If CLDX by itself with limited resources, I would try to run either single arm in HG/TN with ORR as primary endpoint for accelerated approval, with confirmation randomized trial in HG but stratify TN status; or small randomized trial in HG/TN with PFS as primary endpoint for accelerated approval, with same confirmation trial mentioned above to get better read on HG population.
I might be able to agree with that (-g-) on the basis of: some trial costs are fixed cost (e.g. the first patient enrolled is 100 times the cost of the next patient) and enrolling a broader population might result in a faster enrollment for the subgroup you really want.
But just to be clear - the ph i data for melanoma for high expressers is actually substantially better IMO than the data in HG wo TN. Much more promising than in the HG wo TN patients, so every dollar spent on the HG wo TN in refractory BC... .