If survival probability for the VB-111 arm is 50% at 12 months, 40% at 15 months, and 30% at 18 months, then 75% events won't be reached before 18 months have elapsed for patients in the VB-1111 arm of the trial.
Thus, I think the 189 events is a figure of speech. For argument sake, assume VB-111 show stellar results in Globe trial with 80% survival at *24* months for the VB-111 arm, then 189 events won't occur until 2019 or 2020. Under these hypothetical conditions, it would be irresponsible to delay approval for VB-111 until late 2019 (when 189 events occur). Several year would have been lost for no good reason.
The Globe trial is intended to show an improvement of VB-111+Avastin over just Avastin. Given Avastin's past OS results, the Kaplan-Meier survival curve should be able to show convincing results after 12 months (subsequent to trial enrollment completion).
The FDA knows that delaying approval in order to obtain more precise results needs to be counter-balanced with benefits of early approval when current results are already positive (but not as precise as they could be).
We cannot know for sure until the results are published.
Both rGBM and prOC are very aggressive cancers with abysmal survival numbers. P2 results show benefits of VB-111 in both indications. However, there is a risk that P3 results will disappoint. This risk is largely mitigated by the "all-comers" filter (or non-filter) of the two P2 trials.
