One follow-up...

Brought up the point in the previous post that the ASSIST trials may have had an interim analysis conducted. For completeness, let's be clear that I didn't make this point up out of thin air, and here are a couple of relevant passages from past conference calls to get everyone up to speed:

February 19 2004

Michael Wick

"We expect to complete enrollment later this year. Since this (ASSIST-1) is an event-driven trial, the timeframe for having initial results is not entirely predictable. We designed the trial to have the opportunity for accelerated approval, although we are fully prepared to complete this trial."

April 29 2004

Cynthia Butitta

"We design these trials (ASSIST-1 and -2) with interim looks to provide opportunities for accelerated approval, although we are prepared to complete the trials and anticipate filing the NDA in the second half of '05."


Interim looks and accelerated approval is company-speak for an interim analysis that is looking at an endpoint that is a surrogate of survival (time to progression, progression free survival, response rate), with the hopes of showing a large advantage early on in the life of the trial, without having to go the distance for the survival analysis.

Telik IR recently made it clear to me that although these interim analyses were designed, they were never undertaken. The issue was, to quote, moot.

I do find it funny to design a trial with an interim analysis and not go through with it. I could see a scenario where a change in the standard of care for that indication may render an upcoming interim analysis of a surrogate endpoint "moot". For example, you're studying your drug in 3rd line NSCLC in a trial that has an interim analysis with a time-to-progression (TTP) endpoint, and a final overall survival endpoint. While your trial is running and before your interim analysis triggers, a competitor's drug in the same 3rd line NSCLC indication shows a survival advantage. At that point, a positive TTP finding on the interim analysis may not have as much impact as one hoped, and you'd be better off with survival data in hand. Under that scenario, you could forego an interim analysis if you felt that the impact of the endpoint / findings would not be significant.

I could see this for ASSIST-2. In this setting where Tarceva has survival data in hand, you wouldn't have a selling point. You buy a bunch of oncologists lunch during a seminar in the hope of telling them how your drug improved a surrogate and why they should pass up the drug that improved survival?

(And trust me, unless you're a company representative with a very direct, concise and nuance-free message, MDs don't pay attention. MDs focus more on the cheap, free pens and notepads by the food table than to the message that the representative is conveying. Unless they know the rep... then they start talking about their kids and tonight's little league baseball game.)

(Before any MDs start flaming this blog, go to a couple of more afternoon seminars and see how much time the *other* MDs spend with the Caduet and Lipitor trinkets than with the representative and his or her story. If I'm right, then do the honorable thing and grab a trinket off the table for me... we spilled bacteria on one of our Lipitor tape dispensers and could use another.)

(Excuse me while I take a sip of coffee from my Plavix mug.)

Back to topic, I don't see why they would pass up this opportunity for ASSIST-1. Especially since it is written into an SPA, and there have not been any substantial improvements in the standard of care for ~3rd line ovarian. This one confuses me. The two easiest explanations are that they are being very / overly conservative, or are simply chicken. I can't really decide.

The third option, for the conspiracy buffs, is to not take them at their word and believe that they did, in fact, do the interim analyses. Although I might be a fan of this one for other companies, I've always had concise questions for Telik IR and Mrs. DeGuzman has always given me concise (curt?) answers in response. There is not much to be inferred when the speaker says "they weren't done".

So that's that. As always, you decide.

(But I think I heard a quality in her voice that was clearly restrained enthusiasm, like they knew that... just kidding. Don't you hate it when investors overzealously read into every inflexion of the speaker during a talk? Drives me nuts. How come no one ever hears a restrained quality of impending doom in the CEO's voice during a conference call?)

===

Spelling and grammar will be corrected later. Maybe.