Joining us from the company, we have Derek Maetzold, CEO. Frank Stokes, CFO. Great to have you guys here. Thank you for coming.
Thanks for the invitation, Mason. Thanks for Stephens for putting on [audio distortion] .
Absolutely. Maybe just to start off high level here for those who aren't familiar with the story, could you just give us a quick overview of Castle, the focus of the company, maybe a quick overview of the portfolio?
Sure. I'll start with the focus and business and me, Frank, and will fill in, third quarter financials and guidance this year, et cetera. So we are a high-growth molecular diagnostics company that focuses on developing and offering proprietary high-value tests. So what's a high-value test in our mind? In our mind, a high-value test is one where the use of our test will have a clinician change a significant management difference as they're looking to treat their patients. So, for example, I take our TissueCypher test for use in patients with a disease called Barrett's esophagus disease. Barrett's esophagus is the only known precursor to esophageal adenocarcinoma. It is diagnosed in roughly 430,000 patients each year by an endoscopic biopsy procedure. We think that our addressable marketplace for those patients is around 415,000 patients per year. So where does our test come into play?
The standard of care for Barrett's esophagus disease is if patients have a certain likelihood of progressing to cancer, then you intervene. Then you intervene with either FDA-approved ablation tools like Medtronic's Barrx RFA device or a surgical procedure called endoscopic eradication therapy of some sort. And they do that because in those patients, the risk of the procedures versus the benefit of stopping progression to cancer clearly outweighs the risk. But patients who are at a lower risk of perceived progression to cancer, we basically try and survey. So what does that mean in Barrett's disease? It means you go back every one, two, three, or five years and get re-scoped and re-biopsied to see if your Barrett's lesion has actually gotten worse, visually speaking, or under pathology slides. And that's kind of been the standard of care.
Now, how the marketplace before the introduction of our test viewed identifying people at higher risk or lower risk was basically having pathology look at part of the pinch biopsy under microscope and say, "I can see high-grade dysplasia here or low-grade dysplasia." And based upon population numbers, those are kind of patients where we either would try and ablate every patient with high-grade dysplasia, maybe have a strong conversation with low-grade. But if you're below that, like non-dysplastic Barrett's esophagus disease, which is Barrett's disease but no dysplasia present on the slide that the pathologist reviews, you would do this watchful waiting, active surveillance every one, two, three, five years. Sounds good in principle, except because you're ablating all the higher-risk people on grading, it turns out, unfortunately, that most of the people who go on to get esophageal cancer are those that you don't ablate, the non-dysplastic patients.
Even though their population risk is lower, we think out of the 430,000 or so patients diagnosed each year, probably 415,000 fall in that low risk or that non-dysplastic grade of Barrett's disease. What our test does is it enables a gastroenterologist to say, "Hey, Barrett's disease, maybe it's non-dysplastic. I want to see if this is a bad [audio distortion].
Good morning, everyone.
Bad-acting non-dysplastic disease. And so we'll order a TissueCypher test. And we'll go back and say, "This is actually, even though it looks very benign under the microscope, our biological test, molecular test, says it's actually high risk." And if it's high risk, our high-risk chance of progressing is actually higher than the low-grade dysplasia you're already intervening with. And so it lets gastroenterologists identify a group of patients who they'd otherwise just say, "Come back in three or five years. Thank goodness it's non-dysplastic," and say, "Actually, biologically, this is a very aggressive Barrett's lesion.
Let's consider ablating so we can stop progression." On the other hand, there are also some patients in that small group of low-grade dysplasia patients or indefinite patients where you would say, "Based on other health factors or maybe even the patient's own concern about that procedure, do we want to run the Castle TissueCypher test and look at maybe not intervening today, but watching, maybe doing annual re-scope?" So very actionable, great impact on patient care. Sometimes we're eliminating unnecessary procedures. Sometimes we're actually adding procedures because it actually will hopefully save a patient's life. So that's kind of the principal of business overall. We operate in four major areas. Most of our revenue comes from dermatology with two tests, one for evaluating the risk of progression for people with invasive melanoma, one for squamous cell carcinoma, and then a smaller one for helping with differential [audio distortion] melanoma.
And then we have the gastroenterology franchise with TissueCypher today that has a sole product there, uveal melanoma, and then mental health conditions. Frank, do you want to kind of perform?
Yeah. Third quarter, we generated about $86 million in revenue. We increased our guidance to $320 million-$330 million for the year, and we had about $23 million adjusted EBITDA.
Got it. Thank you, guys, for that. So maybe we'll start with TissueCypher. It's been a great growth driver for you guys. And given the traction, I think we all are trying to figure out what that ramp looks like. So maybe if we could just run through a few factors that I think are relevant here, the first being the sales team. You launched your melanoma tests in 2013. By 2018, I think you were already around 10% penetrated. You just exited Q3. I think you're around 30% or at least close to it. Could you talk about how the sales team scaled during that time and then how that compares to TissueCypher?
Sure. So we were a very small, limited-capitalized [audio distortion] So I don't know if I would use the melanoma market penetration growth as a surrogate for TissueCypher. I think that was too slow a growth rate because we had single-digit representatives, 12 or 15 or so in 2015, 2016, 2017, so really quite measured. Here, on TissueCypher, we started out with, I think it was 16 or 15 territories in January 2022, got our feet wet the first nine months or so, increased to around 24 representatives, kind of watched that through 2023, increased to 40 people in the field in the second quarter of this year. Out of a few territories here and there over the course of the year. So we have a much more rapid expansion of educational sales representative TissueCypher than we ever did with melanoma early on.
I think if the 415,400,000 patient number doesn't have seasonality in it on an annual basis, then we probably exited third quarter maybe 5.5% penetration. So it's been more rapid than melanoma was years ago. I think the excellent question to us, what you're asking here is sort of maybe one is what's maybe the right scale to get to from a full penetration for your sales team? And then how does it impact, I guess, market penetration growth? So we think there's roughly 10,000 targetable gastroenterologists. And there is a small group of gastroenterologists who specialize in the ablation of these lesions I was talking about earlier. They certainly will be higher volume ordering customers because they're kind of the internal referral base. But every gastroenterologist does upper endoscopies. So we think 10,000 patients or doctors is roughly our audience.
Gastroenterologists compared to dermatology is a little more larger group, a little more PE ratio oriented. So one thing that we do know is that if you have a sales representative walk into a given practice in dermatology at a certain location, maybe there's five, 3+ some NPs and PAs operating out of a location. In gastroenterology, it's usually 20 to 30 to 40 to 50 in kind of general vicinity. So how efficient can our sales reps be if they can tackle more customers on a single sales call or single sales day, I would say? So based upon that, we're sitting probably in the mid-70s to 80 sales reps for dermatology. And that's roughly 10,000 plus a couple of thousand additional people. I think today, that is, I think we don't need to go that high for gastroenterology.
We think maybe the optimal ends up being 60, 65 people. So getting to 40 and letting these territories kind of mature in after training and then looking at what that size looks like next year is kind of what we're going to get to. So we're getting close there. So ramp-wise, I think that going from 0%- 5.5% in roughly two and a half years is a nice demonstration of the value of our product. And because our product has such a clear and easy utility for gastroenterologists, I expect this ramp to be quicker, certainly, than melanoma is because you can actually act or not act on an intervention that can stop the patient from dying of cancer.
Or you can actually take a patient who maybe you would normally see back every year and say, "Hey, based upon the results here, non-dysplastic disease, low-risk TissueCypher test, maybe we come back every three or five years." That's a great healthcare savings. It's a great reduction in sort of the hassle factor of that patient having to go and think about that every year coming back for upper GI scope. So it's a great activity there.
Got it. So yeah, as we kind of think about what to use as an analog here, I mean, from a sales team standpoint, you're starting off bigger. From a marketplace standpoint, larger patient population, but the same amount of docs, you get more leverage there. The practices themselves are less fragmented. How about from the standpoint of GIs versus derms and their maybe familiarity with using genomics in practice in patient care?
So I think gastroenterologists are ahead of the curve probably because of Cologuard being around for so long. Whereas in dermatology, I think back, we were the first sort of molecular genomic test that a dermatologist could order or should order on their patients, I believe. Is that correct? I think that's correct, yeah. So we had to start off by saying, "Hey, using biology on top of clinical and pathology factors actually improves decision-making." Whereas I think gastroenterologists were already seeing the use of genomics or the use of molecular testing to actually improve the relevancy clinically and pathologically.
Okay. So I mean, you're already on pace to do this, but it's moving much faster than melanoma in terms of initial penetration. A lot of factors to that. As you kind of think about getting to the point of 25%, 30%, I don't want to hold you to a timeline or anything like that, but are there any limiting factors, whether it's certain patient groups or maybe it's not getting utilized that make up a larger part of the market? I mean, is there any type of moderators that are out there that maybe we're not aware of? Like I said, there are a lot of factors here that suggest it can continue growing like it is.
I think most of it is basic block and tackle. Creating awareness, initial usage, converting that to trial and adoption. Now, there is one barrier in gastroenterology that's a little different than dermatology, which might impact uptake to a certain extent, although I'm not sure we'd see that based upon a larger patient base of patients. And that is that my belief, my understanding is that the vast majority of gastroenterologists, when they see patients in the office and they have their staff around them, nurses, medical assistants, et cetera, that staff is not the same staff that's actually in the ambulatory care center where the endoscopies are being done. And so the staff doesn't cross back and forth.
What we are finding is that in order to have a gastroenterologist say, "I've got Derek Maetzold under sedation for upper GI, if it comes out to be Barrett's esophagus, I'd like to have a TissueCypher test ordered on Derek's Barrett's lesion." He tells his surgical nurse that. That person doesn't necessarily tell the person over here in the office to order the test. And so there is a bit of a learning curve to figure out who's going to be the point person accountable for ensuring that the doctor's wishes are followed through on. Whereas in dermatology, you're taking a biopsy in the actual office where you're seeing patients, so you've got a much shorter continuity of care issue. That's the only structural thing I see here that I would say there's a throttling factor. But again, that's just education.
That's saying, "Are you really bought into having this test help your patients get better outcomes and improved outcomes?" And if so, let's figure out who's that person, where do they sit, and make sure that we train them [audio distortion] .
Got it. And going back to the commercial team, I think you said 40 in April. I think you've added some since then. How are you thinking about the timeline until the next expansion? And is this more of you add a couple here or there, or could it be another group, 10+ , something like that?
Our philosophy historically has been not wanting to disrupt more than 50% of our relationships in any kind of sales expansion. We could have gone from 24%- 55% or 60% in April this year, May of this year, but that's a lot of potential handoff disruption. It went to 40% and we're kind of adding on here. I would think that in the first half of next year, we'll be up to kind of the 60%, 65% is pretty comfortable.
Okay. And then just from the standpoint of how many docs are ordering it, the mix of growth, if we want to call it that, same store sales versus new ordering docs?
I don't think we disclose number of ordering doctors here in the third quarter, so I won't comment there. But in terms of the one is that we still have the majority of gastroenterologists who can convert to customers. So that's for certain new store sales. Same store sales, if you sort of measure that or define that as being a gastroenterologist started ordering our tests on a patient January 1st of the year, and they are ordering July 1st, so six months of sort of use, we are seeing, I think it's around 3.4, 3.5 orders per month of that individual clinician. So I don't know if that represents a peak or that still is an early adoption curvature, but that's kind of what we're seeing basically today.
Got it. How about the competitive landscape? Can you just update us?
Sure. So there is one small private company called Previse that launched a test maybe two years ago, a year and a half ago maybe. Maybe a year and a half ago. Very small private undercapitalized company. We haven't seen much from a marketplace standpoint, but that's the only competitor we can see out there today.
Okay. How do we think about the future product portfolio of GI? Is that a focus at all? Is it something that you're considering just given the scale that you already have there from a commercial team?
Yeah. We would definitely like to have more products in the GI bag at some point. When you look at the benefit we've seen to melanoma and SCC, having both of those together, it's just an easy step to leverage that sales force and leverage that investment. What I can't tell you is when the right time is and when the right product is, but the growth is so substantial right now, it would be tough to distract the team with something new. But at some point, we'll definitely take advantage [audio distortion] .
Okay. Maybe moving to the inflammatory skin disease test. Can you just remind us the market opportunity, patient population, timeline for data?
So our large prospective protocol that we initiated a couple of years ago essentially says we will enroll patients who are either moving from topicals only to systemic, which most of the time is injectable biologics, but there are a couple of oral therapies out there. So systemic therapy with moderate to severe psoriasis or atopic dermatitis. So the whole spectrum of sort of itchy scaly skin. And so we have patients in this large prospective study who are both atopic dermatitis patients and also psoriasis patients. Now, a helper expectation here as we kind of see our discovery and development data maturing is to say, "Can we take this to a clinician?" And let's just take atopic dermatitis as an example. Make it simple. Say, "Your normal routine is to probably start those patients on Dupixent, Regeneron Dupixent as the market leader in the first systemic therapy use typically.
What if we told you that patients likely to have a poor response, or at least not a non-satisfactory response to that therapy based upon whatever their atopic dermatitis biological mechanisms are?" And if we told you that and said, "But they might have a better response to JAK inhibitors or maybe some of the future atopic dermatitis drugs coming down the pathway, would you be interested in helping that patient rather than having them start with what you always start with, and then in three or six months having them switch, maybe even switch again or add back in topicals to maybe go right to the therapy that has a better chance of getting a better response for your patient?"
That marker research we've done has come back very, very satisfactory in that clinicians say, "Well, it makes sense if I can go ahead and get a patient to the right therapy sooner rather than later. That's fantastic." I also, by the way, as a dermatologist, can think about that from a sort of staff use. So right now, most practices who do medical dermatology are going to have at least one individual who we'd call a biologic coordinator. So basically, you want to start me in Dupixent. We've got to have somebody in that practice assigned to get prior authorizations working through to get maybe initial starting samples out of Regeneron. There's a body that's dedicated and paid for, by the way, out of the dermatologist's practice. He doesn't get anybody to pay him back on that or her back on that.
That's out of the profit margin to say, "Again, if you have to have that biologic coordinator see the same patient three times in a given year because they're just not on the right drug, what if that was reduced to one or 1.2 times per year per patient?" That's a tremendous cost savings to you and efficiency of your practice going forward. So there's many areas of health and opportunity. So that's sort of what we hope to get to. Now, we had committed to by the end of this year to go ahead and release some of these top-line perspective on kind of where we are, what we're seeing, what we're not seeing. So that still is coming on board probably later in December, is my guess right now.
Okay. And then maybe to get a little bit more specific there, from a selection standpoint, is it specific drugs, class of drugs, a specific drug in a class?
So I think there are several profiles that we have on the wall. Can we get to a test, or maybe it's a series of tests? Maybe it's one test with a series of signatures. That could be drug-specific. Or can we get to a test or a series of signatures within a specific test that actually can find people who are maybe grouping by a therapy class response? So that needs to be worked through. But I think there are several ways to have an attractive test that would meet the clinical need of getting a patient on the right therapy or therapy class earlier rather than later.
Okay. And from a cost point standpoint, is it the same as melanoma? Are there differences?
Our knowledge to date is that, let's take our squamous cell carcinoma as a surrogate. We have right now around three-quarters of clinicians who are using our test for squamous cell carcinoma, our clinicians who order our melanoma tests. So good crossover three to four times. And the reason why it's not closer to 90% maybe is that there are some Mohs surgeons. All Mohs surgeons usually see high-risk squamous cell as part of the practice. But only a proportion of Mohs surgeons see invasive melanoma. So while they would be an ordering customer for squamous cell, they would not naturally see a patient who's appropriate for use of melanoma. In the case of psoriasis, atopic dermatitis, inflammatory skin disease management, our belief is that the majority of our skin cancer doctors are just medical dermatologists. So they will be seeing other skin diseases like inflammatory skin diseases as well.
So I think there'll be high crossover. There are though some practices where you have a large group of dermatologists. We might have a specialist or an expert who deals mainly in atopic dermatitis, psoriasis. Now, those are probably the more tricky cases. So that may or may not be a customer for us necessarily. But I think we'll see substantial overlap.
Okay. And then I think you've targeted end of 2025 for a potential launch. At that point, are you submitting to MolDX at that point? Is there an existing LCD out there that this can be filed under? And what type of options do you have or paths can you take with commercial payers?
Those are excellent questions there. I'll try to unpack those carefully here. So one is that I think that's, assuming that we are successful in developing validation, the most important aspect to launch would be who's going to pay for this test. One, there is no existing LCD from MolDX that I'm aware of that would kind of fit into this would go into the LCD too. So that's probably not there yet. Two is that the population is not necessarily a heavier Medicare population. It's probably only 25%-30%. And that depends on our final profile, of course. So it's not going to be a 50%-70% Medicare population. So that's less important, still less important.
I think our question about launch, should we be successful in our developing validation studies, would be to say, "Do we want to make this test available to patients under UHC insurance plans if UHC says go take a hike?" I'm not sure we want to do that. Do we think there's an opportunity to have a pharmaceutical firm who's second or third or fourth line say, "Well, gee, if you can find people who are non-responders to," use my earlier example, "Dupixent," and that moves me up in market share, do you want to help us cover our cost of part of the equation? If the answer is no, that's fine. We won't perhaps go ahead and make that available.
So I think reimbursement and timeline and planning is the most important element of what I would call the launch of this test here to see that by the time we would build substantial volume and revenue, what can we do sort of to make this a material driver for us in 2028, 2029, 2030? I think that we should not count on any reasonable material revenue in 2025 or 2026 or even 2027, perhaps. Probably a little too early in terms of good reimbursement setting.
Okay. That makes sense. Maybe moving to the melanoma test. A competitor read out a study relatively recently. Maybe if you could, you could just talk about your interpretation of those results, how it compares to maybe DecisionDx-Melanoma and how these tests are used.
Sure. So we have one US competitor, which is a test called Merlin, I guess you would say. It was developed by a Dutch company called Skyline, a small private company. They invested in a prospective study a couple of years ago, and the primary data presentation, I guess, was in October, I guess, about October, about a month ago now. And what they presented was the outcome of their study looking at the accuracy of their test in predicting patients who would have low risk or a low likelihood of having a positive sentinel lymph node. So what's that mean? What's the threshold? So in melanoma, there's a treatment pathway where you go into your dermatologist, he or she sees a mole or something that looks like could be melanoma, biopsies it, goes to pathology. Pathology says, "Yes, it's melanoma.
Yes, it's invasive, meaning it goes through the epidermal layer. They measure the thickness that it goes in, and based upon thickness and also the presence or absence of ulceration or other kinds of adverse features like mitotic figures, et cetera, you would be eligible for being considered for a sentinel lymph node biopsy procedure, surgical procedure, where you would go under general anesthesia, typically get a radioactive dye and a colored dye, inject into the melanoma site, tracer up with a Geiger counter where those molecules are located in the sentinel lymph nodes. It goes here, either here, or maybe the armpit, and then go in there and surgically remove those lymph nodes and look for just a melanoma cell.
Guideline-wise, for I don't know, 20, 30 years, the various guidelines and slides have said if you have a chance of having a positive sentinel lymph node, so even one melanoma cell in your lymph nodes that's higher than 5%, then you should consider or undergo that procedure. If it's less than 5%, we'd recommend that you just avoid that procedure because the risk of complications outweighs the benefit of finding it. That's kind of the 5% threshold. So that was the a priori goal of this study, and they hit 7.1% in their low-risk group. So they basically failed to achieve the primary endpoint. That's disappointing for them. It's good that that data is out there because it can hopefully inform about a use of that test for ruling out a procedure when the risk is actually higher than the rule-out threshold of 5%.
Got it. Thanks for that. So maybe just touching on the SCC test. I know we're all still waiting on an update for Novitas, and there's limited visibility there. Palmetto issued a non-coverage decision. Some key studies may have been excluded there. As we just kind of think about the future of this test, the reconsideration pathway, I mean, how would you rank your confidence over time of this eventually if Novitas were to make a non-coverage decision to obtain coverage through Palmetto at some point?
I think, so the Palmetto LCD that was finalized, I guess, in July of this year, missed or ignored, I guess, or didn't evaluate, we would say, maybe the software didn't say that, six key publications that came out between September of 2023 and May of 2024. And now we've had another one come out in August of 2024. And the key studies are. I've mentioned those things as one is there was an article published in January 2024 by Wysong et al., which addressed every one of the criticisms about why we shouldn't cover this test in the draft LCD. That obviously wasn't reviewed because it was published after the comment period closed in the summer of 2023.
That's an important article and sample of addressing what they felt as a lack of evidence meeting the criteria of how do you add value on top of what's already available as part of staging. The second element, which is I think probably more important, is that although we submitted data to MolDX in a written submission and presented data during the open comment meeting about the use of our test to help guide decisions around adjuvant radiation therapy, by regulations, they should not really review that as part of the evidence for the LCD until they see it in a peer-reviewed publication. So they have the data. In the LCD, they made a note of saying, "Well, we have some promising data." So that was good. That means they read it, which is positive. But they couldn't necessarily review it because it wasn't published yet.
In May of this year, we published our first study, which was a backbone of that data that we submitted the summer before by Dr. Arron et al. That study was over 900 patients that was focused on evaluating the effectiveness of adjuvant radiation therapy in patients with cutaneous squamous cell carcinoma and the use of our test to correctly, safely identify patients who would be eligible for ART. But not everybody benefits. What that study found was that if you had a low-risk Class 1A test result, which is our low-risk result for our test, not only did you have a low chance of metastasizing, but we could see no benefit from use of adjuvant radiation therapy compared with propensity-matched patients who received adjuvant radiation therapy. No benefit clinically. That happened to be actually the majority of patients in this 900+ patient study.
On the other hand, if you happen to have a higher-risk class 2B test result, we found that those patients had a greater than 50% reduction in the likelihood of metastasis at five years compared to not having ART. So clearly, a population where you'd say, "Wow, the benefit to risk ratio of adjuvant radiation therapy says these people should get this, encourage them no matter what." The class 1 patients, there's no benefit here. Now, a patient may still want to go forward. A doctor may still want to push it. That's okay. But you have data showing no benefit here. Now, why is it important? The ARIN study was the single largest study ever published to date in squamous cell carcinoma studying the selection process of ART.
We had a second study published in August of this year from a separate cohort, which was developed by the Brigham and Women's Hospital in Boston and Cleveland Clinic in Ohio, and there, those two centers essentially performed the exact same study. They took a large cohort of patients, all of which would have been ART eligible under one of the guideline inclusion criteria. We ran our test results, and what we saw was the exact same thing. People with a Class 1 test result, no discernible benefit of ART. People with a Class 2B test result, about a 50%+ or higher impact.
Those are two great clinical utility studies, which I believe, as we always interact with our Medicare contractors as best we can, this is data which would say, "Wow, you have patients and clinicians who are getting radiation therapy recommendations and going through it or not going through it based upon clinical and pathological factors that aren't really very good. They're squishy, and they aren't very accurate." If you overlay our test here, you can not only find people who definitely should go forward, but you can rule out most patients. Now, the net cost savings of just looking at radiation therapy, adding in the cost of our test at the Medicare rate, by the way, and if that was employed by patients who received radiation in the 2021 or 2022 time period over the course of one year, that would result in a $900 million+ savings to Medicare.
That's a tremendous impact on healthcare reduction and complication reduction and also getting people who will benefit in a direct pathway.
Okay. And.
So I would hope with all of that, that would be data and work with our contractors to say, "Hey, you didn't see this. This is so important clinically to be able to take patients who are in these broad guidelines of going towards adjuvant radiation therapy and saying, 'Let's take a break, run a molecular test, and see if you'll actually get a benefit or not.' If you don't, that's a whole different conversation plus the benefit and cost savings.
Got it, and we'll ask one more here. Right now, that test has run out of Pittsburgh. If you needed to or wanted to move it to the Phoenix lab, what would that entail?
We have replicated the ability from just a business interruption standpoint. Most of our tests in both Phoenix and Pittsburgh can be run back and forth. It's relatively, so it'd be the technologists, for instance, running our DecisionDx-Melanoma test, MyPath Melanoma test out of Phoenix and pick up pretty quickly the next day and run squamous cell and vice versa, by the way. That was mainly for business continuity perspective.
Got it. Any questions from the audience before we wrap this up?
Can you speak to the (how do I phrase it?) IDgenetix and lack of interest in paying you for that, but it's not really material. Can you flesh all that out for us?
Can I repeat that for you?
Yeah.
So the question was, can we comment on IDgenetix and maybe the UnitedHealthcare reimbursement meeting in general?
Yeah.
Okay. And.
Materiality.
And materiality.
Yeah. So the mechanics there, Adam Moore, United did have a positive policy, but we were not a network provider for them, so we weren't under contract. So we were not getting paid on a substantial amount of our United claims for IDgenetix. So when that went to a negative policy, it wasn't a material hit to our revenue. Now, it is a, it's material to the landscape in terms of how payers view these tests. And it's a statement that says that the largest payer, commercial payer, isn't quite concerned about whether their members are cared for adequately or not. And when you look at the reasons they justified for the non-coverage policy, I bet they've never had a positive coverage policy on the same amount of data. It was much less data than they've ever required for positive coverage.
I can't prove that empirically, but when you read it, you just sort of, it's hard to see it there, so we have, as other pharmacogenomic companies, we've struggled with that commercial stance, and I think the reasons are probably that if you have a depressed patient and they're on a med and they're not responding, the incremental costs are not borne by the payer. United doesn't pay for them. It's the family. It's the employer. Guy's not going to work five days a month. He's having trouble at home, not leaving the house. These costs don't hit the MLR at United, and so I think that that's probably a big driver here is that paying for these tests doesn't lower their costs any, so there's an ambivalence about the result, but that commercial payer approach has been stubbornly resistant for some time now.
And obviously, with Medicare patients, there's a lot more polypharmacy. And so there's a lot more variability. And so certainly, Medicare early on saw the benefit and has continued coverage. But disappointing to see a commercial payer really use a thin argument to turn it down. But there were a couple of estimates by analysts that were higher than what we actually would expect to be the impact for next year.
So the impact on Castle is not material from a revenue standpoint. The impact on patients, unfortunately, is material.
Anyone else? All right. We can stop there. Thank you, guys.