Hello, my name is Christine Jaworski. I'm a professor of dermatology at Case Western Reserve University, and I'm on faculty at MetroHealth Medical Center in Cleveland. I practice dermatology and dermatopathology. The topic for my presentation is the application of molecular techniques in clinical practice. I would like to limit that discussion to pigmented lesions. I have no conflicts of interest to disclose. In our travels, I'm sure we've all had the opportunity to wait at a train station where the signage tells us to please mind the gap. That, of course, is the gap between the platform and the oncoming or departing train. Perhaps in clinical practice, it's a reminder for us to mind the gap between the real, that is the knowledge base that exists, and the ideal. In this way, I'd like to limit my discussion to three separate tests. One is immunoproxity staining. One is tape stripping analysis. And another is the Castle Decision DX test. In each of these, I'll discuss the how-to or the application, the utility, of course, at this point in time, and the cost. We'll start with immunoproxity staining. This technique has been available since the 1980s. It's useful in detecting antigens in various tumors and also highlighting their presence and location in paraffin-embedded sections. There are many that are available. And new ones are constantly being developed. The technique involves stacking or layering of antibodies. If my green LEGO board represents the epidermis and the black LEGOs represent melanocytes, which we would like to detect, we prepare the tissue. We apply a primary antibody to the antigen. In this case, for example, we'd use anti-S100 generated in rabbit. The reaction takes place. The excess antibody is rinsed off. Then a second antibody or linking antibody is applied. This antibody is different. It's an anti-rabbit antibody generated in a mouse. And it has this structure, a tail. What that does, it allows to tag reaction sites for detection with a red chromogen. In this way, we can detect melanocytes in tissue. They're localized in the epidermis, within the stroma, superficially and deep. And they show reactivity in the nucleus as well as dendritic extensions. In general, in nevi and dysplastic nevi, use of immunoproxidase stains is unnecessary. There are exceptions to every rule. One would be that of rare epithelioid nevi, where we would use immunoproxidase to demonstrate loss of BAP1. BAP1 is a housekeeping gene that is responsible for tumor suppression. This example, kindly given to me by Dr. Somak, shows an asymmetric nodule and a nevus that's inflamed. Higher power shows areas such as this, where there are large epithelioid melanocytes interspersed with smaller ones and a relatively intense host response. If we stain this tissue for BAP1, what it shows us is loss of BAP1 staining in regions such as that circled in red and preservation in adjacent areas. So what does that mean? It suggests that the patient and perhaps the family are at increased risk of melanoma, uveal melanoma, mesothelioma, and renal cell carcinoma. As we look at melanoma in situ, that is the intraepidermal component of melanomas, specifically superficial spreading types and lentigoma lignotypes, immunoproxidase stains are not often necessary. Sometimes, Melan-A or MART1 has been used. However, it overestimates melanocyte density. This finding was well-documented in a paper in the American Journal of Dermatopathology almost 20 years ago by Dr. Cerrone and colleagues that showed that Melan-A is not a helpful marker to distinguish between melanoma in situ on sun-damaged skin and a pigmented actinic keratosis. One example would be this tumor in which there are large numbers of melanocytes in the epidermis. Here we see the Melan-A or MART1 stain marking nuclei and also dendritic extensions. And of course, we know dendrites will feed melanosopes into adjacent keratinocytes. Prane stains will mark melanomas. However, they're also found to mark benign nevi and non-melanoma skin cancers. MITF and SOX10 have a high specificity for intraepidermal melanocytes. So here we see a companion panel of the MART1 next to the SOX10. And the SOX10 is a much cleaner stain in that it marks melanocyte nuclei and not the dendrites. This tumor, in fact, was a squamous cell carcinoma that was poorly differentiated and it was colonized by melanocytes, but this was not a melanoma. There is a nice review article on the intraepidermal reactivity of melanocytes with immunohistochemical stains in the Journal of Cutaneous Pathology in 2022. In terms of invasive melanomas, that is the intradermal components of tumors, immunoproxidase stains are helpful to assess the depth of a tumor when there's an intense infiltrate like a halo reaction, and also in desmoplastic melanomas. The example that I showed earlier was, in fact, this desmoplastic melanoma that on higher power shows a confluent proliferation of melanocytes along the dermal-epidermal junction, almost dissecting the epidermis off of the dermis. There's a brisk host response of monotonous cells as the patient had underlying CLL. Beneath this area, there were fusiform cells that were in the dermal epidermal junction in this area. There were fusiform cells which had hyperchromatic nuclei associated with sparse lymphocytic infiltrates. This was the deep dermal component of the melanoma, which was assessed with S100 stain, which allowed measuring the total depth of the tumor. Finally, immunoproxidase stains are helpful in nevoid melanoma. This is a subset of tumors that mimics the growth pattern of a nevus, but in fact is a melanoma. These tumors often are investigated with Preyman P16 stains, which are sometimes helpful, but not always. This example shows an asymmetrically raised nodule. Higher power shows nests of small melanocytes and the surrounding areas with large atypical melanocytes. P16 stains should uniformly mark nevomelanocytes. In this instance, it strongly marks some cells here within the superficial dermis. There is less strong marking here off to the left and even spottier marking off to the right. The companion Prem stain marks more than 80% of melanocytes throughout the dermis. On higher power, we see strong marking of the nucleus in cells in this area. And although the staining is not as strong in the adjacent area, still there's positive nuclear staining. This was a nevoid melanoma. So to summarize the utility of immunoperoxidase stains, judicious and not reflexive use makes it a good tool. It's useful to assess tumors in a dense inflammatory infiltrate where there's a halo response to melanoma, for example. It's useful to measure the depth of the tumor in a desmoplastic melanoma and also in nevoid melanomas. In one other instance, immunoperoxidase stains are helpful and that would be to detect BRF, KIT and NRES mutations for potential systemic therapy of melanoma. The cost per stain is $100. Moving along to tape stripping analysis. The DermTag test is the only non-invasive tape stripping assay available in the US. And because it's non-invasive, it's alluring to do it. The RNA and DNA are extracted via the tape from superficial layers of the skin and sent for genomic analysis. The results are then generated based on RNA markers, LINC518, PRAME and the DNA genomic marker TERT. The stated purpose of the test is to identify high-risk lesions at their earliest stages by detecting melanoma associated genomic markers. The result comes as a report like this. The result in this instance was positive and the clinician was directed to do a surgical biopsy and get histopathologic evaluation. My first encounter with this test was a negative result yielded from a 71-year-old man's cheek. The biopsy, however, that came with it showed a melanoma in situ of the lentigoma lignotype. This prompted a phone call to the clinician to ask why a biopsy was done if the DermTag result was negative. The answer was that the patient had a lot of sun exposure, he had a melanoma elsewhere on his body and the patient was a lawyer. That prompted us to catalog 44 cases which were biopsied and came with a DermTag analysis. Our data includes lesions that were sampled either because of a positive DermTag result or as in the case I just mentioned, because of clinical concern. The histopathology was reviewed by at least two dermatopathologists independently. We cannot comment on the sensitivity or specificity of the test because negative DermTag results without a biopsy were not included and our results were published in the November issue of the Blue Journal of 2023. Of our 44 cases, 11 were melanomas. If we take a look at the cross sections, four of these were invasive melanomas. They were at least 0.8 millimeters deep up to 1.2 millimeters depth. There were five correctly identified melanoma in situ. There was one incorrectly or not identified melanoma in situ, the case I described earlier. And there was one atypical intrapidermal melanocytic proliferation which was correctly identified as positive. Our results included 19 nevi. Of these, two were benign and correctly identified as negative. One was a halo nevus on a child and another an abdominal lesion on an adult. There were also four nevi with moderate atypia, three with severe atypia. Not everyone would excise a moderately atypical nevus, but for purposes of discussion, this was included as the correct result. There were however, 10 nevi that were benign and were incorrectly identified as positive. There were also 14 non-melanocytic lesions. Of these, three were correctly identified as negative. There were however, 11 that were identified as positive. They were not melanocytic. So if we take all comers, the test missed an in situ melanoma. It reported over 50% of nevi as positive. They did not need an excision and it reported 78.6% of non-melanocytic lesions as positive. If we stratify the data into non-melanocytic versus melanocytic lesions alone, the DermTest test incorrectly identifies over 75% of non-melanocytic lesions as positive. PRAME often comes up in the literature. We discussed it in immunohistochemistry testing for nevoid melanomas. It's also one of the markers examined in the Dermtech test. And as Dr. Lebois was quoted by Dr. Hayman saying, that PRAME is an acronym for Preferentially Expressed Antigen in Melanoma, but it can also mean probably right and also make errors. I would agree with that statement. Like any test or immunohistochemical stain or marker, it has to be used in conjunction with clinical and pathologic correlation. After our study was completed, I received another Dermtech result paired with a biopsy. The result was positive. It guided the clinician to do a biopsy and send for histopathology. The histology showed a broad lentiginous compound nevus, which had no atypia within the epidermis or in the dermis. So to summarize the literature, there are a lot of published studies that indicate utility to the test. However, they have a conflict of interest. The test itself costs $760 per tape-stripped lesion. There is an additional charge that's applied for doing the test in the office. There was a systematic review of the literature available for tape-stripping of pigmented lesions in Skin and Research Technology in 2023. One of the authors is also the lead author in this article in the Blue Journal from Denmark of RNA analysis of tape-stripping of a different sort. They wanted to rule out melanoma among suspicious lesions to reduce unnecessary surgery. They tape-stripped 200 lesions before surgical excision. 11 genes were assessed by their RNA measurements, and their results were based on the oncogenes PRAME and KIT. One take-home message from the study is that the longer the tape storage time, the lower capacity to rule out melanoma. So let's take a look at their data. 200 lesions were tape-stripped before biopsy. Of their 200 lesions, 73 were melanomas. In fact, they were in situ and invasive melanomas. 41 lesions were non-melanocytic, and they were correctly identified as negative. There were also 86 lesions that were not atypical or melanoma, but were identified as positive. If we include all of their positive tests for melanoma and incorrectly diagnosed as melanoma, there is a false positive rate of 54%, because 86 of 159 lesions were not melanoma. If, on the other hand, we look at the lesions that were non-melanomas, the test incorrectly identified 86 as risky for melanoma, and only correctly identified 41 as not melanomas. That gives it a false positive rate of 68%. The lesions that were identified as positive included various types of nevi, several keratoses, actinic and seborrheic, basal cell carcinomas, hemangiomas, a solar lentaigo, dermatofibroma, and a trichoepithelioma. The goal of this study, of course, was to rule out melanoma among suspicious lesions to reduce unnecessary surgery. However, they were unable to do this in 86 of 200 cases, or almost 50% of the lesions that were sampled. If we take the two studies side-by-side, each shows a high rate of false positive results in non-melanomas. If, in our study, we include the one melanoma that was missed, inaccuracies existed in 50% of the tests done. So, to summarize, tape stripping is a technique, actually, that has been applied to analysis of skin since the 1930s. It has been used in laboratory tests to induce contact dermatitis in animals and also to assess skin permeability studies in atopic dermatitis. Thus far, for detection of melanomas, it shows significant inaccuracies. For now, it over-reports benign pigmented and non-pigmented lesions as risky for melanoma and, in fact, leads to over-biopsy. And lastly, we move on to the CASEL Decision DX test. This is a test that's done on paraffin-embedded tissue from blocks of melanoma that have already been diagnosed. The tissue itself is sent for reverse transcriptase PCR on extracted RNA to determine levels of 31 genes that are associated with melanoma metastasis. Patients' risks are reported as low risk, that is, the lowest risk for metastasis in five years, and the best survival, or an intermediate risk group listed as 1B or 2A and as high risk groups, 2B. This is an example of a Decision DX report. This tumor, for example, was Class II, subclass 2A. We go down the table to see the recurrence-free and the distant metastasis-free survival rates. And on the second page, there are percentages of risk of positive sentinel nodes depending on the age of the patient. More recently, Decision DX has incorporated some AGCC criteria for predicting melanoma-specific survival and distant metastasis-free survival. Those characteristics include Breslow thickness, the age of the patient, presence or absence of ulceration, the location of the tumor, trunk versus extremity, nodal status, and mitotic rate. Again, one would look at the class to look at the risk rate for survival and metastasis. The second page presents the likelihood of positive sentinel nodes as a percentage. The test assigns a risk category for predicting metastasis based on the genetic profile. This is sometimes used instead of sentinel node analysis. However, I've even seen it used with sentinel node analysis. Obtaining the test presumes that the genetic profile alone determines tumor behavior. I think we've all encountered examples in our practices where a patient has a deep melanoma and one would predict that the tumor would metastasize. However, it does not, and the patient has a good outcome. Alternatively, I recently saw a report of a patient who had a melanoma 0.7 mm. It was reported that the tumor metastasized and yet she developed a metastasis within two years. So observations suggest that more than the genetic profile alone determines outcome. This is one facet of a multi-faceted problem. What other factors might be at play? Is it perhaps a tumor type? We know that desmoplastic melanomas don't behave the same way that superficial spreading type melanomas do. We know that there is a risk of metastasis Perhaps the tumor volume has an impact. We do measure tumor depth and that's a rough gauge of volume but doesn't exactly represent the volume. What about the host response? The degree of cytotoxic T cells impacting the outcome of the tumor? And regression? So there are many factors at play. The pros of the test are that DecisionJX has a tremendous genetic data bank for future reference. Also, you can add on BRAF, NRAS, and KIT to assess for mutations by just checking the box on the requisition. We've discussed the pros of the test. What are the cons? Well, the test reflects the tumor at the time of diagnosis. Tumors can acquire mutations so that doesn't necessarily guarantee the same outcome. The cost per test is $7,193. I have seen a range up to $85. The add-on tests carry an additional cost. The other con is that the result assigns a risk category of a group, not of the individual. Patients, as we've seen, even a 1A can develop metastasis despite being classified as low risk. There are many articles in the literature about the performance of gene expression profile tests. This was a meta-analysis of gene expression profile tests from a group at Sloan Kettering, which showed that for localized cutaneous melanoma, particularly thin melanomas, the information was not helpful and did not change prognosis. This was followed by a viewpoint by Dr. Kovarik, Chu, and Adamson, which also say that the evidence to support clinical use on gene expression testing in thin melanomas remains thin. To summarize the molecular techniques we've discussed in melanocytic lesions, we looked at immunoproxity staining on tissue and paraffin blocks. It marks cell types. The test is reliable, not necessarily always helpful. It can overestimate melanoma in situ, but can be helpful in some instances of dermal melanomas. The cost is $100 per stain. One can also assess mutation status by identifying tumors with mutations of BRAF, KIT, and NRAS. The DermTech test is a tape-stripping or in vivo test that reports risks of lesions. It identifies non-melanocytic and benign lesions as risky. It also costs $760 per test. The DecisionDx test takes sections from paraffin-embedded tissue and classifies tumors into low, intermediate, or high risk for metastasis and five-year survival. Perhaps it may be useful in the intermediate stage, 1B to 2A. This may change and will be determined over time. The test costs $7,193 per test. So how do we apply this in clinical practice? I generally avoid biopsy of dark pigmented lesions that are less than 6 mm in diameter. I find that this puts patients on the treadmill of finding mild atypia and then the path of multiple unnecessary subsequent biopsies, sort of the well-intended misadventure. I do biopsy ugly duckling nevi. I find that patients do often have signature nevi, and when there is one atypical standout lesion, that is the one to biopsy. I do go by the ABCD aberrations, including the asymmetry of a pigmented lesion, the border, the color, and the diameter. And I also pay attention to patient-reported changes. In terms of the techniques we discussed, immunohistochemistry is useful judiciously, not as a reflexive tool. I generally avoid use in in situ melanomas. I find it helpful in some invasive melanomas, not all, but those that have a halo response, desmoplastic melanomas, nevoid melanomas, and for checking for mutations for BRAF kit NRAS if requested by the oncologist. The DermTech test is costly, and I find the results misleading. The Castle Decision DX or genetic profiling test perhaps may have utility in intermediate thickness tumors, perhaps those that are between 0.8 and 2 mm depth. We know generally the ones that are less than 0.8 mm patients tend to do well. The ones that are greater than 2 mm have a more guarded prognosis. Molecular analytic techniques will continue to evolve over time. The questions will remain, and they include how much benefit to our patients will they be, and what cost. It's important to step back and think what impact they will have. What do we do it for? Will it improve the therapeutic outcome or the quality of life to our patients? In the immortal words of Mark Twain, what gets us into trouble is not what we don't know, it's what we know for sure that just ain't so. With that, I would like to acknowledge the assistance of my colleagues in assembling this material for you. I thank you, the audience, for your time and attention, and for minding the gap.

molecular techniques

clinical practice

pigmented lesions

immunoproxidase staining

tape stripping analysis

Castle Decision DX test