HMB-45 antibody, which recognizes the gp100 protein, originally produced by Gown and Vogel in 1980s, defines an oncofetal premelanosomal antigen, positive in fetal melanocytes and melanoma but typically negative in adult resting melanocytes [13]. This marker is valuable in determining the biologic behavior of maturation. In DN, there is a progressive morphologic change in the melanocytes with increased depth. This maturation (senescence) is captured by the antibody HMB-45, which highlights melanocytes in a top-heavy pattern, with loss of staining with increasing depth into the dermis. Overall sensitivity is approximately 85 % but this significantly decreases in spindle-cell or desmoplastic melanomas [14]. In contrast, primary cutaneous melanoma usually expresses gp100 in a patchy pattern throughout the dermal component. HMB-45 can also be used to highlight the intraepidermal component and can label confluence/lentiginous proliferation, and suprabasilar spread of melanocytes, which are features characteristic of melanoma.

Mart-1, also known as Melan-A, is a small cytoplasmic protein, not localized to premelanosomes, initially identified as a target for cytotoxic T-cells [15] and expressed in adult resting melanocytes as well as melanoma. Staining for this melanocyte differentiation antigen has an overall sensitivity of ~85 %, greatest in large-cell, undifferentiated malignancies [16, 17]. Anti-Mart-1 antibodies are positive not just in melanocytes but in adrenocortical adenomas/carcinomas as well as sex-cord stromal tumors of ovary, which may be a pitfall in cases of metastasis of these malignancies to the skin [18]. Anti-Mart-1 antibodies can label confluence/lentiginous proliferation, and suprabasilar spread of melanocytes, highlighting their extent, which can help distinguish DN from melanoma. Also, as a potential pitfall, there is labeling of melanophages by anti-Mart-1 antibodies, which may represent melanocytic antigens that have been phagocytized by macrophages.

MIB-1, also known as Ki-67, is a proliferation marker of cycling cells. The practice of co-labeling the nuclear Ki-67 stain with a cytoplasmic melanocytic marker such as Mart-1/Melan-A, greatly improves the identification of proliferating melanocytes. In DN, Ki-67-positive cells are few (usually <5 %) and are typically located close to the dermoepidermal junction and adnexal epithelium, or in the superficial dermis, but are absent in the deeper portion of the lesion. In contrast, melanomas have a random pattern of immunoreactivity (average ~16 % in “hot spots”), with proliferating cells present at all levels of the lesion, especially at depth, indicating a lack of maturation/senescence [19].

The p16-INK4A product of CDKN2A is a cyclin-dependent kinase inhibitor, which has critical functions at the G₁-S checkpoint of the cell cycle, This enzyme blocks the cell cycle at the G1-S checkpoint by inhibiting CDK (cyclin-dependent kinases), including CDK4, and cyclins such as cyclin D1. This suppresses the proliferation of cells with damaged DNA or with activated oncogenes and is also activated when cells are old or crowded. The p16-INK4A protein is frequently inactivated in human tumors, including melanoma, and inherited mutations are associated with increased melanoma susceptibility [20, 21]. Common acquired nevi show minimal loss of p16-INK4A, while allelic loss of this locus is common in DN and in primary and metastatic melanomas [22]. The expression pattern can be nuclear or cytoplasmic.

Clark and colleagues proposed that failure of melanocytes to differentiate is necessary for dysplasia [23]. MITF is a nuclear transcription factor that regulates development, differentiation, and survival of melanocytes [24]. MITF plays a key role in the pathway leading to melanin production. Specifically, signaling initiated by alpha-MSH binding to the MCR1 transmembrane receptor results in MITF activation and subsequent transcription of genes necessary for melanin synthesis, including the key enzyme, tyrosinase [25], as well as other melanocyte-specific genes such as MART1 and SILV (silver homolog). MITF expression can also result in cell-cycle arrest by the induction of p16-INK4A [26, 27]. MITF is amplified or mutated in ~10 % of primary cutaneous melanoma and ~20 % of metastatic melanoma [28, 29]. MITF amplification occurs in tumors with poor prognosis, being associated with resistance to therapy [28]. There is paradoxically a decrease in genes regulated by MITF, including SILV, TRPM1 (melastatin), and MART1 in certain melanoma subsets and this is thought to accompany the progression from nevus to melanoma, as well as to be a poor prognostic indicator [30, 31].

5-Hydroxymethylcytosine is a recently described marker that correlates with the level of dysplasia [32]. This proves very useful in challenging lesions, including distinguishing DN from melanoma. Tumor cells in various human cancers exhibit global hypomethylation as well as selective hypermethylation at promoter regions of tumor suppressors, resulting in gene silencing and malignant transformation [33]. Progressive loss of 5-hydroxymethylcytosine was noted in one study to be associated with increasing levels of dysplasia, with the common acquired nevi expressing the marker to the strongest extent and near-complete loss in melanoma. Specifically, 5-hydroxymethylcytosine staining was highest in normal resident basal layer melanocytes, with 100 % staining darkly. Common acquired (non-dysplastic) nevi and low-grade DN (defined as those with mild or focally moderate cytologic atypia) showed 60 % of melanocytes staining darkly in this study [32]. High-grade DN (defined as those with diffusely moderate or severe atypia) ranged from 90 % lightly stained to 10 % negatively stained melanocytes. 5-Hydroxymethylcytosine exhibited near total loss in melanoma, being associated particularly with poor prognosis in superficial spreading melanoma and nodular melanoma [34]. In addition, increased nuclear size, a feature of dysplasia, had an inverse correlation with the expression of 5-hydroxymethylcytosine while being directly proportional to the degree of dysplasia [32]. Interestingly, DN showed darker staining in deep aspects of the neoplasm than the superficial aspect, likely highlighting maturation. In melanoma arising within a nevus, where delineating the extent of the melanoma for Breslow depth determination may prove challenging, there was a strong staining of 5-hydroxymethylcytosine levels in the nevus component and loss in the melanoma component [32].

SOX (Sry-HMG-box) proteins are a family of transcription factors involved in regulating a variety of biologic events including lineage restriction and terminal differentiation, through a precise pattern of expression that is cell-type specific [35]. SOX10 plays a key role in the transcriptional control of MITF, which is the master regulatory gene for melanogenesis [36]. However, SOX9, SOX18, and SOX5 have also been implicated in regulating aspects of the melanocyte life cycle. SOX9 (a nuclear stain), and SOX10 (a perinuclear and cytoplasmic stain) have been reported to be expressed in various stages of melanoma progression and in established melanoma cell lines [37–39]. Studies showed that SOX10-positive melanocytes were present in 31 % of nevi, 43 % of primary melanoma, and 50 % of metastatic melanoma [38, 39]. However, SOX9 expression was observed in a majority (~75 %) of the melanocytic neoplasms, with moderate decrease as the severity of melanoma progressed [40]. SOX9 expression has been shown to reduce proliferation of multiple melanoma cell lines [37]. The SOX protein expression in DN have not yet been fully elucidated and the combination of SOX 9 and SOX 10 expression patterns by immunohistochemistry may prove useful in better delineating the spectrum or grades of atypia that characterize DN and melanoma.

Angiogenesis and microvascular density (MVD) are important characteristics in tumorigenesis, with roles in the multifactorial transition from benign to malignant states [41]. These features have been shown to affect the prognosis of malignant tumors and skin neoplasms including melanoma [42]. Benign nevi (dysplastic and non-dysplastic) have similar MVD and mean major diameters of blood vessels. However, these parameters, in addition to total vascular area, are significantly increased in melanoma [43]. Therefore melanomas, unlike DN, have a greater number of vessels over a larger area, providing evidence for correlation of malignancy with increased vascularity.

Survivin is an antiapoptotic protein that has been detected in DN by immunohistochemistry. In one study, the majority of DN with a nuclear and cytoplasmic staining pattern for this marker had severe dysplasia [44]. A lack of nuclear staining was specifically described in benign melanocytic nevi, while 67 % of melanoma showed a positive nuclear stain, with an average index of 7 % [19].

Phosphohistone H3 (PHH3) is a proliferative marker that highlights cells specifically in the M-phase of the cell cycle. A study by Nasr et al. showed a lack of PHH3 expression in the dermis of compound and dysplastic nevi, however an average of positivity of 25 cells/10 hpf was noted in melanomas [19]. Use of PHH3 expression can be of value in identifying the “hot spot” of greatest mitotic index within a tumor.

Confocal microscopy uses point illumination via a pinhole to eliminate out-of-focus signals. The pinhole is conjugate to the focal point of the lens, allowing for optimal resolution [45]. Melanin offers the strongest contrast due to a high refractive index; therefore the cytoplasm of melanocytes is intensely white (Fig. 9.4). However, keratin has a lower refractive index and therefore less contrast, so keratinocyte cytoplasms appear darker. Nuclei appear dark and dermal collagen fibers appear very bright [46]. In vivo reflectance confocal microscopy (RCM) is a noninvasive tool that generates stacks of optical horizontal (z-axis) sections within the depth of intact living tissue. This proves to be a useful tool for studying the skin surface, and was first used in human skin in 1995 [47]. RCM enables visualization of the skin layers to a cellular level resolution (0.5–1.0 μm in the lateral dimension and 4–5 μm axially). The imaging depth is limited to 200–300 μm corresponding to depth at the dermoepidermal junction and papillary dermis, using the current commercially available RCM model [48]. An advantage of RCM is that the technology allows a section of skin to be assessed without processing (which may introduce artifact), and re-examination in order to evaluate dynamic changes over time. Characteristic architectural and cytologic features have been described, with histopathologic and dermoscopic correlates (Table 9.2).

In vivo RCM plays an important role in the characterization of the superficial aspect of melanoma, where a large number of characteristic findings are visible. However, Breslow depth determination, which has prognostic significance, is currently not possible. In the radial growth phase of melanoma, the epidermal pattern can be a disarranged honeycombed or cobblestoned pattern (i.e. an irregular epidermal pattern due to irregularly shaped keratinocytes) [49]. There can be suprabasilar/pagetoid migration of malignant melanocytes, evident as bright cells in superficial layers (Fig. 9.4). Atypical melanocytes are also noted along DEJ and in superficial layers, forming a nested and confluent proliferation (Fig. 9.3). In addition, atypical nucleated cells tend to infiltrate the dermal papillae and correspond histopathologically to nested melanocytic proliferation in upper dermis that invade and cause disarray of rete ridges. In almost 50 % of melanoma with microinvasion, regression is present, with an inflammatory infiltrate, and this is visualized by small bright inflammatory cells and plump bright cells (melanophages) with coarse bright collagen fibers [50]. In hypopigmented and amelanotic melanoma, confocal microscopy may still show features of melanoma due to melanin refractivity [49].

Biopsy of melanocytic lesions for histopathologic assessment is the gold standard but it interferes with the natural evolution of the lesion in vivo. Therefore histopathology-based assessments of the dynamics of the any given DN, either as a benign entity or as a precursor to melanoma, are limited. The advantage of in vivo observation in real time of tumor at the bedside is opening the clinical application of RCM for evaluation of melanocytic lesions and monitoring evolution, which is a necessary component to better understanding DN. Although limited by a small sample size, Pellacani et al. first demonstrated that histopathologic criteria i.e. architectural and cytologic features outlined by the Duke grading system had significant RCM correlates [51]. DN viewed by confocal microscopy were characterized in this study predominantly by a ringed pattern in association with a meshwork pattern, in addition to atypical junctional cells in the center of the lesion and irregular junctional nests with short interconnections (Fig. 9.3). In general, DN had cytologic atypia and atypical junctional nests, i.e. an irregular pattern with short interconnections and/or with nonhomogenous cellularity. However, pagetoid spread, widespread cytologic atypia at the junction, and non-edged papillae were suggestive of melanoma [51].

There were characteristic findings for common acquired nevi, melanoma, as well as the different grades of DN (Table 9.3). Suprabasal melanocytes most directly correlated with the level of dysplasia, i.e., 0 % in common acquired nevus and nevus with mild dysplasia 7 % in nevus with moderate dysplasia, 40 % and 100 % in nevus with severe dysplasia and melanoma, respectively [51]. Marked architectural disorder was observed in all but one melanoma (which showed severe cytologic atypia), in all severe DN, in 3 of 15 moderate DN, and in one common acquired nevus. Marked cytologic atypia was observed in some DN and melanoma but not in common acquired nevi or mildly dysplastic nevi [51].