It is a certainty that the change in MOS was due to some censors being resolved.
Most of these will just be events that happened in the entervening 3 months bewteen data snapshots.
It is certainly quite possible that some of those early censors also got resolved. Clearing 3 or 4 of those would certainly remove any issue in my mind on that subject.
They will probably not release this information prior to the full data being presented.
P.S., to another poster who replied to you, censors are NOT deleted from the data. This trial was the full ITT (intent to treat) set. a patient would count once they were enrolled even if they were shot by a gun prior to receiving treatment. Even in an analysis that uses a modified ITT population censors are not deleted. The patients not in the data are simply deleted, but that is not a censor.