There is no "strategy" in experimental oncology ...
Even in clinical oncology, oncologists often try to modify the given setup of a "standard" protocol.
If it would be my experiment, I would continue the treatment as long as the animals support it.
Once it would remain 4-5 mice in a given group, I would euthanize all the mice, with pieces of tumors put in liqhuid nitrogen, and other pieces for histopathological analyses. This would help you in knowing which are the cell types resistant to your treatment.
Once you would have an idea about the type of "resistant cells" facing your compound, you can perform a new experiment with your compound + a second one aiming to combat the above mentioned resistant cells.