Response to Comments: MolDX: Pigmented Lesion Assay

DermTech, Inc. appreciates the opportunity to comment on the draft Local Coverage Determination (LCD) MolDX: Pigmented Lesion Assay (DL38151) to be harmonized with MolDX Palmetto’s DL38051 as we understand from your proposed LCD document and from the recent meeting in San Diego where we had the opportunity to present. We are writing to express our support for the positive draft LCD.

We offer comments 1-4 below also recently provided to MolDX Palmetto within their comment period (ended on June 20, 2019). We attach a revised version of draft LCD DL38051 which is identical in the test specific sections to Noridian’s DL38151. We did not create a second revised version of the draft LCD to facilitate review and harmonization efforts. Proposed changes correlate to the comments made. We highlighted the proposed changes in the draft LCD for your convenience and also attach a most recently published manuscript with Pigmented Lesion Assay (PLA) 12-month follow-up and registry study data.

Comment 1: Long-term outcomes data now available. The longer-term outcomes data mentioned under ‘Coverage Criteria’ are now published in Ferris et al., DOJ 25(5), May 2019. We added information from two studies described in this manuscript highlighted in yellow to Table 1 and to the appropriate ‘Summary of Evidence’ sections. Most notably, there were no missed melanomas in the 12-month follow-up utility study of 734 PLA negative cases. The most recently published manuscript also includes registry data from 1575 PLA cases corroborating that clinicians (62 providers located throughout the US and including board certified dermatologists, primary care physicians, physician assistants and nurse practitioners) followed the guidance of the test (99.9% of PLA negative lesions were clinically monitored and not biopsied; 96.5% of PLA positive lesions were appropriately biopsied).

Comment 2: Coverage criteria / language. We propose to employ the language currently used and underlined under ‘Specific Coverage Criteria’ (…‘The Pigmented Lesion Assay is indicated for use on melanocytic skin lesions with one or more clinical or historical characteristics suggestive of melanoma … when a clinician trained in the clinical diagnosis of skin cancer is considering the need for biopsy to rule out melanoma’) throughout the LCD. This language is supported by the largest data set available to date (please see Comment 1 above). Sixty-two US providers (board certified dermatologists, primary care physicians, physician assistants and nurse practitioners) generated this data set.
Most recent analyses suggest that there is no difference the percentage of PLA positive and PLA negative cases between dermatologists and other clinicians involved in pigmented lesions management in this large real-world US registry setting (there are 9% PLA positive lesions in cases from either group). These findings confirm that the rate of concordance between test finding and management reported was consistent with each clinician group. Dermatologists and other providers trained in the clinical diagnosis of skin cancer use the PLA in the same conservative fashion and as intended. Additional evidence furthermore demonstrates that capturing suspicious pigmented lesions during primary care visits is important to finding melanomas early. Higher densities of primary care physicians may be linked to increased diagnosis of early-stage melanoma (Flemming et al., PLOS ONE, 2018; https://doi.org/10.1371/journal.pone.0200097).

Comment 3: Second evaluation of a given lesion by PLA. The draft LCD states that ‘The test may not be ordered for the same lesion a second time.’ We propose a change to: ‘The test may not be ordered for the same lesion a second time unless lesion changes require another evaluation.’ Such a second evaluation, while rare, is nevertheless a desirable option for clinicians to have available. Data sets from an additional ongoing study indicate that only 1.4% of over 1,000 PLA negative lesions were subjected to additional PLA testing due to clinical changes within a 2-year follow-up period. In less than 0.3% did the PLA status change, which also triggered a change in management that included surgical biopsies and closer monitoring. These findings demonstrate that rare additional PLA testing on changing lesions can pick up new molecular risk factor findings to further guide pigmented lesion management.

Comment 4: Number of tests per date of service. The draft LCD currently states that ‘only one test may be used per patient per clinical encounter’. We propose to change this language to: ‘Only up to four tests may be used per patient per clinical encounter’. This conservative proposal attempts to address potential concerns of over-utilization we do not see in our data. Only about 13% of PLA tests are used on more than one lesion (please see Table 1 below for details), while multiple surgical biopsies on the same date of service are seen in 42% of cases based on 2016 Medicare Physician Utilization Databases. Our data sets do suggest that clinicians are faced with situations where they have more than one suspicious lesion presented occasionally, and they believe the test is medically necessary for more than one lesion in such situations. Insofar as we get samples and requests for multiple lesions on a single date of service, it is our policy to process and report those samples. While the number is small, it would be financially burdensome for patients if these additional tests were considered non-covered, and we were required to obtain an Advance Beneficiary Notice in order to be made financially whole for performing the tests (it can furthermore raise compliance issues if we were to give away additional tests to patients while at the same time billing the first test to Medicare). It may be appropriate to establish a limit consistent with or close to current practice for biopsies that enables for instance one CPT 11102 biopsy plus up to 6 more CPT 11103 biopsies (based on most recent 2019 Practitioner Services MUE Values; spreadsheet within https://www.cms.gov/Medicare/Coding/NationalCorrectCodInitEd/MUE.html ).

This seems especially appropriate given the findings that the PLA test can improve diagnostic performance and save burdensome procedures. Table 1 below demonstrates that covering up to 4 PLA tests per date of service (3 less than for e.g. shave biopsies under CPT 11102 and 11103) addresses the needs of over 99% of patients benefiting from PLA testing.

Table1. Percentage of patients with 1-6 PLA tests per date of service.
We commend the Contractor Medicare Directors and the MolDX program for their thoughtful review of the Pigmented Lesion Assay and for providing clinicians with access to an objective and highly performing non-invasive test to help rule out primary cutaneous melanoma and guide biopsy decisions of melanocytic skin lesions suggestive of melanoma. We reiterate the importance of the proposed coverage for Medicare beneficiaries.

We will respond to each comment using the format 1.x for clarity in numbering.

1.1 Thank you for informing us of the new publication. We will update the LCD to reflect the publication of new evidence.

1.2 At present, the performance of this test suggests that it operates best as a rule out test for melanoma. As such, we will not expand coverage of the test at this point to better identify melanoma by general primary care providers. We do not require that providers be board-certified in dermatology, though we will require that providers have sufficient expertise that they consider themselves to be dermatologists or dermatopathologists.

1.3 We do not believe that there is sufficient data at this time to cover testing on a repeat lesion. The values given in the comment, 1.4% of over 1,000, suggest a total of approximately 14 cases. Additional evidence on the generalizability and reproducibility of outcomes in repeat testing would be required to support that repeat testing is reasonable and necessary.

1.4 We will cover up to 2 PLA tests per date of service without an appeal. Given the data provided above, this coverage allowance appears to cover what is ordered for nearly all patients. We will defer to the laboratory as to whether to bill patients for the small number of services that will not be covered or to appeal them.

I am a board certified dermatologist, US Navy residency trained, in private practice for 40 years. I have examined nearly half a million patients in my practice, many of whom have several pigmented lesions to be evaluated. Each lesion is evaluated visually for abnormalities which might suggest histopathologic dysplasia. The abnormal lesions have traditionally been surgically biopsied and then examined microscopically for evidence of dysplasia or malignancy.

My experience has been that different excellent histopathologists will often differ in their opinions about questionable lesions. Some will say normal while others will say dysplastic. To err on the safe side, I will routinely re-excise the lesion if any histopathologist thinks the lesion is dysplastic. My experience is that about 25% of biopsied lesions are.called dysplastic by one or more histopathologists. In other words about 75% of biopsied lesions are called normal.

The benefit of the pigmented lesions Assay is that it will identify lesions read as normal by histopathology, that actually have genetic markers of dysplasia or frank malignancy. This will allow me to further excise premalignant and malignant lesions that would otherwise not be properly treated. There will also be lesions that demonstrate no genetic markers that visually and histologically are morphologically premalignant or malignant. These lesions would therefore not need to be re-excised. Without the Pigmented Lesion Assay, these lesions would routinely be re-excised surgically based only on visual evaluation clinically and histopathologically. This is a great deal of unnecessary surgery often resulting in multiple scars on patients; especially young men and women.

Thank you for the comment. We appreciate the input from a highly experienced practicing dermatologist and receiving input regarding how the test may be realistically used and how that use may meaningfully affect patients.

I wholeheartedly agree with your recent positive draft LCD (DL38051) on Dermtech’s pigmented lesion assay (PLA) and would like to take the opportunity to share my rationale and experience for doing so.

Having used the PLA for about 3 years now, the test offers clear advantages to patients and clinicians alike in guiding biopsy decisions to rule out melanoma and reducing the chance of missing a melanoma from about 17% to less than 1% based on NPV comparisons between the PLA and the gold standard of dermatopathology (Gerami et al., JAAD 2017; Elmore et al., BMJ 2017; Lott et al., JAMA Dermatol. 2017; Malvehy et al., BJD 2017).

If I have a high suspicion of Melanoma, I go directly to standard biopsy. I perform PLA when I have a mild or moderate suspicion. To date I have performed over 400 PLA assays. These have yielded 5 melanomas, 5 melanoma-in-situ, and 3 atypical pigmented lesions with excision recommended. These 13 lesions may have been put on my “photograph and follow” list and additional time would likely have passed before discovery. How humbling!! In these 13 cases, I am profoundly grateful the PLA helped me make a rapid diagnosis. In the other 387+ cases, I’m grateful I didn’t leave 387+ scars. I’d like to share the following two striking cases with you as additional illustration.

Patient 1 was a 101-year-old gentleman with a clinically somewhat atypical melanocytic lesion on his back. Based on the patient’s age, living conditions, difficulties to leave his home setting and preference to avoid a surgical procedure due to known wound healing issues, I offered the PLA as a non-invasive way to obtain additional information on this lesion and hopefully avoid the need for a surgical biopsy. The PLA gene expression test detected both LINC and PRAME now justifying a shave biopsy that established the diagnosis of inflamed melanoma in situ.

Patient 2, at the other end of the age spectrum was a young female who was concerned about a pigmented lesion on her neck. Given the clinical appearance of this lesion, I am not sure I would have supported a surgical biopsy. I considered photo documentation and monitoring but ended up offering the PLA. The again double positive lesion (LINC and PRAME detected) was a 1.1mm superficial spreading melanoma histopathologically.

There is no doubt in my mind (wearing either my clinical or my dermatopathology hat) that the PLA has a special and meaningful place in modern dermatology, allows me to offer superior care and saves lives. As a board certified dermatopathologist I particularly appreciate molecular risk factor guidance on changes that may not have a morphological correlate yet.

Thank you for the comment. We appreciate the input from a practicing dermatologist who has experience with the test, and we appreciate the input regarding how the test gets realistically used in practice, including both where it is helpful and where biopsy remains more appropriate.

The management of atypical pigmented lesions typically involves ruling out melanoma through visual assessment, with or without the addition of dermoscopy, followed by a surgical biopsy and histopathological slide review. The non-invasive Pigmented Lesion Assay is a significant and useful addition to the current toolbox, because its NPV of 99% allows physicians to potentially biopsy fewer lesions and as well as miss fewer melanomas. The Pigmented Lesion Assay is a technology that has been validated in multiple peer-reviewed papers for non-invasive genomic testing of the skin.

Integrating this technology into pigmented lesion diagnosis also has the potential to lead to a material reduction in cost while improving pigmented lesion management. The use of the test appears to also be beneficial in patients who experience clinical changes to a previously assessed pigmented lesion. While over-use of a new technology is always a theoretical concern, you could consider coordinating the number of questionable lesions the test can be used on with the use of surgical biopsies in such lesions (one CPT 11102 plus some additional number of CPT 11103 procedures for ‘shave biopsy’ codes utilized).

I hope that you will consider coverage for this test so that our patients can benefit from its ability to enhance clinical diagnosis.

Thank you for the comment. We appreciate the input discussing how the test relates not only to biopsy-based diagnosis, but also to existing clinical skin assessment. We have addressed the issue of testing more than one lesion at a time in comment 1 above.

My experience as originator and director of the Melanoma Center, and later as Director Emeritus, at the Washington Cancer Institute, Washington Hospital Center in Washington DC from 1996 to 2015 includes accruing approximately 2000 melanoma patients. My average patient had 100 pigmented nevi. Many patients had over 200 nevi; for example, one had 1300 melanocytic lesions. We found that upon long-term monitoring of this high- risk group of patients with multiple nevi that about 2/3 of melanomas arose de novo from previously uninvolved skin and only about 1/3 arose from pre-existing nevi. We concluded that it does not make good medical sense to remove large numbers of benign nevi in order to attempt to decrease the risk of melanoma formation. Many of our patients at the Melanoma Center had a history of multiple biopsies of benign nevi, including those designated as mildly or moderately atypical, that left scars that the patients objected to. Patients also objected to the pain and the healing times of the surgery, as well as the anxiety-producing wait to hear the results of the pathology report. Patients also objected to the re-excisions that some physicians routinely performed if the pathology report indicated that the surgical margins were positive of these benign lesions. In our clinic, we had two patients with histories of 500 biopsies of benign lesions performed in previous medical offices. Since 1996, we have relied upon dermoscopic evaluation of our patients' nevi to determine which lesions to biopsy. The routine use of dermoscopy has significantly increased our diagnostic ability which has, as a consequence, reduced the number of unnecessary biopsies. With this experience we welcome any new technique that will further improve our diagnostic abilities, thus improving our specificity and reducing the incidence of unnecessary biopsies.

I continue to care for many of these high-risk patients with large numbers of nevi at the Dermatologic Surgery Center of Washington, an institution I joined in July, 2015. For more than 3 years, we have performed routine PLA testing of clinically and dermoscopically atypical lesions to improve our clinical performance. As such, we are able to diagnose smaller melanomas at an earlier stage. We also can confirm that with PLA testing, clinically, somewhat atypical, lesions are indeed benign, thereby avoiding unnecessary surgical procedures. Almost all of the melanomas we have diagnosed since I added the Pigmented Lesion Assay to my diagnostic routine have been small (less than 6mm) melanomas in situ. Only about 10% of all the pigmented lesions that I have tested were positive by PLA. Thus, 90% of my patients have been saved from an unnecessary surgical biopsy.

It makes sense to me that gene expression changes occur prior to the development of a histopathologically diagnosable melanoma. Therefore, it is likely that PLA positive lesions that are diagnosed on biopsy as being severely atypical are not false positives, but may represent pre-melanomas prior to their evolution into pathologically typical melanoma lesions. It is my opinion that the PLA test is sensitive enough to capture early lesions such as melanoma in situ and what I refer to as pre-melanomas while at the same dramatically reducing the biopsy rates of benign nevi.

Regarding the issue of false negative lesions, part of the problem could be the difficulty in the differential diagnosis between melanoma in situ and atypical nevi by pathologists. For example, pathologists may have different thresholds for the number of cells exhibiting pagetoid spread leading to their variation in making a diagnosis of MIS rather than an atypical nevus. Similarly, variations in evaluation of other pathologic criteria, such as cytologic dysplasia, may lead to different diagnoses being made.

In the situation in which the PLA test provides a negative result and the lesion is not biopsied, the patients in my clinic are then monitored on a regular schedule and are also informed to return to clinic immediately if any change occurs in the lesion prior to their next scheduled visit. At each follow-up examination, we always focus on lesions that we have previously tested with the PLA and that were reported to be negative. An additional clinical point is that it is more effective to monitor previously PLA –negative intact, non- biopsied lesions than to monitor partially-excised benign lesions that have been biopsied at other centers.

We have found that PLA testing is sensitive, reliable, and is a useful adjunct to dermoscopy enabling earlier diagnoses of melanomas and, most importantly, resulting in fewer missed melanomas. In my opinion, this test can be an asset to any clinician involved in the diagnosis of skin cancer, not just experts in dermoscopy. I sincerely hope and recommend that Medicare patients will not be excluded from the potentially lifesaving benefits of PLA testing that can lead to earlier diagnosis of melanoma and at the same time saves money by reducing unnecessary biopsies of benign lesions.

These benefits should also not be limited by restricting the use of the test to one lesion per date of service or to once per lesion. We found it very helpful to re-test PLA negative lesions that change their clinical appearance. The vast majority of PLA tested lesions I monitor did not change within 12 months. When a set of clinically changing, previously negative lesions were re-tested with PLA, about 80% of them remained negative. However, I have had a small sub-set (3 lesions out of hundreds tested in my office), that progressed at the molecular level by acquiring detectable levels of the PLA target genes. I conclude that the ability to re-test is rarely needed, but nevertheless is important to have available.

Thank you for the detailed comment. We appreciate the input from a highly experienced practicing dermatologist, especially a description of how this test may be used in a population who could particularly benefit from better diagnostic tools. At present, we are unaware of adequate evidence to support the use of repeat testing, though testing more than one lesion may be possible as described more fully in the response to comment 1.

I am triple board certified in anatomic and clinical pathology in addition to dermatopathology. I evaluate pigmented skin lesions routinely and know from personal experience that assessing pigmented lesions and ruling out melanoma can be frustrating due to the lack of well-defined histologic criteria, philosophical differences between institutions (e.g., those that believe in dysplasia and those that do not), and apparent differences in personal thresholds on when to call a lesion melanoma leading to disagreement. These issues remain, for the most part, invisible to the patient and ordering clinician as they view pathology as the gold standard offering the absolute truth. Frequent disagreement, however, was shown in recent studies (Elmore et al., BMJ, 2017 and Lott et al., JAMA Dermatol. 2018) where board certified dermatopathologists showed significant variability in the interpretation of melanocytic lesions in a large number of sample cases provided in the studies. Furthermore, the study showed significant intra-observervariability where an individual pathologist disagreed with themselves when shown the same sample case at a different time point.

New and better ways to assess pigmented lesions are clearly needed and gene expression tests like the Pigmented Lesion Assay (PLA) can be of great value not only to the clinician, but also to the dermatopathologist. While I believe this view is shared by a growing number of dermatology colleagues who recognize the value of the PLA (e.g., resulting in a dramatic decreased need for disfiguring biopsies and excisions), I know that this view is not shared by all in the dermatopathology community which may include reasons such as conflicts of interest and desires to protect established revenue streams. The PLA offers a consistent approach to the evaluation of melanocytic lesions, despite where a dermatologist or dermatopathologist has trained and their own personal biases, and I hope to see the PLA benefitting Medicare patients soon.

Thank you for the comment. It is informative to received input from a physician who has firm understanding about how melanomas are currently diagnosed, and who is able to systematically address the limitations of the current diagnostic standard. We are reassured to receive a comment noting agreement with our coverage decision.

I am currently a private practice dermatologist and dermatopathologist and have been in practice since 1980. Over my course of clinical practice and experience, the identification of atypical pigmented lesions is one of the primary obligations and duties in an active dermatology practice.

Over the years, thousands of pigmented lesions have been biopsied by me to determine their benign or malignant potential. Obviously, most of these lesions (up to 90%) are fortunately benign. However, to ascertain their potential biologic behavior, one is required to perform a surgical procedure to obtain tissue and then submit this for histopathologic review.

With the advent of the DermTech Pigmented Lesion Assay (PLA), the need to surgically biopsy almost all of these atypically appearing nevi has finally come to an end in my opinion. The sensitivity and specificity of the PLA has been proven and precludes avoidable and time consuming surgical procedures of biopsying many atypical pigmented lesions that don't need it.

During the past decade, numerous special immunohistochemical stains have been used to help identify and categorize atypical nevi; however, these immunohistochemical stains are expensive and often only minimally helpful for determining a lesion's biologic potential. With the DermTech gene expression test, a more accurate molecular signature can be utilized.

The cost of surgical biopsies as well as the use of expanded immunostains for interpretation also have become quite exorbitant. Increasingly, indeterminate lesions that have not been previously tested with the DermTech method have undergone genomic testing with tissue obtained from the blocks and these types of tests may run into thousands of dollars (very different from DermTech's hundreds of dollars).

With the use of the PLA test, gene expression markers can be very simply, quickly, and effectively utilized to rule out melanoma. The overall cost of PLA testing with regard to biopsy and subsequent histopathology examination is significantly reduced and it is just as or even more accurate compared to histopathologic examination. This type of test will also avoid numerous unnecessary scars.

In summary, I feel that the use of PLA testing by clinicians trained in the clinical diagnosis of skin cancer should be available to patients with all insurances, including Medicare. It avoids unnecessary biopsies and enables better diagnostic interpretation. The PLA is an objective molecular test that shows significant specificity and sensitivity for the examination of pigmented lesions. I feel that this is a true breakthrough which is clearly more accurate than conventional microscopy.

Thank you for the comment. We appreciate receiving input from experienced clinicians, especially those with expertise in both the clinical and the pathological diagnosis of melanoma. We are reassured to receive input agreeing with our coverage decision.

Using the PLA to rule out melanoma by guiding biopsy decisions with molecular risk factors is a useful tool that can save lives and cut down on unnecessary biopsies. I find the non-invasive test useful, because its negative predictive value of over 99% allows me to miss fewer melanomas or find them earlier than histopathology, which is associated with a NPV in the low 80% range. PLA positive tests (roughly 10% of all cases tested by PLA) are biopsied whereas PLA negative tests (about 90%) are followed clinically and not biopsied further supporting the clinical utility of the test to a variety of clinicians (dermatologists, primary care physicians and physician assistants) involved in skin cancer management. All these groups were part of an impressive registry study published most recently (Ferris et al., DOJ, May 2019). Having had quite a bit coding experience and exposure, I feel that it may make sense to not limit the PLA use to one lesion per date of service but instead harmonize it to what we recently implemented for biopsies (e.g. one CPT 11102 biopsy plus up to 6 more CPT 11103 biopsies based on 2019 Practitioner Services MUE Values). Better care at lower cost should not be limited to 1 lesion. I sincerely hope Medicare patients will benefit from this test in the near future.

Thank you for the comment. We will increase the number of PLA tests allowed as discussed in response to comment 1.

I am writing in response to your draft LCD DL38051 on Dermtech’s PLA assay. The PLA helps with biopsy decisions on skin lesions clinically suspicious of melanoma and can reduce biopsy rates by about 88%. One benchmark often raised in debates on avoidable biopsies and over diagnosis is the number needed to biopsy (NNB) to detect a melanoma. Our own recent work (Anderson et al., JAMA Dermatology, published online April 18, 2018;) demonstrated that the number needed to biopsy per diagnosed melanoma was 39 for physician assistants and 25 for dermatologists (p=0.007). The ability to reduce the number of biopsies performed on benign lesions is expected to dramatically improve patient outcome. The Dermtech PLA will allow us to do this.

Thank you for the comment and the discussion of relatively new evidence. It appears that this is in agreement with our coverage policy.

I am regularly impressed by the performance of Dermtech’s Pigmented Lesion Assay, a non-invasive gene expression rule-out test I have been using in my practice now for over two years. I credit Dermtech’s Pigmented Lesion Assay with saving the life of one of my patients. Because I could offer this patient a non-invasive assessment for her melanoma of unusual appearance, her melanoma was caught just before reaching the depth linked to a much worse prognosis. Also and overall, the cost of this patient’s healthcare was greatly reduced. If her biopsy had been delayed even a few months, lymph node involvement and possible distant organ metastasis would have been a concern which would have added hundreds of thousands of dollars to the cost and greatly increased morbidity.

The test’s sensitivity of 91% (Gerami and team, Journal of the American Academy of Dermatology 2017; 76:114-120.e2) compares favorably to the performance of pathology in a fashion that may not be appreciated by many of our colleagues or expected by our patients. After all, we chose the pathologists we work with carefully, try to only work with the best and not infrequently have years of history together. However, recent publications on the true performance of dermatopathology were eye opening to me. For these reasons, I am very grateful to offer my patients access to a non-invasive test that demonstrated a sensitivity of 91% and a negative predictive value above 99%. Our patients deserve Medicare coverage of this non-invasive test that allows clinicians to biopsy less while missing fewer melanomas. The test should be available for more than one lesion per date of service (like surgical biopsies).

Thank you for the comment, specifically for the perspective of how this test compares with existing standards of care.