If the subgroup is large and efficacy is strong enough
like a "cure", it means OS KM cures have long and diverging
tail and should drive the P value into SS, hope you are
right and this should be good results street will like to see,
and logic is it should not take so many month more to get
the clear picture, CVM already knows HLA is the biomarker
to look for and the HLA data should be in the data base, IMO.