I don’t find it counter-intuitive that a targeted drug could accelerate tumor growth (thereby shortening PFS) in certain patients who are especially susceptible to alternative growth pathways triggered by the targeted drug being administered.
I too minorly favor an explanation that includes some aspect of the above (even at the expense of my intuition). But then the interesting question arises as to why this doesn't show up in in-vitro results or mouse models? Two possibilities:
a) Companies are not inclined to publish negative data - e.g. looking at SNTA pre-clinical data on cell lines and synergy of their drug with many other drugs it is probable that some cell lines were not being reported with some combos.
b) There is some in-vitro effect? (e.g. in vitro doesn't look at mets per se - and perhaps the effect is to increase met capability.)
Lots of interesting questions - but from an investor's perspective it is simpler:
Given this odd, but relatively common behavior in this class of drugs it isn't necessary to know the whys. But it would nonetheless be wise of companies testing such products to ensure they are doing a reasonable set of subgroup analyses in ph i/ii to ensure that they can remove the damaging-their-HR patients in ph iii.