I think we're moving into an era when we need to consider that everybody is different, and a lot of cancers are different. This isn't like curing a bacteria, where (mostly) one-size fits all.
I wonder: is it possible to bring a drug to market that is ineffective and sometimes fatal to most people without a genetic marker, but is safe and will cure cancer for those with the marker?
Indications can be as specific as "indicated for tumors with specific mutations"[1] or indicated "prostate cancer when all other options have failed"[2]
[1]https://en.wikipedia.org/wiki/Pembrolizumab [2]https://www.fda.gov/drugs/informationondrugs/approveddrugs/u...
We already do this. A large class of both antipsychotics and antidepressants require certain genes in order to metabolize properly and have the intended effect, instead of just giving a bunch of side effects with no benefit.
These genes are not universal and not evenly distributed; people of certain ethnic backgrounds are dramatically less likely to have the requisite genes, which means that most antipsychotics and antidepressants literally don't work on them.