So let me try to walk through that and clarify a little bit. As you know, we have three cohorts, with cohort 1 being a certain dose and a cohort 2 and 3, so there was from 500 cells per microliter to 2,000 cells per microliter. So we have cohort 1, 2 and 3. And last time we mentioned, because we had gotten so good with all the different centers coming online, and we had patients, we had an opportunity to get a few extra patients, and we did. So we got 3 extra patients enrolled, so that's how we get to the total of 12 patients. We really haven't shared a whole lot of data from those additional three patients that got treated over the last quarter, so that 12 now going to AAO in October is more meaningful than the 8 or so patients that we had shared at ARVO. So we'll have a larger number, significantly larger compared to 8. We'll have 12. And what we think is important is that continued watching or seeing the structural changes that we saw, right? Reduction of drusen we think could be pretty meaningful, and a lot of KOLs agree. The improvement of the ellipsoid zone, the RPE layer, maintaining a lot of those changes for a longer period of time is also pretty meaningful, because we saw some of them at month 4, month 5, and that was pretty good that we can already see some improvement at the month 3, 4, 5, but does that sustain for 8 or 9 months, or 12 months. Now, in either case, there is a product, right? But the idea is would this be a really long-lasting effect, or would there be a shorter effect? Those are things that we can see with some of this data that is longer-term. Some of the earlier patients, they're heading out to almost 2 years, 2-plus years. But I think more important for me is to see that the patients that are now at even 9 months that we had shown 4 months of data, are we able to see similar or continuing improvement. That's important in the first 3 cohorts. The fourth cohort, which we just started treating recently, now, we think if we have 3 or 4 months' worth of data -- I'm sorry, 3 or 4-month time point data, given that the 3 or 4-month time point data and the previous cohort was pretty meaningful, it'd be interesting to see. Whatever we have, I can't promise, and it's hard for me to now make those commitments. But we think we'll have some. Whatever we have we'll share at AAO, but, more meaningfully, the larger cohort of -- I'm sorry, larger number of patients from cohort 4, the data will be more complete end of the year, beginning of next year.