Biotech Values

[iwfal]

SNTA -

Also, he insinuated that the data in these subpopulations is way better than mere doubling of the PFS even though he admits that he has not seen the data.

First a minor correction based upon listening to the cc - the tweak being that he outright said that he himself was only inferring (no direct knowledge) better than doubling of PFS in the subgroups. And the inference came from the excitement in the team in process of doing the interim analysis. I.e. we are one more step away from the real data.

Further, with this understanding of the state of knowledge, I find his statement puzzling - e.g.:

a) They have enrolled probably around 130 patients at this point (they announced 100 several months ago)

b) KRAS subpop was anticipated to be 90% powered for 2x efficacy (e.g. HR=0.5) - implies about 90 events and for PFS expect about 67% of enrollees to event (the rest being LTFU) implying about 130 patients. (Note that this also lines up with KRAS wt being about 70% of nsclc - 60% of 240 is 144.)

c) So even at only 100 patients enrolled and finished with PFS I'd expect 60 KRAS wt - and with 2/3s not LTFU gives 40 events. In order to beat 0.01 p value with 40 events they would need 'only' HR=0.45. I.e. better than 2x, but not much.

I.e. I could easily see an interim PFS comparison only somewhat better than 2x causing a move to ph iii.