Perhaps this is how we'll eventually manage cancer. By frequent testing and moving from drug to drug as cancer sheds one mutation and adapts other one, turning cancer into manageable state.

Yep, I think manage is the right way to put it.

Fundamentally, viewing cancer as whole bunch of different adaptive adversaries as opposed to a single disease is probably correct. Under that view, consider 'cancer' as the mathematical outcome of the genes in the cells in your body attempting to revert to their fundamentally selfish nature and greedily replicate as much as possible. Every gene in your genome got there because its ancestor genes managed to replicate (though perhaps with a mutation or 2 along the way).

So perhaps it's helpful to think of the organism as a giant prison of small bits of self replicating code (genes). Several checks are in place to ensure that no bit of code replicates beyond sane bounds. However every once in a while a guard get's knocked out or cellblock door switch is left unattended, or a prisoner disguises itself, or groups up with another to overpower or sidestep an existing control mechanism. And then it's off to the races. Higher levels of control can attempt to rectify the situation, but they can fail as well. Given that your body has 100 trillion cells, and about 20 thousand genes. You are talking about a 2 quintillion protein synthesizing programs executing, all of them are there for one reason, they had ancestors that were able to reproduce themselves. They don't have any notion of a greater good, or a serving their purpose or any of that. They can't be reasoned with, they are an adaptive chemical reaction, that either doesn't replicate, or does (and perhaps so much so that it takes down the whole system).

So a lot of cancer treatment are going to revolve around, at our level of technology:

- the ability to target specific classes of cells with certain noisy markers, which may or may not be reachable and with a lot of collateral damage and which contain the errant genes and killing them. (e.g. clusters of cells showing up as a visible tumor in an x-ray, cells having certain chemical signatures)

- the ability to deprive (all or better yet classes of) cells of what they need to replicate to slow them down and potentially eventually have them die off taking the cancerous genes with them.

- the ability to shore-up or reactivate compromised internal defenses.

As well as prevention methods such as:

- avoiding weakening systems that protect against replication (e.g. )

- avoiding encouraging massive amounts of mutation (e.g. repetitive inflammation or free radicals or radiation exposure like UV rays or cesium-137 )

...etc

Are any one of these things going to cure cancer? No, they aren't anymore than telling C programmer to use strncpy instead of strcpy is going to stop all buffer overflow attacks, which in turn prevents remote code execution, which in turn prevents botnets. Or marking all attachments to emails as un-executable is going to stop all computer viruses.

But every bit helps, and saves countless lives, gradually this wack-a-mole approach has shown results.

In the future, perhaps it'll be possible to solve cancer by scanning every cell in the body and selectively eliminating cells that fail some basic sanity checks. We aren't at that point yet, but if we were we would have a pretty effective cancer treatment. It would be adding another (very effective) layer on top of all the other systems walking the day to day beat and slowing stuff down. I'd hesitate to call it a cure though, because if it was halted it would just be a matter of time until one of the selfish genes blindly do their replication thing a little to well for everyone's good.