$SLS In a trial with 2 arms of 63, 32 deaths sets the median in each arm. Technically the median is set with the death that crosses the KM 50% boundry. However, that only comes into play with dropouts, censors, which would make the median number lower, it can NEVER be higher than 32.
Therefore, at IA with 60 deaths in Dec of 2024 the median for BAT was set. 15 months later GPS median still has not been set. BUT if it was set today, the resulting HR would still be under .636.
No one wants the trial halted and then end up with a fail and a waste of time and money, but I haven't seen one credible math model to explain how REGAL could fail at this late of a date. Older patients that are AML CR2 status just don't live this long in any great quantity and you would need 25% alive to set the GPS mOS today.