Can I prove it failed? Of course not. Statistical models can not come close to proof, and barely reach the level of suggestion.
whether the 248 PFS occurred in 2017 or way before 2016 like they claim
I am not sure what posts assert 248 was reached way prior to 2016. I have posted that I think (not know) it was reached way back IN 2016.
The hypothesis that the trial failed PFS back in '15 was based on a presumed IA (the second one LP mentioned, that she never took back like the first). This would not have required 248 PFS events.
If the PFS failed way before 2016, how is that possible there are 98 or 100 still alive.
I assume you are talking about the difference between the 248 mark and now.
And that is easy to address. Any of the following, alone or in combination:
. DCVax improves survival but not PFS. . pseudo progression post 3 months confounding the results. . everybody living longer due to a better supportive care / rescue . everybody living longer due to experimental treatments. . an inclusion criterion, or change in definition of PFS, causing all to go longer w/o progression.
marzan, the majority of the 100 alive in June were PFS. Around 12/16 or 1/17 there were around 83 PFS. Based on the death rate of 2 per month since June I am assuming an equal rate of about 2 progressions per months from 12/16 or 1/17. In June we would have therefore ended up with about 71-73 PFS among the 100 survivors. Even if a few patients were lost (LTFU) the PFS constitute a clear majority among the 100 alive.
This would not prove that PFS was a success since we don't know whether the difference between controls and trial was SS but it does show that a very large proportion of the patients (maybe a subtype(s)?) had delayed progression and an extended survival and it would strongly suggest that DCVax-L was responsible.