>>Millions of people have died from cancer. I'd like to think the cost of treating it (successfully or not) is reasonably well known. Yes, there are huge variations from case to case but over large populations you get statistically meaningful answers. If you're trying to tell me cancer treatment hasn't been analyzed in exactly this fashion, then that's disgraceful.

Yes, but it's more bad paraphrase by you. Of course people know what it costs to treat cancer via currently available means and the iterative chemotherapy and other treatments you describe. The point is that knowledge changes and not every advance is published in advance so costs can change quickly whether funders are ready or not. There can be massive pressure when expensive drugs like Kadcyla come on the market costing >$100K per patient: perhaps the underwriters can come to you to find out when this is going to happen and what it will cost so they can build it into their pricing. After that you can let them know about the whole class of PARP inhibitors that are expected to be extremely expensive. The FDA approved one of those just a few days ago so now watch to see how it is rationed.