update in favor of systemic issues in oncology and FDA.

Thanks for the great recap, some of this looks very promising! I remember reading your personal blog several years back, you (and James Watson) were both optimistic about the possibility of leveraging the Warburg effect in cancer therapies. Why didn't much progress happen on that front? Is it that we haven't invested the necessary resources there and instead got excited by alternative promising directions? Or did we come to learn of new information that reduced the feasibility of that strategy?

And another thought, it seems like there is some need for rapid experimentation on voluntary patients to optimize protocols (e.g. how many days to fast, whether to do oncolytic virus therapy and immune checkpoint inhibitors together etc.), and the risk-averse hyper-bureacratization around clinical trial protocols seems like a huge impediment to progress for cancer specifically.