I don't believe you're interpreting that quite right. That projection was to see how many more trials could have been stopped earlier without increasing overall fp risk and determined that 92% could have been terminated EARLIER than they were with planned interim analysis and for 36% it would have been so much earlier that they were still in accrual (just like the dcvax-l trial).
While the trial is sized to be CAPABLE of 0.02@82% it does not NEED 0.02. 0.049 is still statistically significant and remember, the success criteria is simply "significantly higher progression free survival." Six months came into play because they needed some kind of expectation around which to design/size the trial, not because six months is needed (and for that matter, not because six months is enough). Depending on number of patients and variance among them you could have a 2 year median PFS advantage with p=0.05 or a 2 month median PFS advantage with p=0.0001.
The p-value takes into account the number of events and inversely correlated variance between them. Granted, the error rate implied by the p-value ("chance that this is what happened but it happened due to coincidence or confounding variable") is slightly different from the false-positive rate of calculating the outcome and its p-value ("chance that this isn't really what happened" [e.g. 'you should have measured after 67 events instead of 66']). Still, with a study this size, there's a meaningful convergence on the p-value. This would be more concerning if we had an interim analysis planned for every 10 events (essentially a series of phase I trials) but we don't.
The acceptable false positive rate depends on ethical considerations on each side. The DMC has to be certain benefit outweighs risk. *First, what's the consequence of a false positive - if the treatment has serious side-effects, a false positive rate of 10% at interim is unacceptable because if we turn out to be in that 10% scenario, people have been harmed or killed as a result. We'll call that the Gliadel scenario. If the treatment does not have serious side-effects but is a huge burden in some non-medical way (like costing an order of magnitude more than existing treatments thus creating an either-or decision), we'll call this the Provenge scenario, maybe 10% FP rate is acceptable but 15% is not. If the treatment does not have serious medical or non-medical side-effects and has no opportunity cost (is used on top of SoC not instead-of), we'll call this the DCVax-L scenario, maybe a 15% FP rate is still acceptable because even if the outcome turns out false, who cares? Nobody's any worse-off as a result.
*Second, what's the consequence of a false negative (or ignoring a true positive for X amount of time in favor of maturing data). If this is a safety trial of cosmetics, nobody gives a damn if it takes another six months to hit the market. If the drug is an incremental improvement over an existing effective therapy in a grave indication, or is for a serious indication less severe than gbm (like hodgkin lymphoma), people want it on the market as soon as reasonable, but can live with a short delay while investigators make sure. If the treatment has orphan drug status in a highly and rapidly lethal indication with largely unmet need, the consequence of a temporary (six-month) false-negative is the death of thousands of mothers and brothers and husbands and daughters...
Finally, this analysis is taking place with 60% of the events designed into the 0.02@82% endpoint. In other words, while it may be 'interim,' it's also "mostly there". Requiring a .005 this time might make more sense if this one were at 25% and the next one at 50% and a third at 75%, but that's not what's happening.
Now, I grant that accrual phase positive ethical preempts are not the most common outcome and people probably shouldn't plan on it, but, I think your speculated 0.0001 requirement is much more stringent than any clear-headed DMC would mandate under these specific circumstances. If people found out (after the end of the study and publication of raw data) that it was so obviously a victory this early and yet delayed for months, the DMC members would be flooded each year with birthday cards addressed to the patients who just missed treatment because of their decision. That won't happen, because these folks are doctors themselves with a passion for helping people.
Anyway, we'll know in a few days what happens, one way or another. I do find it heartening that even if we 'just' get a continuation, the 88 event trigger is really close behind on the calendar.
News 
Market Data 
Discover 
