Ok, following with the totally naive speculation. What if it were not a place, but a specific environmental factor or a specific trait? Cancers mutate quickly, and resources are limited, so cells that are more successful might starve those which lag behind, as it happens in animal populations. And once the cancer has "fallen into the trap" of specializing for something, that something could be targeted, or the favouring environmental conditions removed.
This is the clonal evolution model of cancer, and is how many solid tumors are comprised and treated. The danger is that the selective advantage for the cancer cell is often by masking itself as a healthy cell to the immune system, which complicates treatments but this interaction can still be targeted (see PD-1).
Yes, the idle speculation about a "honeypot" method was of artificially encouraging the cancer to evolve in a certain direction, even boosting its growth, until it has reliably mutated in a way that makes it susceptible to some other intervention, like being starved or targeted by some particular marker.
