Why am I so sure DNDN will fail here? Because just like with AGIX, there is no clear evidence that the drug actually works:
First, Provenge has never hit a primary endpoint, a big no-no for the FDA as the primary endpoint is what the trial is designed to test. And while there was a survival benefit in the D9901 trial, it was only a secondary endpoint and therefore was never designed to conclude on whether or not Provenge extends survival.
I've heard arguments that survival is the hardest endpoint out there so if they can show a survival benefit, how can the FDA not approve it? The answer is the simple reason that because of the design of the trial, Provenge might have had very little to do with that survival benefit. After disease progression, patients on placebo were given the option of going on Provenge or not. 68% of placebo patients then crossed over and received Provenge but not chemotherapy. If you were on Provenge and progressed, you had the option of receiving chemotherapy right away. This by itself could account for the entire survival benefit seen between the two arms as it is generally better to get chemotherapy sooner rather than later. The company has even disclosed data that showed that Provenge patients who subsequently received Taxotere had median survival of 34.5 months compared to 25.4 months for placebo patients (most of whom received Provenge before receiving chemo). Considering Provenge only showed a 4.5 month difference, this 9.1 month difference is pretty important.
Given these confounding factors it seems clear that the survival benefit is in question and will very clearly be questioned in the FDA briefing documents which will be released on March 27th. So let's move on to the more important statistic, the one that was unaffected by patient crossover, time to tumor progression. Here Provenge showed only a 1.7 week difference (11.7 versus 10.0), a difference that was not significant. And given the imbalance in the arms with factors like Gleason score (a higher Gleason represents more aggressive disease and DNDN itself has referred to it as "one of the most important prognostic factors") where Provenge had a major advantage which could easily explain the meager 1.7 week benefit. The Gleason score breakdown was:
39.0% of Provenge patients and 44.4% of placebo patients had Gleason scores of 8 or higher. These are the patients with the most aggressive disease and worst prognosis. 34.1% of Provenge patients and 40% of placebo patients had Gleason scores of 7, which is an intermediate type of score. 26.8% of Provenge patients and 15.6% of placebo patients had Gleason scores of 6 or less.
In other words, the Provenge arm had an 11.2% advantage in patients with the best prognosis. And in a small 127 person trial in only takes imbalances of a few patients here and there to manufacture a benefit.
The big question is, what is the company thinking, going in front of the FDA with this sloppy data? Well that was not initially the plan. For those of you following DNDN for years as I have, you will remember that initially they were going to run a 500 patient confirmatory Phase III under an SPA with the FDA (AGIX and Telik (TELK) also had SPA's by the way, so that just proves that SPA's don't count for anything in terms of the FDA signing off on a product). They did initiate the trial and were guiding for full enrollment in 2004. Then 2005, then ...
Well, they just decided that it might take years to fully enroll this trial (a very bad sign as doctors appear completely unexcited) so why not just file with what they have. That is not a very good reason in my mind to file with the FDA, because your Phase III isn't enrolling well (and is still not enrolled). It equates to a biotech hail mary pass. And while we are on old management promises, remember how they were in final discussions with 2 or 3 potential partners for Provenge in 2004? What happened with that? I guess the experts in Large Pharma didn't believe the data either.
Besides not trusting management because of promises they have made in the past, I also think the CEO pays himself much too much to be an honest guy. Just like Michael Wick, the CEO of TELK, and Russ Medford, CEO of AGIX, CEO Mitch Gold pays himself a hefty salary of $673,000 (salary + bonus) with an additional $309,000 in restricted stock awards. In other words, DNDN's CEO pays himself over $1 million a year as he erodes shareholder value and can't even enroll a 500 patient Phase III within 5 years.
So what happens with DNDN stock? I think the FDA will do the prudent thing and recommend that they wait for the confirmatory 9902B trial results to come out (DNDN is currently guiding for year end 2007 full enrollment, they really mean it this time) as that would more likely be able to definitively prove whether or not Provenge is a real drug or not. This would effectively mean a 2-3 year delay. The stock should then trade at no more than $2, given DNDN will have around $1 per share in cash (about $80m total as of March 31 by my calculations) and the bulls will likely give it some residual value unless the FDA is as unmerciful as they were with GNTA, in which case it would trade to $1.
And as they are burning through cash faster than a Weimar Republic pensioner (sorry. had a Dennis Miller moment for a second there), they are likely to have to raise cash again in the next six months. At $1 they would dilute shareholders 100% for 1 years worth of cash, at $2, around 50%. So this stock would likely be dead in the water for some time.
In conclusion, yes if it works it could be huge, but most likely it won't and if you are left holding this stock, you will end up with far less money than what you put into it.
Disclosure: Author has a short position in DNDN
DNDN 1-yr chart