Median overall survival is a very commonly used barometer of a drug's efficacy, however, it's not ideal. The preferred way to measure survival is by calculating a hazard ratio, which compares survival of all patients on the drug being studied compared to survival of all patients on placebo or some comparative treatment.
A hazard ratio of 1 means survival, or the risk of death, is the same for both arms of the study. A lower hazard ration (less than 1) equates to an improvement in survival, or a reduction in the risk of death.
As I said above, J&J reported a survival hazard ratio of 0.79 for the Zytiga pre-chemo study, which means patients treated with Zytiga had a 21% reduction in the risk of death compared to placebo. (1 - 0.79 = .21)
I feel smarter already. Thanks. So, if Xtandi brings home a survival hazard ratio of less than 0.79 in the PREVAIL study, it's a win?
Call the Xtandi bogey a hazard ratio of 0.75 or less, which would be meaningfully better than Zytiga. An Xtandi hazard ratio of 0.75 would also imply a median overall survival benefit of 8-10 months compared to placebo (assuming the placebo arm comes in with a median overall survival in the range of 26-30 months.) That would be a nice improvement over the five-month median survival benefit seen with Zytiga.
One more important point to keep in mind: J&J's Zytiga must be given with a steroid, while Xtandi does not. Urologists, who treat a majority of prostate cancer patients, hate giving steroids. This is a big advantage for Xtandi.
What if Xtandi doesn't reach median overall survival when the results are announced?
Hmm. I think you're smarter than you're letting on. You raise a good point. We may not get median overall survival in the Xtandi arm but that would be a plus, not a minus, as long as the hazard ratio is strong.
You've written 700-plus words on Xtandi so far and no mention at all of biostatistics and p values. How is that possible?
Okay, now you're mocking me. We definitely want to see a statistically significant p value attached to the Xtandi results, which would be another point of differentiation from Zytiga. You may not remember, the Zytiga survival benefit was not statistically significant, although to be fair, the statistical miss did not deter FDA from approving the drug anyway for pre-chemo prostate cancer patients. Same can be said for Xtandi in the unlikely event that the survival benefit comes back with an errant p value, although it won't make investors very happy.
When will Medivation and Astellas announced results from the PREVAIL study?
Official guidance is "before the end of the year," but I suspect the end of September or October is more likely. That's just a hunch.
How can you write a story about prostate cancer and not mention Dendreon (DNDN) or Provenge?
-- Reported by Adam Feuerstein in Boston.