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

Good afternoon, and welcome to Castle Biosciences DecisionDx-SCC Launch Conference Call. As a reminder, today's call is being recorded. We will begin today's call with opening remarks and introductions, followed by a question-and-answer session. I would like to turn the call over to Frank Stokes, Chief Financial Officer. Please go ahead.

Thank you, operator. Good afternoon, everyone. Welcome to Castle Biosciences DecisionDx-SCC Launch Conference Call. As with 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. I'll now direct you to our agenda today. We appreciate your interest, and we will go through several presentations and allow for questions at the end. In addition to Derek and me on the call today, we have the additional Castle Biosciences attendees listed on the slide here. They are available for questions if needed but won't have speaking parts during our presentation. We're excited today to be joined by Dr. Ashley Wysong from the University of Nebraska Medical Center in Omaha. Dr. Wysong is the Founding Chair and William W. Bruce MD Distinguished Chair of the University of Nebraska Medical Center Department of Dermatology. She earned her medical degree at the Duke University School of Medicine, where she was also Valedictorian. Dr. Wysong completed her MS in epidemiology and residency in dermatology at Stanford University, where she also served as Chief Resident. Her fellowship in procedural dermatology and Mohs micrographic surgery was completed at Scripps Clinic under the direction of Dr. Hugh Greenway in Mohs surgery, Dr. Victor Ross in laser and cosmetic dermatology and Dr. Leland Housman in veins. I now direct you to the next page of the deck, and I'll turn it over to Derek Maetzold, Castle's President and CEO.

Thank you, Frank. Good afternoon, everyone. Before we get started here, I think it's worth taking a moment or 2 to recognize the fact that Castle is now moving into its third chapter of hopefully a multiple chapter life, and it's appropriate to recognize those individuals that have gotten us to where we are today. Chapter one, when we launched the company in the latter part of 2008, we were focused on providing solutions, diagnostic test products to meet unmet medical needs for patients diagnosed with rare cancers. DecisionDx-UM for patients diagnosed with UV melanoma is a legacy of that initial focus. Chapter 2 began with the initiation of our own internal R&D program. The first program that we worked on is what we now call DecisionDx-Melanoma for use in early-stage cutaneous melanoma. Chapter 3, in my view, represents our move from offering a single product or a single solution to serve patients diagnosed with skin cancer to having the opportunity by year-end to serve 3 areas of unmet clinical need: that of patients with early-stage cutaneous melanoma with our DecisionDx-Melanoma test; early-stage but high-risk cutaneous squamous cell carcinoma with the DecisionDx test, SCC, which is the focus of our discussion today; and a third test for use in patients who were biopsied to rule out melanoma but in which the diagnosis remains suspicious or difficult. As you know, we have completed validation of our suspicious pigmental lesion test, and we are on track and expect it to be commercially available in the fourth quarter of 2020. Now how did we get here? I think there are 3 groups that we should be recognizing. First and foremost are the patients who agreed to participate in our clinical research programs. Without their tumor specimens and their associated clinical outcomes data, we could not perform the work necessary to get us from an idea to a clinically validated test. Second is our collaborators like Dr. Wysong, who you'll meet in a few minutes. In the case of our DecisionDx-SCC test, we now have over 92 collaborating centers who have contributed samples and data on over 1,400 patients. The data that Dr. Wysong will be walking through today is focused on the 420-patient cohort that were part of the clinical validation study. And the third group that needs to be recognized and thanked are our employees, from our strategic individuals who framed out the clinical needs; to our R&D folks who thought through the study protocol approach, including gene selection and analysis approach; to our clinical research operations group who contacted, contracted and successfully worked with our 92 collaborating centers; to our laboratory group, including IT, who developed assay procedures during the discovery phase of the program that were able to move and translate over into our clinically validated test; and of course, to our commercial and reimbursement teams who pulled together what we are seeing as a successful introduction of a much needed test, one that will dramatically improve the accuracy of the likelihood of individual patient diagnosed with localized but high-risk SCC will or will not metastasize. To these 3 groups, I personally thank you on behalf of the patients that we will serve from this day forward. We would not be here today without your decision to support the development of our SCC test. And I thank all of you, our patients, our collaborating centers and our employees. So some of you may not be totally familiar with the Castle story. Let me spend a couple of minutes on this slide here about who we are, what we do and where DecisionDx-SCC fits into the Castle equation. First and foremost, we are a diagnostics company that focuses on developing solutions for unmet needs in patients diagnosed with dermatologic cancers. Thus, we innovate. Once we identified a test that adds substantial value to clinical decision-making, we then commit to investing year after year in clinical research to confirm the accuracy of our tests as well as to build out evidence on how to get the best value out of our test, how do clinicians and patients best get value out of our test. Thus, we develop evidence. In fact, between January 2015 and August 2020, there were over 26 peer-reviewed publications that supported the use of our DecisionDx-Melanoma test and there are more to come. That's going to be the hurdle we set for ourselves for DecisionDx-SCC. We believe that this commitment to evidence translates into appropriate adoption by our clinical customers and reimbursed by Medicare and commercial payers. We believe that the output of focusing on these first 2 blocks is what has put us into such a strong financial position. Next slide. As a reminder of our position today, I wanted to spend 1 minute here on our 2Q 2020 financials. This represented, obviously, second quarter performance and performance through 2Q, although you could argue it was 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 '19. We also saw a continued lift in gross margin that we expected, which came in at 83%. And we ended 2Q 2020 in a very strong cash position, with cash and cash equivalents of nearly $180 million and adjusted in period cash -- operating cash flow of $3.3 million. Our team has executed extremely well over the last couple of years, from the lead up to our IPO in July 2019 through to today. I'll just highlight for you that we set out to achieve certain milestones in early 2020. We have already hit 2 of them and are on track to hit the rest of them, not only hit them on time but early. As I alluded to on the earlier -- on the prior slides, I expect us to turn the page at the end of this year, December 31, 2020, having moved from a single dermatologic product, DecisionDx-Melanoma, to having 3 clinically actionable proprietary gene expression profile tests available for clinical use. This is a great way to end a great year. This slide turns the columns on the prior page into rows and add some perspective. I'm only going to comment on the last 2 rows. We currently have 2 primary customer groups for our DecisionDx-Melanoma test: medical dermatologists, including Mohs surgeons like Dr. Wysong; and surgeons, some general, some surgical oncologists, some plastic surgeons, who perform [ skin ] biopsy, surgical procedures and wide local excisions on patients with early-stage melanoma. The dermatology audience, as you may recall from earlier calls, is responsible for about 80% of our test orders. We do call on 2 additional groups of physicians: dermatopathologists so that they and their teams understand why they're referring dermatologists, our joint customer is ordering our DecisionDx-Melanoma test; and medical oncologists to locate those in colleges who will manage a DecisionDx-Melanoma high-risk patient as a high-risk patient. I mentioned these points because we will be in a unique position in that the primary target for our DecisionDx-SCC test is that same medical dermatologist, including Mohs surgeons who we 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 DecisionDx-Melanoma test, 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 one or more high-risk factors. The call point for our test for suspicious pigmental lesions also overlaps well with our current call base. Specifically, we received tissue specimens for clinical testing from over 1,900 dermatopathologists in the last 12 months. Again, our initial launch targets will be these same dermatopathologists with whom we have developed and had relationships over the last several years. So a couple of brief comments on what the DecisionDx-SCC is and the workflow. Our test has been validated for use in a patient who has been diagnosed with invasive cutaneous squamous cell carcinoma or squamous cell carcinoma of the skin and one or more risk factors. As you may recall, there are well over 1 million invasive SCCs diagnosed per year. But we currently estimate that only 20% or 200,000 patients present with one or more risk factors on an annual basis. This is our target patient population. Once diagnosed, a clinician can order our test, DecisionDx-SCC, via fax or uploading an order form into our secure HIPAA-compliant portal. From there, our client services team in Phoenix takes over. That will reach out to the respective dermatopathology lab where the tissue specimen is located or resides and request sections from that tumor specimen. These are then received in our laboratory, and we begin the RNA extraction/purification process to prepare for the expression of 40 genes: 34 that are discriminatory between a low and high risk for metastasis and 6 control genes. The output of this QPCR step, the expression levels of those 40 genes are then fed into our proprietary, trade secret output -- algorithm. And an output is generated that we report as Class 1, Class 2A or Class 2B, with Class 1 representing the lowest risk of metastasis, Class 2A moderate risk and Class 2B a high risk of metastasis. Of note is the fact that, well these were all patients with one or more risk factors, which put the patients in the high-risk category, according to NCCN guidelines. Approximately 50% of the time, these patients had a Class 1 or low-risk test score, meaning that they have a biologically low risk of metastasis despite having high-risk pathology features. This percent is obviously significant as it can allow a patient and their clinicians to discuss the option of watching and waiting as opposed to proceeding forward interventions meant for patients with true high-risk SCC. On a final note, I think it's important to emphasize that while the order form for DecisionDx-SCC will naturally look a bit different from the order form for DecisionDx-Melanoma test, all current customers will see this process as identical to that existing test. The same goes for the dermatopathologist laboratories that we'll be working with to secure tissues in a rapid fashion. And the same goes for our internal laboratory processes from accessioning to wet lab processes and reporting. In other words, low-risk launch, low risk for confusion, high likelihood of satisfied customers. So how did we get to where we are? At the discovery phase, we employ several principles as a company. The first principle starts with a core belief that we should only work in a specimen type that will fit right into the current workup of a patient. Thus, while it may be scientifically pure to start with fresh frozen specimens, we believe that whatever we discover must work, in this case, in FFPE tissue or formalin fixed impair formed tissue, as that is how all SCC biopsies are preserved. So our starting material throughout the entire discovery to develop the validation process or program has been to limit our work on FFPE tissue, the kind of tissue we'll be testing clinically. We use 2 approaches to identify candidate genes for DecisionDx-SCC development: a targeted approach and a global approach that employee deep learning. Some of you may commonly call this as AI. From these efforts, we identified 140-gene candidates and then move to development stage. Our development stage involves selecting a training cohort of 122 samples with known long-term outcomes then starting again with deep learning with neural networks. We were able to reduce the genes of interest to 34 and finalized and locked the predictive algorithm for independent validation. Dr. Wysong will walk you through the validation cohort data. Over the last 12 months, we've been able to successfully complete a number of studies and have them published. As you see in this slide here, we have generated peer-reviewed publications that support both the clinical validation of the assay and the intended clinical use of the assay in the last 12 months. With that, I'll turn it over to Dr. Wysong, introduce her as the Chair of the Department of Dermatology, University of Nebraska Medical Center in Omaha, Nebraska. Dr. Wysong?

Thank you so much, Derek. And excited to be talking to you guys today about our newly developed test for cutaneous squamous cell carcinoma. And so as you all have heard, cutaneous squamous cell carcinoma is one of the most common cancers that we treat in the United States and really globally. Approximately 1 million patients are diagnosed with squamous cell carcinoma each year in the United States. And as Derek had mentioned, approximately 1/5 of those or 200,000 patients are identified as high risk based on multiple different factors. We take those risk factors then and look at our national guidelines to ultimately develop treatment recommendations based on these risk factors. But what we've learned and by treating these every single day, I recognized the downfall that traditional clinical pathologic-based risk factor systems suffer from what we think of as low-positive predictive value, meaning that many more patients are actually classified as high risk than they are and then those that actually go on to develop metastatic disease. And so what this means is we often can lead to overtreatment for these types of patients or even undertreatment for a substantial number of patients that are identified as low risk but really are at higher risk for metastasis. And so moving on to Slide 18. This is one of my favorite slides to really look at when we try to assess the overall impact of cutaneous squamous cell carcinoma. So what we're seeing here in blue is a typical cancer -- the cancer numbers that are maintained by the SEER national cancer registry. And when you actually stack up squamous cell carcinoma over the top 5 cancers that are measured by the SEER database, which includes breast cancer, lung cancer, prostate cancer, colon cancer and melanoma, we see that there are more cutaneous squamous cell carcinomas than all the rest of those combined, which is pretty impressive when you think about it across the board. And so moving on to Slide 19. I want to take you through what the typical patient journey looks like for a patient that gets a new diagnosis of squamous cell carcinoma and really where DecisionDx-SCC will play a role. So the typical patient, as Derek had mentioned, is seen for a suspicious lesion on the skin, often by the patient themselves, well -- or honestly a spouse or family member or potentially when going into see a primary care provider. The patient is then typically referred to a dermatologist to perform skin exam. And as mentioned previously, around 80% of those patients are seen primarily by a dermatologist. The patient undergoes a typical history and physical examination. And when we're concerned about squamous cell carcinoma, we will often perform a regional lymph node examination. And the true diagnostic component that comes in will be the skin biopsy that happens. So when the skin biopsy comes back with the diagnosis of squamous cell carcinoma, the typical dermatopathologists will look for some "risk factors" that we see in squamous cell carcinoma, which we're going to talk about. This is where it will really trigger a board-certified dermatologist or a surgeon, primary care physician or a Mohs surgeon to really start thinking about DecisionDx and what we can do now that we know we have a diagnosis of squamous cell carcinoma that is considered high risk with one or more risk factors. So moving on to Slide 20. Honestly, like all cancers, SEC management plans are guided by the risk of metastasis and ultimately utilizing tumor staging. Traditional tumor staging is what we call the TNM model or tumor factors, nodal and then metastatic disease. Now as we've kind of discussed previously, the reliance really on the clinical or pathologic risk factors, whether or not you incorporate them into staging, can be really challenging for patients who have high-risk factors but do very, very well. Because we can, unfortunately, overstage those patients and treat them when they may not need a more complicated treatment. So it really makes it difficult to personalize or to really target or trigger a patient's management plan based on a risk factor alone. So moving on to really the clinical problem. So I think as many of you know, and we have talked previously about the DecisionDx-Melanoma prognostic study, melanoma has been traditionally viewed as the most deadly form of skin cancer. And this is absolutely true relative to the overall incidence of melanoma. So I mentioned we treat about 1 million squamous cell carcinomas a year in the U.S. and we treat around 80,000 melanomas. Now as the incidence of squamous cell continues to grow rapidly with the aging baby boomer population and in general, as individuals are living longer, we see that the number of SCC patients that are actually dying from squamous cell carcinoma is now surpassing that of melanoma. And so this really concerns myself as well as all of us in the field when we're thinking about the management of squamous cell carcinoma because it's now obviously a very significant clinical issue. So this is really where DecisionDx-SCC comes in. It's obviously now has been -- gone through discovery phase, development phase and now validation to predict the metastatic risk for an individual SCC patient that has one or more risk factors. And this is really important. And so our studies overall have found that the DecisionDx-SCC gene expression profiling test is truly the most powerful predictor both in univariate and multivariate analysis, which we're going to go through. And so ultimately, we know that incorporating DecisionDx-SCC in tumor biology, along with our clinical and pathologic risk factors, really provides a superior classification to utilizing either one of these alone. So I kind of -- one of the reasons that I've been really excited about DecisionDx-SCC is that this truly is a development program that was initiated based upon the recommendation of clinicians. In fact, a little sidebar, I remember sitting at a melanoma meeting in San Diego several years back and saying to one of the Castle reps. "Man, this melanoma test is amazing. We really need to look at developing this in squamous cell carcinoma." And the reason why is because given the poor predictive value of the clinical and pathologic factors alone, we knew very early on that if we could develop a similar test for high-risk SCC like what's developed for melanoma, it really has the potential for a major impact in our skin cancer population. So moving on to Slide 23. This really goes over the typical risk factors that we see in squamous cell carcinoma, and DecisionDx is -- intended uses for patients that have one or more of these risk factors listed on this slide. And so as mentioned by Derek, DecisionDx-SCC is a 40-gene expression profiling test that actually uses a patient's individual tumor biology to go on and predict that specific patient's individual risk of squamous cell carcinoma metastasis. We know that squamous cell carcinoma metastasis is the biggest predictor for death in these patients. So this is extremely important to be able to identify. And ultimately, the test result, as mentioned by Derek, stratifies patients into a Class 1A, Class 2A or Class 2B risk category and ultimately is meant to be used as a prognostic method for determining the risk of metastasis and ultimately, management in association with the other risk factors. So now 4 peer-reviewed publications have demonstrated that DecisionDx-SCC is an independent predictor of metastatic risk. And ultimately, the integrating DecisionDx-SCC with current prognostic methods and clinical risk factors actually adds positive predictive value to clinician decisions regarding staging and management, which is really important. So moving on to Slide 24. This actually highlights our overall validation study. So Derek mentioned the discovery and development phases, and this is the independently performed validation study. It involved 420 participants, over 33 centers. All patients met NCCN criteria for being high risk in squamous cell carcinoma. 65% of the tumors were on the head and neck. And 1/4 of participants were immunosuppressed, which is what we're seeing in a lot of our patient populations, particularly as the number of organ transplants and other immunosuppressive medications are going up. In our cohort, overall, we saw a 15% metastasis rate. And so this is ultimately what we want to see when we're developing a prognostic test. We want to almost oversample for a higher rate of metastasis, so we can ultimately develop a prognostic test and validate its use in this study. And the overall median time to metastasis for this patient population was 0.9 years, with 95% of them metastasizing by 2.5 -- 2.7 years. So moving on to Slide #25, which is really the money slide when we're looking at the validation study. This is a Kaplan-Meier analysis for metastasis-free survival for the validation study in the 420 patients with high-risk SCC. In the validation study, metastasis was defined as a regional or distant metastatic event. And importantly, local recurrence of the tumor was actually excluded from this definition. We think that local recurrence can have other factors involved rather than just tumor biology alone, including the way that the patient was treated. And so really, as you can see here on the Kaplan-Meier curve, there is a strong separation between Class 1, 2A and 2B results. And as a reminder, the overall risk of metastasis in the patient cohort was 50%. But as you can see here, patients with a Class 1 result have less than half the metastatic risk seen in the overall study population. Half, which is really important for -- to identify these patients who may otherwise be overtreated based on having one high-risk factor. When you look at Class 2A, these patients actually behave similarly to those with most adverse traditional clinical pathologic risk factors such as deep invasion or poor differentiation. So the 2 ways, similar to the strongest clinical pathologic risk factors is about where they're performing from a biologic standpoint. Now Class 2B. In this case, these patients have an over 50% risk of metastasis, and it's really identifying this most worrisome group. And the other really important component is this has the highest positive predictive value, higher than anything that we've seen before. So overall, what's kind of important to know with this large validation study of high-risk SCC is that half of the patients will end up being Class 1B -- Class 1, which is the low biologic risk. Approximately 40%, less than half, will end up as Class 2A, that moderate biologic risk. And about 1 in 18 will have a Class 2B result with that high-risk of over 50% metastasis. And so moving on to Slide 26. When we actually look at univariate and multivariate analysis involving Class 2A and Class 2B gene expression profiling results, we see that overall, Class 2A and Class 2B performed beautifully in univariate/multivariate analysis compared to other common high-risk factors that we think about in squamous cell carcinoma. So kind of as a baseline, an SCC with deep invasion is about 2.1x more likely to metastasize than an SCC that does not have deep invasion. But when you add a Class 2A result to that, it really shifts to more like 4.8x more likely to metastasize. If you get a Class 2B result in a patient with one high-risk factor, it's going to shift that to 14.5x more likely to metastasize. So you can really see the power of that gene expression profiling and the tumor biology being added to classic clinical pathologic risk factors. So moving on to Slide 27. How do we envision using DecisionDx-SCC? So ultimately, again, we have a group of patients, SCC patients with one or more risk factors, which is about 200,000 patients a year. We utilize DecisionDx-SCC and add that to our traditional prognostic risk factors to ultimately come up with a low-, moderate- and high-intensity treatment plan. And then ultimately also, it's really going to, I think, help us with developing follow-up and surveillance plans. So for example, someone with a high-intensity or a Class IIb result, in addition to surgery, which is a baseline for pretty much all cutaneous squamous cell carcinomas that are operable, these patients may need nodal imaging and staging in addition to a multidisciplinary approach utilizing radiation oncology, medical oncology and even surgical oncology, depending on the patient and individual tumor. Surveillance plans also might be increased significantly. Whereas a typical low-risk squamous carcinoma might be seen every 6 months, a patient with a high-intensity or Class 2B result could be seen as frequently as 4 to 12 times a year, meaning monthly, for a period of up to 3 years. And they will often undergo baseline CT or nodal examinations and may do so typically every 3 to 6 months for 2 years. So this is a drastic difference in treatment plan as well as surveillance, depending on our DecisionDx-SCC tumor biology result. So moving on to Slide 28. I honestly am just so excited about the integration of DecisionDx-SCC into the clinical care of my patients. And I have to tell you, I called the company immediately when I heard it was being released and said, "I need the order form." And this is the reason why. Because when you really think about the way that we treat cutaneous squamous cell carcinoma and really all cancers in the United States today, it's a combination of the patient factors, the things that the patients were born with: their age, their underlying immunosuppression, other things that might put them at higher risk for developing squamous cell carcinoma. Then we take our typical clinical and pathologic tumor characteristics that are used commonly in the staging systems we use today. But the real power in my mind comes with that personalized precision medicine approach when we're able to integrate tumor biology into our daily decision-making. And at this point, I'll go ahead and pass it back to Derek for a summary of today's discussion.

Thank you, Dr. Wysong. I'm on Slide 30. So to preempt a couple of expected questions, which I think will be how does the launch look so far, which we may not answer, by the way. I thought it was worth sharing with you a little bit about what we did do with preparing up through August 31, 2020. First I, again, have nothing but praise for our employee group, our consulting groups, our industry partners in pulling off a well-coordinated training and launch preparation. We were far enough ahead of the game in training earlier this spring that we were actually able to successfully pivot and use some of the COVID related to stay-at-home time to accelerate disease state training in the second quarter. That made for a much more efficient training session at last week's virtual national sales meeting, where we completed training. And we were also able to successfully train a number of speakers who will help launch this program from a peer-to-peer opportunity beginning Monday, August 31 going forward when it was available. And finally, of course, we were able to develop and complete, I think, a very nice orchestrated effort from both website activity, slide decks for speakers and our sales force and our medical liaison group to use an order form and a report form that we believe will be self-explanatory to dermatologists and Mohs surgeons, who will be explaining the results of our test and the implications on our patient care going forward. So congratulations on a really, really Herculean effort from a team standpoint. So in closing, the Castle team, together with our patients and collaborators were successful in efficiently discovering and validating our DecisionDx-SCC test for use in patients diagnosed with invasive squamous cell carcinoma of the skin and one or more high-risk factors. We are launching into a market that we believe contains 200,000-plus patients per year and assuming achievement of an average solid tumor reimbursement level of $4,200 represents a U.S. TAM of approximately $820 million. But we've already developed and published evidence, 4 studies in the last 12 months. We are well on our way to building out additional evidence. And finally, we're able to launch this test to the very same clinicians who adopted our DecisionDx-Melanoma tests. Our hope will be that these same customers will already be oriented towards knowing the value of biology of their patient's tumor through a test provided by Castle Biosciences, with the end result being that patients will have more accurate, more appropriate treatment plans. Operator, we will now take questions.

[Operator Instructions] And our first question will come from the line of Sung Ji Nam from BTIG.

Congratulations on the launch. Just a couple of clarification questions for Dr. Wysong and then a follow-up. Dr. Wysong, was curious about how this test might be used in conjunction with the existing AJCC and Brigham staging system, if they will be used concurrently. And I'm just kind of curious as to how much do the clinicians currently rely on the AJCC and Brigham systems given the limitations there?

That is a fantastic question. And actually, if you read our manuscript, we did multivariate analyses comparing the Castle Class 2A and 2B results against both AJCC and Brigham and Women's staging. And so class -- the DecisionDx-SCC test outperformed both Brigham and Women's as well as AJCC and added an increase in positive predictive value overall, which is really, really exciting. That being said, I don't think in any way, shape or form that gene expression profiling is meant to replace staging. It's really meant to be in partnership and an adjuvant to clinical staging. And so really, clinical staging, again, is reliant on pathologic and clinical risk factors. But I think we've known as physician scientists for quite some time that the missing piece there is the tumor biology, just based on anecdotally what we see in our patients, but also just in terms of what makes sense for metastatic risk. So overall, we're very excited to be able to integrate the gene expression profiling tumor biology data with what we know about our current AJCC and Brigham and Women's staging.

Great. That's very helpful. And just another clarification question. In terms of the 20% of the squamous cell carcinoma patients being classified as high risk, how robust is the -- is that -- the criteria used, the NCCN and others? I'm just kind of curious if there is a significant cohort among the other 80% that end up metastasizing as well.

Yes. That's a great question. There are cases. Obviously, we know of patients that don't have one of the NCCN, AJCC or Brigham and Women's high-risk features but do go on to metastasize. And honestly, that's a study that I would -- that's a group that I would love to be able to study further. But that being said, 20% of 1 million, that's a pretty high number. We also don't want to get into the situation where we're overtreating patients that were otherwise low risk. But I think there is a subset of patients, I would say probably less than 5% overall in that 80% group that might go on to metastasize. And what we find, at least clinically, is those are often patients that have multiply recurrent disease based on persistent tumor that was left behind or patients that might have underlying immunosuppression or other patient-specific risk factors that could be impacting that as well. So tumor biology would be extremely helpful in those types of patients. And so those are -- that's a group that we would love to study further and figure out how to integrate this kind of testing.

Great. And then lastly from me, this is not related to the product launch. But kind of curious given -- in light of the pandemic, I'm kind of curious to hear your observations in terms of patient flow during the pandemic, maybe at the start of the pandemic, compared to during the summer months. And as you look out into the fall, do you anticipate -- I'm just kind of curious as to kind of -- what kind of trends you're seeing in terms of patient flow. And there are concerns about patients delaying their diagnosis and treatment in oncology in general. So kind of curious what your thoughts are there.

Yes. Great question. Absolutely. So really through the months of March and April across the country, obviously, we saw a delay of care with people appropriately being unwilling to come out during the pandemic and the high-risk times. In addition, what were considered elective surgeries were postponed during that time. Now we were still treating high-risk squamous cell patients in melanoma patients, what we would call medically necessary patients during that time. But nonetheless, I think, particularly through the months of May, June, July, we saw a pretty significant increase in volume of that backlog. I would say, at least in our -- throughout most of the -- what we're seeing in the Midwest and what I'm hearing from most of my colleagues across the rest of the country is most of us are back to about 90% of pre-COVID volume to 110%. We're a little over 100% of pre-COVID volume because we're still just a little bit backlogged. But I think what we've learned, and I will tell you, the University of Nebraska is kind of ground central for emerging and infectious diseases, and so we've really been on the front lines of this. As you may remember from the footage, we received the Diamond Princess crews, the very first COVID patients here in the U.S. So we've really been involved with this from every level. But what we're seeing is we've gotten a lot better at learning how to safely treat our patients. And particularly when it comes to skin cancer patients, the vast majority of those are done in the outpatient setting, under local anesthesia and very safely in the time of COVID. So I really don't see there being major shifts over the next 12 to 18 months in terms of delay of care, particularly when it comes to kind of more aggressive and high-risk squamous cell carcinoma. And I did just want to also follow up and answer your question. I pulled up the exact number, so you can have it. We did a multivariate Cox regression analysis of risk for metastasis, utilizing the 40 GEP validation cases and multivariate regression with AJCC as well as Brigham and Women's T stage. And so when you put AJCC 8 into the model with gene expression profiling, what we found is that the high-risk AJCC 8, which is considered to be T3 or T4 tumors, had a hazard ratio of 2.68 for developing metastasis compared to the Class 2B result, which had a 9.55 hazard ratio for metastasis. Very similar results that we saw when putting that multivariate model with Brigham and Women's staging, which looking at high-risk BWH staging, which would be considered T2B and T3 tumors. T2B and T3 tumors were -- had a hazard ratio of 2.03 for metastasis compared to 8.72 for a Class 2B DecisionDx-SCC results. So really, we're finding that the increased ability to predict metastasis utilizing tumor biology and the DecisionDx test.

Our next question comes from the line of Puneet Souda from SVB Leerink.

Derek and Dr. Wysong, so the first, let me cover a few questions, Derek, with you and then with Dr. Wysong. Just on the initial centers, how are you targeting? Are you specifically looking at certain Mohs surgery, academic centers first and then expanding broadly? Just walk us through how are you thinking geographically and the number of salespeople that are getting deployed in those geography.

Yes. So the inputs to our targeting efforts was a combination of existing customers and call points with our DecisionDx-Melanoma test. We were able to secure some third-party billing data that was able to identify relative volume, the physicians who use that billing software system for the diagnosis of SCC. So we overlaid that against our known data sets. And then we also looked at some Medicare claims largely gets cutaneous melanoma because SCC is a little more dicier in some of the day we think about of CMS and overlay that -- I think the field laps the virtual natural sales meeting last Friday, I think, with a solid list of people who we have data on were undoubtedly missing some out there. But they're going to go out -- they have been going out beginning Monday, in a targeted fashion to -- of course, first priority is to regain the sort of pre-COVID growth of a DecisionDx-Melanoma test and then introduce the DecisionDx-SCC test. I think it's natural that one goes to where clinicians have already reviewed cutaneous melanoma and thought about the value of incorporating tumor biologies Dr. Wysong said into what they already have and complement to help them make a more accurate and more informed decision. So my belief is that the early sort of adopters, if you want to call that, would be those who have already thought intellectually through and believe that adding the tumor biology data afforded to them through our DecisionDx-Melanoma test, that individual should be already oriented towards the value of getting more information to make a better decision in squamous cell.

Okay. That's helpful. If I could sort of -- if you can give us a view about the potential adoption and trajectory here for the assay, should we assume that this is going to be similar to cutaneous melanoma? Obviously, those were different time points. And your experience with gene expression profiling and adoption of and recognition of Castle in dermatology practices was different back in the day. So considering a different time point today, how should we think about adoption of the assay over the next couple of months into 2021?

Yes. So one, I want to give you no quantitative data you'll hold me to later. Second, of course, is that when we completed our initial validation for our DecisionDx-Melanoma test, I think we have 3 people who we would call sales representatives in the company. And now we have 32. So far different. And of course, we had -- nobody knew who the name was or what even a gene expression profile test meant in dermatology. So we had no name recognition, no targeting data and no awareness of kind of what goes on in breast cancer for the average kind of dermatologists out there. And Dr. Wysong probably is one of the first doors we actually knocked into in San Diego there so many years ago. But anyways, I anticipate that we will try and hold ourselves to reporting reports out the door as opposed to orders in. That would mean just having the test available here at the 31st of August, we'd be able to report some tests that are ordered in the next 2 weeks before the end of September, but most of them will end up being fourth quarter report. So I'm not sure if we'll adjust on a one-off basis for the third quarter. But my personal perception is that, without knowing impact of COVID on openness to a longer educational discussion with our field forces, is it going to take us all the way through the end of the year to really have most current customers be aware of the test? Or is it going to be compressed? I don't have an answer to that. So my -- our internal expectation here is to really kind of watch the sort of remainder of the year as a single unit as opposed to the third and fourth quarter. And I think by the time we close out the fourth quarter, we'll have a good early indication of what we've sort of seen in terms of numbers of doctors evaluating our test, i.e., ordering it once. And then I think the most important point from there, probably looking at our second quarter performance against the fourth quarter and that would at least show us some velocity growth numbers that should help us project what we look like in '22 and '23. So that's my current thinking about the data. So I know I avoided your question as best as I could here, but I really don't have a handle on what we'll be seeing. And I think that's dependent upon really having time in front of clinicians like Dr. Wysong and working through here's what we saw as an issue clinically. This is not the 80% of the people out there, it's the 20%. Here's the data that we developed. And here's how we believe your peers are thinking about employing biology with the clinical and pathologic features. And it'll -- that will either be a 1 or 2 sales call and educational effort or it's going to be 3 or 4. And so I would really discard anything that we see in the third quarter. Fourth quarter might give us equaling, but really I think it's fourth quarter compared to 2Q '21.

Okay. That's still very helpful, Derek. Dr. Wysong, on the current guidelines, just wanted to get a sense of if you look at the current Brigham and Women's guidelines, what are some of the blind spots and how well is the recognition that those blind spots exist in the current guidelines? I mean how well is that recognition among the Mohs surgeons and just the broader surgeons into the community setting? Any view there?

Yes. Great question. I would say, among Mohs surgeons, for sure, the awareness is very high. We all see these patients every day that maybe have one high-risk factor. And they're kind of on the edge, and we don't really know what to do with them. And what I always say is it's extremely hard to predict which one of these patients is going to be fine, go home. We just watch and wait or someone that's going to rapidly metastasize in 3 months. And yes, the factors that have come out over and over again for Brigham and Women's staging in AJCC include depth of invasion, perineural invasion, size greater than 2 centimeters. Those are kind of the key ones. Now -- and we know that those are independent predictors of metastasis. Those have been shown in the literature across the board. But when we actually -- and our new paper that's about ready to come out, when we actually compare individual risk factors, again, to tumor biology, tumor biology really outperforms any one of those individual risk factors when it comes to predicting metastasis. And it just makes sense, right? Because there are -- when we take one slice of a tumor, whether it is a size or the depth of invasion, there are all kinds of ways that might be misdiagnosed. So depth, for example, is really challenging to actually measure in squamous cell carcinoma. Number one, the vast majority of tumors are transected on biopsy. Number two, the vast majority of these tumors are treated by Mohs surgery, which allows for complete margin assessment, but does what's called on FAS or horizontal sectioning. So it's really challenging to get a final depth of invasion. So there are a lot of what you would call like blind spots to staging because we don't always get the final -- we can't always tell for sure whether or not they're the final depth of invasion. Perineural invasion is another extremely challenging one because unless a tumor is sent off for exhaustive vertical sectioning, we can often miss perineural invasion. And then there's lots of challenges about the size of the nerve, how many nerves are involved, where -- what the depth of the nerve is that's involved. And so it's not a black and white risk factor in and of itself. So I think across Mohs surgeons, to directly answer your question, I think we recognize that there are a lot of holes in traditional staging, and that comes up over and over again. I would say on the population that probably needs a little bit more education is going to be the general dermatology population. However, again, and/or the nondermatology population that's going to need to have a little bit more education around risk factors. But again, I would say the majority of SCCs, particularly in the head and neck area, are going to go to a Mohs surgeon or a surgical oncologist for treatment, primarily the Mohs surgeons. But further education of our entire colleague group is going to be important for sure.

Okay. That's very insightful. And then I'm wondering if you look at the cutaneous melanoma, AJCC existing guidelines that was fairly ingrained in, compared to that, how would you say are the current guidelines in squamous cell carcinoma, how well are they ingrained? And is it -- what is it going to take to -- in order to sort of push it over the chasm, if I was to say, in order for broader adoption into the marketplace among the dermatologists and across the board? Is it data? Is it -- is there more to this? Can you elaborate on that?

Great question. So I think there's been a lot of movement around high-risk squamous cell carcinoma in the last 5 years alone. And I think there's increasingly -- or just an acknowledgment of the fact that we have not done a very good job at assessing these typical clinical risk -- clinical or pathologic risk factors. So CAP is a group that comes out for the National American College of Pathologists. And they are now recommending actually doing the typical measurements, Breslow's depth for squamous cell carcinoma, looking for perineural invasion. The latest NCCN guidelines -- I'm sorry, yes, the latest NCCN guidelines also look -- when you look at pathology for squamous cell are recommending that. Now the adoption of that, I would say, is a different thing. I would say most major academic medical centers at this point are doing complete staging for squamous cell carcinomas, particularly if there's any risk factor of any kind. I would say, larger adoption around staging for squamous cell carcinoma is definitely behind melanoma nationally, but I think we're getting there, both in terms of just ongoing education of colleagues as well as just more and more data coming out about the importance of staging squamous cell carcinoma. And I really do think that DecisionDx coming out is really going to help us, again, look at -- well now we've got a prognostic method that can really help us do something about it. So it's even more reason to be staging, to be thinking about these single -- identifying these risk factors so that we can ultimately go on and better predict these patients that are going to be developing metastasis and ultimately poor outcomes.