Replies to post #142699 on Biotech Values

[iwfal]

CLDX -

I think it's not that meaningful to do that kind of statistical analysis due to really small sample size, including HG/TN group as well. I would conduct a randomized trial by enrolling patients with HG, but stratify by TN status.

Interesting discussion - if you ran a small 20 person ph ii and got pfs of 0.85 would you take it to the next level? I wouldn't since you can state that it is unlikely to have a pfs below about 0.65 (WAG) - and with a pfs of even, say, 0.70 it is unlikely to get a great OS due to the general tendency for OS HR to be a deterioration of pfs HR. (Note that I did say I might carry it forward in this population, even with the poor pfs results, if the OS data looked good - i.e. the results show that this is one of the drugs that maps pfs results to OS fairly well). E.g. I'd say the math WAGS looks a little like: 50% chance PFS is 0.85 or greater (i.e. worthless), 35% chance pfs HR is between 0.65 and 0.85 in which case the chance of an OS benefit is perhaps 20%, and 15% chance pfs HR is less than 0.65 in which case the chance of an OS benefit climbs to 60%. Altogether about 15% chance of success in that population - and note I am probably erring on the high side for the above estimate.

Also note - As a small company with scarce resources I especially wouldn't take this forward if I had had a parallel ph ii that had showed stellar efficacy. Or even if I had to finish a ph ii in a different indication where the ph i had been promising.

PS It might be interesting to know what the ph i data says on the topic of HG wo TN responses.