Summary Of Evidence

Background

Invasive and in situ cutaneous melanoma is a type of skin cancer that is diagnosed in over 178,000 patients annually in the United States. Over 9,300 people die from cutaneous melanoma in the US per year.¹ Detecting melanomas at their earliest stages (melanoma in situ (MIS) / Stage 1) impacts disease outcome and patient survival. The 5-year relative survival rate from diagnosis for localized, early melanoma is over 98%, but less than 20% for melanoma that has spread to distant sites.² The generally well accepted approach to assessing pigmented lesions includes visual inspection followed by surgical biopsy and histopathologic analysis of the biopsied tissue.^3-12 One large study assessing dermatologists’ biopsy decisions using existing decision making tools is approximately 25.³ In another study, roughly 24 biopsies were needed to diagnose 1 invasive melanoma, and roughly 12 biopsies were needed to detect either invasive melanomas or MIS.¹³ In summary, this approach results in many biopsies that do not lead to a melanoma diagnosis. Guidelines from the American Academy of Dermatology recommend that a prebiopsy photograph be taken to help with clinical / pathologic correlation.¹⁴

Additionally, the diagnostic yield of early stage melanoma on biopsied tissue is limited. Histopathologic assessment of early stage biopsied melanoma tissue is challenging and has significant discordance between pathologists.^15-17 It also appears that underinterpretation is more common than over-interpretation of a patient who has had a biopsy,^16,17 which is tantamount to missed diagnoses. Additionally, while fellowship-trained or board certified dermatopathologists tend to have a higher accuracy than other pathologists, even among this group, under-interpretation is highly prevalent.¹⁶

In summary, conventional melanoma care may lead to both biopsies of non-malignant lesions, and even in those patients who do have a biopsy the diagnosis of a malignancy may be missed. As such, there is potential clinical utility for a test that can either spare a patient the need for a biopsy.

The Pigmented Lesion Assay (PLA) was developed to address fill the niche of reducing the biopsy rate of non-malignant lesions.

PLA Test Description

It is a gene expression test using samples collected via adhesive patches providing a non-invasive alternative to the surgical biopsy pathway in the assessment of pigmented skin lesions.^18-25 The test is positive if LINC00518 and/or PRAME (2 genes known to be overexpressed in melanoma) are detected.^29-30 The PLA is based on a platform technology for non-invasive genomic testing of the skin that allows the analysis of samples collected with an adhesive patch.²⁴ Four patches are placed on a lesion. For each patch, the margin of the lesion is outlined by the clinician. This outlined tissue is dissected away from the surrounding tissue by the processing laboratory, and RNA is extracted only from the lesional tissue. In contrast to histopathologic sectioning, the adhesive patch method of tissue sampling allows the collection of tissue from the entire the lesion in the plane of the skin surface. Further, genomic information obtained by adhesive patch sampling of the stratum corneum contains information from deeper epidermal cells.

Bioplausibility

The PLA has been validated against hotspot driver mutations in melanoma (e.g., BRAF other than V600E, NRAS, and the TERT promoter) that are associated with disease progression and histopathologic findings, such as mitotic counts and ulceration.^26-33

Analytical and Clinical Validation

The analytical and clinical performance of the PLA is supported by multiple investigational studies as discussed below.

As noted above, trained pathologists may disagree over their assessments of pigmented skin lesions, and comparisons of pathologists opinions to each other or to a consensus group have been used as reference standard in studies. The clinical validity for the PLA has been assessed both using pathologist opinion and longitudinal patient outcomes.

Following early work identifying that the expression of LINC and PRAME can accurately classify pigmented lesions using a simple 2-gene detection methodology, a classifier method based on these 2 genes was used in an independent test set.²⁰

The performance metrics of the PLA were validated by Gerami et al. against consensus panel histopathologic assessment clearly demonstrating the test’s clinical validity in the assessment of early stage pigmented lesions.²⁰ In this study, samples were collected prospectively from multiple dermatology practices and centers, in patients 18 years of age or older, and from pigmented lesions that were suspicious for melanoma, meeting 1 or more ABCDE criteria. Clinically obvious or frank melanomas were excluded. Lesions were simultaneously sampled using the adhesive patch and surgically biopsied. Biopsy specimens underwent pathologic diagnosis from 3 independent dermatopathologists, and lesions that received a concordant diagnosis from all 3 dermatopathologists were enrolled in the study. Overall, 11% of lesions sampled had a discordant pathological read and were excluded, creating a reference set upon which there was diagnostic agreement among pathologists. A blinded evaluation of these concordant biopsy samples was performed against the PLA result. An initial training set of 157 lesions was tested and demonstrated a 91% sensitivity, 53% specificity. An independent validation set was subsequently studied that included 398 pigmented lesion samples (87 melanomas, 253 atypical pigmented lesions, 53 non-melanocytic lesions). All melanomas enrolled in the study were classified as very early stage and were either MIS or Stage 1 with a median Breslow thickness <0.5 mm. The PLA demonstrated a sensitivity of 91% and a specificity of 69%.

A separate study assessing the clinical performance of the PLA in patients who received longitudinal follow-up was also done. This study included available outcomes and clinical management decisions for PLA- and PLA+ cases at 4 US dermatology practices using the PLA commercially.¹⁸ Cases were reviewed with a minimum of 6 months to 9 months follow-up, with 273 samples of this ongoing effort having 12 months follow up. Serial dermatoscopy studies indicate that melanomas have detectable visual changes within 3 months and recommended surveillance guidelines are 3-6 months.³⁴ For the 381 lesions evaluated in this study, the sensitivity was 95% and the specificity was 91%. While the sensitivity in this study is similar to that found in the histopathologic validation²⁰, the specificity is higher.

Table 1 below summarizes the publications regarding analytical and clinical validation.

Summary of Performance and Utility (Table 1)

Clinical Utility

A review of over 20,000 commercial PLA results indicated that 88% of reported PLA tests were negative and 12% were positive²². This combined with the finding in a 2017 study of 18,715 surgical biopsies of pigmented lesions showing that 83% of the lesions biopsied were either benign or mildly atypical lesions¹³, suggests that if the test has sufficient clinical performance to rule out melanoma (i.e., adequate sensitivity and negative predictive value), and treating clinicians use the test results as intended, it should result in significantly fewer unnecessary biopsies without compromising melanoma outcomes. Clinical performance is reviewed above. Here clinical decision making following the use of the test is reviewed.

In the Ferris longitudinal follow-up study mentioned above,¹⁸ 99% of the 330 PLA negative lesions were managed by dermatologists with surveillance. Three of the PLA- lesions that were biopsied in the follow-up period were done so at the patient’s insistence. One PLA- lesion was simultaneously surgically biopsied (not the intended use of the test) and adhesive patch sampled and was diagnosed as MIS. There were 0 missed melanomas found in the follow-up period. Of 51 PLA+ test results, 100% were managed by dermatologists with a surgical biopsy. Nineteen (37%) of these cases were MIS / Stage 1 invasive melanomas with a thickness of <0.5 mm and demonstrating a Number needed to biopsy (NNB) of 2.7 (51/19).

In an additional utility study by Ferris et al., 45 dermatologists who regularly evaluate pigmented lesions, assessed 60 cases containing dermatoscopic and lesional images (8 melanoma and 52 nevi with known pathologic concordance) with full patient and lesion history.¹⁸ The photographic/dermatoscopic analysis design of this study provided information nearly identical to the dermatologist’s primary clinical visual assessment used to make biopsy decisions and is therefore more relevant than typical decision impact studies that involve select case information review with and without a test result. Cases/images were initially presented without PLA results, and the dermatologists were asked to make a biopsy decision for suspicion of melanoma. The 60 cases were then shuffled and presented again, including the PLA test data. Again, dermatologists were asked to make a biopsy decision for suspicion of melanoma. Outcomes included changes in biopsy decisions, specificity, and sensitivity. Biopsy decisions increased from 750 to 1331. Assuming correctness of the reference diagnosis, the specificity of the biopsy decision increased by 1.8-fold with the PLA (32%-56%, p<0.001). The sensitivity also improved to approximately 99% (p=0.01) with the PLA, even with significant increases in specificity.

Most recently, a 2019 study by Ferris reviewed 12 month management decisions and outcomes for patients testing using the PLA.³⁵ The study involved retrospective chart reviews of 734 lesions that were PLA(-) and a registry of 175 pigmented lesions tested using the PLA. Among the 734 PLA(-) lesions, 13 were biopsied within one year. Of these 13 biopsied lesions, 11 were nevi with various degrees of atypia, 1 was a basal cell carcinoma and 1 was a squamous cell carcinoma. None were melanomas. In the registry cohort, 1433 of 1575 total lesions were PLA(-), and in follow-up only 2 had a surgical evaluation within a year. One of these had a scoop excision and was found to be a melanocytic nevus. The other was a squamous cell carcinoma removed by Mohs surgery. Of the 142 PLA(+) lesions in the registry cohort 96.5% were biopsied.

Analysis of Evidence

Level of Evidence

Quality of evidence – Moderate
Strength of evidence – Moderate
Weight of evidence – Low

The diagnosis of pigmented skin lesions is challenging both for dermatologists examining a lesion in vivo in a clinical setting and for pathologists examining tissue specimens following biopsy. As such, pigmented skin lesions may be biopsied in many cases where the biopsy did not need to be done. Additionally, once a skin lesion has been biopsied, diagnostic concordance among pathologists for a given lesion is limited.

This test is not actually a replacement for histopathological examination, but rather a tool to help clinicians decide if a patient requires histopathologic examination. As such, evidence reviewed regarding alternative diagnostic approaches was reviewed for both clinical examination of a lesion prior to a biopsy and for histopathologic examination post-biopsy. The diagnostic accuracy of a melanoma for a patient is driven by both of these things; however, we did not find adequate literature examining the joint probability of correct diagnosis based on both clinical examination and histopathology to allow us to use a comparison for the PLA. 

Given the existing evidence for the PLA and alternative diagnostic approaches, the evidence is sufficient to indicate that the pigmented lesion assay provides adequate sensitivity and negative predictive value for malignant pigment lesions to be used as a clinical decision tool in select pigmented lesions where there is a question as to whether or not a biopsy is needed.

The area of pigmented skin lesion diagnosis is rapidly changing with new developments to better aid in diagnosis, and standards of care may change, and better outcomes may be definitively demonstrated as additional diagnostic evidence emerges. Additionally, the field of melanoma diagnosis currently has a paucity of studies comparing diagnostic methods with long term (i.e., >5 years) outcomes. As such, MolDX will continue to monitor the evidence for not only the PLA test, but also for diagnostic performance or outcomes of alternative diagnosis and management approaches. Such developments may have an impact on this coverage policy.

Additionally, while this coverage decision is presently based on only the published evidence discussed above, the test developer notes plans to publish longer term outcome data. This data will be considered when available and may result in changes to the coverage policy.

This Medicare contractor will provide limited coverage for the Pigmented Lesion Assay / PLA (DermTech, Inc., La Jolla, CA), a Ribonucleic Acid (RNA) gene expression test conducted on skin samples obtained noninvasively via adhesive patches.

The PLA is indicated only for use on pigmented skin lesions, for which a diagnosis of melanoma is being considered. The test may only be ordered by clinicians who evaluate pigmented skin lesions and perform biopsies. The test is covered for use as a source of information on whether or not to perform a biopsy.

The specific characteristics that the lesion must have are as follows:

• The lesion must meet 1 or more ABCDE criteria (Asymmetry, Border, Color, Diameter, Evolving)
• Primary melanocytic skin lesions between 5mm and 19mm
• Lesions where the skin is intact (i.e., non-ulcerated or non-bleeding lesions)
• Lesions that do not contain a scar or were previously biopsied
• Lesions not located in areas of psoriasis, eczema or similar skin conditions
• Lesions not already clinically diagnosed as melanoma, or for which the clinical suspicion is sufficiently high that the treating clinician believes melanoma is a more likely diagnosis than not
• Lesions in areas other than palms of hands, soles of feet, nails, mucous membranes and hair covered areas that cannot be trimmed

Additional coverage requirements:

• The ordering clinician must also have a plan at the time of ordering the test to continue to monitor the skin lesion for changes if the test is negative. The record must also contain a photograph of the lesion at the time that the PLA is ordered to allow for appropriate evaluation in subsequent follow-up.
• Records must clearly support that the ordering clinician has the knowledge, skills, and experience to evaluate and biopsy pigmented skin lesions. If this information is not contained with the chart of the beneficiary to whom a service is being rendered, it must be supported by other readily available documentation, such as credentialing documentation, or documentation of training in the performance of such tasks. Such documentation should be provided if there are documentation requests.
• The ordering physician must clearly document the lesion site on the patient’s body.
• The test may not be ordered for the same lesion a second time.
• Only 1 test may be used per patient per clinical encounter, in most cases. In roughly 10% of patients, a second test may be indicated for the same clinical encounter. For rare cases where more than 2 tests are indicated in a single clinical encounter, an appeal with supporting documentation may be submitted for additional tests.

The PLA is not intended to be used as a screening test in patients without melanocytic skin lesions. It is also not covered as an adjunctive test in lesions that are considered to already warrant a biopsy. The PLA is a decision tool for atypical melanocytic lesions prior to the decision to biopsy.

Specific Coverage Criteria

The PLA is indicated for use on melanocytic skin lesions with 1 or more clinical or historical characteristics suggestive of melanoma, including 1 or more ABCDE criteria when a clinician trained in the clinical diagnosis of skin cancer is considering the need for biopsy to rule out melanoma. The PLA should not be used on clinically obvious melanoma. The PLA result is 1 element of the overall clinical assessment, and should be used in combination with clinical and historical signs of melanoma to obtain additional information prior to a decision to biopsy.

(PLA positive lesions [LINC (LINC00518) and/or PRAME (preferentially expressed antigen of melanoma) detected] should be considered for biopsy. The biopsy decision of a PLA negative lesion should be based on the remainder of the entire clinical context.)

The PLA is indicated only for use on:

• Primary melanocytic skin lesions between 5mm and 19mm
• Lesions where the skin is intact (i.e., non-ulcerated or non-bleeding lesions)
• Lesions that do not contain a scar or were previously biopsied
• Lesions not located in areas of psoriasis, eczema or similar skin conditions
• Lesions not clinically diagnosed as melanoma
• Lesions in areas other than palms of hands, soles of feet, nails, mucous membranes and hair covered areas that cannot be trimmed

The PLA is not intended to be used as a screening test in patients without melanocytic skin lesions. It is also not covered as an adjunctive test in lesions that are considered to already warrant a biopsy. The PLA is a decision tool for atypical melanocytic lesions prior to the decision to biopsy.

The evaluation with the PLA is limited to order by a physician or other qualified healthcare professional.