Replies to post #157735 on Biotech Values

[mcbio]

What I'm saying is that this serum protein is probably just a reflection of patient health within the treatment cohort. So even though these people have failed X number of therapies, they're not in identical health nor is it a homogenous disease, and the baseline AAG measurement likely reflects that.

I'm pretty sure ARRY has said this is not the case at all and AAG levels do not correlate to patient health.

If I was ARRY I wouldn't exclude based on AAG. From the literature it doesn't appear that you can make a reliable, prospective stratification because many papers are just retroactively binning the serum levels into quartiles and quintiles. Dunno if you want to try and sort through all that and risk your clinical development plan on it.

Yeah, I'm not sure what ARRY is going to do for pivotals. They have discussed the various options but seem to be awaiting further data this year to inform pivotal trial decisions. Why not just run one pivotal where you don't stratify based on AAG and another where you do?

[jq1234]

I agree. I made similar comment when they first released the data before ASCO last year that AAG isn't a biomarker. Not only AAG doesn't relate to specific disease but also change too often, thus the nominal increase of response rate from 22% to 32% by using AAG baseline data point as "biomarker" retrospectively reflected something else in the trial population.