MART-1-labeled melanocyte density and distribution in actinic keratosis and squamous cell cancer in situ: Pagetoid melanocytes are a potential source of misdiagnosis as melanoma in situ.

BACKGROUND: Actinic keratosis (AK) and squamous cell carcinoma in-situ (SCCIS) within or near melanoma in situ (MIS) can complicate diagnosis due to overlapping clinical and microscopic features. This study aimed to describe basilar melanocyte density and pagetoid spread in AK and SCCIS for improved diagnostic accuracy. METHODS: A total of 22 AK and 22 SCCIS biopsies containing a margin of uninvolved epidermis were immunostained with MART-1 (melanoma antigen recognized by T-cells 1). The basilar melanocyte:keratinocyte ratio and the number and distribution of pagetoid melanocytes were compared in AK, SCCIS, and uninvolved epidermis. An in-vitro human skin model was created to assess the impact of keratinocyte atypia on melanocyte distribution. RESULTS: The median basilar melanocyte:keratinocyte ratio in SCCIS (1:11.49) was lower than in uninvolved epidermis (1:5.59, P = 0.0011), and the ratio in AK (1:6.94) was similar to uninvolved epidermis (P = 0.987). Pagetoid melanocytes were absent in perilesional skin but common in AK (21/22, P < 0.0001) and SCCIS (22/22, P < 0.0001). Pagetoid melanocytes at or above the mid-spinous layer were more common in SCCIS (21/22) vs AK (7/22, P < 0.0001). Pagetoid melanocytes were present in the in-vitro skin model made with neoplastic but not normal keratinocytes. CONCLUSIONS: Pagetoid melanocytes in AK and SCCIS should be interpreted with caution to avoid overdiagnosis of MIS.

A quantitative comparison between SOX10 and MART-1 immunostaining to detect melanocytic hyperplasia in chronically sun-damaged skin.

Histologic differentiation of melanoma in situ (MIS) from solar keratosis on chronically sun-damaged skin is challenging. The first-line immunostain is usually MART-1/Melan-A, which can exaggerate the epidermal melanocytes, causing a diagnostic pitfall for MIS. By comparing MART-1 and SOX10 immunostaining, we scored the percentage of epidermal melanocytes per 2-mm diameter fields in pigmented actinic keratosis (n = 16), lichenoid keratosis (n = 7), junctional melanocytic nevus (n = 6), keratosis with atypical melanocytic proliferation (n = 17) and MIS (n = 10). These cases represented an older population (68 years median age) and the head and neck (50%) was the most common anatomic site. MART-1 score was significantly higher than SOX10 (P value <.05) in solar keratoses, but showed no difference in detecting melanocytic proliferations, demonstrating their equal detection rate of melanocytes. The sensitivity of both MART-1 and SOX10 was 100%, while their specificities were 17% and 96%, respectively. These results show that SOX10 is more specific than MART-1 in distinguishing epidermal melanocytes on sun-damaged skin by avoiding overdiagnosis of melanoma.

Loss of Conventional Melanocytic Markers in Malignant Melanoma and Lymph Node Metastasis; an Uncommon but Dangerous Pitfall.

Although malignant melanomas exhibit a wide range of immunophenotypes, concurrent loss of all 3 conventional melanocytic markers (S-100, Melan-A, and HMB-45) is relatively rare. We report a case of primary malignant melanoma with lymph node metastasis, both exhibiting loss of immunoreactivity for conventional melanocytic markers, while aberrantly expressing epithelial antigenicity (pancytokeratin, CAM 5.2).

Novel mouse model for primary uveal melanoma: a pilot study.

BACKGROUND: To establish a mouse model with the aim of studying the tumour biology and metastasis formation of uveal melanoma. METHODS: Two human primary uveal melanoma cell lines (UMT2 and UMT42) were injected into the choroid of BALB/c nude mouse eyes. Intraocular tumour growth and metastasis formation in the liver and lungs were assessed after 13 to 22 weeks. Formalin-fixed, paraffin-embedded material was processed via haematoxylin and eosin staining for histological examination and periodic acid Schiff staining to search for extravascular matrix patterns. Immunohistochemistry for Melan A, CD34 and Ki67 was performed to assess the expression of a melanocytic lineage marker, angiogenesis and proliferative activity. RESULTS: All eyes injected with UMT2 cells, but only 25% of eyes treated with UMT42, developed intraocular tumour growth. The morphology of intraocular melanomas resembled that of primary tumours and showed signs of malignancy, including retinal invasion, optic nerve invasion and scleral penetration with extraocular tumour growth. UMT2 tumours formed extravascular matrix patterns exclusively. Most of the tumour cells expressed Melan A. Intratumoural angiogenesis was detected in both tumour entities. Proliferative activity was verified in all but one tumour. However, no metastases appeared in the liver or lungs. CONCLUSIONS: The mouse model presented with the UMT2 cell line allows for investigations of tumour biology of the primary UM because of the high degree of similarity between the tumours generated in the mouse eyes and the corresponding primary human UM. Unfortunately, the model is not suitable for investigations of metastasis formation.

Effects of Long-term Serial Passaging on the Characteristics and Properties of Cell Lines Derived From Uveal Melanoma Primary Tumors.

PURPOSE: Development of liver metastasis remains the most common cause of mortality in uveal melanoma (UM). A few cell lines cultured from primary UM tumors have been used widely to investigate the pathobiology of UM. However, the translation of basic knowledge to the clinic for the treatment of the metastatic disease has remained incremental at best. In this study, we examined whether the properties of UM cell lines at various passages were similar to their corresponding primary tumors. METHODS: Gene expression profiling by microarray was performed on UM primary tumors and derived cell lines cultured at varying passages. Expression of UM protein markers was monitored by immunohistochemical analyses and Western blotting. The in vivo tumorigenic properties of UM cultures were evaluated using athymic nude mice. RESULTS: Cell passaging severely reduced the expression of genes encoding markers typical of UM, including those of the prognostic gene signature. Marked differences between gene expression profiles of primary tumors and cell lines could be linked to the infiltrating immune and stromal cells in situ. In addition, the tumorigenic properties of UM cell lines also increased with cell passaging in culture as evaluated by their subcutaneous injection into athymic mice. CONCLUSIONS: Together, these findings demonstrate that the short-term UM primary cultures exhibit molecular features that resemble the respective surgical material and, thus, represent the best model for in vitro-assessed cancer treatments.

Cysts in angiomyolipoma with epithelial cysts may be consisted of entrapped and dilated renal tubules: report of a case with additional immunohistochemical evidence to the pre-existing literature.

Angiomyolipoma with epithelial cysts (AMLEC) is a distinctive variant of angiomyolipoma characterized by grossly apparent epithelial cysts and a cellular, müllerian-like subepithelial stroma. Some authors suspect that the epithelial cysts mainly represent dilated entrapped native renal collecting duct epithelium, while other authors think that they represented true epithelial differentiation of the AML. Recently, it has been reported that obvious immunolabeling of melanocytic markers such as Melan A and HMB45, which are often immunolabeled in classical angiomyolipoma, are present in epithelial cysts in cases of AMLEC. Here, we report the case of a 43-year-old Japanese woman with AMLEC, and attempt to elucidate the significance of melanocytic marker immunolabeling in the cyst's epithelium. In the present case, Melan A was labeled in the cyst's epithelium, and was thought to reflect its labeling in renal tubules existing in the renal parenchyma outside the tumor. This finding may indicate that the cyst epithelium is derived from entrapped renal tubules inside the AML. Non-immunolabeling of the estrogen and progesterone receptors in the cyst epithelium may also suggest that the cyst epithelium is not neoplastic, in contrast to their labeling in neoplastic cells existing in cyst wall. Further examination, such as molecular analysis, is needed to determine whether these epithelial cysts is neoplastic or non-neoplastic.

[Adrenocortical oncocytoma].

The paper characterizes adrenocortical oncocytoma, a rare adrenal tumor, accompanied by Cushing's syndrome and estrogen and androgen production and provides histological and immunohistochemical features. The authors describe their observation of a 33-year-old female woman. It is shown that estimation of the malignant potential of adrenocortical oncocytomas requires a special approach and must be done using the Lin-Weiss-Bisceglia criteria.

The relationship of melanocytic differentiation with prognostic markers in medullary thyroid carcinomas.

INTRODUCTION: Medullary thyroid carcinoma (MTC) makes up 5-10% of thyroid malignancies. Small cell, squamous, giant cell or melanocytic differentiation can rarely be seen in MTCs. It is important to determine those with the potential to act aggressively such as cases with melanocytic differentiation at the time of diagnosis. MATERIALS AND METHOD: A total of 46 MTC cases diagnosed at four different centers between 2002 and 2013 were included in the study. Immunohistochemical (IHC) staining with Melan-A and HMB-45 was performed in all cases. RESULTS: Six of the 46 MTC cases were medullary microcarcinomas and three were multicentric medullary carcinomas. There were 34 females and 12 males with a mean age at onset of 51.4 years and mean tumor diameter of 23.2mm. Lymph node metastasis (LNM) was found in 13 of the 38 cases that had data regarding the lymph nodes. Immunohistochemically, Melan A staining was seen in four cases. HMB45 staining was seen in four cases. A statistically significant relationship was found between LNM and diameter, Melan A expression (p=0.02, p=0.03 respectively) but there was no significant relationship with HMB45 expression (p=0.07). General survival data were present for 35 of the 46 cases. All cases without lymph node metastasis survived (21/21) while 8 of 11 cases with lymph node metastasis survived among cases with survival data; one case that was diffuse-strong positive for both HMB45 and Melan A was lost due to distant organ metastasis six months after the diagnosis. DISCUSSION: Should the possibility of melanocytic differentiation be evaluated in cases where melanocytic differentiation is not reflected in the morphology (lack of pigment) in MTCs? We did not come across a study on the subject in the English literature. The effect of melanocytic differentiation on the prognosis in MTCs should be investigated in larger series.

Immunoexpression of S100A4 in canine skin melanomas and correlation with histopathological parameters.

BACKGROUND: Melanoma is one of the most common skin neoplasms in humans and dogs. The tumor microenvironment in melanoma comprises cancer cells and stromal cells that interact to accelerate tumor progression. Several prognostic markers for melanomas have been studied in many human tumors, including fibroblast-specific protein 1 (S100A4). S100A4 is a member of the S100 family of calcium-binding proteins in stromal cells. HYPOTHESIS/OBJECTIVES: The objective of this study was to describe the immunohistochemical patterns of S100A4 in stroma and neoplastic cells of canine skin melanomas and correlate them with some histological parameters. ANIMALS AND METHODS: Forty-eight samples (38 pigmented and 10 non-pigmented melanomas) were first selected and their nature confirmed using S100, Melan A and vimentin. All cases were examined by immunohistochemistry using S100A4 to correlate expression, histotype, and level of invasion. RESULTS: All the tumors, including 10 non-pigmented, were positive for S100, Melan A, vimentin and negative for cytokeratin AE1/AE3 (consistent with melanomas). The 48 melanomas were classified as epithelioid (n = 21), spindle (n = 14), and mixed (n = 13). S100A4 was preferentially expressed in epithelioid and spindle cell types compared with mixed melanomas and S100A4 expression was not associated with level of invasion (Clark's levels IV to V). CONCLUSION: S100A4 expression in melanoma samples varied among histotypes but not between levels of invasion.

Comparative genomic hybridization in a case of melanoma that loses expression of S100, HMB45, Melan A and tyrosinase in metastasis.

We recently reported three cases of metastatic melanoma that does not express S100, HMB45, Melan A and Tyrosinase. A concurrent cutaneous scalp primary melanoma was identified later in one of the cases, which showed strong expression of these markers. The difference in immunophenotype between the primary melanoma and its metastasis in the parotid gland in this case raised the question of the biological significance of the expression of these markers and metastatic potential. To address this question, we utilized microarray comparative genomic hybridization (aCGH) to compare the cytogenetic features between the primary and metastatic melanoma. We observed chromosomal gains including 6p, entire chromosome 7, and 8q11.1-q24.3 in both primary and metastatic tumors. However, the metastatic lesion showed unique additional copy of chromosomal 7q, and loss of chromosome 9p24.3-q13 and chromosome 4, which included Melan A encoding gene region in 9p24.1. The above findings suggest the unique cytogenetic changes in the parotid lesion are most likely related to the metastatic behavior, as well as responsible for loss of multiple melanocytic marker expression in the metastatic melanoma for this case.

Evaluation of the role of routine melan-A immunohistochemistry for exclusion of microinvasion in 120 cases of lentigo maligna.

BACKGROUND: To assess the usefulness of routine melan-A immunohistochemistry (IHC) for exclusion of microinvasion in lentigo maligna (LM). METHODS: One hundred and twenty cases of LM from our archives were reviewed by 2 authors with S100 protein and melan-A IHC using a red chromogen. RESULTS: Melan-A was useful to confirm the diagnosis of LM in early lesions and to differentiate these from chronically sun-damaged skin. The presence of scattered melan-A-positive cells was noted in the dermis in 72 of 120 cases (melanophages in 36 cases, nonspecific cells different to melanophages in 16 cases, and a dual cell population in 20 cases). The significance of these cells was uncertain. Only 3 cases suspicious for microinvasion were identified: 2 on haematoxylin and eosin and 1 on S100. CONCLUSIONS: We recommend use of melan-A to confirm the diagnosis in early lesions of LM and in the differential diagnosis from melanocytic hyperplasia in chronically sun-exposed skin. We do not recommend routine use of melan-A to identify or exclude microinvasion. However, it may have a role, in conjunction with S100, in cases with suspicious features for early invasion on haematoxylin and eosin sections.

Balloon cell melanoma of the anal canal: a wolf in sheep's clothing?

Balloon cell melanoma (BCM) is a rare histologic variant of cutaneous malignant melanoma with exceptional reports of occurrences at non-cutaneous sites. Herein we present a case of primary amelanotic BCM of anal canal, a heretofore undescribed location. Histologically, the tumor was characterized by sheets of pale cells that bore striking resemblance to foamy macrophages. Presence of rare atypical mitoses confirmed the malignant nature of the cells. Neoplastic cells were immunoreactive for S100, Melan-A, and focally for HMB-45 while were negative for myogenic, gastrointestinal stromal tumor, epithelial and neuroendocrine markers. Resemblance to foamy macrophages, bland cytology and absence of pigment imparts this tumor a deceptively benign histological appearance making it prone to diagnostic pitfalls. Awareness of this rare entity and judicious employment of immunohistochemistry is imperative in segregating it from its diverse mimics.

Analysis of the processing of seven human tumor antigens by intermediate proteasomes.

We recently described two proteasome subtypes that are intermediate between the standard proteasome and the immunoproteasome. They contain only one (ß5i) or two (ß1i and ß5i) of the three inducible catalytic subunits of the immunoproteasome. They are present in tumor cells and abundant in normal human tissues. We described two tumor antigenic peptides that are uniquely produced by these intermediate proteasomes. In this work, we studied the production by intermediate proteasomes of tumor antigenic peptides known to be produced exclusively by the immunoproteasome (MAGE-A3(114-122), MAGE-C2(42-50), MAGE-C2(336-344)) or the standard proteasome (Melan-A(26-35), tyrosinase(369-377), gp100(209-217)). We observed that intermediate proteasomes efficiently produced the former peptides, but not the latter. Two peptides from the first group were equally produced by both intermediate proteasomes, whereas MAGE-C2(336-344) was only produced by intermediate proteasome ß1i-ß5i. Those results explain the recognition of tumor cells devoid of immunoproteasome by CTL recognizing peptides not produced by the standard proteasome. We also describe a third antigenic peptide that is produced exclusively by an intermediate proteasome: peptide MAGE-C2(191-200) is produced only by intermediate proteasome ß1i-ß5i. Analyzing in vitro digests, we observed that the lack of production by a given proteasome usually results from destruction of the antigenic peptide by internal cleavage. Interestingly, we observed that the immunoproteasome and the intermediate proteasomes fail to cleave between hydrophobic residues, despite a higher chymotrypsin-like activity measured on fluorogenic substrates. Altogether, our results indicate that the repertoire of peptides produced by intermediate proteasomes largely matches the repertoire produced by the immunoproteasome, but also contains additional peptides.

Human macrophages and dendritic cells can equally present MART-1 antigen to CD8(+) T cells after phagocytosis of gamma-irradiated melanoma cells.

Dendritic cells (DC) can achieve cross-presentation of naturally-occurring tumor-associated antigens after phagocytosis and processing of dying tumor cells. They have been used in different clinical settings to vaccinate cancer patients. We have previously used gamma-irradiated MART-1 expressing melanoma cells as a source of antigens to vaccinate melanoma patients by injecting irradiated cells with BCG and GM-CSF or to load immature DC and use them as a vaccine. Other clinical trials have used IFN-gamma activated macrophage killer cells (MAK) to treat cancer patients. However, the clinical use of MAK has been based on their direct tumoricidal activity rather than on their ability to act as antigen-presenting cells to stimulate an adaptive antitumor response. Thus, in the present work, we compared the fate of MART-1 after phagocytosis of gamma-irradiated cells by clinical grade DC or MAK as well as the ability of these cells to cross present MART-1 to CD8(+) T cells. Using a high affinity antibody against MART-1, 2A9, which specifically stains melanoma tumors, melanoma cell lines and normal melanocytes, the expression level of MART-1 in melanoma cell lines could be related to their ability to stimulate IFN-gamma production by a MART-1 specific HLA-A*0201-restricted CD8(+) T cell clone. Confocal microscopy with Alexa Fluor®(647)-labelled 2A9 also showed that MART-1 could be detected in tumor cells attached and/or fused to phagocytes and even inside these cells as early as 1 h and up to 24 h or 48 h after initiation of co-cultures between gamma-irradiated melanoma cells and MAK or DC, respectively. Interestingly, MART-1 was cross-presented to MART-1 specific T cells by both MAK and DC co-cultured with melanoma gamma-irradiated cells for different time-points. Thus, naturally occurring MART-1 melanoma antigen can be taken-up from dying melanoma cells into DC or MAK and both cell types can induce specific CD8(+) T cell cross-presentation thereafter.

Melan-A/Mart-1- or HMB-45-positive melanocytes are not present in calcifying cystic odontogenic tumors (calcifying odontogenic cysts): a study in 13 Caucasian patients.

Melanin pigment and melanocytes may be found in odontogenic cysts and tumors, particularly calcifying cystic odontogenic tumor (CCOT). In the present study we investigated the immunohistochemical expression of the Melan-A/Mart-1 and HMB-45 antigens in 13 Caucasians patients with CCOT. Melan-A/Mart-1- and HMB-45-positive melanocytes were not seen in any of the cases. Our findings are in agreement with the assumption that pigmentation in odontogenic lesions may be a racial phenomenon.

Aberrant Melan-A expression in atypical fibroxanthoma and undifferentiated pleomorphic sarcoma of the skin.

BACKGROUND: Atypical fibroxanthoma (AFX) is a distinctive clinicopathologic entity presenting on sun-damaged skin of the elderly. Its behavior is benign if strict diagnostic criteria are applied. Tumors showing invasion of deeper structures or perineural/lymphovascular invasion are best regarded as undifferentiated pleomorphic sarcoma of the skin. The diagnosis requires immunohistochemical studies to exclude melanoma, squamous cell carcinoma, angiosarcoma and leiomyosarcoma. METHODS: Two AFX and one undifferentiated pleomorphic sarcoma showing aberrant expression of Melan-A were identified. Clinical data were obtained and histopathological features, immunohistochemical profile and electron microscopy were assessed. RESULTS: All tumors arose on sun-damaged skin of elderly males. Two AFX showed pushing growth into superficial subcutis only. The undifferentiated pleomorphic sarcoma was characterized by infiltrative growth into galea as well as perineural invasion. Multifocal expression of Melan-A and MART-1 was largely limited to tumor giant cells in the absence of S100 or HMB-45 labeling. No melanosomes or premelanosomes were identified by electron microscopy. CONCLUSIONS: Aberrant expression of Melan-A and MART-1 in AFX and undifferentiated pleomorphic sarcoma of the skin represents an important diagnostic pitfall with potential for misdiagnosis as melanoma.

Identification of an HLA-DPB1*0501 restricted Melan-A/MART-1 epitope recognized by CD4+ T lymphocytes: prevalence for immunotherapy in Asian populations.

CD4 T lymphocytes play a central role in orchestrating an efficient antitumor immune response. Much effort has been devoted in the identification of major histocompatibility complex class II eptiopes from different tumor-associated antigens. Melan-A/MART-1 is expressed specifically in normal melanocytes and tumor cells of 75% to 100% of melanoma patients. Melan-A/MART-1 is considered as an attractive target for cancer immunotherapy. In the past, several human leukocyte antigen (HLA) class II restricted epitopes have been identified and characterized, including Melan-A/MART-11-20 (HLA-DR11 restricted), Melan-A/MART-125-36 (HLA-DQ6 and HLA-DR3 restricted), Melan-A/MART-127-40 (HLA-DR1 restricted), Melan-A/MART-151-73 (HLA-DR4 restricted), Melan-A/MART-191-110 (HLA-DR52 restricted), and Melan-A/MART-1100-111 (HLA-DR1 restricted). Owing to the infrequent expression of the above HLA class II alleles in Asian populations, immunotherapy using these defined Melan-A/MART-1 peptides could potentially only benefit a very small percentage of Asian melanoma patients. In this study, we established several CD4 T-cell clones by in vitro stimulation of peripheral blood mononuclear cells from a healthy donor by a peptide pool of 28 to 30 amino acid long peptides spanning the entire Melan-A/MART-1 protein. These CD4 T-cell clones recognized a peptide that is embedded within Melan-A/MART-121-50, in a HLA-DPB1*0501 restricted manner. Finally, we demonstrated that this epitope is naturally processed and presented by dendritic cells. HLA-DPB1*0501 is frequently expressed in Asian population (44.9% to 73.1%). Therefore, this epitope could provide a new tool and could significantly increase the percentage of melanoma patients that can benefit from cancer immunotherapy.

An immunohistochemical comparison between MiTF and MART-1 with Azure blue counterstaining in the setting of solar lentigo and melanoma in situ.

BACKGROUND: Evaluation of cutaneous pigmented lesions can be diagnostically challenging and represents an activity often supplemented by immunohistochemistry. Immunohistochemical studies typically employ 3,3'-diaminobenzidine (DAB) resulting in brown staining of both melanocytes and melanin. Difficulty may thus arise in distinguishing different cell types in heavily melanized lesions. Azure blue counterstaining has been used in conjunction with melanoma antigen recognized by T-cells (MART-1) to differentiate melanocytes from melanin by highlighting the latter blue-green. Microphthalmia transcription factor (MiTF) represents an alternative immunomarker that shows nuclear reactivity, which facilitates ease of interpretation. METHODS: Twenty examples of solar lentigo and melanoma in situ (MIS) were independently evaluated utilizing MiTF and MART-1/Azure blue for melanocyte quantification. Melanocyte counts were averaged over five high-power fields (×400) to obtain a mean melanocytic count. RESULTS: There was no significant difference in the mean melanocytic count between MART-1/Azure blue and MiTF as assessed in the solar lentigo group and as assessed independently in the MIS group. MiTF nuclear staining facilitated interpretation and required less laboratory preparation, as an additional counterstain was not necessary. CONCLUSIONS: MiTF is as effective as MART-1/Azure blue in identifying melanocytes in the context of solar lentigo or MIS. On the basis of our results, we favor expanding the use of MiTF as an immunohistochemical marker, as it provides an efficient alternative to MART-1 with Azure blue counterstaining in the evaluation of cutaneous pigmented lesions.

Immunohistochemical study of 40 cases of longitudinal melanonychia.

The etiology of longitudinal melanonychia (LM) is difficult to establish by clinical and dermoscopic examinations alone. Microscopic examination of the nail matrix remains crucial. Two groups of LM may be identified: melanocytic activation (melanic pigmentation of the matrix epithelium without any increase in the density of melanocytes) and melanocytic proliferation (lentigo, nevus, or melanoma). The histological examination is challenging, and immunohistochemical investigations can be helpful. The objective of this study was to analyze the immunohistochemical findings with routinely used markers in melanocytic tumors-S-100 protein, HMB-45, and Melan-A-in LM. A series of 40 cases were analyzed: 10 activations, 4 lentigines, 7 nevi, 12 in situ melanomas, and 7 invasive melanomas. The sensitivity of S-100 protein is weak in benign and malignant intraepithelial melanocytes of the nail matrix, and if this marker is performed alone, it may be wrongly reassuring. However, the use of S-100 protein is essential to differentiate invasive melanoma, lacking an intraepithelial component, and particularly desmoplastic melanoma, from epithelial and mesenchymal tumors. HMB-45 and Melan-A are more sensitive than S-100 protein for the evaluation of intraepithelial melanocytic proliferation of the nail apparatus, with HMB-45 being the most intense marker. In the dermal component, HMB-45 and Melan-A were less sensitive than S-100 protein. In conclusion, we recommend that the panel of antibodies used for histological evaluation of LM should include HMB-45 and/or Melan-A and S-100 protein only if an invasive melanoma is suspected.