>>published data on bromodomain inhibitors in cancer are not convincing
Agreed - this is very much a theoretical target at this stage, not a confirmed one.
Conceptually, I think of oncology treatments as falling into 3 classes:
1) Conventional therapies aimed at fast-growing cells. These therapies produce tumor shrinkage as an obvious measure.
2) Targeted therapies - these also have clear objective endpoints.
3) Therapies aimed at cancer stem cells, metastasis, and generally keeping the cancer in check. These all have more difficult clinical paths, as endpoints are murkier and OS is generally going to be the only acceptable endpoint. So trials will be long and expensive except in conditions where untreated OS is very short.
Bromodomain inhibitors very likely fall into category 3.