The use of immunohistochemistry when evaluating melanocytic lesions may become necessary when the distinction between various types of benign melanocytic nevi and melanoma is not readily apparent upon examination of hematoxylin and eosin (H&E) stained sections. The standard antibodies used in dermatopathology practice for these types of lesions include S100, HMB-45 (anti-gp100), anti-MART-1, and MIB1 (anti-Ki67) [57–59].

HMB-45 and anti-MART-1 label intraepidermal melanocytes and thus the extent of upward pagetoid upward migration of melanocytes in the epidermis may be highlighted with either of these antibodies. Benign melanocytic nevi will demonstrate maturation sequence or loss of HMB-45 labeling with progressive descent into the dermis (Fig. 6.3). Exceptions to this pattern of labeling with HMB-45 in benign lesions include blue nevi and subset of Spitz nevi. These two lesions may demonstrate a diffuse labeling pattern with respect to HMB-45 (Fig. 6.3). In contrast, melanomas will consistently demonstrate a patchy pattern of labeling with HMB-45 (Fig. 6.3).

Evaluation of the proliferative rate of the dermal component of melanocytic tumor is measured by Ki-67. This marker also aids in the distinction between benign and malignant melanocytic proliferations. Benign melanocytic nevi (ordinary, Spitz, and dysplastic) demonstrate less than 1–10 % labeling with MIB1 (anti-Ki-67). However, melanomas demonstrated a proliferative rate up to 16.4 % with MIB1 and absence of orderly pattern of labeling in the dermis [60]. Double labeling with MART-1 (cytoplasmic, red reactivity)/Ki-67 (nuclear, brown reactivity) enhances the evaluation of proliferative index in melanocytes, and is particularly useful in lesions with a high number of background lymphocytes (Fig. 6.4) [59].

Macrophages may be reactive with anti-MART-1 and may be interpreted as invasive melanoma cells [61]. Melanocytic nevi uniformly label with anti-MART-1 whereas desmoplastic melanoma may demonstrate weak, variable expression for MART-1 antibody.

Chromosomal abnormalities have been recognized in melanoma for decades [62–68]. The advent of CGH enabled a systematic characterization of melanoma at the genomic level. In traditional CGH, DNA derived from a tumor of interest is labeled with a discrete fluorochrome and mixed together with an equimolar amount of reference DNA labeled with a different fluorochrome. This mixture is hybridized onto denatured metaphase chromosomal spreads in the conventional method or onto an array containing serially arranged portions of DNA representing the entire genome in the array platform. Chromosomal gains in the tumor sample are represented by increased tumor-specific fluorochrome at a specific locus while losses in the tumor exhibit increased reference-specific fluorochrome at that site.

A critical development in our understanding of the pathogenesis of melanoma came from the work of Bastian and colleagues who initially utilized CGH to describe chromosomal gains and losses in primary melanomas from 32 patients [69]. They identified that primary melanomas were typified by multiple gains and losses in discrete regions of the genome. Subsequently, CGH was applied to a series of melanomas (n = 132) and melanocytic nevi (n = 54) to demonstrate a distinct pattern of chromosomal aberrations in melanomas not present in nevi. Whereas greater than 95 % of melanomas demonstrated some degree of chromosomal copy number aberration, only a rare subset of melanocytic nevi (13 %, all of which were Spitz nevi) demonstrated a single isolated gain of the entire short arm of chromosome 11—a change not seen in any of the melanomas analyzed [70]. Furthermore, central to the distinction between melanomas and nevi (of varying sorts) as demonstrated by CGH is the fact that melanomas contain gains and/or losses in partial regions of the chromosome, whereas when nevi contain alterations, they typically involve whole chromosomes or whole chromosomal arms. CGH not only distinguishes between melanomas and nevi, but also demonstrated discrete, reproducible differences among melanomas of varying subtypes. In a landmark study, Bastian and colleagues further used CGH to demonstrate that melanomas of varying subtypes (grouped according to whether they arose on or sites with or without chronic sun exposure) demonstrated distinct patterns of genomic alterations, including the frequency of oncogene mutations as well as differences in chromosomal aberrations. These findings argue that melanoma is a heterogeneous disease, arising through different mechanisms, which are a function—at least in part—of their anatomic location and the degree of antecedent ultraviolet light exposure [22].

Because CGH simultaneously scans the entire genome for all possible copy number alterations, it has considerable promise as a diagnostic tool in melanocytic tumors. However, there are many practical issues that limit the utility of CGH in a routine diagnostic setting. First, it requires a relatively large amount of pure tumor cells which can be difficult to obtain in some lesions and could compromise a thorough histopathologic assessment of a lesion. Lesions with relatively few tumor cells or lesions with a high degree of admixed inflammatory or stromal cells are also less amenable to CGH analysis because the DNA obtained from the non-tumoral cells can obscure the chromosomal abnormalities present in the background tumor cells. In addition, the abnormality must be present in a sufficient number of tumor cells to be detectable in the analysis. Finally, the DNA obtained must be suitable for subsequent enzymatic manipulations for fluorescent labeling reactions [71, 72].

The limitations of conventional and array CGH necessitated the development of an assay that exploits the differences among melanocytic tumors described by CGH, but which is more amenable to routine clinical practice. FISH offers important advantages over CGH as a routine and practical clinical assay. FISH utilizes fluorescently labeled probes corresponding to discrete chromosomal locations that are hybridized to routine (5µm) formalin fixed paraffin embedded tissue sections.The number of fluorescent signals per nucleus (described as “dots”) is counted, and aberrations are described as a percentage of tumor nuclei carrying greater or fewer than two signals. FISH offers the ability to evaluate copy number aberrations at a cellular level and further, to correlate any changes in copy number at particular locus (corresponding to a specific fluorescent probe) with cellular morphology under conventional light microscopy .

The FISH assay for melanoma evolved directly from the CGH studies of Bastian and colleagues [73]. A retrospective analysis of the original CGH data from the series of 186 of primary melanomas and nevi identified 13 loci spanning 8 different chromosomes which in varying combinations best discriminated between melanomas and nevi. FISH probes corresponding to these regions were chosen and configured into different panels of probe combinations and applied to a series of melanomas and nevi to determine the optimum combination. Four probes were identified that appeared to yield the most effective distinction between melanomas and nevi: RREB1 or 6p25 (red dot), CEN6 or centromere 6 (aqua dot), MYB or 6q23 (yellow dot) and CCND1 or 11q13 (green dot; Fig. 6.5). The centromere 6 probe is used to quantitate relative gains (or loss) of the other loci on this chromosome. Further studies generated the following algorithm that best discriminated between melanomas and nevi: (1) more than 38 % of the tumor nuclei contain greater than 2 signals for 11q13 or, (2) more than 29 % of tumor nuclei contain greater than 2 signals for 6p25 or, (3) more than 55 % of the tumor nuclei contain more 6p25 signals than centromere 6 signals, or (4) more than 40 % of the tumor nuclei contain fewer 6q23 signals than centromere 6 signals. Positivity for any of these parameters was determined as a positive test for melanoma. Numerous studies have subsequently tested the applicability of this FISH assay as a proof of principle in the distinction between histopathologically unambiguous melanomas and nevi (Table 6.1) [74–77]. Additional studies have described the applicability of the FISH assay to different histopathological settings, including lentiginous junctional melanoma of the elderly [76]; blue nevus-like metastatic melanoma versus epithelioid blue nevus [78]; metastatic melanoma to lymph node versus intranodal nevus [79]; pagetoid melanocytosis as a distinctive subtype of melanoma [80]; nevoid melanoma versus benign melanocytic nevus [81]; conjunctival melanomas versus conjunctival nevi [82]; desmoplastic melanomas versus sclerosing melanocytic nevi [83]; and blue nevus-like melanoma versus cellular blue nevi [84]. From these preliminary studies, the overall sensitivity of FISH in distinguishing among melanomas and nevi of the indicated subtypes from various anatomic locations was ~ 80 % with an overall specificity of > 90 %. However, it is important not to overestimate the value of FISH in a routine diagnostic setting since an important and often understated limitation inherent to the above studies was their reliance on histopathologically unambiguous melanomas and nevi to determine the utility of FISH. In morphologically unambiguous lesions, an experienced dermatopathologist would rarely—if ever—exclusively rely on a FISH assay for diagnostic purposes. An important question for the FISH assay thus became whether the FISH assay was applicable to histopathologically ambiguous lesions. In their original study, Gerami et al. applied the FISH assay to 27 ambiguous lesions—6 of which developed metastases and 21 of which were “event free” with a follow-up period of at least 5 years [73]. Of these cases, all 6 primary tumors that eventually developed metastases were FISH-positive, while only 6 of the 21 cases without metastases were FISH-positive. The latter observation was interpreted as melanomas that were either cured by excision or inadequately followed to detect their eventual metastases, while the former was interpreted as a high degree of sensitivity of FISH in identifying bona fide melanomas .