There is a hell of a lot more to know before we can say the data is valuable or not.
. How many biomarkers were looked at? If one was to search 20 random variables one would be expected one to be stat sig.
. Did the biomarker expressing subgroup do substantially better on Bavi than placebo? Note that the stat sig in the PR does not imply this.
. Is the biomarker something that looks like it is real? Something that Bavi could reasonably be expected to work better in?
. If Bavi was relatively better in the expressing group, it must have been absolutely worse in the non expressing group (since the combined population saw no benefit). This would be odd, as nobody here thinks Bavi is harmful. And such oddness leads one to suspect random chance.
Yeah, IF these concerns are addressed Oct 10. And if, as you say, the population is significant. THEN one could make a case for running another trial in that population. A P2 though, unless they want another crapshoot.
Bingo. It would be stunning it they didn't find a correlation with SOME marker after data mining. What matters is how many patients, how significant it is and does it make any kind of sense. Regardless of what the answers to those issues are, it means nothing until phase I or phase II trial is run based on that biomarker. What is also important is that they give data on the mOS and PFS from Sunrise for the overall population. If they don't give that basic information, then PPHM won't even have a shred of credibility left.
As far as being excited about ESMO, this is what generates excitement.