So here are the range of radiation doses from lutetium PSMA-617. So I'm presenting the data both in terms of the gray per gigabecquerel delivered as well as the gray theoretically delivered if you give a patient all six cycles of radiotherapy. And this was pulled from a handful of papers, but in particular Zach Ells's meta-analysis of the published dosimetry data for lutetium PSMA was very helpful in preparing this table. So what do you see? You see that the kidneys in particular of interest can have a dose that ranges from four to 62 gray over the course of six cycles. That's a very wide range of 14 fold range from your kind of min to your max. Parotid and submandibular glands get quite a bit higher dose than the kidneys on the order of 13 to 115 gray. Lacrimal glands were discussed earlier today. Some of the estimates indicate that it can get upwards of 170 gray. Bone marrow, we see a wide range again. Usually it's quite low, around 0.2 gray per cycle, but it can be higher, closer to one gray per cycle. And then for tumors, we see a 300 fold range from the lowest reported doses in tumors to the highest reported doses in tumors. So there's an immense interpatient and interlesion variability with respect to the absorbed dose from this therapy. So we're operating in a world where you have some patients that are getting right at the median that you report out for clinical trials, but you have a lot of patients that are getting above and below that.
And so with such interpatient variability, what does dosimetry let you do? It lets you know whether somebody is an outlier or whether they're sort of near the middle of the distribution. And what you do with that information is kind of up to you. So if it impacts management, that's because you've decided it does because of some evidence that's out there. So what evidence do we have? What is our experience from external beam and from other studies and radiopharmaceuticals? Well, the best data that's out there for kidneys with beta-emitting radiopharmaceuticals is summarized in this quite prominent plot from Barry Wessel's in the MIRD pamphlet. And what this is showing is external beam normal tissue complication probability as a function of kidney biologic effective dose, which is corrected for dose rate. And you also see on top of that as a dotted line, which is the same normal tissue complication probability curve calculated for Yttrium-90 DOTATOC using the formula that's shown in the top right, as well as a few assumptions about DNA repair rate and alpha-beta ratio for kidneys. And what you find is that for Yttrium-90 DOTATOC, the data lines up quite nicely for kidneys with external beam. Now, there's been a lot that's been discussed about how Yttrium-90, the beta particle range is quite a bit longer, and so the dose distribution within the kidney is quite uniform. And so that is a good explanation for why these do agree. So in the case of lutetium-177 where your beta range is closer to a millimeter instead of a centimeter, the range of the beta particle likely reduces the glomerular dose a little bit when it's localized to the proximal tubules.
Now, I think that if you take a nephron model and you kind of linearize it and you look at the range of the beta particle and you calculate, well, what is the actual dose of the glomerulus? You're going to become overconfident that the dose to the glomerulus is low. Because if you look at a model of the tubules and the glomeruli in three dimensions, they're kind of bundled up like a ball. And so you may have localization to the tubules, but you could still have a very proximal glomerulus. And so I think that there is still a significant dose that's being delivered to the glomerulus. With that said, we do expect that the threshold for renal toxicity for lutetium-177 is going to be shifted a little bit to the right. We don't know where. We haven't actually explored high doses, renal doses in a systematic way for lutetium in either neuroendocrine tumors or in prostate cancer. But if we did, I expect the curve would be shifted a little bit. For salivary glands, we have external beam data. We know that as measured by a 75% reduction in salivary flow, about a 20% rate of xerostomia at 26 gray to the salivary glands, and about a 50% rate at 40 gray. So how does that compare to what we're seeing with lutetium PSMA-617? We see that rates of around 25 to 50% for xerostomia, depending on the study.
And in terms of the vision trial, you get on average a dose of 28 gray over six cycles. But again, the range is huge. And it is very likely, at least in my opinion, that the patients that are getting the higher salivary gland doses beyond 50 gray, beyond 60 gray, are probably more likely to have severe xerostomia, but we don't know exactly what that dose-response relationship looks like. So at the present moment, would you make any patient management decisions based on a particular parotid dose? It'd be hard to convince my attending physicians to do so. So that's where I kind of fall on the salivary glands. With the lacrimal glands, it's quite interesting because we really see very low rates of dry eye in the range of three to 10%, depending on the study you look at. And that's despite having doses that would be considered ablative for the lacrimal glands for external beam. So what does that tell you? That tells you that the lacrimal glands are unique in their ability to respond to low dose rates. So I think it's worth knowing that the linear-quadratic model is just a model and that sometimes you can fit things with just a quadratic or just a linear depending on the dose rate. We also know that lower dose rates tend to be associated with lower alpha-beta ratios. And so I think fundamentally this is an issue where we're extrapolating so far from the dose rates of external beam that we see totally different responses there to the very low dose rates.
The other thing is that these are too small to measure on CT. So even if you want to do patient-specific dosimetry, you're really only measuring the activity in the lacrimal glands, not the activity divided by the mass, which is really how you get to dose. So it wouldn't be patient-specific in any meaningful way. For the bone marrow, we see no more than one gray per cycle from Pluvicto. Dose-effect relationship has not really been demonstrated for lutetium PSMA-617, but I haven't seen many people publishing that it doesn't exist. So it very well could exist in a way that's sort of similar to the data we've seen with radioiodine. But in general, the doses are so low, you give on order of 0.2 gray per cycle that we really don't see a lot of acute bone marrow toxicity. And more to the point, the patients with the really heavy metastatic disease are both the patients with the highest bone marrow dose, as well as being the most difficult to measure the bone marrow dose. Because you can't find a region in the spongiosa that's free of metastatic disease to do an image-based dose assessment. And so the most challenging cases are the ones where you would want it the most, and it's also the most difficult to measure accurately. So I would say that the prevailing wisdom in this setting is take a watch-and-see approach. When we have the benefit of fractionated therapies, you can see how the marrow responds to the first, second, third cycle and make decisions based on the clinical biomarkers.
Okay. Tumor dose-effect data. So we know that higher tumor doses are associated with improved outcomes. So better PSA reduction following treatment, there's a moderate correlation between tumor, whole-body tumor dosimetry and patient outcomes. We know that patients with "total body marrow dose", and I'm putting that in quotation marks because I have some slides to talk about that in a minute, greater than seven to 10 gray have improved PFS and OS. That's great, but that doesn't tell you whether the patients that are in the low dose cohort, whether they still benefited from therapy or not, right? So you would have to do a study in that low dose range to see whether therapy was futile or if it actually did help them too. So the other things we know is that there are other prognostic biomarkers that we can get from simple things like imaging. So we know that post cycle one PSA response is also predictive of outcomes.
It's not just radiation dose. We know that post cycle two SPECT/CT in comparison to post cycle one SPECT/CT tells you something about radiologic response, and that also is associated with prognostic value. And similarly, we know that higher SUVs on the baseline PET also are prognostic of impacts. So with all three of those, does dosimetry really add a lot of useful information at this time for tumor dosimetry? Well, I would argue that it gives you the benefit of knowing what doses are to individual lesions. And if a patient has painful metastatic disease, meaning there's a particular vertebra that's bothering a patient, knowing what the dose is to that lesion may have particular clinical benefit when you're talking to the patient, you say it's very likely that you'll have symptomatic relief. But otherwise, it can be hard to make an argument that tumor dosimetry is clinically relevant today. I do want to highlight one article from Milan Grkovski and colleagues mostly at Memorial Sloan Kettering. This was published earlier in the year, and this is just some of that correlation data that I was alluding to. So we know that at very high tumor absorbed doses, you have patients that are much more likely to have good PSA response, but in the low dose range, it's quite noisy and it's hard to know what to do with the dosimetry information. Okay.
Now for the controversial statement. So the controversial statement is that I don't think total body PSMA lesion dosimetry has been done accurately yet, and it's debatable as to whether or not it's been done in a reproducible way at all. And the reason I say that is that the gold standard for dosimetry with partial-volume correction is what's called the dual-contour approach, where you seeing the lesion on CT contour the lesion, and then you expand that morphologically to account for the spill-out. So why is that the gold standard? Because spatial resolution in SPECT is spatially variant. Patients move during studies, so you can have blurring because of that. You could have misregistration. And so by separately quantifying the activity of a lesion, as well as the mass of a lesion, you end up getting a relatively reproducible dose estimate across different reconstruction. So this is an example where I reconstructed the same patient data in three slightly different ways, kind of representing the kind of variability you would see among scanners and institutions, and you get roughly the same unbiased dose estimate if you can see the lesion to contour it. Now, what most people do, because you can't usually see the lesions on CT, is you use some sort of thresholding approach.
A 40% threshold about a tumor will universally overestimate volume and also underestimate activity. So what you see here is that the 40% threshold gives very unreproducible effects. So if you apply a recovery coefficient curve, you improve the bias, but the reproducibility gets even worse. So in summary, does dosimetry help with metastatic castration-resistant prostate cancers? It doesn't really impact management. For hormone-sensitive and pre-taxane patients, I would argue that the dosimetry is a longitudinal monitoring tool, and it may eventually be used to limit cumulative therapy in these patients, and therefore it's worth collecting the data today, even if we don't know what the stopping point should be yet. So I would argue that that's really important. Looking ahead, if your goal is to cure cancer, I think that dosimetry has to be a part of the mission, because otherwise, how do you ensure that patients don't have toxicity later in their life if they're living a long time? So does it matter? It should, but often it does not, but there are several good reasons to believe that it will grow in importance for PSMA-targeted therapies over time. The end.