Replies to post #250523 on Biotech Values

[iwfal]

Have you seen the final trial design? It’s just one cohort. The 2L setting was eliminated because of uptake of Trodelvy in that setting. The trial is overenrolled and seems adequately powered to me.

Agree they have disclosed that the PDL+/Second Line cohort was shut down early, and most of the enrollment has been in the other cohort (first line only, no selection for PDL status). But, for the record, it should be noted that Clinical Trials still shows the original trial design in this regard - no way to tell one arm got shut down early.

There wasn’t a material difference in 1L vs 2L efficacy in the P2. It was numerically lower but very small numbers. The bigger variable is the larger PDL (-) population in the P3

This is incorrect:

1) The Forrest Plot in the published ph2 was: 1L vs 2+L OS efficacy HR was 0.46 vs 0.22. And although the arms are not huge (for 2+L it was 25 vs 13, for 1L both arms are about 50% bigger (43 vs 21)). Regardless, this is a big split (albeit not stat sig), and the company chose to go ONLY with the weaker group in the Ph3 cohort that is the upcoming interim.

2) The PDL+ vs PDL- in the published poster for the ph2 was: PDL+ vs PDL- OS efficacy HR was 0.34 vs 0.48. This is a substantially smaller gap than for 1L vs 2+L. And the arms in the PDL+ vs PDL- are effectively **even smaller** than for 1L vs 2+L, so more uncertain (for PDL- it was 26 vs 10 and uncertainty is mostly driven by the smaller arm. For the PDL+ it was 32 vs 17). Finally, unlike the 1L vs 2+L distinction... they did not discontinue one of the PDL groups in the ph3 cohort that will be the the basis of the upcoming interim. So they will, by definition, have closer to the same ratio of PDL+ vs PDL- seen in the ph2 than is true of 1L vs 2+L (but see further discussions on this below)

Further commentary:
a) The 1L vs 2+L distinction for different efficacies in cancer treatment is very old and well founded. I.e. you should expect to often see big shifts in efficacy, although the direction of the split is generally uncertain prior to trials.

b) As best we can tell, the split in PDL status that they saw in the ph2 did have more PDL+ more than typical. They saw about 57% PDL+. And in other TNBC trials it's 30 to 45%, so the ph2 data appears high. But now we have to deal with big uncertainties in 'expected rates of PDL+ vs PDL-'... e.g. the rate of positive is likely much higher in primary tumors than in metastases, and even higher in lymph nodes. It's also likely that it differs by how many chemo regimens have occurred although I haven't seen lit on that. All told, yes, its reasonable to assume that they will have a somewhat lower rate of PDL+ in the ph3, but there is a lot of uncertainty on that. And, regardless, it's not like the 1L vs 2+L where they just removed the better performing group (for whatever reason).

b) I am not sure where this focus in the on-line investor community on PDL status comes from, but IMO it's not particularly warranted. There are 100 of these prognostics, most of which never get published. The unknown unknowns. My guess is the ultimate source of the PDL+ concern is too much of the on-line investment community bought into the narrative upon which the company focused on for a long time (i.e. that it's an IO drug. My comment: maybe, but if so it's weak. My guess is it works by sparing the immune system better? Certainly this is fairly plausible given the history with multiple of their trials showing sparing of blood components).