Replies to post #243168 on Avid Bioservices Inc (CDMO)

[exwannabe]

The 2:1 ratio I’m expecting to see in the first look-in, is a derivative of the MOS data that we saw in the phase2 NSCLC trial. The MOS of the Doce arm was 5.6 months, just short of 6 months. The Bavi arm MOS was 11.7, just short of 12 months.

If Doce events were occurring at a mean of 6 months (rounded), and Bavi events were occurring at a mean of 12 months (rounded), a look-in on that trial, at some point, should have seen approximately 2 Doce events for each Bavi event. The Bavi MOS was twice that of the Doce MOS. That’s where my 2:1 derivative comes from.

Even if I accept your P2 numbers (which I do not), you are making a fundamental math error here.

First, mean is not median.

Second (and more importantly), your math fails to account for the simple fact that as the trial progresses there will (hopefully) be more patients at risk in the bavi arm, thus more deaths there than in the placebo arm.

As a simple reducto to absurdium, consider the event count at the final look. Let us suppose the end is set at 500 total events, clearly the ratio will not be 2:1. Perhaps 270:220 would be an exceptionally good result.

BTW: Opdivo targets T-cells, not cancer cells.

[Protector]

sunstar, I agree you must use "Herbst et al, 2010" to compare/simulate/guestimate SUNRISE related things.

That is the mistake I made in my simulations, I use PPHM 2nd ln NSCLC results and that makes a BIG difference (9-12 months) because 5.6 vs 9.9 months for the Control Arm alone is almost DOUBLE.

Furthermore Herbst et al, 2010 has only 1% of the patients that are not of the same ECOG, so that is a good basis to compare with. And of course we have always to keep in mind that a Median is not the same thing as an average. It is the middle value of an ordered list (by survival time in this case) of which we, sometimes, also get the window (first and last values).

But the REAL problem in dealing with the 1st and 2nd look-in is the Bavituximab arm. We know about bavituximab's latency and saw on the charts (of past trials) that it doesn't kick in immediately. The separation takes some time to kick-in. Now we have a controlled clinical trial with much healthier patients. That means Bavituximab has the time to kick in WITHOUT loosing as many patients that will event BEFORE Bavituximab starts working as i PII because there they were ECOG 3 or 4 (much sicker).

That may lead to SURPRISING Effects because in a median table a few eventing early doesn't weight in. A median table can have big gaps, say some eventing at 5 months and the next one at 10 month, then one at 11 months and the MOS would still be 10 months with a 5-11 window. And that a median is used vs an average is a good thing for Bavituximab in this case IF healthier patients suffer less or not at all from the Bavi latency.

The latency seem to be about 2 weeks. The Yervoy trial was designed to give Bavituximab alone during 2 weeks and then start the cycles. I think that is a PPHM move to overcome the latency and see what Bavi's effect is WITHOUT the damage of, in this case IO therapy such as Yervoy. I think in chemo, using lower dosage chemo, would even improve the results if combined with Bavi because the chemo is the cause of a lot of collateral damages leading to complications and side effects (mainly infections leading to ...abc).