Cutaneous melanoma is the fifth-most common cancer in the US and accounts for the majority of skin cancer-related deaths. Among cancers treated by surgical oncologists, there are few whose management has changed so drastically in the last decade and in which so much progress has been made. In 2010, the average overall survival for Stage IV patients was <10% but now nears 50%.1-4 Much of this progress is ultimately a downstream effect from the technical advances in genomic sequencing technologies that began in the late 1990s and early 2000s. Next-generation sequencing (NGS) technologies have redefined what assays can be performed, and information can garnered with patient specimens and has markedly reduced the cost and time required for such studies—making these assays valuable in both research and clinical settings. Herein, we discuss ways in which the genomic analysis of melanoma has been utilized to improve patient care and discuss additional exciting applications being developed.
Impact on Melanoma Classification
Classically, there were four defined histological subtypes of cutaneous melanomas—superficial spreading, nodular, lentigo maligna, and acral lentiginous. However, these histologic subtypes are not prognostic and do not factor into staging by the 8th edition of the American Joint Committee on Cancer (AJCC) guidelines.1 Multiple international genomic sequencing initiatives established for cancers have offered potentially more clinically meaningful classification schema.
The Cancer Genome Atlas (TCGA) program was the first to complete a large whole-genome sequencing study for melanoma. Their analysis identified four major subtypes of cutaneous melanomas: BRAF, making up approximately 52%; RAS, making up approximately 28%; NF1, making up approximately 14%; and triple wild-type, accounting for the remainder. Mutations identified in the triple wild-type include CKIT, BAP1, CDKN2A, HRAS, NRAS, GNAQ, GNA11, IDH, and others.5
Unsurprisingly, these subtypes correlate only partially to the histologic classification schema. The superficial spreading and nodular melanomas found on areas of the body with less sun exposure, such as the trunk and limbs, are most likely to carry BRAF mutations, whereas nodular melanomas in chronically exposed skin are more likely to carry RAS mutations. Melanomas unrelated to sun exposure, such as acral and mucosal melanoma, are more likely to carry other mutations, such as C-KIT.6 NTRK fusions are relatively uncommon, most often found in Spitzoid melanomas (and related benign lesions) but also more rarely in acral melanoma and even occasionally in non-acral cutaneous and mucosal melanoma.7
Considerations for Therapy
For localized early-stage melanoma, the mainstay of treatment remains surgery; for advanced disease, though, surgery alone is not enough. Previously, chemotherapy, such as dacarbazine, and high dose IL-2 were used but were not effective and poorly tolerated. However, in 2011, the first targeted therapies for BRAF-mutant melanomas and the first immune checkpoint inhibitors were approved by the FDA and have since become first-line therapies. These agents have dramatically improved the outlook for patients with advanced locoregional and metastatic melanoma.4
Targeted Therapies
Approximately half of melanomas carry the BRAF mutation, of which the V600E mutation is the most common, making up approximately 74% to 86% of BRAF mutations.8 First approved in 2011, BRAF inhibitor vemurafenib demonstrated a 63% relative reduction in risk of death and 74% reduction in the risk of death or disease progression compared to dacarbazine in BRAF V600E mutated melanoma.9 Downstream kinases activated by BRAF, MEK1 and MEK2 function in the same mitogen-activated protein kinase (MAPK) pathway; activation of these pathways leads to rapid development of resistance. Combining BRAF and MEK inhibitors, such as dabrafenib and trametinib, have demonstrated significantly increased survival compared with BRAF inhibitors alone, and combinatorial strategies are now standard.10 Since 2011, several BRAF and MEK inhibitors and combinatorial regimens have been trialed. US Food and Drug Agency (FDA)-approved agents include vemurafenib (Zelboraf®), dabrafenib (Tafinlar®), and encorafenib (Braftovi®) for BRAF inhibition and trametinib (Mekinist®), cobimetinib (Cotellic®), and binimetinib (Mektovi®) for MEK inhibition.
NGS technology-based multiple-gene panels have greatly facilitated the testing for so-called actionable-mutations or "drug-able" targets. These are now recommended in many patients with advanced or metastatic melanoma. In addition to BRAF and MEK inhibitors, other targeted therapies are available for use in patients with melanoma. The tyrosine kinase inhibitors imatinib, nilotinib, and sunitinib have all been trialed in advanced melanoma with a CKIT mutation.11 The inhibitors entrectinib and larotrectinib also are FDA approved for advanced metastatic or unresectable solid tumors that harbor NTRK gene derangement.7,12 As the number of clinically efficacious targeted agents are developed, these panels will be further refined and more patients will have additional therapeutic options.4
Immunotherapies
Immunotherapy relies on the immune system's ability to recognize and mount an immune response against "non-self" tumor. This relies on the presence of neoantigens, a direct result of genetic mutations in the tumor. These mutant proteins are processed by antigen-presenting cells and presented to naïve T cells eliciting a vigorous immune response. Unfortunately, these T cells eventually become less active secondary to the expression of immune checkpoints in place to protect the body from chronic immune stimulation. Checkpoint inhibitors can both broaden the T-cell response and/or reignite an "exhausted" T-cell response and include anti-CTLA-4 and anti-PD-1/PD-L1 monoclonal antibodies.13
Tumor-mutational burden (TMB), defined as the number of mutations per coding area of the tumor genome, functions as an index of antigenicity and partially explains tumor response to immune checkpoint inhibitors.13 As melanomas result from UV exposure and genetic alteration, they have a characteristically high TMB. It is perhaps not surprising that checkpoint inhibitors are especially efficacious in patients with melanoma.14 FDA-approved in 2011, Ipilimumab (Yervoy®), a human immunoglobulin G (IgG) monoclonal antibody targeting CTLA-4, was the first checkpoint inhibitor used in melanoma patients. Ultimately, Phase III trials showed efficacy of Ipilimumab in extending the overall survival of patients with metastatic melanoma and recurrence-free survival in patients with advanced locoregional disease following surgical resection.4,13 Anti-PD-1 (nivolumab, Opdivo®; pembrolizumab, Keytruda®) and anti-PD-L1 (atelizumab, Tecentriq®) agents have also demonstrated efficacy in the adjuvant and metastatic settings.4 These agents are now being used in the neoadjuvant setting and there are ongoing trials of their use in patients with high risk Stage II melanoma.4
Combined CTLA-4 and PD-1 blockade has demonstrated the most efficacy, with improved treatment outcomes compared to either treatment alone.2,15 In part, these findings are thought to be due to the differing mechanisms of action between the two treatments, with anti-PD-1 treatment targeting the effector phase and anti-CTLA-4, the priming phase.8 Similarly, more recently, combination therapy with lymphocyte-activation gene 3 (LAG-3) and PD-1 has been shown to effective and safe in patients with metastatic or otherwise unresectable melanoma.16 Other adjuncts to checkpoint blockade include personalized neoantigen peptide-based vaccines, which can induce durable T-cell responses in patients with advanced melanoma. These personalized vaccines are yet another result of improved sequencing and bioinformatics technologies.17
Future Opportunities
No doubt, these sequencing technologies have led to substantial improvements in our treatment of patients with advanced and metastatic stage melanoma. However, these technologies and other research tools are also transforming additional areas in the care of patients with melanoma.
Diagnosis
The diagnosis of cutaneous melanoma relies heavily upon histopathologic examination, but the difference between a benign nevus from a malignant melanoma is not black and white. Moreover, overdiagnosis leads to overtreatment and increased cost and patient harm, while under-diagnosis leads to poor patient outcomes. Identification of common genomic alterations associated with invasive melanoma may help guide diagnosis. The combination of individual mutations present, for one, is informative. Driver mutations like BRAF (and NTRK in Spitzoid melanoma) are often identified in early melanomas and not infrequently in benign lesions, as well. In contrast, the detection of other mutations in genes like CKIT or the TERT promoter are rare in benign lesions and more consistent with a diagnosis of malignant melanoma. Moreover, copy number aberrations and TMB increases along the progression from benign nevus to invasive melanoma. We predict that the further characterization of the genomic/transcriptomic status of both benign and malignant skin lesions via the implementation of spatial genomics/transcriptomics technologies (which combine high-resolution microscopy with state-of-the-art genomic technologies) will help us further refine our diagnostic criteria for melanoma.18,19
Prognosis
Determining the early-stage melanomas that will progress to advanced and/or metastatic disease also is currently difficult. We rely heavily upon the AJCC staging, which incorporates a number of prognostic factors including tumor thickness, presence of ulceration, mitotic rate, and the presence of regional lymph node, non-nodal locoregional, and distant metastases into staging. The presence of disease in the sentinel node is especially prognostic; thus, all melanoma patients with a Breslow depth exceeding 0.8 mm and/or ulceration are recommended to undergo a surgical sentinel node biopsy.1
Several gene expression profiling (GEP) tests have become available in the context of melanoma, such as Decision-Dx Melanoma (Castle Biosciences, USA), Melagenix (NelaCare, Germany), and SkylineDx (Rotterdam, Netherlands).20 In several retrospective studies, gene expression profiling (GEP) has been associated with recurrence-free survival and distant metastasis-free survival.4,21 However, the question remains as to whether GEP data is more informative than known prognostic factors such as mitotic rate and lymphovascular invasion. In some instances, GEP assay data already incorporates current staging tumor characteristics (such as the ratio of keratinocyte and melanoma genes, which ultimately reflects tumor thickness). While these strategies hold promise, GEP has yet to be adopted by national guidelines, including those established by the National Comprehensive Care Network and the American Academy of Dermatology.22
Predicting Immunotherapy Responses
As immunotherapy still ultimately fails most patients with metastatic melanoma upfront, and toxicity can be severe, predicting those patients that will respond is paramount. TMB does not uniquely define response. The immune infiltrate within a given lesion also has profound influence on the response to checkpoint inhibitors.13 Single-cell analyses (flow cytometry, immunohistochemistry) currently are utilized to assess the tumor immune microenvironment. Single cell transcriptomic technologies (single cell RNAseq or scRNAseq) have also been instrumental in this regard, especially in defining the activation state and TCR specificity of tumor-infiltrating T-cells.23 Spatial transcriptomics can provide even more critical information regarding the tumor immune microenvironment. NanostringDx technology was used to define an association of B cells in the context of TLS with response to checkpoint inhibitors.24 In the future, both scRNAseq and spatial transcriptomic technologies may be available for initial biopsies of primary and metastastic melanoma and may help guide our therapeutic choices. Another known effector of immunotherapy response is the gut microbiome.25 Assessing the microbiome of patients by advanced sequencing technologies may eventually be used in the clinic to predict the patients who might respond to immunotherapy.
Surveillance
Circulating tumor DNA (ctDNA) is one example of a liquid biopsy that enable one to detect the presence of tumor within peripheral blood. ctDNA requires sequencing of tumor and germline tissue to delineate differences between tumor and germline DNA and create a personalized "tumor signature" that can be quantified in blood. The presence of ctDNA at baseline and following treatment have been independently associated with decreased metastasis-free survival. Moreover, a rise in levels during a surveillance period is associated with recurrence. As such, ctDNA can be useful in guiding therapeutic decision-making, and we suspect the use of ctDNA or other liquid biopsies will become the norm for melanoma surveillance.26,27
Conclusion
So many recent innovations in melanoma treatment have their basis in improved sequencing technologies—innovations from the late 1990s and early 2000s. The staggering improvement in melanoma outcomes since that time portends a bright future for melanoma patients as we further utilize these technologies and their downstream applications to improve the diagnosis and treatment
