I'm not understanding what you are getting at here. Why is it strange for the FDA to ask for an updated interim analysis?
The company passed their pre-defined interim OS endpoint by a very large margin - but was effectively told by the FDA that it wasn't good enough. They had to finish the complete trial. I.e. interims do not count.
The surprises are:
a) The FDA violating statistical common sense.
b) The company not announcing an interim passed with many zeros.
c) The FDA (which in other circumstances guides companies to do x-over) effectively nixing x-over here when no one was looking.