So there must be something, even in the original data, that the European authorities liked
Well there was a very strong HR in the PFS data (less than 0.5 in the FR(++) group), and if you stratify the patients based on platinum sensitivity and other factors the survival numbers evened out even prior to the latest update (in fact there was a clear advantage in the FR(++) group). Also small numbers of non-drug patients and an exceedingly long survival in the non-drug arm.
As I commented a few weeks back on SI, I thought that they have a shot with the EMA with the current data, but not with the FDA.