So, that would seem to me to still leave in play the potential that a large majority of the G+ patients may have been in the triple negative group.
Possibly. But I don't see anything in the poster that says 'we preferentially tested all the triple negative patients for G+ status'. And even if it were true that that is what they did there are clearly still some G+ patients that are not triple negative - i.e. the ORR in G+ patients is likely to be meaningfully less than 22%.
PS I'll be interested in how they cast this ph ii in the IR Day on Monday. (I'd call this a good ph ii - to find out the best subpopulation in which to conduct the ph iii. But it will be interesting to see if they try to pitch it as more than that.)