In 1948, Dr. Sophie Spitz detailed the clinical and histopathologic observation of 13 melanocytic lesions which she termed “juvenile melanomas”, (1) Though these lesions greatly resembled melanomas histologically, the young age of the patients and the clinical course captured the attention of Dr. Spitz, concluding that they “probably would not behave as malignant tumors.” (1) This correct observation sparked a larger interest in these lesions, marking the importance of distinguishing them from melanoma and thus avoiding overtreatment in these young patients. Later observations of these “juvenile melanomas” in adults shifted the concept of Spitz nevi beyond purely pediatric lesions (2). Subsequently, cumulative experience allowed for description of the epithelioid and spindle cell melanocyte morphology as a distinct features of these lesions, laying the ground for a wider recognition of these lesions in daily practice by pathologists globally (3).

These histologic features have also inspired these lesions’ nomenclature which has evolved over time. There is indeed a relatively long list of synonyms, such as “spindle and epithelioid cell nevus”, “spindle-cell nevus”, “epithelioid cell nevus” and “nevus of large spindle and/or epithelioid cells”. The now most commonly used name of “Spitz Nevus” (SN) was established in 1967 by the Pathology Committee of the Queensland Melanoma Project, using an eponym honoring and marking the legacy of the late Dr. Spitz, prematurely passed away a decade earlier (4).

With the better characterization of Spitz nevi as distinct melanocytic neoplasms, the first variant, the pigmented spindle cell nevus, was then recognized and described in 1975 (5). Subsequently, several more variants of SN, including the desmoplastic, pseudogranulomatous, and a variant with halo reaction were described in 1978 (6).

Given the difficulty of differentiating between benign SN and melanoma, along with multiple reports of benign SN displaying atypical features, the concept of an intermediate lesion emerged (7). Reed et al. classified these borderline lesions as “minimal deviation melanomas” (5). The terminology and designation of “Atypical Spitz Nevus”, currently in use, emerged more recently, in the 1990s (8). Around the same time, initial efforts through molecular cytogenetic analysis culminated, in 1999, with the characterization of Spitz nevi with 11p gain, by Bastian and colleagues (9).

Following the discovery of BRAF mutations in cancers, including melanoma, it was found that benign Spitz nevi do not contain actually BRAF mutations, marking an important distinction of these tumors with other melanocytic lesions (10–12). More recently, Wiesner and colleagues found that kinase fusions and translocations are commonly found across the spectrum of Spitz neoplasms (13). In-frame kinase fusions and translocations of ROS1, NTRK1, ALK, BRAF, and RET were described, and will be described later in this review (13).

In 2016, NTRK3 kinase fusions were added to the repertoire of kinase fusions found in Spitz melanoma and the effort to find additional molecular defects continue to this day, as an estimated 20-50% of Spitz tumors remain with no known genetic mutations (14, 15). Most recently, MAP3K8 translocations have been identified as an additional genetic driver (16, 17).

A historical timeline of the evolution of our understanding of Spitz tumors is presented in Figure 1 .

Thus, the current classification of Spitz tumor includes (a) benign Spitz nevi (SN), (b) atypical Spitz nevus/tumor (AST) as tumor of intermediate malignant potential, and (c) Spitz melanoma, a distinct subgroup of Spitzoid melanomas (SM). In defining “Spitz”, vs. “spitzoid”, the WHO 2018 guidelines define “Spitz” lesions as being associated with at least a single genetic alteration and “spitzoid” lesions as lesions with morphological resemblance to SN (18). Currently, only a little over half of spitzoid tumors are associated with a known genomic rearrangement or an HRAS mutation, and are therefore formally defined as “Spitz”, creating a problem of categorizing the remaining lesions (15). Given recent advancements in discovering new genomic aberrations, it is likely that more of these remaining spitzoid lesions will eventually be classified as “Spitz” based on a currently unknown genomic aberration.

Detailed epidemiologic studies of Spitz tumors are difficult to perform, and few are available in the literature. First, these lesions are rare, with only 1-2% of all melanocytic lesions in all ages diagnosed as SN (18). Furthermore, even with the assistance of ancillary studies, there is generally poor concordance between dermatopathologists on the diagnosis of these lesions (19–22). However, correlations between lesion morphology with recently characterized genetic oncologic drivers can assist in identifying these lesions, facilitating accuracy of future epidemiologic studies (23).

A few estimated incidences of Spitz nevi have been reported. In 1977, Weedon and Little estimated an incidence of 1.4 cases per 100,000 individuals in Queensland, Australia (24). An incidence of 7 cases of Spitz Nevus per 100,000 individuals in Connecticut, United States has also been reported (25). More recently, an incidence of 10 cases per 100,000 individuals in the region of Emilia, Italy has been published (26). As a general estimate, 1-2% of all nevi removed from both children and adults are diagnosed as Spitz nevi (18).

The incidences of AST and SM are also unknown, though both tend to be less common than SN (18). A recent study on the yearly incidence of SN, AST, and SM in the Netherlands between 1999-2014 notably demonstrates the yearly number of SN increasing from 525 cases to 751 cases during this period; AST from 9 to 153 cases; SM from 8 to 40 cases (27). However, the general increase in incidence as well as the significant increase of AST in this study’s cases may not be due to a real increase but rather may be attributable to better recognition and clinical/histopathological/genetic classification of these lesions (27).

Currently, while the pathogenesis and etiology of SN, AST, and SM are unknown, the discovery of kinase translocations and other mutations hints to potential pathways involved in their pathogenesis. In fact, kinase fusions lead to a loss of a 3’ regulatory domain, leading to constant activation of the kinase domain portion of the fusion; this molecular event is likely responsible for cellular proliferation in these lesions (28). Described translocations include those involving ROS1, ALK, RET, BRAF, NTRK1, MET, and NTRK3 (28). We will describe later, in more detail, each of these translocations.

Though initially termed “juvenile melanomas”, Spitz tumors occur in individuals of all ages. The mean age of diagnosis of SN in studies involving both pediatric and adult populations ranges from late-teens to early 30s, with one study reporting a mean age of 22 years (29–31). Given the later categorization of AST separate from SN and the difficult nosologic interpretation, very few studies report an average age of AST diagnosis exclusively. A mean age of 20.8 years in 55 cases of AST, has been reported, which falls within the general range of mean age of SN diagnosis (32). The mean age of SM diagnosis does tend to be higher, as expected, with a mean age of 55 years in a 54 patient study (31).

Other authors have exclusively studied pediatric populations. In a series of 512 SN and 107 AST in this younger population, the median ages for each were 7.4 and 7.2, respectively (33). Though data is limited due to the relatively rare nature of SM in a pediatric population, the mean age tends to be higher at 12.9 years (34).

SN, AST, and SM occur in all races, though the majority of cases have been reported in white patients (33, 35). SN and AST occur in both sexes; studies are split between demonstrating an equal distribution versus greater incidence in females (29–31, 36–38). Though more cases of SN and AST occurring in females are reported in the literature overall, potentially increased cosmetic attention and motivation to remove the lesion may explain, at least in part, this predominance (18, 31, 36). The WHO 2018 classification mentions Spitz Melanoma as more common in men than women; in other studies, SM appears to exhibit a relatively equal distribution between sexes, though this conclusion is limited by small sample sizes (18, 31, 35, 39).

SN, AST, and SM can occur anywhere in the body. The lower extremities tend to be the most common site of occurrence followed by either the trunk or upper extremities for all three types of Spitz tumors (29, 31, 36, 40). A study comparing 110 cases of SN with 55 AST showcase this localization; for SN, 47.3% were localized to the lower extremity, 20% the upper extremity, 19.9% the trunk, 4.5% the head and neck, and for AST, 40% were localized to the lower extremity, 23.6% the trunk, 16.4% the upper limbs, 5.5% the head and neck (32). A similar localization is seen in SM, with 50% located on the lower extremities, 22.2% located on the trunk, 20.4% the upper extremities, and 7.4% the head and neck (31). An illustration of Spitz tumor localization is presented in Figure 2 .

Some studies of exclusively pediatric populations have observed higher incidences of SN and AST located in the head and neck (33, 35). In a recent study, 41.97% of 243 SN cases, 40.91% of 22 AST cases were biopsied from the head and neck (35). Due to the rarity of SM in an exclusively pediatric population, there is no consensus on where these lesions tend to manifest (33–35). Therefore, clinicians should consider potential differences in SN and AST localization in a pediatric population and remain vigilant in including SM as a differential at all sites.

Sophie Spitz described varying clinical appearances of these Spitz lesions. Most lesions were less than 1 cm. in diameter, all were elevated, and had a wide range of coloration from pink to black (1).

Conventional Spitz nevi are on average <6 mm. in diameter (18). This assertion is mostly supported in the literature; though one study does describe a mean of 76 mm (31, 38, 41). SN are classically described as dome-shaped, but may present as flat or polypoid lesions (18, 26). A pink to red color is most commonly seen in SN; flesh-colored, brown, and black lesions have also been observed (35, 37, 42).

AST are often in a range between 5 and 10 mm in diameter; a larger diameter in conjunction with appropriate histopathologic features can be a distinguishing clinical feature between AST and SN (43). Shape-wise, AST may either present as a plaque or a nodule (18). Like SN above, AST can also present in a variety of colors ranging from pink to black and like SM may be multi-colored (18).

SM are typically larger than AST, with a mean of 1.05 cm, however, SM <6mm can also occur (31, 41). SM tumors are more likely than SN to be nodular, but may also present as flat or only slightly elevated (18, 44). SM also exhibit a spectrum of color; analyzing 15 pediatric SM cases, there is a mean number of 2.7 colors per lesion, with 80% containing red/pink, 53.3% containing dark brown, 35.7% containing grey, and 33.3% containing light brown (34).

Representative clinical pictures of Spitz lesions are illustrated in Figure 3 .

The majority of SN are symmetrical, well-circumscribed, and pigmented (18, 30, 36, 37, 59). SN are divided into three histopathologic subtypes: compound, junctional, and intradermal, based on melanocyte location. Compound SN tends to be the most common subtype and contain both junctional or intradermal elements ( Figure 4A ) (30, 31, 36, 40). SN are synonymous with “spindle-cell and/or epithelioid cell nevus”, and as the name implies, SN are made up of large epithelioid and/or spindle melanocytes, the predominance of which varies (29, 30, 36, 38, 59). These cells are usually found in collections of nests at the dermo-epidermal junction, and are usually oriented perpendicular to the epidermis, in a configuration frequently described as “raining down” or “hanging” from the epidermis, like bunches of bananas ( Figure 4A , Inset) (42, 50, 59). Epidermal hyperplasia is commonly observed in SN (36). Kamino bodies are often located at the periphery of melanocytic nests within the epidermis ( Figure 4A , Inset) (50).

Although Kamino bodies are an important distinguishing diagnostic feature of SN, they are not necessarily found at the same 60% frequency Kamino et al. reported (60). In the literature, detection of Kamino bodies in SN lesions range from being found in 11% to 67% of lesions (29, 30, 36, 41). Presence of lymphocytic inflammatory infiltrate and mild to heavy pigmentation of SN are quite common (24, 30, 36, 37, 41). Maturation is also detected at high frequencies in compound and intradermal SN (30, 36, 38). Ulceration and abnormal mitotic activity are rare in SN compared to AST and SM (24, 37, 41, 61). Pagetoid spread is also relatively uncommon in SN, and when observed, is typically focal, involves the lower epidermis, located at the center of the lesion, and frequently present as cluster of cells rather than single cells (18, 30, 36, 59).

Multinucleated or mononucleated giant cells were described in the basal layer of the epidermis and superficial dermis by Dr. Spitz, and were originally considered a defining characteristic separating SN from SM (1). Despite this assertion, several studies report only finding these giant cells in <30% of lesions (29, 36, 59). However, despite the low frequency in general, two studies do also find that the majority of these giant cells were found in SN, especially the ones composed of primarily epithelioid cells (29, 36).

Encompassing both low and high-grade tumors, AST is a more contentious diagnosis as it has characteristics of both SN and SM with uncertain malignant potential; many diagnoses are based on at least one or more abnormal histopathological findings. AST tend to be more poorly circumscribed and asymmetrical compared to SN (18, 43, 61). Comparing 50 pediatric AST cases with 20 pediatric SN cases, AST diagnosis was found to be significantly correlated with increased mitotic rate, presence of deep and atypical mitoses, nuclear pleomorphism, high nuclear/cytoplasmic ratio, absence of maturation, solid growth, and presence of asymmetry (61). Kamino bodies are even more infrequent in AST and may be smaller in size as well (43). Pagetoid spread in AST tends to be peripheral rather than focal, involves the upper epidermis, and shows distinct single-cell or small nested patterns (18). AST histology is represented in Figures 5 C–E ). Clonal p16 loss and heterogeneous dermal expression of HMB45 are also immunophenotypic features that AST may display (62).

There is no consensus in morphologically distinguishing AST from SM, with the designation more dependent on associated molecular pathology and retrospective diagnoses based on metastases development (18, 43). Although the ESP, EORTC, and EURACAN guidelines list morphological criterion on differentiating between a low and high grade melanocytoma, they also acknowledge AST and SM are commonly grouped into an intermediate grade and a clear cutoff of number of genetic alterations between the two has not been determined; TERT promoter alterations may help favor a malignant diagnosis (62) MAP3K8 fusions are more common in AST and SM than SN; recently, a morphologic correlation of epithelioid morphology, high grade cytologic atypia, presence of multinucleated giant cells, and p16 loss with AST harboring MAP3K8 fusions was described (17, 23). An example of AST histology is represented in Figures 5 C, D ).

Based on the current WHO 2018 guidelines, histopathologic diagnosis of Spitz melanoma or “malignant Spitz tumours” is based on abnormal histopathological findings (18). Compared to SN, Spitz Melanoma and SM are much larger, >1 cm, and like AST, display asymmetry, poor circumscription, ulceration, irregular nesting patterns, lack of maturation, increased mitotic rate (>6mm²), and deep or atypical mitoses (18, 63, 64). These lesions also display high rates of cytological atypia (41, 63, 64). Despite this criterion, these rare tumors do bear a striking histopathologic resemblance to benign SN and AST, as these lesions may also display abnormal characteristics, and misdiagnoses of SM for SN/AST and vice versa based on morphology is not an uncommon occurrence (9, 19, 37).

Molecular pathology is often used to confirm suspicion based on morphologic assessment. Genetic alterations associated with Spitz melanoma include kinase fusions, PTEN mutations, homozygous loss of 9p21, and HRAS mutations (18, 65). Because of the wide variance in genetic alterations in lesions morphologically categorized under SM, three subclasses of SM have been proposed: Spitz melanoma, spitzoid melanoma without identified MAPK activating mutations, and spitzoid melanoma non-spitz MAPK activating mutations (65). Spitz melanoma is defined as a morphologically diagnosed SM that also exhibits typical Spitz tumor-associated genetic alterations involving HRAS, ALK, NTRK1, BRAF, MAP3K8, and CDKN2A (65). When a non-Spitz mutation, such as a BRAF mutation, is identified, the lesion should be classified as SM; an example is illustrated in Figure 6 .

Some Spitz melanomas retain histopathologic features of SN such as loss of p16 in those with HRAS mutations, fascicular melanocyte nests and increased dermal mitoses in those with ALK fusions, and dermal rosette-like structures in those with NRTK1 fusions (65). Spitz melanomas with MAP3K8 fusions also exhibit similar morphology to AST with MAP3K8 mutations: epithelioid morphology, high grade cytologic atypia, multinucleated giant cells, and p16 loss (17, 23). SM categorized under both spitzoid melanoma without identified MAPK activating mutations and spitzoid melanoma non-spitz MAPK activating mutations display more dermal solar elastosis and a higher number of somatic mutations (65). Further characterization of Spitz melanoma and SM and correlation of subclasses with histopathologic characteristics may play a role in more accurate future diagnoses.

Immunohistochemistry (IHC) utilizes either monoclonal or polyclonal antibodies to detect specific proteins within a given tissue and is sometimes a useful tool in the diagnosis of melanocytic tumors (15, 67). IHC can detect cell cycle and/or apoptosis regulators such as Ki67, PHH3, Cyclin D1, and p16 and can be used in conjunction with melanocytic markers such as HMB45, S100/S100A6, MITF, Mart-1/MelanA, and SOX10 for assessing the general architecture of the tumor (15, 68–72).

One of the most commonly used markers, the Ki-67 protein (MIB-1), is expressed during the interphase and mitotic phases of the cell cycle (G1, G2, S, M) and is absent in the G0 phase, making it a useful marker to differentiate an actively proliferating melanoma from a quiescent melanocytic nevus (69). Compared with melanoma (where usually the ki67 proliferation index is >10% of melanocytes), SN tend to have a much lower Ki-67 proliferative index (<2% of melanocytes) and a lower dermal/epidermal count ratio (0.25 SN vs. 0.94 MM) (68, 73–75). Within Spitz tumors, SN and AST share similar expression patterns of Ki67 in the superficial dermis or at the junction, without any expression usually at the base of the lesion; AST tend to display a higher proliferative index (2-10% of melanocytes) (68). SM tends to have a high and diffuse expression throughout the entire lesion and like other melanomas tends to display a higher proliferative index (>10% of melanocytes) (68).

PHH3 (Phosphorylation of histone H3) is another helpful marker of mitotic activity that is associated with chromatin condensation in the G2 and M phases of the cell cycle (76). Although helpful in distinguishing melanoma from a benign melanocytic nevus, PHH3 does not demonstrate any advantage over and is less robust than Ki-67 (76, 77). This marker holds a similar expression pattern and proliferative index as Ki-67 for SN, AST, and SM (68). Another potentially interesting marker is Cyclin D1 which is overexpressed in more advanced lesions and may be useful as a potential differentiator of Spitz lesions; however, there are discordant and limited data in the literature (68, 78).

Loss or inactivation of CDKN2A on chromosome 9p is one of the most common molecular events in cutaneous melanoma, occurring in 50% of cases (79, 80). CDKN2A encodes the p16 protein, a cyclin-dependent kinase inhibitor that is vital to tumor suppression and whose absence promote cell cycle progression (80). SN stains strongly and diffusely for p16, but sometimes a mosaic pattern is also interpreted as a sign of benignity; on the contrary, complete absence or absence in sizable areas of the tumor, due to homozygous loss of 9p21, is an indication of loss of p16 expression and is a useful ancillary study to suspect melanoma (68, 81, 82). AST can present with either homozygous or heterozygous loss of 9p21 leading to some degree of variability in p16 positivity (80, 83).

HMB-45, Melan A, S100, MITF, SOX10, and tyrosinase are commonly used to determine melanocytic differentiation and aid in the diagnosis of melanocytic lesions (15, 84, 85). Although HMB-45 is not a particularly helpful as a lineage-specific marker, its staining pattern may help distinguish SN from SM. SN displays an organized, stratified HMB-45 staining pattern in contrast to highly variable staining frequency and distribution in melanoma, though heterogenous HMB-45 staining patterns in SN have been observed (69, 71). Commonly SN display a progressively diminished HMB-45 staining pattern; AST display either diminished or variable staining in the dermis; deep staining of the dermis is common in SM (18).

Melan-A, SOX10 and S100 tend to display identical staining patterns in melanocytic lesions, including SN (86). Subtypes of the family of S100 proteins may be of some help in distinguish different melanocytic lesions; in a study of 42 SN cases, all of these lesions strongly expressed the S100A6 protein compared with only 33% of 105 melanoma cases (72). Moreover, SN may display a diffuse pattern of S100A6 whereas melanoma appears to display a patchy distribution (21, 72). Another marker, MITF, appears to have a somewhat lower intensity in SN (18, 87, 88).

While IHC can be useful, no singular IHC stain can definitively diagnose a given Spitz lesion. Combinations of these different IHC stains may assist in algorithmically differentiating between SN, AST and SM and supplement histopathological differential diagnoses. In this context, a recent study proposes that if IHC of a lesion displays deep dermal cell proliferation through Ki67 in conjunction with complete loss of p16 and/or HMB45, the lesion is less likely a SN and more likely SM/AST; this staining pattern should prompt additional molecular investigations, such as FISH studies (80). Garola and Singh propose utilizing a combined score of p16, Ki-67, and HMB-45 (PKH score) to differentiate between SN, AST, and SM: a PKH score <4 in this study appears to be significantly associated with SN and AST, while a PKH score >4 is associated with a diagnosis of SM (70).

Array comparative genomic hybridization (aCGH) identifies copy number alterations by comparing the entire genome of a given lesion against reference DNA (15, 68). Copy number alterations are typically absent in SN, however, gains of 11p have been noted in desmoplastic SN and gains of 7q have also been reported (15, 68, 89). An AST with more benign features tend to have no more than one to two gains/losses of chromosomes, while a more atypical AST tend to have gains/losses of several chromosomes (15). No definitive single gains or losses should be used to define the malignant potential of a Spitz tumor (15). Most melanomas, including SM tend to display multiple copy number alterations (15). The limitation of aCGH is the inability to screen for other genomic aberrations, such as point mutations or balanced translocations; moreover, the input quantity of tissue can be another limitation, with false negative results when a specimen contains <50% tumor cells (15). Furthermore, aCGH is expensive, time-consuming, and labor-intensive technology, making it impractical to perform on all borderline Spitz tumors (15, 68).

Fluorescence in situ hybridization utilizes DNA probes labeled with fluorescent molecules to detect DNA sequences on metaphase chromosomes and/or interphase nuclei (90). With the identification of copy number gains and losses through aCGH, standardized FISH assay to assist in the diagnosis of melanoma have been developed and represent a viable assay to clarify the nature of problematic Spitz lesions (15, 90). The original standard melanoma FISH assay, currently still in use, test the following targets: CCND1 (11q13), RREB (6p25), MYB (6q23) and centromere 6 (15, 90). Efforts have been made to adapt this standard melanoma FISH panel to be more specific to spitzoid lesions. With the association of 9p21 homozygous deletions of CDKN2A with Spitz melanoma, a five probe assays (6p25, 8q24, 11q13, 9p21, CEN9) and a six probe assays (RREB1, CCND1, MYB, EP6, CDKN2A, and CEP9) have been developed (15, 80). FISH was found to be useful in stratifying risk in ASTs, with homozygous 9p21 deletions associated with significant aggressive behavior and intermediate risk of aggressive behavior in cases with 6p25 or 11q13 gains compared to FISH negative or 6q23 deletions (80, 91).

Though FISH is a more cost-effective measure in comparison to aCGH, FISH analysis of Spitz lesions tends to have a higher false negative rate (15, 68, 92). This is in part due to interobserver variability, thus FISH should be performed only by highly trained labs. Moreover, FISH analysis is also limited to only 4-6 genomic loci, therefore microdeletions or insertions and genomic rearrangements typically go unrecognized and/or undetected (15, 68). Therefore, a negative FISH test does not necessarily exclude the diagnosis of melanoma or, specifically a malignant spitzoid lesion. Furthermore, tetraploidy found in 5-10% of benign SN can confound FISH results with false positives for melanoma (68, 93).

Next generation sequencing (NGS) is a powerful sequencing tool that allows parallel sequencing of millions of small DNA fragments at a time and mapping of these sequenced fragments against the human reference genome allows detection of several abnormalities (94). Compared with aCGH and FISH, NGS assays are able to assess a wider variety of genomic aberrations including point mutations, microdeletions/insertions, genomic arrangements, and copy number alterations (15). From assessing only a selected panel of genes to the entire genome, NGS is easily scalable, making it an appealing option for assisting in the diagnosis of Spitz lesions. As the cost of NGS will decrease over time, it may become a more viable alternative to other molecular studies.

In a pilot study of a new NGS-based test for melanocytic tumors, 16 benign nevi (including 7 spitzoid melanocytic lesions, 4 “epithelioid” nevi, 5 blue nevi variants) were compared against 15 melanomas (95). NGS findings in SN included BRAF, MET, NTRK1, and ROS fusions and for SM included EML4 fusion, BRAF, MAP2K1, and TERT mutations (95). Comparing FISH results against NGS for one SN and 3 SM cases, NGS concurred with positive FISH tests for 2 SM cases, showed malignancy in one FISH negative SM case, and was benign in one FISH positive SN case (95). In the same study, NGS was also able to correctly diagnose all malignant lesions (95). Thus, NGS may be able to mitigate the problem of false negatives, as we sometime see in FISH studies, and it may help to clarify the malignant potential of a melanocytic spitzoid lesion.

NGS was recently used to reclassify SN and SM from melanomas of Spitzoid features (MSF) after the WHO’s classification of skin tumors proposed elimination of lesions with BRAF and NRAS mutations from the categories of SN and SM (96). Out of 80 SN and 26 SM, 12 were reclassified as MSF based on the presence of BRAF and NRAS mutations detected by NGS and 81% of SN and 76% of SM post-reclassification displayed chimeric fusions and/or truncations (96). The majority of MSFs contained RAF, NRAS, NF1, and TERT promoter mutations and were associated with worse recurrence-free survival and metastatic potential than SN and SM (96). With the incorporation of genomic data being highly predictive of recurrence and the reclassification of lesions previously diagnosed as SN or SM through NGS, there is utility in using NGS to clarify diagnoses of SN and SM from MSF, with important implication for patient management (96). As mentioned above, though costs are declining, NGS is still significantly more expensive than aCGH and/or FISH studies and thus is not yet widely available (15).

Overall, as we learn about additional molecular determinants of Spitz lesions, the use of all these molecular studies, in conjunction with standardized protocols, will become increasingly relevant in the correct interpretation and classification of SN, AST, and SM.

As described above, as techniques such as comparative genomic hybridization (CGH) and fluorescence in situ hybridization (FISH) have become commonplace, a paradigm shift has occurred in the evaluation of Spitz lesions (30). The advent of NGS has allowed for the sequencing of a growing number of Spitz lesions, uncovering relevant mutational profiles (Table 3).

Polypoid SN is a rare variant of SN that displays concerning features resembling polypoid melanoma (127–129). These lesions are pedunculated with large epithelioid or spindled melanocytes with prominent atypia including large nuclei and prominent nucleoli (127–129). Unlike polypoid melanoma, polypoid SN are typically monomorphic, negative for aberrant chromosome copy numbers, and mitotic figures are much more difficult to detect (127–129). Polypoid SN also demonstrate a mesenchymal component with increased stromal collagen, similar to the desmoplastic variant of SN and other benign melanocytic lesions (127, 129). A case of polypoid AST demonstrating asymmetry and moderate nuclear polymorphism was reported recently that demonstrated a homozygous deletion of 9p21 and a CLIP2-BRAF fusion (130).

Agminated SN are eruptions of multiple (>2) conventional SN in a disseminated or clustered “agminate” pattern, though cases of a dermatomal distribution have been reported (131–135). Three different categories have been described: “Grouped on grossly normal skin,” “Grouped on hypopigmented macule,” and “Grouped on congenital hyperpigmented macule.” (135) The disseminated pattern is rare and is more commonly found in adults; the agminated pattern, in particular agminated arising on hyperpigmented skin, is more common and is associated with children (133, 134).

The Pagetoid SN was first described by Busam and Barnhill in 1995 and is characterized by a prominent pagetoid distribution of epithelioid melanocytes typically located in the spinous cell layer of the lesion (136, 137). The melanocytes are individually distributed (136, 138). Melanoderma, mild dermal inflammation, epidermal hyperplasia, eosinophilic bodies can be seen in Pagetoid SN (138). Clinically, these lesions are small (diameter <6mm), symmetrical, usually located on an extremity, and are well-circumscribed (136–138). Young children are associated with Pagetoid SN, however, one case series found an average age of 34 years; due to the rarity of this variant, further studies are needed to clarify this demographic (136–138).

Superficial spreading melanoma in situ is a common differential for these lesions (137). However, the distribution of melanocytes in superficial spreading melanoma in situ is less orderly and confined than Pagetoid SN and are clinically larger and asymmetrical (137). FISH positivity for 6p25, 6q23, Cep6, and 11q13 may be helpful in diagnosis of superficial spreading melanoma in situ over Pagetoid SN (139).

Dysplastic/SPARK SN are a variant of SN that displays characteristics of both classic SN and Clark’s nevi (dysplastic nevi) (140, 141). In 1978, Clark et al. found that moles that exhibited “B-K mole syndrome”, now known as “dysplastic nevi” displayed atypical melanocytic hyperplasia, lymphocytic infiltration, and angiogenesis, increasing the risk of melanoma (140, 142). Analyzing 2164 compound melanocytic nevi, 67 (3.1%) lesions demonstrated epidermal and dermal features of SN (141). A clinical image of Spark SN can be found in Figure 3F .

Histopathologically, either spindle and/or epithelioid melanocyte nests are located in both a hyperplastic epidermis and dermis in an architectural pattern reminiscent of Clark’s nevi, including dermal fibroplasia, denser lymphocytic infiltrate, and cytologic atypia (140, 141) ( Figures 5A, B ). The junctional component in compound lesions displays lateral extension and the spindle cells, if present, are typically parallel to the epidermis (140, 141). Kamino bodies may also be found in Dysplastic/SPARK SN (140). Clinically, Dysplastic/Spark SN are small (<1 cm), symmetric, well-circumscribed and display a female predominance (140). Based on a study of 27 cases, mean age at diagnosis is 33 years old, with a range of 6-64 years; these lesions do not recur or metastasize (140).

Initially, the prominent stromal sclerosis, decreased cellularity, and increased deposition of collagen characteristic of Desmoplastic Spitz Nevus were thought to be age-related change in the natural history of the SN (5, 143, 144). However, better characterization of these lesions have led to the contestation of its classification as either a subtype of Spitz tumor or as a separate entity (5, 6, 144–146).

Histopathologically, Desmoplastic SN are often intradermal or compound with little junctional activity and composed of either large spindle and/or epithelioid melanocytes arranged in small nests or solitarily distributed between thick collagen bundles (6, 137, 144). The melanocytes display a variety of phenotypes including type A/B nevus cells, ovoid melanocytes, dendritic melanocytes, and Spitzoid melanocytes (144). ( Figure 4C ). Many of these lesions are amelanotic though variable levels of melanin have been observed and mitotic figures are uncommon (137, 145, 146).

Clinically, Desmoplastic SN are asymptomatic lesions that may present on the head and neck or the extremities (137, 144, 145). They average 3.5 mm in diameter and are typically symmetric, wedge-shaped, and well-circumscribed (144, 147). Because of the striking fibrosis and sparse melanin found in these lesions, they are often confused for fibromas and desmoplastic melanoma (137).

Although desmoplastic melanoma is high on the differential, there are several distinguishing characteristics. Desmoplastic melanomas are typically larger, asymmetrical without sharp circumscription, and located in sun-exposed areas (137, 147). Desmoplastic melanomas have deeper involvement, with the melanocytic nests located deeper than those found in Desmoplastic SN (137, 145). The melanocytes in Desmoplastic melanomas display a fibroblastic-like appearance with hyperchromic nuclei (145). Small lymphocyte aggregates in the dermis are found in both Desmoplastic SN and Desmoplastic melanoma, however, these are typically accompanied with dermal mitotic figures in Desmoplastic melanoma (137, 147). HRAS expression and/or low Ki-67 positivity on immunohistochemistry or isolated gain of 11p make a diagnosis of Desmoplastic SN more likely (137, 147).

Hyalinized SN is a variant of SN and may even be considered a variant of the Desmoplastic SN (137). Extensive and prominent stromal hyalinization distinguish hyalinized SN as its own entity (148, 149). These lesions may be composed of spindle or epithelioid melanocytes that may be arranged in nests, a single-file arrangement, or as solitary cells in the setting of their characteristic hyalinized stroma (148, 149). Clinically, these lesions are small, symmetric, tan-pink, raised, and can be found anywhere on the body (148, 149).

Angiomatoid Spitz Nevus is a rare benign variant of SN with less than 10 cases reported (150, 151). These lesions are intradermal with no epidermal involvement and are characterized by stromal fibrosis, sparse proliferation of spitzoid melanocytes, and have abundant small, thick-walled blood vessels (137, 150, 151). Because of the numerous blood vessels present in this lesion, it may be confused for an angioma; additionally, the hypervascular stroma may imply a diagnosis of regressed melanoma (137).

Because of shared clinical and histopathologic features, Angiomatoid Spitz Nevus may be considered a subtype of Desmoplastic SN (137, 150). The presence of solitary melanocytes rather than nests, stromal fibrosis, low cellular density, and asymptomatic clinical presentation are common between both types of lesions (150, 151). High vascularity and lack of spindle cells with hyperchromatic nuclei distinguish Angiomatoid SN from Desmoplastic SN (150).

Although combined nevi are commonly associated with blue nevi, combined SN is a relatively rare variant of combined nevi displaying populations of spitzoid melanocytes combined with one or more other distinct nevomelanocytes (152–154). Typically, Combined SN are composed of either large epithelioid or spindle spitzoid melanocytes combined with either banal nevus, dysplastic nevus, or blue nevus cells (152–154). Clinically, combined SN occur in patients of all ages and display a female predominance (152, 153). These lesions are small, symmetrical, and well-circumscribed (152).

Histopathologically, the spitzoid melanocytes are typically located more superficially and are accompanied by other characteristics of SN including epidermal hyperplasia, clefts between nests of melanocytes, and lymphocytic infiltrates (152, 153). The other nevomelanocytic populations may be located laterally to the spitzoid melanocytes, above the spitzoid melanocytes, or may be intermingled with spitzoid melanocytes throughout the lesion (152). Though combined SN is uncommon, more atypia, including cytologic atypia, mitotic activity, and deep dermal infiltration accompanied with increased cellularity are found in these lesions than in other common types of combined nevi (153). Dermal sclerosis may also be observed (153).

With this combination of relatively high rates of atypia and dermal sclerosis, invasive melanoma is high on the differential for combined SN (152, 153). Careful identification of deep dermal mitosis is important in differentiating between combined SN and melanoma (153). Melanoma will typically be larger, asymmetrical, and display poor circumscription (152).

Myxoid Spitz Nevus is an exceedingly rare variant of SN, with fewer than 10 cases published in the literature (36, 155–157). First described in 1990, this variant presents with a myxoid stroma and mucin deposits may be present in the stroma between either isolated or nests of melanocytes (155–157). Clinically, these lesions are small, symmetrical, and may be compound or junctional (157). One study detected Kamino bodies in four out of six cases of Myxoid SN (157). These lesions bear a strong resemblance to malignant myxoid melanomas and care must be taken to rule out malignancy (157, 158).

SN with Halo Reaction is a rare variant of SN. Generally, a “halo reaction” is a finding of dense lymphocytic infiltrates throughout a part of or the entirety of a lesion histologically but may lack a clinical halo of depigmentation visible to the naked eye (159). A dermal inflammatory reaction consisting of either lymphocytes or eosinophils or plasma cells infiltrating the symmetrically distributed nests of epithelioid and/or spindled melanocytes is characteristic of SN with Halo Reaction (6, 159, 160). The halo reaction in SN with Halo Reaction may occur in conventional SN or in combined SN; though there may be a slight predominance of halo reaction in combined SN (159–161).

In conventional SN with halo reaction, characteristics of SN are preserved including epidermal hyperplasia, Kamino bodies, and clefts between melanocytes and keratinocytes (159). The “halo reaction” is symmetrically distributed throughout the whole thickness of the lesion, stopping at the dermal-epidermal junction (159). These lesions are pigmented, symmetrical, sharply circumscribed, wedge-shaped, and commonly do not present with clinical halos although the possibility does remain (159–161). Few mitotic figures are observed (159).

In compound SN with halo reaction, a conventional SN phenotype was commonly seen combined with common compound or intradermal nevus or with superficial congenital-type nevi (159, 160). In general, the “halo response” is confined to the SN component of the lesions, with the lymphocytes symmetrically distributed within the SN component (159). Deep mitotic figures are absent in these lesions (159, 161). Although compound SN with halo reaction may appear asymmetrical, raising concern for melanoma, a distinguishing characteristic in compound SN with halo reaction is that each distinct nevomelanocytic component should appear symmetrical when analyzed separately (159).

Pigmented Epithelioid SN (PESN)are a variant of SN characterized by a heavily pigmented lesion composed entirely of epithelioid cells; it may also be considered the epithelioid counterpart to the Pigmented Spindle Cell Nevus of Reed (162, 163). The PESN is much rarer than the PSCN, with fewer than 10 cases in the literature (162, 163). Striking epidermal hyperplasia is characteristic of these lesions along with nests of epithelioid melanocytes present at the dermo-epidermal junction (162, 163). Morphologically, PESN appears similar to the epithelioid blue nevus; however, blue nevi do not typically feature a prominent junctional component (163).

Additionally, the PESN must be distinguished from the Pigmented Epithelioid Melanocytoma (PEM); both lesions present with abundant heavily pigmented epithelioid cells with junctional nests (163–165). No specific genetic alterations have been associated with PESN yet; PEM is associated with fusions in PRKCA and inactivation of PRKAR1A (164, 165). Complicating the matter, PRKAR1A-inactiated PEM have recently been found to harbor genomic anomalies associated with other melanocytic tumors, including Spitz tumors, therefore, PRKAR1A inactivation may be a secondary genetic event in a broad group of nevi, including Spitz nevi (165). Hence, the question of whether the PESN is its own entity or some variant of PEM represents an area for further investigation.

Tubular SN is a rare variant of SN featuring prominent tubular or microcystic structures in the setting of aggregates of cuboidal epithelioid cells (166). These lesions do not have mitotic figures, cellular atypia, or inflammation (166). Tubular SN stains positive for the S-100 and NKI/C3 melanocytic markers on immunohistochemical analysis (166). Clinically, tubular SN resembles SN as a small, pigmented or pink, symmetric lesion (166). Although accepted as a variant of SN, there is conjecture that the “tubular” structures observed in this lesion are merely an artifact of formalin fixation, and therefore, tubular SN should not be considered a separate variant (167).

Pseudogranulomatous SN is characterized by a dense lymphocytic infiltrate surrounding nests of epithelial melanocytes located at the dermal-epidermal junction (6, 36, 168). The epidermal hyperplasia characteristic of SN is often accompanied with collarette (6, 168). Pyogenic granuloma is high on the differential for these lesions and clinically, they present as red, soft, sessile papules in young patients (168). Pseudogranulomatous SN stain positive for S-100 and HMB-45 and CD45 and CD3 can be detected in the lymphocytic infiltrate (168).

Recurrence is rare in SN, however it is known that melanocytic nevi may recur post-excision (169). The reported mean time to recurrence is 13-17 months and the average age reported is 14.7 years (169, 170). Stern’s analysis of 16 recurrent SN cases found that recurrent lesions were intradermal, featured prominent fibrotic stroma, and the clusters of melanocytes extended deeper into the deep reticular dermis and/or the subcutaneous fat (170). Four general types of recurrent SN have been characterized: a pseudomelanoma-like lesion with intraepidermal involvement, a typical SN lesion arising on a scar, a melanoma-like lesion with nodular growth pattern, and a desmoplastic SN-like lesion (169). Greater cytologic atypia, deeper mitoses, asymmetry, pagetoid spread, and lack of maturation may also be observed in recurrent SN (169). Clinically, these lesions may be larger than the original SN and extend beyond the biopsy scar; they may be either nodular or elevated (169). Immunohistochemical staining of recurrent SN for Ki-67 and HMB-45 may display identical reactivity patterns to SN (low proliferation Ki-67 and strong dermal HMB-45 staining), and may be helpful in distinguishing between recurrent SN and melanoma (169).

The Reed Nevus or pigmented spindle cell nevus (PSCN) was first described by Reed et al. in 1975 (5). Its designation as a separate entity vs. a SN variant remains controversial, but is currently considered by the 2018 WHO Classification to be a “distinct variant of Spitz naevus.” (18) Like SN, PSCN are small (mean 3 mm), benign lesions that primarily occur on the lower extremities of young adults (mean 25 years), with a predilection for women (18) (171–173). A clinical image of PSCN can be found in Figure 3B . PSCN are well-circumscribed and symmetrical papules or nodules with dark brown, bluish-black, or black coloration ( Figure 4B ) (18, 171). Unlike SN which can consist of primarily epithelioid cells, PSCN are characterized by spindle-shaped, uniform, finely pigmented cells that are either vertically oriented, whorled, or concentric fascicles ( Figure 4B , inset) (18, 171, 173). Epidermal hyperplasia is also common (171, 173). Pigmented Kamino bodies may be observed in the lower epidermis and were found in 12% of lesions (18, 171).

In addition to symmetry and good circumscription, compared with superficial spreading melanoma, PSCN have a low mitotic count, less lymphocytic inflammation nest uniformity, more epidermal hyperplasia and pigmentation, and moderate/marked presence of melanophages (174). As mentioned above, PSCN tend to exhibit a dermoscopic starburst pattern, that may later evolve and stabilize into a homogenous or reticular pattern (56, 175). A superficial black network, with a reticular appearance, due to increased melanin in these lesions is another dermoscopic feature common in PSCN (56). In a recent analysis of 23 PSCN, 57% of PSCN exhibited NTRK3 fusions compared to 3% of SN, AST, and SM combined (121). Complete excision of these lesions as treatment is recommended due to potential risk of melanoma misdiagnosis (171).

Like PSCN, atypical PSCN exhibit characteristic fascicles of uniform, spindle-shaped pigmented cells (176). However, they may exhibit asymmetry, poor circumscription, increased lymphocytic inflammation, ulceration, melanocytic hyperplasia spreading peripherally along the basal layer of the epidermis, abnormal mitotic activity, and other cytological atypia (18, 176). Higher numbers of epithelioid cells may also be seen in atypical PSCN (176).

The plexiform spindle cell naevus (PLXSCN) is a variation of the PSCN with a distinct wedge-shaped plexiform architecture of pigmented spindle cell fascicles found in the reticular dermis differentiating them from other spitzoid lesions (176–178). Reported mean age is 33.1 years, with a relatively equal gender distribution and are also most commonly found on the extremities (177). These lesions are slightly elevated and tend to be blue, blue-brown, or blue-black, often being mistaken for blue nevi or melanoma (177, 178). These lesions average 3mm and tend to also be symmetrical, well-circumscribed with low mitotic activity and mild cytologic atypia (177). Atypical PLXSCN exhibit more cytological atypia including greater mitotic rate, increased nodular confluence, and larger size (177).

The discovery that inactivating mutations of BAP1 (BRCA1 associated protein-1) lead to a familial predisposition for development of melanocytic tumors resembling AST sparked investigation of the loss of BAP1 in Spitz lesions (192, 193). This subset of AST featured loss of BAP1, often in conjunction with BRAFV600E mutations, and spitzoid morphology including epithelioid melanocytes in sheet-like growth with large vesicular nuclei (193, 194). However, many histopathologic characteristics of SN are not found in these lesions including epidermal hyperplasia, junctional supranest clefting, spindled melanocytes, and Kamino bodies (193, 194). As mentioned earlier, BRAF mutations are not considered a distinct hallmark of Spitz tumors. Therefore, since the molecular profile of BIMT is markedly distinct from those of conventional SN and can be associated with other common nevi, BIMT is now considered a separate entity and no more considered as a Spitz lesion (18, 194, 195).

Some melanomas may display a small component of spitzoid features including granular pigmented or glassy eosinophilic cytoplasm (137). Other MMs may arise in or be associated with a conventional SN (137). As both types of tumors typically exhibit the same behavior as melanoma, defining these lesions as SM is not recommended (137).

Nevoid melanoma refers to melanoma displaying morphology of conventional or congenital benign melanocytic lesions (137). These lesions may display some overlapping morphologic features with SM such as epidermal hyperplasia (137). However, NRAS and BRAF activating mutations are characteristic of these lesions which rules out a diagnosis of SM (137).

Currently, melanocytic lesions featuring MAP2K1 mutations are being studied to determine whether categorization as “Spitz” is appropriate (196, 197). These lesions range from benign to malignant and exhibit a diversity of phenotypes (196). Melanocytic Neoplasms with MAP2K1 lesions exhibit some spitzoid cytomorphology including large melanocytes featuring vesicular nuclei; some lesions also feature a prominent desmoplastic stroma as seen in Desmoplastic SN (196, 197). However, cases reported in the literature are few and other genetic features of these lesions, for example, lack of HRAS mutation or co-existing BRAF mutation, may argue against a diagnosis of a Spitz lesion (196, 197). Therefore, despite spitzoid features, it is currently difficult to definitively correlate a morphological diagnosis of a Spitz tumor with a MAP2K1 mutation.

Three cases of RASGRF1-rearranged spitzoid neoplasms have recently been identified (198). All three lesions displayed spitzoid morphology including nests of spitzoid epithelial/spindled melanocytes and epidermal hyperplasia; fusion partners included: CD63, EHBP1, and ABCC2 (198). Given that RASGRF1 is a guanine nucleotide exchange factor gene rather than a kinase-encoding gene, although these lesions are spitzoid in morphology, it is still too early to determine whether a RASGRF1 genomic aberration classifies this spitzoid lesion as a Spitz lesion (198).