Castle Biosciences, Inc. (CSTL) Earnings Call Transcript & Summary

Good afternoon, and welcome to Castle Biosciences DecisionDx DiffDx-Melanoma investor webcast and conference call. As a reminder, today's call is being recorded. [Operator Instructions] I would like to turn the call over to Derek Maetzold, President and CEO. Please go ahead.

Thank you, operator. Good afternoon, everyone. Welcome to Castle Biosciences DecisionDx DiffDx-Melanoma investor webcast and conference call. As most presentations, the following discussion contains forward-looking statements, and our actual results may differ materially from those discussed here. Additional information concerning factors that could cause such a difference can be found in the disclaimer slide included in the presentation and the company's quarterly report on Form 10-Q for the quarter ended June 30, 2020, and the annual report on Form 10-K for the year ended December 31, 2019. We have a fairly packed agenda this afternoon. I will open up and set the table of where we are today, turn the microphone over to one of our long-time collaborators, Dr. Sarah Estrada. We will then hear from Dr. Matt Goldberg, and I will close out with a brief summary, and we can take questions at that point in time. We have been blessed since the early days of Castle at being able to attract and maintain a leadership team that's been able to manage through a time of limited resources to modest scaling in the latter part of last decade. Through to the last 15 months or so as a public company, we've experienced significant growth. I look to the many people on this management slide, who are with us since the early days, to our latest addition, Dr. Matt Goldberg, who joined us in August 2020. While Matt is new to the Castle team, he brings with him what we expect, experienced yet open to accelerating knowledge in areas where he has less experience and a team-oriented leadership style. We look forward to adding additional names to this list going forward. It is a sincere pleasure to introduce Dr. Sarah Estrada. I'll let you read this slide as I add just a couple of comments. Dr. Estrada is an experienced fellowship trained dermatopathologist who has been collaborating with Castle for the past 6 years or so, and has reviewed for us approximately 30,000 cases of melanoma during that time period, not to mention the work she's done with us in collaborating on the development and validation of our DecisionDx-SCC tests and now on our DecisionDx DiffDx-Melanoma test, the subject of today's call. Turning to Slide #7. I shared this slide during the DecisionDx-SCC launch call, which was just a few short weeks ago, and it's worth covering a couple of highlights. We are innovators in the dermatologic cancer market. We develop innovative products that solve significant unmet clinical needs. We support validity of these products by not only developing an adequate baseline of evidence but have continued to drive additional evidence to support appropriate clinician adoption as well as reimbursement. Finally, because we started our business in the rare cancer market prior to pivoting to the dermatological cancer market, we built our business from this perspective of serving smaller cancer markets. This means that the decisions we made led to thoughtful investment decisions and have placed us in a unique part of the molecular diagnostics world, being able to get to cash flow profitability and growth on a revenue base of $51 million and change in 2019. Now we are accelerating investments to drive both penetration of our proprietary test products as well as support new pipeline initiatives, and the team and I expect to deploy some of the capital that we have on hand to make these investments. In other words, we have very strong financials. Turning to Slide #8. On our second quarter 2020 financial report. As it relates to our financials, let's look at our performance through the second quarter of 2020, which is really through the first quarter of the COVID pandemic. As a reminder, we achieved $12.7 million in revenue in 2Q 2020, which represented growth over 2Q 2019 and $30.1 million for the first 6 months versus $19.5 million for the first 6 months of 2019. We also saw a continued lift in gross margin, which came in at 83%, and we ended 2Q 2020 in a very strong cash position, with cash and cash equivalents balance of nearly $180 million and adjusted operating cash flow of $3.3 million. We entered 2020 with 1 dermatologic product in the marketplace, DecisionDx-Melanoma. We estimate the U.S.-only TAM to be around $540 million based upon a fully seasoned average reimbursement rate of $4,200 per test, which is near the average target reimbursement rate for solid tumor test like ours. We expect to exit 2020 in just a few short weeks with 3 dermatologic products on the marketplace, DecisionDx-Melanoma, DecisionDx-SCC for patients with squamous cell carcinoma of the skin and 1 or more high-risk features. We launched this test in late August 2020 and believe that the U.S.-only patient population is 200,000 patients, which translates to a calculated U.S. TAM of around $820 million using the same fully seasoned average reimbursement rate. The test that we are here for today, the DecisionDx DiffDx-Melanoma test is for use in patients with a suspicious or difficult-to-diagnose pigmented lesion. We estimate that there are approximately 300,000 of these suspicious pigmented lesion biopsies. For this test, we were launching behind Myriad, which has already established an initial price of around $2,000. Using this lower average sales price, we calculated a U.S. TAM of $600 million. So we entered 2020 with a U.S. TAM for our marketed dermatologic product of around $540 million, and we will exit 2020 with a U.S. TAM from our marketed dermatologic products, that's plural now, of around $2 billion. That is really fantastic progress. Slide #10. This slide is really all about alignment and execution. As I mentioned on the previous slide, we entered 2020 with a focus on 2 primary customer groups for our DecisionDx-Melanoma test. Medical dermatologists, and surgeons, some general, some surgical oncologists, some plastic surgeons who perform a Sentinel biopsy surgical procedure and wide local excisions on patients with early-stage Melanoma. As you may recall from earlier calls, dermatologists are responsible for about 80% of our DecisionDx-Melanoma test orders. We do call on 2 additional groups of physicians, not so much for orders, but for awareness of the value of our DecisionDx-Melanoma test. These 2 groups are dermatopathologists so that they and their teams can understand why their referring dermatologists, our joint customers, are ordering our DecisionDx-Melanoma test. The other group of physicians are medical oncologists. We locate those medical oncologists who are interested in using the value of biological information regarding early stage melanoma and who will manage those patients accordingly with a high-risk DecisionDx-Melanoma score. Our target audience for our DecisionDx-SCC test that we launched in late August of this year is the same medical dermatologist, including the subsegment of most surgeons, who we currently call on and work with today. And today, in the last year, over 4,200 of them have reviewed, considered and acted upon the value of our initial dermatology test, DecisionDx-Melanoma, what that value can bring to their patients in the management of their cutaneous melanoma disease. We expect that many of these same clinicians will favorably review our DecisionDx-SCC test for their patients who were diagnosed with 1 or more high-risk factors. The call point for our DiffDx-Melanoma test also overlaps with our current call base. Specifically, we received tissue specimens for clinical testing from over 1,900 dermatopathologists in the last 12 months. Our initial launch ordering targets are these same dermatopathologists, and we will be targeting them based upon the relationship that we've developed with them over the last several years. Once we get past the initial introduction phase, we do expect to expand our base of DiffDx-Melanoma ordering docs to include other dermatopathologists and also dermatologists who are seeking clarity from an uncertain pathology report. In other words, the targets for our 2 launch products, DecisionDx-SCC and DiffDx-Melanoma are our current customers. I'll remind you that the opportunity to launch these 2 high-value proprietary tests right into the target audiences for our DecisionDx-Melanoma test is extremely unique, and we look forward to updating you on our launch performance on both these tests in the coming months and quarters. I'll now turn the presentation over to Dr. Estrada. Sarah?

Okay. Thank you so much for having me here today. It's an honor. All right. So let's start with the natural flow of the series of events that takes place when a patient presents to the dermatology office with a suspicious pigmented lesion. The dermatology provider uses the naked eye or perhaps a dermatoscope, which is a sort of handheld magnifying glass to assess the patient's lesion. Around 2 million times per year, a pigmented lesion looks suspicious enough to take a biopsy. The biopsy is dropped into a bottle with formalin and sent to the pathology lab. This is where the dermatopathologist steps in. If the biopsy looks definitely malignant, the diagnosis is Melanoma. If it's definitely benign, the diagnosis is also quite straightforward. Most lesions fall into one of these relatively simple categories. So where does the problem lie? It lies in the approximately 300,000 melanocytic lesions each year that are not so straightforward, or are deemed difficult to diagnose. Now let's take a look at the Melanoma diagnosis process again, this time focusing on the divergence of treatment options for Melanoma versus a benign lesion. First, if a patient receives a definitive benign diagnosis, thankfully, the issue is resolved, they can move forward with a simple routine follow-up, and this is great news. However, with the diagnosis of Melanoma in situ, which is confined to the very top layer of the skin, or invasive melanoma, which means it's grown deeper into the skin, both will require a wide local excision as the primary treatment. And if it is indeed invasive melanoma, there are additional treatments that have to be considered, including Sentinel lymph node biopsy, imaging and increased follow-up. And this is where Castle's first skin cancer test, DecisionDx-Melanoma, fits in by assessing prognostic features of these invasive Melanomas. So the big takeaway message for this slide is that a definitive diagnosis, whether benign or malignant, informs primary treatment and follow-up decisions such as imaging and observation. Now remember the last slide, where I mentioned that there's a subset of melanocytic lesions, which are difficult to diagnose, the ones that don't clearly fall into the benign or malignant category. Unfortunately, about 15% of melanocytic lesions will fall into this difficult-to-diagnose category, and we just cannot make a definitive diagnosis. This, in turn, makes it difficult for the treating clinician to know how to manage their patient. In general, they are forced to err on the side of caution and treat the lesion as if it's malignant. So now we're going to look at the same diagnostic and treatment pathway again, but focusing on that difficult to diagnose or uncertain malignant potential category. Recall, this is around 300,000 cases every year across the United States. Take a look at the blue box in the middle of the screen, the one that says uncertain malignant potential. It's clear to see, and this is really important, that a step in the wrong direction after this point can lead to unnecessary and life-altering events for our patients. If we over-diagnose the case and call it Melanoma when it's actually benign, a patient may be subject to unnecessary excision, imaging, Sentinel lymph node biopsy and protracted clinical follow-up. If we under-diagnose, a patient could end up with a later stage melanoma that was not appropriately addressed, and this could have devastating consequences. Here is a look at the routine dermatopathologist workflow for a difficult-to-diagnose melanocytic lesion, so you can get a bird's eye view of how we ultimately reach the diagnosis of benign, malignant or uncertain malignant potential. When melanocytic neoplasms are difficult to diagnose, we typically order some additional levels of the lesion, along with between 1 and 3 routine immunohistochemical stains, and we usually share the case at this point with a colleague to get a second set of eyes on it and obtain a trusted second opinion. However, even after incorporating these common diagnostic tools, there are many cases where significant diagnostic uncertainty persists, and we classify these lesions as uncertain malignant potential. It's at this stage of review where DecisionDx DiffDx-Melanoma is designed for use. After we have completed our in-house review with IHC levels and review with local colleagues. We know that there's clearly an unmet clinical need for difficult-to-diagnose melanocytic lesions. We have 300,000 cases each year that cannot be confidently classified as benign or malignant. As dermatopathologists and as dermatologists, we all agree that this should not be an acceptable number, but we remain constrained by the diagnostic tools which are currently available to us. If a melanocytic lesion remains difficult to diagnose after routine measures have been taken in our own labs, a commonly utilized option is to send the case to an expert who specializes in, and has extensive experience, with these types of lesions. Most of the time, they will have the final say on whether a lesion is ultimately called benign or malignant. The issue with this, however, is that even amongst the world's leading expert in melanocytic pathology, discordant diagnoses are rendered in 25% to 43% of these lesions. These numbers simply aren't good enough. We need a tool which can bring diagnostic clarity, reduce the number of our patients that fall in the category of ambiguous diagnoses and place our patients on appropriate treatment plans. The good news is that DiffDx-Melanoma provides the highly accurate tool that we've been waiting for. DecisionDx DiffDx-Melanoma provides a highly accurate and objective evaluation of difficult-to-diagnose melanocytic lesions, leading to a more confident diagnosis and clearer treatment plan for our patients. Recall, under-treatment or delayed identification in Melanoma can lead to tumor spread and increased risk of Melanoma-specific mortality. On the other hand, overtreatment can impact patient quality of life, potentially resulting in adverse events and increased morbidity. Gene expression profiling is the most recent diagnostic method to assess Melanoma. The goal of GEP testing and assessing melanocytic lesions is to provide additional objective information to dermatopathologists to aid in the rendering of a benign or malignant diagnosis. When interpreted in the context of other clinical laboratory and histopathologic information, DiffDx-Melanoma is designed to add diagnostic clarity and confidence for dermatopathologists, while helping dermatologists deliver more informed patient management plans. This is a tool that I am personally looking forward to introducing into my own daily workflow. I work in a practice that sees a lot of melanocytic lesions, and even with 14 years of dermatopathology experience under my belt, not a day goes by that I'm not faced with a melanocytic lesion that really makes me have to stop and think, is this thing benign or malignant? So I get the levels, I get the IHC stains, and sometimes that's not enough to make a decision. Then I can show it to my colleague and get his opinion. And sometimes then, that's not enough. So I sit there trying to decide which side of the line to fall on, benign or malignant. And every time, I have to consider the ramifications of going in the wrong direction. At that point, I might give it an ambiguous diagnosis, knowing the patient is going to be treated as if it's malignant. But if I cannot, in good conscience, make a confident diagnosis, I send it to an expert, knowing that even his or her opinion isn't necessarily the right answer, but it's the best option I have at this point. So I render the diagnosis, issue the report and, frankly, I hope for the best. Sometimes, only time will tell. This sort of decision-making is not particularly satisfying. And to be blunt, it's very stressful. Patients are putting their confidence in us, and sometimes we can't know for sure that we're making the right decision. I work with dermatology residents in training at our program, and I share such conundrums with them on a regular basis. And I always tell them, I've told them for years, the answer for these difficult cases is in the DNA. Until we can look at the genetic profiles of these difficult lesions, we can be diligent and we can do our best, but it's still somewhat of a guessing game. I'm now extremely optimistic that with DecisionDx DiffDx-Melanoma, we can put our trust into GEP testing and infuse confidence into the diagnosis of difficult melanocytic lesions and ultimately set our patients on the right path. I truly think this is going to be a real game changer, and I'm really excited to have the opportunity to integrate this test into my workflow. Thank you.

Thank you, Dr. Estrada. Dr. Goldberg?

Yes. Thank you, Dr. Estrada, for describing the unmet clinical need. And really, here at Castle, when we set out to develop the DecisionDx DiffDx-Melanoma test, our goals were to develop a test with a higher level of accuracy, improved rate of actionable test results, a high rate of technical success, and with the ability to test more lesion subtypes within this group of difficult-to-diagnose melanocytic lesions that Dr. Estrada has really just highlighted that are commonly encountered in dermatopathology practice. And here on Slide 19, to achieve these goals, we leveraged our artificial intelligence-based development process here at Castle, where we apply deep learning and utilize neural networks to select genes and refine the algorithm that form the basis of our gene expression profile test. Now the development and validation efforts were focused on maintaining or improving diagnostic accuracy by lowering the rate of unactionable results that fall in an indeterminant zone, intermediate zone, lowering the technical failure rate of the assay and including more subtypes of lesions with uncertain malignant potential in the training and development cohorts. As a result of these efforts, Castle has achieved the goals that were set out initially. And based on the development and validation study, DecisionDx DiffDx-Melanoma is an objective test with a high level of accuracy, a low rate of intermediate risk results, a high technical success rate and has included a broad range of lesions with uncertain malignant potential that present diagnostic challenges to clinicians currently. Moving on to the next Slide 20. We've laid out here the discovery and development process at Castle. As with our other commercially available products for cutaneous skin cancer, cutaneous squamous cell carcinoma and cutaneous melanoma, the starting point for these research efforts really begins with archival tissue samples. And here, formalin-fixed paraffin-embedded tissue from benign melanocytic lesions and malignant melanomas with concordant diagnoses. In the initial training cohort, there were 200 benign melanocytic nevi and 216 malignant melanomas. And from these tissue samples, we performed RT-PCR to evaluate the gene expression levels. Next, a panel of 76 genes was evaluated in the discovery step. And using the deep learning and neural network modeling approaches, we identified 32 discriminant genes and 3 control genes that comprise the locked 35 gene expression profile test. This 35 gene expression profile test algorithm was then validated on an independent cohort of 503 additional tissue samples with benign and malignant concordant diagnoses. Turning to Slide 21. This diagram here outlines the workflow of the DecisionDx DiffDx-Melanoma assay that was used in the validation process and that will be applied for patient samples when the test becomes commercially available. It's important to highlight here that Castle has other commercially available gene expression profile products for cutaneous squamous cell carcinoma and cutaneous melanoma. And really, the process for tissue handling will be identical to the established methods for our other products in skin cancer. The starting point here for the DiffDx-Melanoma test is a diagnostically challenging melanocytic lesion that Dr. Estrada was discussing. And the clinician will place the order for the assay and submit formalin-fixed paraffin-embedded tissue to Castle's laboratory. Once the tissue is received, RNA will be isolated from that sample, RT-PCR will be performed and the proprietary gene expression profile algorithm will be applied. With the result being 1 of 3 discrete results, benign or suggestive of a benign neoplasm; intermediate risk, where a malignancy cannot be excluded; or malignant gene expression profile suggestive of a malignant Melanoma. The next slide, 22, highlights that these development efforts are based on peer-reviewed publications that support both the clinical validation of the assay and the intended clinical use of the assay to support clinician and payer adoption. Now specifically, the discovery and development and validation research is described in the manuscript by Dr. Estrada and her coauthors in the forthcoming article that is currently impressed at the Journal of Cutaneous Medicine SKIN. A clinical utility study on the right of this slide is also forthcoming in an article by Dr. Farberg at all -- that is also impressed at the Journal of Cutaneous Medicine SKIN. Focusing now on the forthcoming development and validation manuscript by Dr. Estrada et al, now on Slide #23, is the consort diagram or a flow diagram for the DecisionDx DiffDx-Melanoma training and validation cohort. So we started with 1,300 benign melanocytic lesions and 1,500 malignant Melanoma samples, and we selected a random study cohort of 951 patients with concordant benign and malignant diagnoses. From this study cohort of patients, we again selected a subset into a training cohort of 416 patients where the 35 gene expression profile algorithm was developed and locked, and then the validation cohort was comprised of an independent set of 503 samples. Importantly, across this entire study cohort, there was a low level of technical failure, with only 3.4% of patients removed due to multi-gene failure. And this really represents a high technical success rate in Castle's development process for DecisionDx DiffDx-Melanoma. On the next slide, #24, are the accuracy metrics that were calculated from the validation cohort of 503 patients based on the ability of the gene expression profile test to accurately assign a gene expression profile result of benign or malignant. Here on the left-hand column are the statistics for the all ages group with a sensitivity of 99.1% and a specificity of 94.3%, which represents a high level of diagnostic accuracy in this validation cohort. The intermediate risk result was importantly only 3.6% of patients, which translates to the DiffDx-Melanoma test providing a clinically actionable result of either benign or malignant in greater than 96% of cases that received a result in the validation study. On the right-hand side of this chart are the statistics for a subset of patients included in the validation cohort who are over age 65. Here, the accuracy metrics are again excellent with a sensitivity of 99.2% and a specificity of 100% in this smaller subset of patients within the validation study. The intermediate risk result was again low with only 3.4% receiving an intermediate risk result in this over 65 population. And again, the technical success rate was excellent across the all ages and the over 65 population, with 96% samples able to receive a result with the gene expression profile test in the validation study. The development of validation data shown here in this chart reveal excellent accuracy metrics and high rates of technical success, demonstrating that the DecisionDx DiffDx-Melanoma is designed and validated to improve diagnostic clarity for the benefit of patient care in those challenging, difficult-to-diagnose melanocytic lesions. So in summary, on Slide 25, after the completion of the discovery and validation process reviewed here, we at Castle have achieved our development goals. Now what this means for patients who receive testing from DecisionDx DiffDx-Melanoma is that based on the validation data, the test provided a definitive result of a benign or malignant gene expression profile in over 96% of lesions that were submitted for testing. Importantly, this result is both objective and highly accurate in order to provide greater diagnostic certainty and reduce ambiguity for difficult-to-diagnose melanocytic lesions. Now our development effort included more of the difficult-to-diagnose melanocytic lesion subtypes that are so important to clinicians. And the test had a high rate of technical success with 96% of lesions able to be tested in the study cohort. And we're also able to anticipate a similar 5 to 7-day turnaround time upon receipt of tissue samples for testing. That's similar to the experience that Castle has with our other commercially available tests, DecisionDx-Melanoma and DecisionDx-SCC through processes developed at our experience laboratory. And importantly, for those lesions that subsequently qualify for testing with DecisionDx-Melanoma with a diagnosis of invasive malignant melanoma, clinicians can then order the DecisionDx-Melanoma test through a tissue sparing approach that will be able to use the same tissue block as was used for the DecisionDx DiffDx-Melanoma assay, and this will be a significant benefit to patients where there's clinical utility for both tests, both DecisionDx DiffDx-Melanoma and DecisionDx-Melanoma. And finally, it's important to emphasize again that the gene expression profile assay be interpreted in the context of the clinical, laboratory and histopathologic information. And when used in this context, DecisionDx DiffDx-Melanoma is designed to add diagnostic clarity and confidence for dermatopathologists, while helping dermatologists deliver more informed patient management plans. And with that, I'll turn it back to Derek.

Thank you, Dr. Goldberg and Dr. Estrada. Turning to Slide #27. I'll pick up where Dr. Goldberg left off just a little bit here. As a quick reminder, with the launch of this test, DiffDx-Melanoma, we are now able to serve our patients upstream of where our DecisionDx-Melanoma test is used. And importantly, when a dermatopathologist chooses to use our DiffDx-Melanoma test to provide a more certain diagnosis, and if that more certain diagnosis is suggestive or is now malignant Melanoma, we are uniquely positioned to rapidly provide the treating dermatologist with information from our other Melanoma test, DecisionDx-Melanoma. When we started down the path as a company, we've always strived to maintain a focus on the patient. If our tests solve treatment decisions for a patient, then we should have our tests ordered, we should get paid properly, and these elements should result in a high-growth revenue company. It is really exciting to be able to assist a patient who has a suspicious pigmented lesion, not only with their diagnosis, but also with their prognosis and treatment plan. Turning to the milestone slide on Slide 28. The Castle team has executed extremely well in the last couple of years, from the lead up to our IPO in July 2019 through to today, and it is really, really satisfying that we are on track to hit all of our 2020 milestones. From the launch of our DecisionDx-SCC test in late August, to our impending launch of our DiffDx-Melanoma test and to the finalization of the expanded LCD and accompanying billing and coding article for our DecisionDx-Melanoma. But we have not publicly announced the additional dermatology test that we are initiating development on. We are confident they will represent significant areas of unmet clinical need, and we look forward to updating you shortly. For '21 and '22, we have not updated our public milestone list since earlier this year. But suffice it to say that we believe we are on track to have draft LCDs issued for both our DecisionDx-SCC test as well as the focus of this conference, our DiffDx-Melanoma test in 2021, with final LCDs being effective roughly 1 year later. Slide #28 is a short glimpse at some of our launch material and our laboratory reports. I'm pleased to say that we were successful to not only achieve the production of these commercial launch materials, but we were able to successfully recruit and hire another great, high-quality sales and medical affairs team who will lead the charge going forward. They have been focused for the last few weeks and months on learning not only our DiffDx-Melanoma test, but our other 2 skin cancer tests as well. And again, we look forward on updating you on progress once the test is made available clinically, and we can track the adoption of this test in the coming months and quarters. As a reminder, on Slide #30, our DiffDx-Melanoma test is focused on the 300,000 biopsies that are taken per year but remain difficult to diagnose. We have completed validation and our 2 initial publications have been accepted and should be available publicly shortly. We are also uniquely positioned to capitalize on our existing relationships within the dermatology communities. Those 1,900 dermatopathologists who we have worked with over the last year and the 4,200 treating clinicians, largely dermatologists who use our DecisionDx-Melanoma test. In summary, on Slide #31, 2020 has been a fantastic year, even with COVID 19 causing a couple of bumps in the road. We entered this year with 1 product in the marketplace to serve our dermatology patients with a U.S.-only TAM of approximately $540 million. And we are on track to exit 2020 with 3 marketed tests and a U.S.-only TAM of approximately $2 billion. That's a nice place to close out the year. And with that, operator, we will take questions.

[Operator Instructions] First question, we have Sung Ji Nam from BTIG.

Congratulations again on the launch. Maybe, Derek, starting off with you mentioned, obviously, Myriad has a product on the market, myPath Melanoma. Just would love to hear your thoughts in terms of how this test compares. Obviously, without a head-to-head study, it's hard to compare. That being said, based on the performance characteristics that you guys are talking about here, obviously, pretty impressive, so I'm just kind of curious what your thoughts are. And would love to actually hear from Dr. Estrada as well if she has any thoughts on how the 2 tests compare against each other.

Sure. I'll give you, I guess, my perspective, which hopefully is unbiased, and then Dr. Estrada can give her unbiased perspective, too. When we look to undertake this program a couple of years ago, we held a number of focus groups with both dermatopathologists and dermatologists to just -- to kind of understand, is there a need that needs to be filled or is the marketplace completely satisfied with what they have available today. And what we were handed was sort of a checklist that Dr. Goldberg went through, which was to say, hey, this is a large number of lesions or biopsies that we're trying to get to the best possible answer. If you could go ahead and develop a gene expression profile test similar to your DecisionDx-Melanoma test, that would meet a couple of goals, so for instance, a high technical success, could we go ahead and develop laboratory processes where maybe we could generate results for more than 95%, 96%, 97% of samples coming in? If we're able to do that, can we reduce the sort of intermediate risk zone from kind of the mid-teens or low teens to maybe even lower single digits and maintain accuracy from a sensitivity and specificity standpoint? The other element that was quite attractive to our more dermatology customers was, would you be able to then fairly quickly provide me information on the DecisionDx-Melanoma test, so if I was confident that this uncertain biopsy was now correctly categorized as a malignant melanoma, I'd be much more comfortable knowing what treatment pathway to go forward. So that's the feedback that we got a couple of years ago. And that's not necessarily a negative on Myriad's test. It just was a request to try and do a bit more and a bit better. There was a fifth element there as well, which was to say there are some lesions that we see, which actually, there isn't a whole lot of good validated data on to see how other gene expression profile tests perform. So do you mind going a little bit wider in terms of different kinds of subtypes. So we set out with those sort of goals in mind. And as you could see from the, I guess, the one performance slide from Dr. Goldberg's presentation, we're able to achieve a very, very high technical success rate of around 96% of the time when samples came in. For our clinical validation study, we were able to generate a test result. We were able to reduce -- or rather achieve an intermediate risk zone. I think it was 3.6% for all ages and 3.4%, I think, for people who are in the Medicare population group, and maintain a high accuracy stats regarding sensitivity and specificity in both all ages and the Medicare-based population. We don't have, as you would expect, any sort of head-to-head comparisons. So all we can look at is what was kind of published in the Myriad studies and just be able to have a conversation openly with our dermatopathology customers and colleagues in the weeks and months going forward, and let them make their own judgment call on do they see this as being something which is adding value and is, therefore, something they should consider ordering? Or is it something that they just see as redundant to what they have already. So I think it will be in the eyes of the customer going forward. I hate to go ahead and say that we developed a superior test because we don't really have, as you said, Sung Ji, head-to-head data. But I think we are quite pleased, I guess, with the check boxes that we're able to check off during the discovery, development and validation of the test. Dr. Estrada?

Yes. I think I can essentially echo what Derek just highlighted. But as a customer, a previous customer of Myriad, I've sent about 50 cases to Myriad over the years. And in general, I'm getting a decent result, but I think the more cases that I've sent, the more the limitations of the test has been magnified. So first of all, Myriad has a technical failure rate of nearly 15%. So that means I get to the end of the line with my routine testing. I can't come to a decision and send it to Myriad and wait a week or so. And basically, it's like a tie game. There's no answer at the end of that. So that can be frustrating. And one of the big benefits of the DiffDx test is that the technical failure rate is much lower than that, in the low single digits, like you mentioned. So second, there's subtypes of melanoma where the Myriad test is less reliable, so there's always going to be certain cases that I don't feel like I can confidently send to Myriad. One of them is desmoplastic melanoma. And in addition, something that I have found working in the Desert Southwest, where we have patients with a lot of skin damage, we commonly see a subtype of melanoma called Lentigo malignant melanoma, and I also am not super confident with the results that I'm getting from Myriad for that subtype of melanoma. In contrast, the tests that we're talking about today has been validated and have very strong results for all subtypes of melanoma, including melanoma in situ, which hasn't been validated by the Myriad test. So that means when I'm looking at a lesion that's confined to the epidermis, which is going to be a significant number of lesions that I find difficult to diagnose, I have a new method that I can send this case to and help get a final diagnosis. So third, there's a significant number of cases with the Myriad tests that fall into an intermediate or indeterminant zone between benign and malignant. And again, it's the same results. At the end of the day, it's a technical failure. It's a tie game, and I don't have an answer. So that can be frustrating. For the Castle DiffDx test combines the 2 algorithms with the AI approach, and that results in a really high accuracy. The sensitivity and specificity of the Castle tests are better than I could have ever hoped for and better than anything that's been on the market, and certainly, light years beyond what we can do just under the microscope and with our stains. And finally, being able to use the same tissue sample from the DecisionDx-Melanoma and use it for the -- or the DiffDx-Melanoma and then sending it for the DecisionDx-Melanoma test, all without having to wait for a block to come back to our lab or increasing turnaround time is really nice, and we also know that we're sparing tissue in case we need additional testing down the road for these patients.

Great. That's super helpful. I have 2 quick follow-up questions. Dr. Estrada, you talked about using this test after the IHC stain step. I was curious, in the future, is there -- do you -- given the performance characteristics again, do you see this potentially kind of even replacing the need for the IHC stain step?

Yes. No, at this point, I don't see it replacing it because there's still a subset of these lesions where we can come to a final decision with just the stains. And we have the stains in-house so we may as well be using them and come to an answer and not have to send it off to the genetic testing.

Okay. Do you know what percentage of the 300,000 might be further elucidated based on IHC then?

Sure. The 300,000 is actually the number that comes after doing the IHC testing. So that's after we've exhausted all the tools that we have in the lab.

Got you. And then just lastly, I look forward to the publications, but just wanted to make sure the malignant samples that you guys used, is there -- are there good representations of all the different stages of melanoma as well? Or I'm just kind of curious how you guys looked at that.

The answer is yes. So we included both in situ and invasive melanomas, broad age range, broad histopath subtype and broad thicknesses. So I think we have, largely speaking, kind of what you did -- what, I guess, Dr. Estrada will be seeing in her dermatopathology lab on a routine basis.

Next question, we have Puneet Souda from SVB Leerink.

Derek, and thanks for the presentation and thanks for Dr. Estrada for going through the workup of the patients here. So first one for you, Derek, on reimbursement, what's your expectation in terms of the $2,000 per test that you mentioned? Could that move higher given the performance of the test here? And would this test be amenable to an ADLT given the sensitivities, specificities in the overall performance that you are getting for the test?

Asked a lot of questions there. I'll try and answer those completely, Puneet. I think we will certainly -- we certainly believe that the value of our test hopefully is higher than $2,000, and that's not to say, hopefully higher. I think we can demonstrate the value at a higher level, as you said, maybe based upon accuracy, performance, et cetera. I think from a modeling standpoint, though, I wouldn't want to be that aggressive and expect that we would end up at a fully seasoned ASP that's higher. I think that's -- we're going to obviously do what we can to get paid fairly and get paid fully. But I think from an external modeling standpoint, $2,000, I think, is a pretty safe, credible way to be at until we kind of see how things mature out over the next couple of years. Regarding ADLT status, that's an interesting question. As you may know, Myriad had applied for ADLT status for their myPath melanoma test, and they were granted that status, I guess, maybe a year ago or so now maybe. We have assumed, even with the -- what we would see as an improved or increased test performance, that -- or even with that, that we would probably still be categorized as a CDLT because, generally speaking, we aren't changing the clinical use of the Myriad test, and they got here first. So my assumption here is that we should assume that Medicare will see us as a CDLT and look over to the Myriad test as the innovator test. But it's an interesting question you asked there, which is to say, well, if you actually are providing what may be perceived by Medicare as a superior profile, will they view that separately. I think that's something we'll be examining and talking with the group at CMS. My assumption, though, is they're probably going to say, one and only for a given sort of clinical answer that's being provided.

Okay. That's very helpful. And you have 3 manuscripts here, and I see the clinical utility data as well. And what is the level of data that you need for the LCD? And then in terms of the -- if we start thinking about the clinical data stack to convince the commercial pairs and to differentiate the assay even more versus the competition, how are you thinking about that in terms of further studies, larger studies? And then I know this is not a screening tool, but it is upstream. So just wondering if there are any thoughts around FDA regulatory approval down the line.

Let's see here, studies for commercial adoption. Regulatory was the third question. What was the first question, Puneet?

Yes. So just first one is you have 3 manuscripts there. So do you need to -- in terms of the clinical data stack and the studies that you need to build out, are you planning further studies or larger studies here to differentiate yourself versus the competition? And the second question is FDA regulatory.

Perfect. Okay, okay. So first of all, I don't -- we didn't necessarily mean to put up a slide that is misinterpreted slightly here. We will have 2 studies out in November. I think 3 mastheads were on that 1 slide, but the first 2 mastheads are from the same Estrada et al publication that goes through the development and the validation of our test. And then the second publication is the far right, which is on clinical utilities. So there aren't 3 that are coming out, just 2 there. I think from a study standpoint, you can see in the initial consort diagram that Dr. Goldberg went through, looking at the flow, we had accrued during the course of working through this first validation study over 2,800 specimens. And we use roughly, I think, 900 or so for the initial development and training set development and validation. So you can bet that we are working on performance that is the -- after this first one is being done. We'll hopefully wrap those up timely to provide more additional data to provide comfort of the continued, hopefully, consistent performance of our assay. We are also initiating additional protocols as we speak to go ahead and build out a stronger, more robust publication track record over the next couple of years. So we will definitely do what we've done with our other tests, which is to say, once you get to validation, your life has only just begun. There's lots of work to do. So that will build out over time. I think from a sort of Medicare coverage perspective, my assumption is that consistent with sort of the 21st Century Cures Act and sort of the adjustments that we've seen Palmetto make and Noridian make in terms of moving from test-specific to disease-specific or clinical use-specific LCDs, that's probably the approach that Palmetto would want to take with this test result. As we sort of benchmark some of the more recent LCDs that posted either in draft or even final, it looks like we have a good database once these studies are published to go ahead and have Palmetto consider an LCD based upon these 2 initial publications in the context of developing a disease-specific or a clinical use-specific LCD. So that feels like that should be adequate based upon the data that we generated so far. But of course, that's the eye of the beholder. I think from a commercial kind of progress going forward, as we've talked about with our other tests or mainly the DecisionDx-Melanoma test, commercial coverage, I think, will be a plotting opportunity going forward with, of course, slow time lines from an adoption standpoint. So that's how I would view that perspective. In terms of FDA regulatory aspect, and this is really only to be used on a patient who has had a biopsy taken because the disease is suspected, or at least has a sign of being a suspected disease or diagnoses. So we should not be in the sort of screening category of concern of FDA oversight and regulation at this point in time because the ordering -- the initial treating clinician -- the dermatologist has already taken a biopsy to try and figure out a diagnosis based upon signs and symptoms, that puts us more into the diagnostic category as opposed to screening product category. So I don't think this poses some kind of a different expectation from what we have today with our other 2 tests.

Okay. Perfect. And for Dr. Estrada, if I could ask a question on, obviously, you have been involved in the field for quite some time, and you're appreciating the role of gene expression profiling in melanocytic lesions and publishing on it. And so -- but when you look at the broader field, your peer groups, peer dermatologists and especially in the broader community setting, how do you think they would view a test like this? And sort of what is the sort of the curve that they need to come up in order to be more open and adopting of this test? What do you think needs to sort of happen there in order for that broader community to sort of bring this test into their clinical practices?

I think it's mostly a matter of being able to clearly communicate the validity of this test and its utilization in a really large number of cases that are putting us in a difficult situation. And I think that as we start to get providers throughout the community to adopt it and especially bigger name providers at well-known academic institutions, then the rest will follow. It tends to be a pattern where Dr. A adopts whatever test Dr B. promotes because he respects Dr. B's opinion. So if we can get Dr. B on board, then Dr. A and several doctors after that are going to follow suit. So I think it's a matter of getting out there, communicating the message and really relaying that this is a game changer, and then getting key physicians in the community to adopt it. And I really do think the rest will follow.

Next question, we have Thomas Peterson from Baird.

First for Derek or Frank. I'm just curious, I know it's obviously well too early to be talking about volume trajectory. But if we think about larger annual patient population, more awareness with Castle in general, are you guys expecting a similar launch trajectory to DecisionDx-Melanoma or squamous? I guess, early days, how should we be thinking about that volume ramp?

Yes. Good to hear from you, Tom. I think, of course, the right answer to say is new product forecasting is tough. But that being said, I think in terms of unit volume, we are benefiting in 2020 from over 5 years of name recognition, I guess, you would say, for the company. Hopefully, that's almost all positive name recognition. By the way, we do interact in the last 12 months through the end of June with 1,900 dermatopathologists who should know our name because we ended up sending request to them for cutaneous melanoma tissue from patients that they diagnosed. So again, hopefully, there are some relationships, some open doors here. We are, as you may recall, though, not launching this test initially with our larger sales force. So we have around 32 sales territories -- not around, we actually have 32 sales territories that are focused on dermatologists and surgeons that are educating doctors around our DecisionDx-Melanoma test and our DecisionDx-SCC test, we made the decision earlier this year actually pre-COVID, and maybe COVID just helps reinforce the decision that having an individual sales representative, a medical science liaison learning 2 products in sort of an 8, 9-week time period, launching 2 products, maintaining a focus on the fact that we still have low penetration for our DecisionDx-Melanoma test sounded like one too many things to do well. And so we purposely hired up a separate group of around 10 or 11 sales individuals who are actually in training with us now as we speak to you, to launch this DiffDx-Melanoma test to the dermatopathology community first. And I think once we get past the initial sort of launch period, now I don't know if that's going to be 3 months, 6 months or 9 months, but I expect sometime by the end of next year, we'll go ahead and fold back to a single sales force of around 45 and probably add another 10 more, so that we exit '21 with maybe 55 or 60 people covering smaller geographies, hopefully confidently communicating the value in all 3 of our tests. So I think the initial quarter or 2 where we really have 10 representatives educating dermatopathologists on this test. I think it should be faster, I would think, given the larger number of patients or samples than our DecisionDx-Melanoma test was 5 years ago, but it's going to also be purposely limited to only really 10 or 11 sales territories. So I wouldn't get overly confident or overly unconfident about what we see in the first couple of quarters post launch. I think once we fold ourselves back into a single sales organization, you now have 45, 55-plus bodies helping to educate our customer base on this most recent DiffDx-Melanoma test will give us a better idea of trajectory in kind of '22 and '23. I know that's a long answer, which essentially avoids your question. I apologize for that. But I think I would expect maybe a slightly slower ramp than what we might see with our squamous cell carcinoma test, only because we have 10 people working on education versus 32. But a year from now, assuming we've made the decision to fold back into 1 larger sales force, I think you can then say now we can take those couple of quarters at that point in time and get a projection going out several years.

Okay. Now that's helpful. I appreciated you guys' commentary on CMS conversation disease to LCD. Could you just remind us what percent of this patient population you think falls in that Medicare range?

Yes, excellent question. I think -- as a reminder, our DecisionDx-Melanoma test for managing patients diagnosed with melanoma, about 45% of the patients are Medicare age, so 65 years of age and older, and that's pretty close to what you expect. I think the average age of diagnosis of melanoma is sitting around 60 years of age, plus or minus a year. So that feels okay. Under the assumption that these difficult-to-diagnose melanomas or the lesions don't show some sort of age bias, like they're only seen in younger patients, not older patients. And I would think we'd expect around 45%, 50%, 40% of the patients who have a difficult-to-diagnose lesion to be in the Medicare age population.

Okay. Super helpful. One more, if I could sneak it in. I think most of mine on the DiffDx have been covered, but just curious, do you guys have any update on the squamous launch here 2 months in. And also if you're seeing any volume, patient restrictions as COVID kind of flares back up here into October?

Yes. So maybe I'll talk about the second one first, maybe. So we have an earnings call coming up in a couple of weeks, we can provide some more specific quantitative data, certainly. I think on the COVID impact and patient flow, we're seeing -- our perception, at least through September, which is the most recent data that we purchased so far, suggests that the diagnoses of melanoma is still recovering, I guess, you would say, not only over pre-COVID levels in January and February, but also over 2019. What we have not seen happen yet, which is really disappointing from a patient perspective, I mean cash will disappoint too, but really as a patient perspective, is that I think there are several thousand patients. We've seen one projection upward of 20,000 patients who should have been diagnosed in the first 9 months of this year who have not been diagnosed, and that's not because melanoma went away, it's because either dermatology practices haven't gotten back to figuring out how to kind of return to kind of full patient throughput, or you've got too many patients who are still concerned about catching COVID in their dermatology office versus dying from what could be melanoma. That's concerning from just a human health perspective because those -- if the true number ends up being around 20,000 diagnosis end up getting deferred into '21, for example, that's tough because they're going to be much thicker, much more worrisome diseases upon diagnosis, and that's not good for patient care. So I think we haven't seen sort of full recovery yet from looking at outsourced third-party data. But hopefully, we'll have a bit more flavor at the earnings call here in a couple of weeks regarding even some responses to October, et cetera. But that's how that goes. So I do think, though, from a modeling standpoint, if you put together 2020 and 2021, and you assume 130,000 patients might be diagnosed each year, so 260,000 total, maybe we end up having 110,000 diagnosed this year and 150,000 next year. But I hope, for patients' sake, that you end up getting 260,000 melanomas diagnosed in this 24-month period because if we don't see that, it's only because people are unfortunately making what I think is a wrong choice from a health care standpoint. So that kind of answers COVID. On the SCC test, we'll provide some data at the earnings call. I would say that we asked our representatives to go out there and sort of -- I don't want to call it a soft launch, I'm not sure what that means exactly. But essentially, our #1 priority is really making sure doctors are educated and aware of our DecisionDx-Melanoma test. And when they get done with that discussion, they then switch over to another test in skin cancer that we can talk about, which is DecisionDx-SCC. Largely speaking to most of our field force, when they're making calls to clinicians or to dermatology practices, they're able to get both of those call cycles in there or both of those product sales in there, which is a good thing. So it seems to me that we're seeing things lining up about how we were hoping for prior to launch time.

I am showing no further questions at this time. I would now like to turn the conference back to you, Mr. Derek Maetzold.

Thank you for joining us today. It's not very often that you get to launch just 1 clinically important test in a given calendar year, it's exceptional that you're able to launch 2 tests within a couple of months period. I want to thank the R&D team for their dedication and focus of not only developing protocols, implementing protocols, our external collaborators who believed in us and wanted to come work with us in developing not only this test, DiffDx-Melanoma, but also our DecisionDx-SCC test. And of course, I thank their patients and our future patients because without them, we wouldn't be here. And that's important to go and keep in mind. And with that, I thank you for your attention, for your participation, and we look forward to catching up with you in a couple of weeks.

Ladies and gentlemen, this concludes today's conference. Thank you again for your participation, and have a wonderful day. You may all disconnect.

Thank you, operator.

You're very much welcome, sir.