Diagnosing melanoma

Dermatologists and dermatopathologists can address the challenges of diagnosing and treating melanomas by using combination stains, a growing assortment of genetic tests and even new prognostic indicators, says Whitney A. High, M.D., J.D., M.Eng., an expert who spoke at the 2016 Fall Clinical Dermatology Conference (October 2016, Las Vegas).

"It has been asserted recently that the most mildly atypical nevi, and even moderately atypical nevi, are of clinical consequence," says Dr. High, an associate professor of dermatology and pathology at the University of Colorado School of Medicine in Denver.

While epidemiologic studies show an association between clinically or histologically dysplastic nevi (HDN) and melanoma, Dr. High explains, the odds of melanoma arising in any individual dysplastic nevus are low. In fact, a retrospective study showed that at a mean follow-up of 17.4 years, among 115 patients with HDN that extended to within 0.2 mm of a sample border, no melanomas occurred at the nevus location or at metastatic locations.¹

However, critics of the aforementioned study noted that a narrow margin is not the same as a frankly involved margin, and that some challenging melanomas may be misidentified as HDN.² In an editorial response to the study, these critics explained that "The gray zone of lesions with higher grade that extend to or closely approach a specimen margin is problematic, as one dermatopathologist's moderately atypical nevus may be another's melanoma."

Atypical Spitz nevi

Dr. High laments, "Perhaps dermatopathologists are not as good at distinguishing atypical nevi from melanoma as we think they are," and this may be another good reason to remove a concerning lesion. He shared images from a recent case, signed out by an experienced dermatopathologist without use of immunohistochemistry (IHC) or step levels as a "moderately atypical nevus," which proved to be melanoma. Retrospective analysis of the earlier sampling, using IHC and genetic testing, revealed the lesion was a melanoma, even from inception.

Dr. High addressed challenging spitzoid melanocytic proliferations and the emerging concept of "atypical Spitz tumors (ASTs)," which is a nosology used to differentiate these lesions of uncertain biologic potential from "classic" Spitz nevi, the latter being more common in children.

He presented a rapidly eruptive lesion on the leg of a healthy 17-year-old girl with no personal or family history of melanoma. IHC revealed evidence of spitzoid morphology and favorable expression of p16, but also some atypical features and dermal mitotic activity. 

"Atypical Spitz tumors place the dermatopathologist in a real bind," Dr. High says.  If a lesion is "overcalled" and treated as melanoma – considering that up to one-third of ASTs may have lymph node involvement, but without the same observed consequences of melanoma – surgical morbidity may quickly ensue.  By the same token, if the lesion is "undercalled" when it is indeed a spitzoid melanoma, this can lead to medicolegal issues and great harm to the patient.

Dr. High next shared a case where an experienced dermatopathologist signed out a lesion located on the right neck of a 23-year-old as a "combined nevus, with features of a common nevus and Spitz nevus," only for the patient to develop right-sided cervical lymphadenopathy and lung metastases, clearly indicative of Spitzoid melanoma, three years later. 

While admitting that in the past, and apparently correctly (per followup), he had himself diagnosed lesions as "combined nevi with spitzoid elements," Dr. High states that this was among "the most dangerous of creatures for dermatopathologists – great care must be taken."

NEXT: Genetic expression profiling

Genetic expression profiling

Dr. High emphasizes that "Rather than using morphologic analysis to assess genetic potential, there is increased promise for using actual genetic testing to assess melanoma." In a recent study, the presence of telomerase reverse transcriptase (TERT) promoter mutations provided the most accurate prediction of aggressive behavior in 56 patients with spitzoid melanocytic neoplasms (p<0.0001).³ At a median follow-up of 32.5 months, four patients – all with TERT promoter mutations – had died of disseminated disease. None of the 52 surviving patients had such mutations.

Dr. High adds that genetic testing may be used for prognostic purposes, not just for diagnosis.  "There is evidence that a new 31-gene expression profiling test (31-GEP) can distinguish high-risk melanoma from low-risk melanoma, at least in many instances." This 31-GEP test predicted metastatic risk in multple cohorts of primary cutaneous melanoma (ROC = 0.93, 0.91, respectively).⁴ Kaplan-Meier analysis showed that five-year disease-free survival (DFS) rates in the development cohort for predicted low-risk and high-risk cases were 100% and 38%, respectively (P < 0.0001). DFS rates for the validation set were 97% and 31% for predicted low-risk and high-risk cases, respectively (P < 0.0001).  

Dr. High speculates that this type of prognostic testing could prove particularly useful in certain situations, such as melanoma of less than 1 mm in depth (or perhaps between 0.76 mm and 1 mm) but with mitotic activity, and those persons who undergo SLN biopsy with negative results. Among SLN-negative cases that were identified as high-risk by the 31-GEP test, a recent report showed that at five years' follow-up, 67% had recurred, and 83% of patients had died.⁵

"Since we really want to know how a melanocytic lesion will behave on a genetic basis," Dr. High says, "it is reasonable to assume that the future of dermatopathology testing modalities will be genetically-based, rather than morphologically based."

With more than 6,000 mutations described in melanoma, he said, attempting to distinguish malignant lesions from benign atypical nevi by histologic examination is more subjective than people realize.

Immunohistochemistry (IHC) can be helpful, but there is no "melanoma stain," he says. Each IHC stain has strengths and limitations. For example, human melanoma black (HMB)-45, demonstrates "zonal" staining, with lesser expression with descent of melanocytes in nevi, and it is generally a sign of benignity.  However, in lesions with what Dr. High called "dusty" melanocytes, such as deep penetrating nevi (DPN) and so-called "clonal" nevi, this result can give unhelpful results.

The tumor suppressor protein p16 is normally expressed in a mosaic pattern, with nuclear and cytoplasmic staining, in most benign melanocytic nevi, including most Spitz nevi, he said. Conversely, loss of p16, particularly loss of nuclear expression of p16, is a concerning feature that is more often associated with melanoma. Loss of p16 expression in melanocytic lesions can be a surrogate of biallelic loss of portions of chromosome 9p21, which in turn can be an indicator of a lesser prognosis in spitzoid melanocytic lesions.

"Broader stains for general badness involve combining a MART-1/Melan A stain (a marker of melanocytes), with Ki67 (a marker of cellular proliferation)," Dr. High says. This combination allows the examiner to gauge the architecture of a melanocytic process, and to note, more specifically, which melanocytes/melanocytic nests have increased proliferation.

Disclosures: Dr. High reports no relevant financial interests.

References

1. Hocker TL, Alikhan A, Comfere NI, Peters MS. Favorable long-term outcomes in patients with histologically dysplastic nevi that approach a specimen border. J Am Acad Dermatol. 2013;68(4):545-51.

2. Elston D, McNiff J, Maize J Sr. Histologically dysplastic nevi that extend to a specimen border. J Am Acad Dermatol. 2013;68(4):682-3.

3. Lee S, Barnhill RL, Dummer R, et al. TERT promoter mutations are predictive of aggressive clinical behavior in patients with spitzoid melanocytic neoplasms.Sci Rep. 2015;5:11200. doi: 10.1038/srep11200.

4. Gerami P, Cook RW, Wilkinson J, et al. Development of a prognostic genetic signature to predict the metastatic risk associated with cutaneous melanoma.Clin CancerRes. 2015;21(1):175-83.

5. Zager JS, Messina J, Sondak VK et al. Performance of a 31-gene expression profile in a previously unreported cohort of 342 cutaneous melanoma patients. Poster 186. Presented at: American Society of Clinical Oncology Annual Meeting. June 3-7, 2016; Chicago.