>>"at 260 one is certain to have median survival data"
Assuming A-7 has efficacy *vastly* superior to control: at 260, all control patients are gone and 130/260 A-7 patients are gone.
Moving away from that, at 260 deaths, if less patients in control arm died, that means more had to die in the A-7 arm (in other words, you'd reach both medians sooner).
Only way this doesn't hold true is if the control has vastly better efficacy than A-7. If 200/260 A-7 patients died, and 60/130 control patients died, you could have 260 deaths without reaching both medians. Extremely unlikely, but hypothetically possible.