Biotech Values

[DewDiligence]

Re: IMCL, ABGX, AMGN (from Yahoo)

>>
Re: will FDA be prudent or reckless?
by: DewDiligence
12/06/05 02:06 pm
Msg: 334150 of 334211

>The FDA will immediately see through the flaws of [Panitumumab] trial 408 and in order to use PFS as a surrogate endpoint, ALL BIAS WILL HAVE TO BE ELIMINATED and no sensitivity analysis on earth can do this.<

You make two misstatements here:

1. PFS may be considered a non-surrogate endpoint in support of an application for full approval as well as a surrogate endpoint in support of a Subpart H accelerated approval. AMGN is not committed to the Subpart H route and it is more likely, IMHO, that they will seek a full approval based on the PFS data in the “408” trial.

2. Sensitivity analyses can eliminate the bias that accrues from the way progression scans are scheduled and performed. One obvious method: record the progression date as the earliest possible date in the active arm (an earlier date than the actual scan date if a patient missed a prior scheduled scanning appointment) and record the progression date as the latest possible date in the placebo arm (a later date than the actual scan date if the progression was based on symptoms observed prior to the next scheduled scan date).

The treatment benefit in the sensitivity analysis described above would necessarily be no greater than the bias-free treatment benefit. Roger Perlmutter said that AMGN performed this sensitivity analysis and the treatment benefit was still statsig by a wide margin.

[Posted as a reply to: Msg 334114 by bubbietubbie]
<<