You know, I've written quite a few posts about censoring on the K-M curve - trying my best to understand how it works - but I haven't really looked much at how censoring would impact the K-M curve for PFS.
Since I believe that when they announced that they had reached those 248 events is also the time that they actually did reach that number - then I correspondingly think then that 83 patients then had not even evented. This too is exactly what Baranuuk aka Bohsie in his seeking alpha article thinks.
And so one can see from your DCVax curve how all those censored patients would potentially plot out on the K-M curve. Since PFS is not confounded by cross-over, meaning both arms legitimately represent what they were intended to represent, one has to think what those 83 patients who are not only still alive - but had actually not evented would do to a K-M curve when the trial is eventually stopped and they are censored onto it.
Baranuuk made a chart showing that if those 83 patients were the last enrolled, their PFS range for these 83 patients by February 2018 (not today) would be from 15.27 to 23.30 months. Bear in mind, that's 25% of the trial population showing a range such as that. See his chart below.
So it's worth considering that perhaps they've perhaps been holding off for the 233 threshold OS events to be reached in order to also demonstrate a particularly impressive K-M curve for the primary endpoint of PFS for the trial's unconfounded treatment arm - which will be measured against the control arm.
And to restate - those patients that had not even evented by February 2018 are demonstrating some impressively high numbers (see chart), even if you assume they were the last 83 patients to enroll. Which is likely not the case. At which point, those numbers only get better.
News 
Market Data 
Discover 
