Elephants
1) OS between the ITT arms. It was a ranodmized trial after all.
From surgery (as 2018 did not have all from randomization data)
mOS Blended : 23.1 m.
mOS Treat:: 22.4 m.
mOS Control (est): 23,8 m.
36m OS blended: 23%
36m OS Treat: 22%
36m OS control: 25%
Why did placebo trend better than treatment?
2) MRD (minimum residue disease)
Virtually no DCVax-L benefit over historical controls in this. OK, but what about the "significant" residue population? DCVax-L did well there. Kool. Alas the P3 required "intent for near total resection" while others allowed partial resections. It is reasonable to assume that the "significant" subgroup had less tumor ln the P3 compared to others, and that would explain the trial results.
3) What happened in 2015.
Many longs argue that RAs saw that placebo was harming patients. How would they see this when the opposite was happening? The.more rational guess is that it was futility based on the PFS.
Fair enough to say that PFS was always bogus, but now they would know that the trial failed the original primary. And knowing that they changed the endpoints. That is a no-n, even if blinded.
4) Why this crap about waiting for a paper then dumping info like this?
Because a top end journal will not accept a paper with the paper in major disagreement with the clincial registry. And they can not update the US registry because that will put them at risk for not disclosing the laughter the FDA replied to the SAP change with.
