Replies to post #54621 on NorthWest Biotherapeutics Inc (NWBO)

[exwannabe]

Even though they had to define subgroups in advance, just in case they choose to use that later, they may not have been required to commit in advance to whether efficacy will be evaluated subgroup by subgroup, or as just one big GBM population.

Not quite.

Predefining subgroups is not sufficient. Most large trials have many such, and results might very considerably by pure luck.

To do what you suggest, they must formally allocate the alpha spend amongst the groups in question. Thus, the trial plan could predefine something like p<.03 as the target for the ITT population, and p<.03 as the target for messenchal subgroup (the numbers need not add to .05 because groups overlap, I do not know what they actually could be though).

If this was to be the case, then they could go to the FDA with whatever works.

Is it possible? Perhaps. We have never heard an explanation of why LP says the trial is powered at P<.02, perhaps the rest of the power was assigned to a subgroup? If this was the case, LP might be keeping it quite for fear it would be taken as a sign of going for a smaller indication.