And as flipper indicated, those ticks indicate either a person alive, or LTFU. Eventually, when they unblind, those ticks are frozen in time. If the patient was LTFU and they don't discover what ultimately happened to the patient, then that tick will remain in the same place. If the patient continued living, their tick mark will move forward to the point they reached when the trial unblinded. If the patient lived and then died at a later date, the tick will disappear and they will be plotted as an event on the curve - with the ability, by the way, to change the median, backwards or forwards, depending upon where they plot. Of course, one patient isn't likely to move the median. But 25 or so, depending upon how much longer they lived, could move the median FORWARD.
But as others have pointed out, it's not really the medians that should matter. It's the long tail. And your chances of living to a certain point in time, based on how long the patients actually did or do live in this trial.
And those ticks on the curve represent the trial's blended data back in March 2017. I would venture to guess that if they found that some of those patients were LTFU back then (9, 12, whatever the number), I'm hopeful they've been able to track down the data on a good number of them by now. As Avii indicates, it doesn't take a rocket scientist to find out when those patients actually died. Oh, and if they do, those numbers will also move up the curve to the real time they passed. Guess what that can do?
Move the median FORWARD.
News 
Market Data 
Discover 
AI
