Replies to post #115155 on Biotech Values

[BTH]

I think the rida situation speaks more to the FDA's posture on the SPAs it grants, and what it wants to do with the PFS endpoint in general.

If you have the entire Ridaforolimus group having a 3 week progression free benefit (on the median), how would, say, a 20 week benefit in a majority of patients on the leiomyosarcoma subset versus a 3 week benefit in the vast majority (or no benefit at all) of osteosarcoma subset be looked at by ODAC or the FDA?

What I am getting at is, I believe there are going to be dramatic differences in the subsets (a few of the STS showing dramatic PFS results - whereas Bone sarcoma, IMO, likely to show little, if any, difference). Has there ever been a situation whereby a drug has been approved in such a heterogeneous indication, in which some subsets show great benefit, and others do not? How would ODAC and the FDA approve, or not approve such an indication? Drug labels? Completely decline the drug or make them do another Phase 3 or 4 trial?

Personally, I believe the subset numbers and OS numbers (in those who showed the most benefit) are going to show a benefit worthy of approval, but, as someone just pointed out (and after re-reading it again), that Board of Director Letter about Berger was disturbing (to say the least), and the guy is still one of the biggest jerks in all of biotech; it's it's likely THEE number one reason the larger investors are questioning Ridaforolimus right now.

Thanks

[jq1234]

Why does it matter if the control arm was another drug or not? If it's a properly run controlled trial that doesn't deprive patients of suitable treatment (presumably the rida control arm is BSC), why should the agency think less of it? If it's the suitable trial for the stage of disease, then...? Tarceva did just fine getting on the market for lung cancer based on a trial versus BSC.

We are talking about comparing trial, and data relevance across setting and indication. You have to be careful what you compare to. You have to put data in correct context. The tendency is to compare to any cancer in any setting, which ususally draws wrong conclusion.

Comparing Rida in maintenance setting for sarcoma with Tarceva in 1st line pancreatic cancer in combination with gemzar against gemzar is comparing apple to orange. Tarceva data in pancreatic cancer might not have met expectation at that time, but there was little argument against approval, because 1. first line vs SOC with both stat significant PFS and OS; 2. pancreatic cancer is much much more agressive cancer, small improvement might not be much in absolute term, but in relative term of the disease it is different. If rida were in 1st line setting against SOC with stat significant PFS and OS, no matter how small, I am not going to make same argument here.

My point is Tarceva in maintenance setting in NSCLC is the appropriate comparison here. ODAC argued against approval even though stat significant PFS and OS, 2 weeks and 1 month respectively. At this point, there is no gurantee rida's 3-week PFS advantage will translate into OS advantage.

Tarceva in 2nd/3rd line NSCLC vs placebo data were impressive, no argument there.

[DewDiligence]

The Rida question of consequence is (IMO) not whether the OS results are statsig, but rather is whether the OS results show at least a strong trend in favor of the Rida arm. If the p-value for the OS comparison turns out to be really bad, I think MRK will abandon the program in this particular indication because the commercial upside in such a case would be de minimis.