Yes. That's a great question, Michael. There are two main measures of density on the mammograms themselves, and then I will discuss the concept of statistical significance and clinical significance, which comes up in every single clinical trial of any drug you would be involved with. So the first, the measurement endpoint. Radiologists have for almost two decades used a A, B, C, D, categorization of approximately quartiles where A is the lowest density and D is the highest density. There is an abundance of research results looking at those density measurements. And for example, the effect on sensitivity of finding cancers, the effect on a 5-year incidence of cancers going forward. So the long history in mammographic density is that A, B, C, D, categorization. Now we have gotten to the point where we now have 5, 10 cleared, measurements, which are quantitative. The machine does it for you with software. Those give you a score from zero to a 100. And so we have less long-term experience with those, but they do map onto what you'd expect where a radiologist calling a breast A, the machine calls at zero to 25, 26 to 50, 51 to 75, 76 to a 100. So the biggest winners are when you cross the thresholds between groups, because that's the highest level. Statistical significance will probably require a pretty small amount, perhaps 3% or 4% will be enough. The breast in these premenopausal women, the breast will be changing less than 1% per year when they get near menopause and you'll see a 5% or 6% drop over a couple years. But -- so statistical significance would be a couple of percent, clinical significance will be larger drops and perhaps changes in the A, B, C, D categorization.