There are three mammalian RAF proto-oncogenes—ARAF, BRAF, and CRAF. However, of these, only activating mutations of BRAF have thus far been shown to be present in a significant percentage of human malignancies, with the highest rate occurring in melanoma [19, 20, 22, 23]. BRAF-mutated melanomas tend to occur at anatomic sites exposed to intermittent, rather than chronic, sun damage. With approximately 70 % of cases harboring such a mutation, BRAF is the most commonly mutated proto-oncogene in cutaneous melanoma. Furthermore, a significant proportion of both benign and dysplastic melanocytic nevi have been shown to harbor mutations of BRAF as well, suggesting a relatively early event in melanomagenesis [19, 22, 23, 34].

Greater than 90 % of BRAF mutations in melanoma result from a single base missense mutation (T → A) at codon 1799 that leads to the substitution of valine in favor of glutamic acid at position 600 of the BRAF protein [35]. This alteration introduces a conformational change in BRAF’s kinase domain, which can lead to a 480-fold increase in kinase activity when compared to wild-type BRAF [36]. This mutated BRAF induces uncontrolled proliferation in melanoma. Ongoing research of this molecule includes its role in inducing senescence of benign melanocytic nevi, the mechanisms involved in bypassing this senescent response in the development of melanoma from such nevi, and the mechanisms for acquired resistance to targeted inhibition of this molecule in patients undergoing anti-BRAF therapy [37–48].

A major breakthrough in the treatment of metastatic melanoma has been the successful treatment of BRAF-mutant melanoma with selective BRAF inhibitors such as vemurafenib and dabrafenib [49]. Since this agent shows tenfold greater activity for V600E-mutant melanoma compared to wild-type melanoma, most current clinical trials require pre-treatment assessment for the presence of this mutation [50]. The mutation is ascertained by means of a polymerase chain reaction (PCR) assay (TaqMan, Applied Biosystems), which involves hybridizing a probe specific to the 1799T->A substitution with DNA isolated from formalin-fixed, paraffin-embedded tumor tissue and determining the presence or absence of amplification after repeated chain-reaction cycles [51]. This assay is now routinely available at most major commercial reference laboratories. There is also a commercially available monoclonal antibody (VE1) that can be used to assess for the BRAF V600E mutation immunohistochemically. This antibody was 100 % sensitive and specific for the mutation compared to the gold standard of molecular analysis in a recent series of 44 cases [52]. There have also been attempts at predicting BRAF mutation status by analyzing various clinical and histologic features of primary tumors. In a recent study of 302 primary melanomas, groupings of various combinations of 17 such factors revealed that phenotypic features such as increased upward scatter, intraepidermal nesting, epidermal thickening, sharp lateral demaracation of the tumor, and the presence of rounded, highly pigmented tumor cells were distinct features of melanomas with BRAF mutation. In this study, classification tree analysis showed that BRAF mutation status could be correctly predicted in up to 82 % of tumors using certain variables [53]. While there is certainly great interest in correlating easily analyzed clinicopathologic variables with mutation status, and active research aimed at this goal, at present, however, the gold standard for assessing tumors for BRAF status remains mutation analysis.

c-KIT, a proto-oncogene that encodes the type III receptor tyrosine kinase KIT, was first identified in 1987 as a result of sequence similarity to the Hardy-Zuckerman 4 feline sarcoma virus oncogene, v-KIT [54]. Upon binding its ligand, stem cell factor, KIT undergoes receptor dimerization, autophosphorylation, and activation of its intracellular tyrosine kinase domain [18, 55]. Once activated, KIT is capable of stimulating downstream signaling pathways, such as MAPK [18, 55, 56].

Activating mutations and/or gene amplification of KIT have been described in distinct subsets of melanoma [57–61]. In one study, aberrations were found in 39 % of mucosal melanomas, 36 % of acral melanomas, and 28 % of melanomas arising in chronically sun-damaged skin (as defined by the presence of solar elastosis on review of histopathology)—anatomic sites at which BRAF mutations occur far less frequently [60].

The most prevalent KIT mutations in melanoma occur in exons 11, 13, and 17 [62]. These mutations are thought to promote the constitutive activation of KIT either through precluding the protein from assuming its default autoinhibited conformation or by promoting its dimerization in the absence of stem cell factor [63, 64]. The precise manner in which constitutive KIT activation promotes melanomagenesis remains unclear. Nonetheless, recognizing activating mutations of KIT as well as copy number gains can be of great consequence, as this subtype of melanoma may respond the receptor tyrosine kinase imatinib [65–67]. Since the initial case report of a dramatic response to imatinib in c-KIT-mutated melanomas [65], interest has been raised in treating patients with acral and mucosal melanoma with this agent, leading to several ongoing phase II clinical trials (ClinicalTrials.gov) of imatinib and other c-KIT blockers such as sunitinib, dasatinib, and sorafenib. These studies require documentation of mutation or amplification of the c-kit gene. Currently, therefore, it seems reasonable that subgroups of patients with metastatic melanoma prone to c-kit mutations, such as primary mucosal or acral lentiginous tumors, be analyzed for KIT mutation status. If mutated, a therapeutic trial with a c-KIT blocking agent has been suggested. Testing, using PCR following by sequencing analysis of exons 11, 13, and 17 can be performed at several reference laboratories, including ARUP, as well as university laboratories such as those at University of Washington and University of Michigan.

The RAS proto-oncogene family is at the initiating point of the MAPK pathway, and theoretically activating mutations of these molecules should provide the affected cell with continuous growth signals in the absence of extracellular stimuli. Indeed, RAS molecules have documented roles in melanomagenesis [18–20, 22, 23, 33]. Of the three members of this family (H-RAS, K-RAS, and N-RAS), only the latter has been shown to sustain activating mutations in melanocytes, as well as nearly one third of all melanomas [19, 23, 33]. The most commonly documented N-RAS aberration in melanoma is a missense mutation at codon 61 (Q61R), which results in the substitution of arginine in place of glutamine and impairs GTP hydrolysis, locking the protein in a state of constitutive activation [68]. However, therapeutic attempts at blocking the farnesylation, which is required for proper function of the molecule, using farnesyltransfer inhibitors, have yielded disappointing results [56, 69].

There has been a resurgent interest in the role of the N-RAS pathway in sustaining melanoma growth in the wake of the phenomenon of acquired resistance to BRAF inhibitors. Specifically, investigations into this phenomenon in cell lines have revealed that hyperactivation of the MEK-ERK1/2 pathway occurs rapidly after treatment with the BRAF inhibitor PLX4032, and this has been noted most prominently in cell lines containing the activating N-RAS mutation, which demonstrated resistance to apoptosis [70, 71]. Therefore, future trials of targeted therapies may require genotyping for N-RAS to stratify treatment.

EGFR is yet another receptor tyrosine kinase that is involved in activation of the MAPK and PI3K/Akt signal transduction pathways. The latter pathway is also activated by RAS and results in decreased apoptosis [19]. The selective expression of this receptor in human melanoma cells harboring an extra copy of chromosome 7 was first studied in 1985 [72]. Numerous studies have since investigated the importance of this receptor in melanomagenesis and/or progression with contradictory results. More recently, however, researchers have shown an association between polysomy 7 and EGFR gene amplification by fluorescent in situ hybridization (FISH), with significantly higher amplification in tumors that metastasized within 2 years of surgical excision of the primary lesion [73]. Implying a poorer prognosis in this latter subset of patients, further studies into the potential benefit of anti-EGFR therapy for those with EGFR alterations have been recommended. However, at this time there appears to be no clinical utility to testing tumors for EGFR overexpression.

The technique of comparative genomic hybridization (CGH), wherein lesional DNA as well as normal control DNA are hybridized to either normal metaphase chromosomes or an array of thousands of defined DNA probes (array CGH) in order to detect copy number changes, has been used to examine differences between melanoma and benign melanocytic lesions. Specifically, gains of chromosomes 1q, 6p, 7p, 7q, 8q, 17q, and 20q and losses of chromosomes 6q, 9p, 9q, 10p, 10q, and 11q have been observed in greater than 95 % of melanomas [74]. Loss of chromosome 9, an early and common aberration in melanomagenesis, typically involves deletion of the distal region of the p arm that encompasses the 9p21 locus [75]. Of note, gain of 6p is a frequent aberration in both primary and malignant melanomas and may confer a poorer outcome [76, 77].

Melanocytic nevi rarely demonstrate chromosomal aberrations, with the exception of Spitz nevi, which may harbor gains on occasion [74, 78–80]. A gain of the entire short arm of chromosome 11 is the most commonly encountered aberration in Spitz nevus and is an abnormality not detected in primary cutaneous melanomas. Although most congenital nevi are readily distinguished from melanoma according to established histopathologic criteria, atypical nodular proliferations that arise in congenital nevi can be problematic [78]. These lesions simulate nodular melanomas microscopically, but tend to regress within a few months. CGH offers diagnostic aid in this area [78, 80]. While cases of atypical nodular proliferations show chromosomal aberrations by CGH, the pattern of involvement is distinct from that created by melanoma. Abnormalities in the former group involve only entire chromosomes. While gains or losses of entire chromosomes can frequently occur in melanoma as well, 95 % of such cases are paralleled by gains or losses of chromosomal fragments.

CGH has also underscored significant genetic differences between acral melanoma and superficial spreading melanoma [81], justifying the distinction between these subtypes by demonstrating disparity in pattern of chromosomal aberration between the two [80, 81]. While acral melanomas have been shown to invariably harbor gene amplifications (defined as greater than a threefold increase in copy number), superficial spreading melanomas, of comparable tumor thickness and patient age, contain an amplicon in less than 15 % of cases. The amplified regions frequently consist of areas containing known oncogenes. Study of the amplicons not harboring obvious candidate genes may provide further insight into the mechanisms of melanomagenesis.

Isolated melanocytes located beyond the microscopically-designated free margins of resection have been shown to harbor these genetic amplifications as well [80, 81]. Researchers believe that these field cells represent unrecognized radial growth of the acral melanoma and are likely responsible for local recurrences. If further studies support such a notion, then CGH may play an instrumental role in shaping future guidelines pertaining to the attainment of clear margins of resection.

A multiprobe fluorescent in situ hybridization (FISH) assay against several of the aforementioned chromosomal abnormalities detected in CGH studies has been developed and tested in several studies with promising results [3, 82]. The assay identifies copy number alterations in chromosomes 6p25 (RREB1), 6q23 (MYB), and 11q13 (cyclin D1) as well as centromere 6. Results of these preliminary studies have demonstrated high sensitivity (~87 %) as well as specificity (~95 %) in distinguishing benign melanocytic nevi from melanomas using this particular panel of probes.

Therefore, this technique plays a potential role in classifying problematic melanocytic lesions. The area of most intense investigation of the applicability of this technique is in the distinction between spitzoid melanoma and benign Spitz nevus. Two recent studies have demonstrated sensitivity and specificity of 43–60 % and 50–80 % for determination of malignancy in ambiguous Spitzid lesions, using metastatic behavior as the gold standard [83, 84]. In another recent, albeit small, study evaluating the efficacy of FISH in distinguishing nodular nevoid melanomas from mitotically active melanocytic nevi with probes targeting these same for four loci, chromosomal aberrations were detected in each of the nevoid melanomas and in none of the melanocytic nevi [85]. Other scenarios for which FISH has shown potential benefit include distinguishing epithelioid blue nevus from blue nevus-like cutaneous melanoma metastasis [86], distinguishing intranodal nevus from metastatic melanoma [87], classifying superficial melanocytic neoplasms with pagetoid melanocytosis [88], and the microstaging of melanoma arising in association with a nevus [89]. Several commercial laboratories now offer this test for use in formalin-fixed, paraffin-embedded tissue using probes developed by Abbott Laboratories (Caris Life Sciences, Neogenomics Laboratories).

There has been abundant research aimed at identifying biomarkers for prognosis as well as treatment response. The widespread availability of immunohistochemical staining makes this technique one of the most widely investigated, because it can characterize patterns of protein expression while maintaining preservation of tissue and cellular architecture. There have been literally thousands of manuscripts devoted to potential molecular markers of melanoma that could identify patients at greatest risk for recurrence, or most likely to benefit from adjuvant therapy. Two recent systematic reviews of published data on immunohistochemistry-based prognostic molecular markers for cutaneous melanoma have been performed [90, 91]. The first of two reviews sought to report on associations between immunohistochemical expression of protein biomarkers and survival outcomes in cases of cutaneous melanoma. Molecular markers were then scored according to the protein’s major biological function, as categorized by a modified Hanahan-Weinberg functional capabilities scheme. This review identified effectors of DNA replication, cyclin-dependent kinase inhibitors, growth-promoting transcription factors, and regulators of tissue invasion and metastasis as having a statistically significant impact on patient outcome. Ki-67, proliferating cell nuclear antigen, metallothionein, Ku70, Ku80, and microtubule-associated protein-2 were consistently associated with disease-free as well as overall survival. Interestingly, cyclins and cyclin-dependent kinases were not associated with melanoma prognosis. More recently, a significant association between Ki-67 expression and melanoma-specific mortality—in addition to dermal mitoses, which remain a significant and independent prognostic factor—has been shown in a large and representative population of patients diagnosed with thin (≤ 1 mm in thickness) cutaneous melanomas [92]. With external validation in a similar population-based sample, it is possible that Ki-67 may be formally adopted into routine clinical practice, either as a prognostic factor for disease recurrence or a predictive factor of positive sentinel lymph node biopsy.

Expression of cyclin-dependent kinase inhibitors was significantly associated with mortality. For example, increased levels of p16 were associated with decreased all-cause and melanoma-specific mortalities, while increased levels of p27, which typically cause cells to arrest in the G₁ phase of the cell cycle, are paradoxically associated with increased mortality. The rationale behind this observation lies in the activation of p27 through cytoplasmic sequestration, rather than protein degradation. Cytoplasmic accumulation of p27 has been associated with metastatic potential [93]. Among proteins involved in growth signaling, statistically significant associations were most frequent among transcription factors, ATF-2 and AP-2α, and transcriptional coactivators, NCOA3/AIB-1. This observation suggests altered transcriptional regulation as a crucial step in regulation of melanoma-specific survival. Increased expression of three cellular adhesion molecules was shown to be associated with statistically significant worsening of disease-free survival: the melanocyte-specific melanoma cell adhesion molecule (MCAM/MUC18), neuron-specific L1 cell adhesion molecule (L1-CAM), and glandular tissue-associated carcinoembryonic antigen-related cell adhesion molecule 1 (CEACAM-1). These molecules are thought to promote tumor infiltration and distant spread through abnormal tumor-stroma interactions [94–97]. Also, elevated levels of osteopontin and tenascin-C—members of the matricellular protein family, which consists of secreted molecules that interface between cell surface receptors and the extracellular matrix [98]—were associated with statistically significant worsening of melanoma-specific mortality and disease-free survival, respectively. Among proteases, increased levels of tissue plasminogen activator and matrix metalloproteinase-2 were statistically significant for mortality outcomes.

There are several major mechanisms involved in inherited forms of BCC: defects in detoxifying proteins, DNA repair, and embryonic signaling pathways. Defects in these pathways cause well-recognized skin cancer syndromes, including xeroderma pigmentosum, Gorlin syndrome, and familial cylindromatosis, and play a role in sporadic BCC as well. One of the most important components that protect the skin against oxidative stress after UV irradiation is the glutathione S-transferase (GST) enzyme system [105]. In humans, GST is expressed in melanomas, but only weakly in BCC [106]. Several GST polymorphisms have been identified in patients with high UV sensitivity and increased risk for BCC [105]. Impaired DNA repair, especially in patients with xeroderma pigmentosum, predisposes to non-melanoma skin cancer, preferentially of the squamous cell carcinoma (SCC) type (discussed later in this chapter). Of the embryonic signaling pathways, the most important is the hedgehog (Hh) signaling pathway.