<"Why is the time to treatment failure a better measure of efficacy than total treatment failures over the 50 day period?">
It is like a Survival trial, say the Provenge 36-month analysis.
The total treatment failures measures only how many failed, which is of course the most important thing. Statistically it is a simple analysis (I can give you a link if you want to try).
The time to treatment failure of course also measures how many failed therapy, but spices it up by factoring in a bias for how long exactly each patient lasted on therapy before failing. That is not a simple analysis (usually thing is to do a "Kaplan-Meier" study on the data).
Maybe the most important thing though is that the pre-specified method was the time to treatment failure.
Which is why you and I were able to buy the stock so incredibly cheap.