The Taxotere ± OGX-011 hazard ratio of 0.60 blows away the efficacy seen with DNDN’s Provenge at the interim analysis of the 9902b study in progress.
Random notes (for my future reference as much as anything else):
1) PR was pretty data-light. If they really had an HR of 0.60 on a 1:1 randomized 82 person trial I would expect them to be stat sig already for any reasonable number of events - and there was no mention of stat sig or even trends. So they probably have had less than 40 events so far.
1a) As a chemotherapy I would expect that as they let the trial run longer without adding new patients to the start of the curve the HR will shrink. And the fact that they don't mention stat sig implies that they don't have too many events yet. This, when combined with the fact that the trial only started 3-1/2 years ago, implies that they probably have very few events in the tail. (Also note that the factoid about the earliest the median could be for the treated group implies an incredibly steep curve down to the median - flat..., flat..., out to 20 months or so and then drop off a cliff to the median. And their July 2007 Interim results indicate the same. See: http://ir.isispharm.com/releasedetail.cfm?ReleaseID=257302 )
2) The PR didn't talk of HR - and the ratio of medians is suspiciously close to the benefit mentioned. It is possible that they have perfectly exponential curves, but the coincidence is enough to beg a question.
3) Anti-sense in general is certainly worthy of a PSB derating. There are some indications of some success (Genta) - but generally pretty weak.
Market Data 
Markets 
AI
