The application of molecular techniques has much improved our knowledge of molecular abnormalities in melanocytic lesions. It is accepted that there are three cell signaling pathways in melanocytes critical for their survival and proliferation, including (A) mitogen-activated protein kinase (MAPK)/extracellular regulated kinase (ERK), (B) phosphoinositide-3 kinase (PI3K)-AKT/PTEN, and (C) receptors such as KIT, melanocortin-1 receptor, (MC1R) and glutamate receptor metabotropic (GRM3) receptor [1]. The most commonly detected mutations in melanoma include alteration in RAF. Of the three isoforms (A, B, and C), BRAF is the primary isoform in the RAS/MAPK pathway and the isoform most susceptible for mutation in several solid tumors. Mutation in the BRAF gene located on chromosome 7 accounts for approximately 80 % of activation mutations in benign nevi and 50–60 % in primary cutaneous melanoma [2–4]. The majority of mutations occur in exon 15 as a single point mutation with substitution from thymine to adenine (T to A) that converts valine to glutamic acid at the 600 position (BRAF V600E) and less frequently other including V600K and V600R [5]. When compared to the wild form, melanoma cells containing the mutation BRAF V600E demonstrate an almost 500-fold increase in activity.

A second group of genes with mutations detected in melanoma is the RAS oncogene family (NRAS, HRAS, and KRAS). Activated RAS recruits RAF, which activates MEK and ERK and promotes cell proliferation, differentiation, and survival. NRAS mutations in melanocytic nevi and melanoma most commonly occur in exons 2 and 3. NRAS mutations occur in 14–20 % of cutaneous melanoma (particularly nodular and lentigo maligna histologic subtypes) [6–8]. Alterations in HRAS are seen in nearly 30 % of Spitz nevi [9, 10]. Increased copy number of chromosome 11p (site of HRAS gene by CGH or FISH)/mutations in HRAS have been reported in over 20 % of Spitz nevi [11]. In contrast, spitzoid melanomas usually lack detectable mutations in HRAS [10].

NRAS and BRAF mutations are mostly exclusive; however, there are small percentages of melanomas which may harbor both NRAS and BRAF mutations (around 1 %) [8, 12, 13].

KIT is a tyrosine kinase growth factor receptor and KIT mutations and or increased copy numbers are seen in 15–20 % of acral-lentiginous/mucosal melanomas and lentigo maligna melanomas (chronic sun damage) [14, 15]. Most mutations in KIT in melanomas occur not only in exon 11, but also in exons 13 and 17 [16, 17].

GNAQ and GNA11 encode for the α-subunit of G-protein-coupled receptor. Benign and malignant melanocytic lesions show mutations at codon 209 with change of glutamine to leucine or proline (Q209L or Q209P). This results in constitutive activation of this G-protein. Such mutations have been detected in 83 % of blue nevi, 50 % of melanoma with blue nevus features, 46 % of uveal melanoma, and in 4 % of lentigo maligna [18].

Another gene mutated in melanocytic lesions is CDKN2A/INK4A in chromosome 9p21 which codes p16INK4a [19]. It is associated with familial melanoma and dysplastic nevus syndrome [20–22]. Within the last 10 years there have been descriptions of application of additional techniques applied to the detection of genetic abnormalities, particularly those described in this paragraph, that seem helpful in the diagnosis and treatment of melanocytic lesions.

In comparative genomic hybridization (CGH), DNA derived from a tumor of interest is labeled with a fluorochrome and mixed with a reference DNA labeled with a different fluorochrome. Chromosomal gains and losses are represented by relative changes of color. Bastian and colleagues initially described the use of CGH to detect multiple discrete gains and losses in primary melanomas [23]. In contrast, only rare nevi (mostly of the Spitz type) showed single isolated gains of the entire short arm of chromosome 11 [24]. Furthermore, in those rare nevi with alterations, these typically involved whole chromosomes or whole chromosomal arms. CGH also confirmed that the histologic subtypes of melanoma (superficial spreading type, acral-lentiginous, lentigo maligna) have different patterns of chromosomal aberrations.

As a drawback, CGH requires a relatively large amount of pure tumor cells. Thus small lesions or those with admixed inflammatory or stromal cells are less amenable to CGH analysis.

Fluorescence in situ hybridization (FISH) has been used as an alternative to CGH. It uses fluorescently labeled probes to be hybridized to formalin-fixed, paraffin-embedded sections. By counting the number of signals per nucleus, aberrations are described as a percentage of tumor nuclei with more or less than two signals. The current FISH assay for melanoma originated from studies of Bastian and colleagues [25]. The original four probes were: centromere 6 (CEN6) as a surrogate of the number of chromosomes, RREB1 (6p25), MYB (6q23), and CCND1 (11q13) (see also below). The cutoffs to determine gains or losses are different depending on the authors. Depending on the studies, the overall sensitivity of FISH in distinguishing between melanomas and nevi is around 80 % with an overall specificity of 90 % [25].

Some other studies have concentrated in ambiguous lesions to try to determine the feasibility of FISH when applied to the diagnosis of “difficult” cases [26, 27]. Gaiser et al. compared the FISH results with histopathologic assessment, array CGH, and clinical follow up. Overall, when comparing FISH results with clinical outcome there was a sensitivity of 60 % and a specificity of 50 % [26]. Vergier et al. characterized 113 ambiguous tumors according to clinical outcome [27]. Expert histopathologic review yielded a sensitivity and specificity of 95 % and 52 %, compared to 43 % and 80 % for FISH. Tetzlaff et al. [28] reported a specificity of 87.5 %, positive predictive value of 62.5 %, and negative predictive value of 80.7 %. Overall, these studies indicated that the combination of expert histopathologic review and FISH generated the best results. Also important is to consider a possible pitfall of FISH: the presence of polyploidy yields potentially false positive FISH results [29]. This is important because a number of Spitz nevi, with benign behavior, display polyploidy, so if the FISH results are misinterpreted the lesion may be labeled as “malignant” (we have encountered several such cases referred to our institution as melanoma). Therefore, whoever analyzes the FISH studies should make every effort to remove from the final counts those cells with increased copy numbers on all probes (including the centromere one) thus consistent with polyploidy. A newer set includes p16 (9p21) instead of MYB and, sometimes, an added MYC (8q24). Analysis of 9p21 appears to provide better distinction between Spitz nevi and spitzoid melanoma, since spitzoid lesions with aggressive behavior (i.e., malignant) show homozygous deletion of 9p21 [30]. Regarding 8q24, it seems that gains of this gene result in an amelanotic phenotype due to downregulation of MITF and tyrosinase [31], and may be associated with poor prognosis [30].

Regarding FISH, in summary, we believe that: (1) FISH is insufficient as an isolated diagnostic assay in the differential diagnosis between melanoma and nevus, (2) a positive FISH should not modify treatment in the absence of histopathologic confirmation, (3) a negative FISH test does not exclude melanoma, and (4) FISH should be applied with caution in desmoplastic/sclerotic (because it may be difficult to collect pure enough DNA from the tumor cells) and spitzoid lesions (for the possibility of polyploidy).

Another technique that may be helpful in the distinction between nevus and melanoma is mass spectroscopy to determine differences in the protein components of melanocytic lesions. The group of Dr. Lazova has applied this technique to spitzoid lesions and has determined that there is a different spectrum of proteins present either in the lesional cells or in the stroma of Spitz nevi and spitzoid melanomas [32]. Interestingly, of the 12 proteins that are differentially expressed between these two types of lesions, two of them are the almost ubiquitous vimentin and actin. Other studies are ongoing to study other types of melanocytic lesions, such as ocular melanoma [33].