Biotech Values

[iwfal]

From PR:

However, the primary endpoint, which evaluates the 95% confidence interval (CI) in the difference in response rates between the active and placebo arms, did not reach the 15% lower bound of the CI that was proposed in the study’s Statistical Analysis Plan submitted to the U.S Food and Drug Administration. Following initial conversations with the FDA on these topline efficacy and safety results, DBV will continue ongoing discussions with the agency, and plans to proceed with the BLA preparation process.

DD Given the FDA approval of PTLA's Bevyxxa (which also missed the primary endpoint), it makes perfect sense for DBVT to file, IMO.

I didn't track the Bevyxxa saga, but in this case there is definitely something surprisingly different about their SAP -

1) I've not seen the use of 95% CI before on a superiority trial - e.g. I've never seen an Cancer trial with OS that says you have to show a 95% CI HR<0.8 (instead all superiority trials of which I am aware are effectively 95% CI HR<1.00).

2) Essentially this means that the p value for the trial has to be astronomically low. p<E-10 or something. Way beyond the limits at which you can be pretty sure something is happening.

3) This would, potentially, make some sense if there were no supporting RCTs that show clinically meaningful benefit on their pre-specified primary endpoint - but there is.

4) Finally, note that in their CC they more or less said they picked a random number they thought they could meet (here they were dumb since they based it on their post-hoc changed endpoint from the ph2 - which was inevitably going to regress substantially towards the mean).

All told I think it would be very surprising if the FDA did NOT approve the treatment.

Interesting question: what minimal clinical efficacy did AIMT have to sign up to? And was it using the same mechanism (95% CI must be xxx% greater than placebo)?