Sure. Sure. I can talk about that. I mean we obviously are in very close contact with participating PIs as well as key opinion leaders. So -- and I can just reiterate. There is really a great enthusiasm for a drug that works across different subtypes of low-risk MDS. That's to start with, right? So for them, there's just the fact that they don't need to subsize whether you RS+ or RS-, makes their clinical and every day work, very easy, right? I mean we were not subtyping for RS positivity prior to REBLOZYL's approval, let's put it that way. So this kind of is one convenience for that, right? And then, I mean, the fact that we have durability of 20 months versus 7 months of our already approved drug that is doing very well on the market is really another convenience, if you will, for patients. You don't have to come back to the clinic for multiple transfusions to which they're used to prior to treatment. And I have to say something and really often discuss this internally. I mean, what really makes investigators excited is the molecular data. It's really -- and this was -- that's maybe even the best reason why the paper ended up in JCO. So it's really the decrease in not only SF3b1. SF3b1 is just an example that we are using here to illustrate the effect. But it's really multiple mutations that we've seen and also cytogenetic rate decrease in these patients that may believe our KOLs, NPIs that imetelstat is really altering the course of disease in low-risk MDS patients. Same goes, Chip, you answer for MDS. I don't know, Jason, if you wanted to cover for MS, but it is the same pretty much is really the molecular data. And I mean and Chad also a little bit to go to your question, right, about the OS. I mean data has been evolving and the agencies has been more receptive to different endpoints, right? We started with SVR as Jakafi was on the market for so long. But as new data with new compounds, right, show that we can improve anemia with one drug. We can improve symptoms, right? New endpoints has been established. And now we come with OS. We had a question, that the crown of the endpoint, I mean, the ultimate endpoint of every clinical trial, right? So when you show such data granted at the moment, it's just from 2 different doses of imetelstat, enthusiasm is high. I don't think we had leads to justify the OS as an endpoint to any of our investigators nor to regulators so far. So we remain enthusiastic. It's long, and that's what makes it a challenge. It's a long study, right? It's a different endpoint. But other than that, there's really no other reason to believe that we have any questions about the data.