John Scarlett
Analyst · BTIG.
Maybe I can just comment that the 8 week -- Aleksandra, I'm sure will have a few other comments. But Tom, the 8-week RBC-TI is kind of the accepted or approved regulatory outcome or regulatory end point. And so you're -- it's really common in virtually all of these studies in lower-risk MDS to use that as the primary outcome measure. I think we have a very robust -- in the Phase II, we saw a very robust outcome measure. And even in the MEDALIST, when you got into the higher transfusion-burdened patients such as we will have in our study in our Phase III as we had in our Phase II, the noise really goes down low -- much lower. You're bringing up a really interesting point. So for example, with the MEDALIST study, which I just used MEDALIST in this case not to harp on anything with regard to luspatercept, but because it is a study with a wide variety of patients in the baseline transfusion burden. So when you looked at the patients -- just at the placebo rates because that's what you're actually kind of getting at, right? So when you looked at the patient's placebo outcomes, in their study as a whole, which included, remember, from 2 units transfusion burden up to actually 20, their placebo response rate was 16%. And I have to say, when we looked at that, we went, "Oh, holy cow, what was that?" But the reality is that when you knock out the patients who have the less than 4 units per 8-week transfusion burden, so the 2 and 3 patients, okay, that placebo rate comes down to 7%. In the MDS 005 study, which Celgene did couple of years ago and was reported, well, yes, a couple of years ago at ASH, that was Revlimid versus placebo. The placebo rate, as I recall, was 4% or 5%. So I think we don't see a signal-to-noise ratio here as long as you have a sufficiently high transfusion burden. And if memory serves me correct, if you then flip the analysis and you look at the patients in the luspatercept study that had only 2 units or 3 units per 8 weeks in just that subgroup alone, their placebo response rate was extremely high. So it -- I think they're -- it's really defined by the baseline transfusion burden. And that signal-to-noise ratio in the patients that we'll be studying, I think, is more than adequately high. The additional durability is the icing on the cake, right? And I think everyone will pay a great deal of attention to the durability because we would expect an exceptionally low -- I expect a very low placebo response rate there. And I think that will not be an issue at all in signal-to-noise ratio. I hope that didn't [indiscernible] too much and confuse it.