Being best in class is great, but it puts a target on your back that everyone is trying to beat.
First up today is Johnson & Johnson
While we know that Zytiga plus prednisone works better than prednisone alone, it's the magnitude of the increase that counts. Johnson & Johnson is saving the number for a medical meeting, but we can take some stabs at whether it might beat Provenge based on what we do know.
A larger difference in survival is required to stop a trial early than is required to show an effect if the trial goes to completion, so the early end suggests a large difference for Zytiga. But Provenge's pivotal study, which squeaked by with a p-value of 0.032, enrolled only about 500 patients while the Zytiga trial had more than 1,000 patients in it. With more patients, you can detect a smaller difference with the same p-value; if you're flipping coins, the difference between 445 heads and 455 tails isn't as significant as the difference between 45 and 55 even though they both have a delta of 10.
Of course, not all the 1,000 patients read out, since the trial was stopped early; if you assume the interim peak was halfway through, the minimum increase that Zytiga could show is probably higher than the 4.1-month increase Provenge demonstrated. If the interim readout was later, Zytiga could, theoretically, show a smaller difference that was statistically significant enough to justify stopping the trial earlier.
Even if Zytiga does show a larger increase in overall survival, which I'd guess is the case, and you ignore all the usual precautions about comparing results with different control groups, the results might not be that bad for Dendreon. The two treatments work in entirely different manners, so doctors could use them sequentially or concurrently, which Dendreon is currently testing. If you can't beat 'em, join 'em: Seems like a reasonable strategy when companies are gunning for you.