First, admit you know five censors came from the difference between the number at risk versus the number surviving at five years, and no censors are evident at one, two, three or four years.
Next, admit you know why they wouldn’t have been able to end the trial at the time of an en masse crossover for ethical reasons, and that not ending the trial (even if they thought they had persuasive numbers), could have been due to the inability in 2015 to convincingly demonstrate (for approval reasons back then (older technology)) PFS versus psPD and because the FDA back then, held the position in writing (later to be changed) PFS was not a sufficient endpoint for approval (See IMUC).
In fact, you believed PFS wasn’t enough back then.
(Besides, they didn’t have automated closed system manufacturing technology back then)