Replies to post #690 on Dendreon Corp fka DNDNQ

[iwfal]

The factors that go into predicting P-11 success based upon unblinding date:

a) Median time to Relapse to 3 ng/ml when on LHRH for 4+ months - very likely (90% chance) to be less than 9 months given all the literature.

b) Distribution of time to relapse - 95% patients relapse by 3x median (nominal assumption)

c) Enrollment distribution - Assume that 'enrollment completion' mentioned in Jun 05 PR is equal to 'randomization completion'. Assume a r^4 enrollment curve (moderately conservative assumption). Only conflicting data is that Jan presentation on CD54 upreg only talked of 104 P-11 treated patients when there should be about 117 - implying either the CD54 data was more than 6 months old or that the above assumption about the definition of enrollment is incorrect.

d) Number of trigger events required - Industry typical in short trials (median time to event measured in months) is 75% of enrollment without much variation (per at least 6 trials I found). So pick 130 as a moderately conservative number for triggering.

e) Number of censored due to lost/withdrawn... - 10%.

Result - If they unblind in mid Oct or later then there is at least a 50% chance of stat sig. Of course that analysis is only as good as its input assumptions - see above.

[walldiver]

The thinking is that since the trial enrollment was completed in June 2005, and the first interim look is triggered by a certain number of PSA threshold events, then the longer the trial drags on without the trigger occurring, the better it looks for Provenge.

Most patients should have "evented" 7 to 10 months after June 2005, based on trials conducted in the past. There has been speculation that the first interim look is triggered by 75% of the patients eventing, which for this trial is having a PSA score rising to 3.0 or higher. Historically, 75% of patients is about right, but it could just as easily be 50% or 2/3. DNDN has never released the # of events needed to trigger the first interim look.

Iwfal is the one who has looked into this more than me, so if I've quoted his figures wrong, I'm sure he will let us know.

Edit after reading Iwfal's post: one concern of ours is if whether or not the June 05 announcement of the trial completion of enrollment is also the date the last patient was randomized. Patients had to undergo Lupron treatment to get their PSA score down into the 0 to 1 range before being randomized to either Provenge or placebo.