Avid Bioservices Inc (CDMO)

[exwannabe]

Let us dig into the Herceptin example compared to the Bavi example.

Herceptin was developed specifically to target the HER2 gene. No big surprise it would work better in such. But at the time the concept of testing patients for a cancer gene in real world practice did not exist. So DNA did the "normal" thing and ran the P2 trials in all patients.

But they critically analysed this and launched the P3s in those over-expressing HER2. And the rest was history.

Was seeing the HER2+ subgroup in the P2 data mining? Possibly, but I would assume they were always looking for this as a key secondary in a P2.

Now Bavi.

Bavi needs B2GP1 to work, that we all know. So is it reasonably possible that Bavi works better in those with higher concentrations of the glycoprotein "glue" that bavi needs? Of course.

Then Bavi runs the P2 in all patients. Reasonable.

But at this point they failed to detect the correlation. Was it present? Did they not look? Was it Fargo? Who knows.

So they ran the P3 in all patients. and SUNSET happened.

My point is that PPHM failed to get the job done in P2, and that cost 3+ years and more than a couple dollars.

BTW, I am not saying the <200 correlation is real. But I can believe it. I strongly think the >240 is a "pure" data-mining where they are seeing pure chance that has no basis in theory.