I'd like to qualify anything negative I say with the disclaimer that I am bullish on halo, it being my largest holding. That said, I have been pretty consistent in saying that my problem with the interim results has to do with the quality of the data. I was ( and still am) looking more for explanation of methodology, or numbers that strengthen significance, than top line results.
As Rod has suggested recently, it is a no-brainer to guess the OS for stage 1 202 would be better than historical AG for treatment of Stage4 PanCan. There has been about a 3-4 month difference between the two so this trend toward OS of 12 mos, when PFS seems to be about 9 mos, in this cohort isn't a big surprise.
I'm not concerned that they don't have stat sig yet on OS, (it might be too early yet) though I would like to see a slide that outlines those numbers. But, what I'd really like to see is a high HA OS cohort that is roughly half of all subjects.
If the OS analysis has the same large number of subjects with unidentified HA status, like in January, then I'd like to see a discussion that shows how the observed OS trend in the identified could hold true for the unidentified group, assuming they had the same ratio of high to low HA. The ORR and. PFS analysis we saw in January had p values attached to them that were a joke if you took into consideration how many subjects were left out. Furthermore, if you looked at the unidentified HA subjects and tried to imagine they had the same ratio of high to low HA, it was hard to imagine a benefit for pegph20 there.
Perhaps we'll get more than this PR after the conference.
