We humbly submit this case report for pre-submission review to the KiKo community.  All comments are welcome, from typos to "you need a major revision".   We will respond to and edit the manuscript based on what you say!  Comments in this post receive 3x KiKoXP points for both commenting and upvotes on your comment.

Daniel Gonzalez, Jonhan Ho, Arivarasan Karunamurthy

Neurofibromas can share histological features with variants of melanomas.  When two patterns are juxtaposed to each other, it can be difficult to discern whether there is one tumor with two morphologies or two distinct tumors.  In the case of melanoma, the scenarios could significantly change staging. We present a case of an acral melanoma overlying an atypical neurofibroma.  An 89-year-old female presented with a toe mass.  Initial biopsy showed two morphologies - a superficial melanoma and an underlying, intertwined spindle cell component, with SOX-10 expression in both components.  FISH was performed separately within the superficial melanocytic component showing regional copy gains with CCND1 and RREB1 probes, while the spindle cell component did not show any copy number aberrations.  Next generation sequencing performed on the amputation specimen showed the superficial melanocytic component harbored a c-KIT mutation while the spindled component harbored a NF1 mutation. Taken together, these molecular results help prove that the superficial melanocytic component and deep spindled component are unrelated and have different origins. 

Pathologists occasionally encounter neoplasms with different morphologies in the same specimen.  Whether the two morphologies represent the same neoplasm with different appearances or distinct, concurrent neoplasms can be difficult to determine. Melanomas not uncommonly exhibit heterogeneous morphologies, including spindle cells and neural differentiation.  When the differential diagnosis includes a co-occurring neurofibroma, immunohistochemistry may not be helpful since melanocytes and neural cells are derived from the neural crest and exhibit similar immunophenotypes.  However an appropriate distinction is critical as measurement of Breslow’s depth may be significantly different between the two scenarios, possibly affecting the tumor stage.

An 89-year-old female presented with a toe mass.  On biopsy, an acral lentiginous melanoma with typical melanocytic morphology was identified in the epidermis (Figures 1, 2).  Scattered atypical melanocytic nests with similar cytology were also identified in the superficial dermis.  In addition, the melanocytic nests were surrounded by  a moderately cellular spindle cell neoplasm with bland cytology arranged in irregular fascicles and sheets with prominent vasculature. Mitotic figures, cytologic atypia or necrosis were not identified in the spindle cell component. Though cytologically different, a  definitive conclusion could not be made immunohistochemically as both superficial epidermal component and deep dermal spindle cell components exhibited SOX-10 positivity. Fluorescent in situ hybridization studies (FISH) studies were performed separately on both the components.  The epidermal/superficial nevoid component showed absolute gains with CCND1 (11q13) and RREB (6p24.3) probes while the spindle cell component was negative for copy number abnormalities. FISH results along with different histomorphological features raised the possibility of a superficially invasive acral melanoma co-occuring over an underlying neural tumor such as a variant of neurofibroma.  The subsequent amputation specimen showed similar histological and immunohistochemical findings (Figures 2-5) within both superficial and deeper components.  Sufficient tissue was available in the excision specimen for further molecular testing, and a targeted next generation DNA & RNA sequencing (NGS) studies was performed on each component using Oncomine v3 panel that includes 161 genes.  The superficial acral lentiginous melanoma component showed hotspot KIT (p.L576P) and TERT promoter (c.1-124C>T) mutations and the dermal spindled cell component did not show a KIT mutation but was positive for TERT promoter (c.1-124C>T) mutation. In addition a pathogenic NF1 (p.Y2240*) was also identified in the dermal spindled component that was not seen in the acral melanoma component.
  

In this case, given the presence of two morphologically distinct components with SOX-10 positivity and variable mutational profile, we considered several possibilities - (1) an invasive acral lentiginous melanoma with a related spindled/desmoplastic melanoma component in the deep dermis, (2) a neurocristic dermal tumor, neurotized nevus, or a perineuriomatous nevus that evolved into an overlying melanoma, (3) two separate melanomas, superficial acral lentiginous melanoma and deeper desmoplastic melanoma occurring together, and (4) an acral lentiginous melanoma and atypical nerve sheath tumor /neurofibroma occurring together. 

Scenarios 1 & 2 are considered with a possibility that both superficial and deeper components are clonally related to each other. However genomically, the superficial component showed 6q24.3 & 11q13 gains along with an activating KIT mutation characteristic for acral melanomas that are not shared with deeper spindled component, which exhibited a NF1 (p.Y2240*) mutation. Tumors can and do transform and KIT is a driver mutation that can initiate clonal expansion but constituently present in even in genomically heterogentic areas and very unlikely to be lost in deeper spindled invasive component. [1, 2].  The dermal component showed a distinct NF1 mutation that is not seen in superficial melanoma and additionally did not show other copy number aberrations, making a related and transformed malignancy less likely [1, 2]. Similar TERT promoter mutation seen in both neoplasms does not imply that they are related because TERT promoter is highly prevalent in melanomas and many soft tissue neoplasms. Additionally it is usually considered as an intermediate event harbored after an initial pathogenic driver mutation, hence their presence in both neoplasms but with two dissimilar driver mutations makes it a greater probability of clonally unrelated [3].    Scenario 2 is also unlikely as neurocristic hamartomas or perineuriomatous nevus commonly occur in head and neck regions and do not typically show KIT or NF1 mutations [4, 5]. BRAF V600E is the commonly identified driver mutation in perineuriomatous nevi and it is not identified in our case. Similarly, the literature on mutation signatures in neurotized nevi is sparse and NF1 mutations have not been reported in these lesions.  Scenarios 3 & 4 considers that these two components are clonally unrelated to each other and it is the likely possibility given the differences in genomic features. A collision of two genetically different melanomas is considered in scenario 3 and it is less favored given the histological assessment of deeper spindled component. While a deeper desmoplastic melanoma is a possibility, absence of overlying & related melanoma in-situ component, presence of relatively bland cytology without high grade atypia and architecturally uniform growth makes it is less likely for a high grade desmoplastic melanoma. NF1 mutation is a common finding in a desmoplastic melanoma however these neoplasms exhibits high burden of co-existing mutations, which is not a feature identified by DNA sequencing in this dermal neoplasm [6, 7].  Scenario 4 considers that the dermal component is a nerve sheath tumor without malignant features. The presence of pathogenic NF1 mutation in the deeper spindle component in combination with histology and immunohistochemical features are compatible with a nerve sheath tumor such as a neurofibroma. However the architecture of the dermal neoplasm is slightly atypical with diffuse sheets that are not of typical neurofibroma, mild cytologic atypia and presence of TERT promoter mutation, led us to conclude that this best considered as an atypical neurofibroma. For these reasons and genetically unrelated mutational signatures, we conclude that scenario 4, an acral lentiginous melanoma co-occurring with an underlying atypical nerve sheath tumor /neurofibroma is best determined diagnosis.  However, while we feel there is ample evidence that the melanoma and the bland spindle component are unrelated, we allow that the exact nature of spindled component is not fully certain. 

Molecular studies can help pathologists distinguish neoplastic processes that otherwise are indistinguishable using traditional histological and immunohistochemical methods.

Figure 1.  Initial biopsy, superficial fragment, H&E.
Figure 2.  Initial biopsy, deep spindled fragment, H&E.
Figure 3.  Amputation, H&E.
Figure 4.  Amputation, S100.
Figure 5.  Amputation,  Melan-A.