Although melanoma represents only 10% of all skin cancer diagnoses, it accounts for at least 65% of all skin cancer–related deaths. The number of new cutaneous melanoma cases projected during 2010 was 68,000—a 23% increase from the 2004 prediction of 55,100 cases. In 2015, the lifetime risk of developing melanoma is estimated to increase to 1 in 50. As the incidence of melanoma continues to rise, now more than ever, clinicians and histopathologists must have familiarity with the various clinical and pathologic features of cutaneous melanoma.
- •
Although melanoma represents only 10% of all skin cancer diagnoses, it accounts for at least 65% of all skin cancer-related deaths.
- •
The most important point in melanoma management is its early recognition by a clinician or patients themselves by using ABCDE melanoma mnemonic.
- •
The use of an adequate biopsy technique that will provide the entire lesion for microscopic evaluation is essential to ensure the correct diagnosis.
- •
The most important stain in microscopic evaluation of melanocytic lesions remains hematoxylin and eosin with special stains used only if melanocytic differentiation is not readily apparent.
- •
The depth of melanoma remains the most important prognostic factor.
Introduction
Although melanoma represents only 10% of all skin cancer diagnoses, it accounts for at least 65% of all skin cancer–related deaths. The number of new cutaneous melanoma cases projected during 2010 was 68,000—a 23% increase from the 2004 prediction of 55,100 cases. In 2015, the lifetime risk of developing melanoma is estimated to increase to 1 in 50. As the incidence of melanoma continues to rise, now more than ever, clinicians and histopathologists must have familiarity with the various clinical and pathologic features of cutaneous melanoma.
Clinical features of melanoma
Cutaneous melanoma may present with many different clinical appearances and histopathologic correlates. Most early lesions of melanoma demonstrate the ABCDE features physicians and some patients are now aware of; however, rare cases of melanoma may demonstrate only some or none of these features. Morphologically, most melanomas appear as patches, plaques, nodules, or tumors; however, rarely, melanomas may be polypoidal with a stalk. Usually, lesions measure greater than 6 mm in diameter at diagnosis but smaller lesions are well recognized.
The most common and well-known types of melanoma include superficial spreading melanoma, nodular melanoma, lentigo maligna, and acral lentiginous melanoma. Rare variants of melanoma include desmoplastic melanoma (DM), amelanotic melanoma, verrucous melanoma, follicular melanoma, and melanomas arising from or mimicking blue nevi. Due to a lack of familiarity, these rare variants pose a greater diagnostic challenge to clinicians and histopathologists. DMs present subtly as an indurated patch, plaque, or papule that is either skin colored or pigmented. These lesions are notoriously difficult to diagnose, appear on sun-exposed skin, and clinically resemble a fibroma or scar.
Amelanotic melanomas, which account for 2% to 8% of all melanomas, contain little or no pigment and are the most difficult variant of melanoma to diagnose. Fortunately, truly amelanotic melanomas completely devoid of pigment are rare and account for less than 2% of all melanomas. These lesions present as pink or skin-colored patches, plaques, or nodules that may ulcerate. Clinicians commonly mistake these for pyogenic granulomas, basal cell carcinomas, nevi, seborrheic keratoses, or fibromas. A potential clue to the correct diagnosis is recognition of faint pigmentation at the periphery of the lesion, which may be easier with dermascopy. Dermascopy may also show small red dots evenly distributed or grouped on a whitish or pink-red background, regression structures, irregular dots/globules, or a blue-whitish veil.
Verrucous melanoma, first described in 1967, presents as hyperkeratotic pigmented lesions commonly on the extremities (71%). Clinically, they mimic benign lesions, including seborrheic keratoses, verrucae, nevi, and Spitz nevi.
Follicular melanoma, first described in 2004, represents a rare form of melanoma that presents as a perifollicular, darkly pigmented papule measuring less than 5 mm. It commonly occurs on the face or neck of older patients (46–82 years of age) with sun-damaged skin. Clinically, it more closely resembles a comedo or pigmented cyst than a melanoma.
Occasionally, melanomas arise in association with blue nevi or arise de novo but clinically resemble blue nevi. This type of melanoma favors the head and neck and may affect children, adults, or the elderly. These lesions tend to evade detection, which leads to a poor prognosis, because they typically grow slowly and predominantly involve the deep dermis. The clinicopathologic correlations of all melanomas remain critical in the effective management of melanoma.
Clinical features of melanoma
Cutaneous melanoma may present with many different clinical appearances and histopathologic correlates. Most early lesions of melanoma demonstrate the ABCDE features physicians and some patients are now aware of; however, rare cases of melanoma may demonstrate only some or none of these features. Morphologically, most melanomas appear as patches, plaques, nodules, or tumors; however, rarely, melanomas may be polypoidal with a stalk. Usually, lesions measure greater than 6 mm in diameter at diagnosis but smaller lesions are well recognized.
The most common and well-known types of melanoma include superficial spreading melanoma, nodular melanoma, lentigo maligna, and acral lentiginous melanoma. Rare variants of melanoma include desmoplastic melanoma (DM), amelanotic melanoma, verrucous melanoma, follicular melanoma, and melanomas arising from or mimicking blue nevi. Due to a lack of familiarity, these rare variants pose a greater diagnostic challenge to clinicians and histopathologists. DMs present subtly as an indurated patch, plaque, or papule that is either skin colored or pigmented. These lesions are notoriously difficult to diagnose, appear on sun-exposed skin, and clinically resemble a fibroma or scar.
Amelanotic melanomas, which account for 2% to 8% of all melanomas, contain little or no pigment and are the most difficult variant of melanoma to diagnose. Fortunately, truly amelanotic melanomas completely devoid of pigment are rare and account for less than 2% of all melanomas. These lesions present as pink or skin-colored patches, plaques, or nodules that may ulcerate. Clinicians commonly mistake these for pyogenic granulomas, basal cell carcinomas, nevi, seborrheic keratoses, or fibromas. A potential clue to the correct diagnosis is recognition of faint pigmentation at the periphery of the lesion, which may be easier with dermascopy. Dermascopy may also show small red dots evenly distributed or grouped on a whitish or pink-red background, regression structures, irregular dots/globules, or a blue-whitish veil.
Verrucous melanoma, first described in 1967, presents as hyperkeratotic pigmented lesions commonly on the extremities (71%). Clinically, they mimic benign lesions, including seborrheic keratoses, verrucae, nevi, and Spitz nevi.
Follicular melanoma, first described in 2004, represents a rare form of melanoma that presents as a perifollicular, darkly pigmented papule measuring less than 5 mm. It commonly occurs on the face or neck of older patients (46–82 years of age) with sun-damaged skin. Clinically, it more closely resembles a comedo or pigmented cyst than a melanoma.
Occasionally, melanomas arise in association with blue nevi or arise de novo but clinically resemble blue nevi. This type of melanoma favors the head and neck and may affect children, adults, or the elderly. These lesions tend to evade detection, which leads to a poor prognosis, because they typically grow slowly and predominantly involve the deep dermis. The clinicopathologic correlations of all melanomas remain critical in the effective management of melanoma.
Histologic diagnosis of melanoma
Although the pathogenesis of melanoma remains unsettled, the most valid and currently accepted pathway was proposed by Ackerman and colleagues in the 1980s. According to their hypothesis, a de novo melanoma arises when an oncogenic stimulus, usually chronic ultraviolet radiation, acts on one or more melanoctyes in the epidermis at the dermoepidermal (DE) junction. These mutated melanoctyes then proliferate at the junction as solitary units that appear histologically as scattered, single hyperchromatic cells with a cleft or halo surrounding them. At this stage clinically, the lesion may appear as nothing or a light tan macule ( Fig. 1 ). With time, these melanocytes multiply and coalesce to form small nests that remain confined to the DE junction. Adnexal involvement, including acrotrichia and acrosyringia, may also occur. At this point clinically, a small tan or brown macule (3–4 mm in diameter) that lacks significant atypia ( Fig. 2 ) can be appreciated.
As the tumor progresses, more nests of atypical melanocytes accumulate at the DE junction. Eventually, spread throughout the epidermis occurs and both singular atypical melanocytes and nests of atypical melanocytes extending into the upper layers of the epidermis are seen. Clinically, this manifests as tan lesions ranging in size from 3 mm to greater than 6 mm in diameter, which contain foci of dark brown or black pigment ( Fig. 3 ).
Dark brown or black coloration reflects the presence of melanin in the upper levels of the epidermis, including the cornified layer. At this point, lesions manifest noticeable abnormalities, including asymmetry, color variation, and irregular borders, which should prompt a biopsy.
With further progression, the epidermis and adnexal structures becomes more involved and occupied with atypical melanocytes arranged either in nests or singularly. For the first time, melanocytes invade the papillary dermis, initially as small foci of neoplastic melanocytes. Eventually, the papillary dermis becomes filled with neoplastic melanocytes and expands. Clinically, these lesions are macular or slightly elevated, greater than 6 mm in diameter, and demonstrate the ABCDEs characteristic of most melanomas. Skin markings are typically obliterated at this point ( Fig. 4 ).
With even further progression, neoplastic melanocytes spread to involve deeper structures, such as the reticular dermis and subcutaneous fat. At this point, ulceration, nodule formation, and the emergence of colors, including red, blue, and white, may occur. The appearance of white, a poor prognostic sign, indicates regression has occurred ( Fig. 5 ). All of these findings (ulceration, nodule formation, and regression) signify more advanced lesions and a poorer prognosis.
The final stage of progression in melanomas, like many other cancers, is metastasis. The primary metastasis in two-thirds of patients is classified as locoregional metastasis, which comprises satellite, in-transit, and regional lymph node metastasis. Regional lymph node metastasis accounts for one-half of all melanoma metastases. Satellite metastasis is defined as the formation of metastatic nodules within 2 cm of the primary tumor. In-transit metastasis develops in the metastatic drainage area en route to the first regional lymph node. The remaining one-third of metastases in melanoma present primarily as distant metastases to virtually any organ.
Metastatic melanoma presents with different clinical and histologic manifestations depending on which site in the body is involved. In cutaneous metastasis, histology reveals aggregates of atypical neoplastic melanocytes in the dermis often without contiguity with the epidermis. These atypical melanocytes may be arranged in nodular aggregates or cords of cells between and among dermal collagen bundles. The superficial papillary dermis may also be involved in so-called epidermotropic metastases. Blood vessels, lymphatics, and nerves may contain neoplastic cells. Clinically, cutaneous metastases of melanoma present as intracutaneous or subcutaneous nodules, ranging in color from that of normal skin to jet-black. Lymph node involvement usually manifests as one or more firm nodules in a lymph node chain that may be fixed to underlying structures. Histologically, the normal lymph node architecture is replaced by atypical neoplastic melanocytes.
The pathway, described previously, applies to melanomas that arise de novo. Approximately 25% of melanomas develop in association with nevi, however. Clark and colleagues recognized this and proposed a pathway to melanoma that begins with precursor lesions, which are nevi containing dysplastic melanocytes. All melanocytic nevi, especially giant congenital nevi, must be monitored regularly for changes signaling the development of melanoma.
Clark and colleagues described a model for the progression of melanoma that involves 2 separate and distinct growth phases: the radial phase followed by the vertical phase. The radial growth phase, which lacks metastatic potential, is characterized by expansion of malignant melanocytes throughout the epidermis and papillary dermis at the periphery of the lesion. Cells in this stage show limited capacity for growth in cell culture and are considered less biologically aggressive than cells in the vertical growth phase, which have the capacity for immortality in cell culture. As the name suggests, melanomas in the vertical phase expand vertically to involve deeper layers of the skin and are more likely to metastasize. Significant exceptions to this model of melanoma progression exist, however. Nodular melanomas have only a short or no radial growth phase, and thin melanomas may metastasize without a vertical growth phase. In addition, acrolentiginous melanoma demonstrates genotypic features of metastatic melanoma while confined to the epidermis. These findings led some to question the validity of assigning a radial and vertical growth phase to melanomas.
Molecular studies suggest some dysplastic nevi (DN) might represent intermediate lesions in a multistep melanoma tumorigenesis pathway. Studies have found common genetic alterations between DN and melanoma, including loss of identical tumor suppressor genes, the presence of similar microsatellite instability patterns, and reduced mismatch repair enzyme activity. Many experts do not consider dysplastic nevi to represent precursors of melanoma but rather markers. A recent study has produced evidence that tends to corroborate this position.
Biopsy technique for melanoma
The accurate histologic diagnosis of melanoma requires a biopsy that appropriately represents the lesion in question. The biopsy must achieve adequate depth, symmetry, and circumscription to allow for microscopic evaluation of as many diagnostic criteria as possible. Obtaining adequate depth and breadth is especially important to ensure capture of the entire process. Once obtained, specimens must remain intact and receive appropriate processing. Fragmented, deteriorated, or damaged specimens are often not interpretable and may lead to misdiagnosis. Additionally, because several entities mimic melanoma both clinically and histologically, truly representative and pristine biopsies are critical for accurate diagnosis.
Biopsy can remove part of the lesion (incisional) or the entire lesion (excisional). Incisional biopsy techniques include superficial shave biopsy, saucerization, and punch biopsy. Superficial shave biopsy should never be used for suspected melanomas because this technique does not sample deeply enough. Excisional biopsy is the preferred method because it encompasses the entire breadth of the lesion as judged clinically.
The most important prognostic factor for melanoma remains the thickness of the lesion (Breslow depth). A nonrepresentative biopsy may provide inaccurate information concerning the true depth of the lesion as well as other important prognostic factors and lead to mismanagement. A retrospective review of 145 cutaneous melanomas compared the accuracy at which saucerization, punch biopsy, or excisional biopsy estimated the Breslow depth found at subsequent therapeutic excision. Saucerization outperformed punch biopsies that were less than 5 mm in diameter for thin melanomas (≤1 mm). Excisional biopsy was found the most accurate method of biopsy. Hsu and Cockerell showed similar results when they reviewed 1123 histologically proved cutaneous melanomas to compare Breslow depth determined from saucerization, punch biopsy, or excisional biopsy to the Breslow depth at re-excision. They found significant diagnostic discrepancy between the re-excision specimens and the initial punch biopsies (86.5% accurate), whereas both excisional biopsy and saucerization demonstrated near 100% accuracy. (Hsu M, Cockerell CJ. Punch biopsy of melanocytic neoplasms: a poorly recognized pitfall in the diagnosis of cutaneous malignant melanoma. Submitted for publication, 2012.) Karimipour and colleagues examined 253 patients with suspected melanoma who underwent subtotal incisional biopsy with retainment of a significant portion of the initial lesion (≥50%). They found a statistically significant ( P = .001) and substantial discrepancy between the initial Breslow depth (mean of 0.66 mm) and the final depth (mean of 1.07 mm) after re-excision. After complete excision, 21% of these patients were upstaged and 10% became eligible for sentinel lymph node (SLN) biopsy. This study underscores the importance of obtaining representative specimens to ensure patients are managed properly.
The American Academy of Dermatology published updated guidelines for the management of primary cutaneous melanoma in 2011. For the biopsy of lesions suspicious for melanoma, they endorse narrow excisional biopsy that encompasses the entire breadth of the lesion with clinically negative margins to a depth sufficient to ensure the lesion is not transected. Obtaining clinical margins of 1 mm to 3 mm during the excisional biopsy is recommended. This can be accomplished using an elliptical or punch excision (with sutures) or saucerization. Saucerization is commonly used when the suspicion for melanoma is low, when the lesion lends itself to complete removal by this technique, or in the setting of suspected lentigo maligna. Incisional biopsy of the most atypical portion is acceptable in lesions that are large, have a low clinical suspicion for melanoma, or have on a facial or acral location. Incisional biopsies, however, can potentially underestimate the true depth of the lesion.
Histopathologic criteria for melanoma
Ackerman and colleagues proposed a unifying concept regarding the histologic classification of melanoma based on both architectural and cytologic features. Architectural characteristics of melanoma include asymmetry, poor circumscription, and the presence of irregularly distributed melanocytes, arranged singly or in nests, occupying the epidermis (pagetoid spread), adnexa, and dermis. Melanocytes exist above the DE junction and sometimes form irregular nests, which may become confluent ( Fig. 6 ). The nests and individual melanocytes lack maturation with progressive descent into the dermis. Additionally, melanin is irregularly distributed in the lesion within the epidermis, dermis, and adnexa. Cytologic characteristics include the presence of atypical melanocytes, necrotic melanocytes, and melanocytes undergoing mitosis.
Pathologic features of different clinical forms of cutaneous melanoma
Superficial Spreading Melanoma
Superficial spreading melanoma commonly demonstrates the characteristic features described previously. In addition, it typically demonstrates prominent spread of large atypical melanocytes with abundant pale cytoplasm throughout the epidermis.
Melanoma in Situ
Virtually all forms of melanoma possess a stage in which the neoplasm is confined to the epidermis. Lentigo maligna refers to a subtype typically affecting older individuals with sun-damaged skin. (Note: Not all is Lentigo Maligna–Melanoma in Situ on Acral Skin would not be called that.) Histologically, atypical melanocytes, arranged singly and in nests, are distributed irregularly at and above the DE junction. These atypical melanocytes are often present within hair follicles and appendageal structures. The epidermis is commonly thin and atrophic with loss of retia. Solar elastosis is usually abundant on lesions that develop on sun-damaged skin. The superficial dermis might contain a lichenoid infiltrate of lymphocytes and mimic benign lichenoid keratosis on low-magnification evaluation.
The histologic evaluation of margin involvement in melanoma in situ on sun-damaged skin may be difficult because of diffuse melanoctyic proliferation caused by longstanding sun damage. As a general guide, diffuse melanocytic proliferation caused by sun damage contains fewer melanocytes per unit area and typically lacks nests of melanocytes. Clinical correlation may be needed if the pathology is equivocal.
Nodular Melanoma
Virtually all melanomas, if left untreated, may eventually progress to involve the dermis and form papules or nodules clinically. Nodular melanoma, however, tends to involve the dermis at an early point in time in contrast to others that tend to remain confined to the epidermis. Nevertheless, some investigators contend that a nodule appearing in a melanoma is a sign of metastatic potential rather than a distinct subtype of melanoma.
Acrolentiginous Melanoma
Relative to the other forms of melanoma, acrolentiginous melanoma predominantly affects the palms and soles and contains a greater number of dendritic melanocytes with dendrites extending into the upper layers of the epidermis. Initially, the degree of cytologic atypia may be minimal. With time, melanocytes become more spindle shaped in appearance and display prominent pleomorphism and pagetoid spread. Eventually, involvement of the dermis and deeper structures occurs.
Desmoplastic Melanoma
DM is notoriously difficult to diagnose clinically, so dermatopathologists must have a high index of suspicion for this lesion. Unfortunately, most lesions present at an advanced stage. Several histologic variants of DM exist. In one form, melanocytes appear delicate and spindle shaped and are arranged in fascicles in the upper dermis. These cells may demonstrate minimal atypia and simulate a neural neoplasm, such as neurofibroma. A myxoid stroma is usually present. A second variant demonstrates prominent collagen in the dermis with admixed spindle cells and, occasionally, epitheliod cells; this lesion resembles a scar or fibroma. Many of these lesions have been previously biopsied and may contain scar tissue admixed with neoplasm. A third variant is compromised of numerous markedly atypical spindle and/or epitheliod cells arranged in sheets or fascicles. This variant contains abundant pleomorphic cells with hyperchromatic and bizarre nuclei. The majority of DM lesions (80%) have a component of melanoma in situ in the overlying dermis. They arise either in association with a preexisting melanoma of another type or de novo. They commonly contain nodular aggregates of lymphocytes scattered throughout the dermis.
Perineural invasion, which can cause recurrence and allow spread along nerves, is more common in DM compared with the other types of melanoma. Immunohistochemsitry is often needed to diagnose DM. Most DM lesions are positive for S-100 protein, p75 neurotrophin receptor, and vimentin. Smooth muscle actin may be positive, whereas HMB-45 and Melan-A are inconsistent.
Verrucous Melanoma
Many investigators consider verrucous melanoma a variant of superficial spreading melanoma. Verrucous melanoma is characterized by an exophytic papilliferous growth pattern. Prominent features include pseudoepitheliomatous hyperplasia and overlying hyperkeratosis. These findings may obscure the underlying melanoma cells that exhibit varying degrees of cellular pleomorphism at the DE junction and beneath it. Blessing and colleagues reviewed 20 cases of verrucous melanoma and found 10% of patients initially received benign histologic diagnoses. Furthermore, greater than 50% of patients were given benign clinical diagnoses before biopsy. The investigators also reported difficulty in assigning Breslow depth given the papilliferous architecture of these lesions. Unfortunately, 8 of the 20 patients developed metastases and 7 died of their disease. This study underscores the difficulty clinicians and pathologists face when encountering verrucous melanoma.
Animal Melanoma
Animal melanoma is a rare dermal-based melanocytic neoplasm comprised of sheets of heavily pigmented epithelioid or spindle cells with numerous melanophages. Epidermal hyperplasia is common and mitoses are infrequent. Abundant pigment makes the cytology difficult to evaluate. Animal melanoma may be confused with heavily pigmented blue nevi although diffuse architecture and deep extension are features of animal melanoma that are consistent with this diagnosis.
Nevoid Melanoma
At scanning magnification, these lesions resemble ordinary compound or dermal nevi. In the past, some investigators favored calling these lesions “minimal deviation melanoma” or “borderline melanoma” because they contain only subtle deviations from a nevus. These terms have fallen out of favor, however, and are no longer used. With close inspection, these lesions contain sufficiently distinct characteristics that portray a malignant nature. Clues to the malignant nature of these lesions include pleomorphism, impaired maturation, asymmetry, and mitoses. In addition, the junctional component may contain melanoma in situ, which is a good clue to the diagnosis. Another characteristic of this variant is the presence of irregularly sized and shaped dermal melanocytic nests. Also commonly seen in these lesions is pseudomaturation, which mimics true maturation. True maturation demonstrates progressive smaller melanocytes and cytoplasmic volume at increased depths of the lesion. Pseudomaturation, like true maturation, exhibits a decrease in cytoplasmic volume with increasing depth; however, nuclear size does not follow this trend and are larger than immature nuclei. Finally, unlike most other types of melanomas, inflammatory cell infiltrates are typically absent in this variant. Pathologists should recognize this entity because although they may appear benign, these lesions are no less aggressive than conventional melanomas.
Malignant Blue Nevus
Malignant blue nevus, a misnomer, refers to either de novo melanoma that simulates a cellular blue nevus or melanoma that arises in association with a blue nevus. Histologically, these lesions resemble animal melanoma in that they are characterized by large, deep proliferations of melanocytes with sheet-like growth patterns. Individual cells are epithelioid or spindle shaped and demonstrate mitoses, necrosis, nuclear atypia, pleomorphism, hyperchromasia, and prominent nucleoli. Pigmented dendritic cells exist in virtually all lesions. Three histologic patterns are recognized. The first is a lesion with an overtly malignant component adjacent to a benign blue nevus component. The second resembles a subtle sarcoma-like presentation (without florid benign and malignant components), which initially suggests a cellular blue nevus. These lesions, however, exhibit large densely cellular fascicles or nodules of spindle cells that on closer inspection express adequate atypia for malignancy. The third pattern also suggests a benign cellular blue nevus but contains atypical features, including large diameter, asymmetry, prominent cellular density, nuclear pleomorphism, and some mitotic activity. This pattern, not as obviously malignant as the first two patterns, is termed biologically indeterminate .
Balloon Cell Melanoma
Balloon cell melanoma is a rare variant of melanoma characterized by the presence of balloon cells, which are epithelioid cells containing abundant eosinophilic or foamy cytoplasm. These balloon cells may present either focally or diffusely throughout the lesion. These lesions typically stain for the usual melanoma markers, including S-100, HMB-45, and Melan-A. Additionally, balloon cell melanoma may simulate benign balloon nevus cells by exhibiting similar cytoplasmic features. The presence of nuclear pleomorphism, atypia, and mitoses and the absence of intervening stroma help distinguish balloon cell melanoma from a balloon cell nevus.
Clear Cell Saoft Part Sarcoma
Clear cell sarcoma (CCS), also called melanoma of soft parts, is a rare variant of a soft tissue sarcoma showing melanocytic differentiation. These deep lesions primarily affect soft tissues in the foot and ankle of young adults. Genetically, these malignant cells express a characteristic translocation of t(12;22)(q13;q12) involving EWS and ATF-1 genes. Histologically, they are characterized by a diffuse neoplasm of atypical, pleomorphic, clear-staining cells arranged in sheets and focally in nests. The absence of a direct connection to the epidermis is a useful feature in distinguishing CCS and melanoma. The presence of melanin pigment is occasionally seen in scattered cells. Neoplastic cells demonstrate immunoreactivity for S-100 protein (100%), HMB-45 antigen (97%), Melan-A (71%), and microphthalmia-associated transcription factor (MITF) (81%). The clinical course of CCS is usually protracted with multiple local recurrences and late metastases.
Spitzoid Melanoma
Spitz nevus and melanoma histologically simulate one another and are sometimes impossible to differentiate. Features that favor the diagnosis of melanoma over Spitz nevus include asymmetry, prominent confluence, high cellular density of melanoctyes, failure of maturation, deep invasion, increased mitotic rate, deep mitoses, lack of Kamino (dull-pink) bodies, and necrosis en masse. In adults, a low threshold should exist to diagnose melanoma if any of these features are present. The reverse is true of similar lesions in children (especially prepubertal children).
Sarcomatoid Melanoma
Sarcomatoid melanomas are rare melanomas with features that simulate a spindle cell or epithelioid cell sarcoma. Sarcomatoid melanoma lesions are extremely poorly differentiated and exhibit diffuse sheets of anaplastic cells with ulceration and necrosis en masse. Because these lesions are poorly differentiated, immunoperoxidase stains are necessary to render an accurate diagnosis.
Spindle Cell Melanoma
Spindle cell melanomas are characterized by a proliferation of spindle-shaped melanocytes arranged in fascicles and sheets. Unlike sarcomatoid and DM variants, these lesions do not demonstrate diffuse and deep involvement. Features that suggest melanoma include involvement of the epidermis, pagetoid spread of melanocytes, nesting, and the presence of melanin pigment. Because these lesions are poorly differentiated, they must be distinguished from other malignant neoplasms of the skin that may demonstrate a spindle cell morphology, including spindle cell squamous carcinoma and atypical fibroxanthoma. As with sarcomatoid melanoma and other poorly differentiated variants of melanoma, immunoperoxidase stains are required to render a precise diagnosis.
Follicular Melanoma
Follicular melanoma, a rare and newly reported variant, involves hair follicles. Atypical melanocytes, arranged in nests or singly, may involve the entire length of the hair follicle, including the sebaceous duct, with extension into the adjacent papillary dermis.
Melanoma in Children
Cutaneous melanoma is rare in childhood, particularly before puberty. Clinically, as with adults, melanoma in children mainly affects the white population. There likely exists a minor female predominance with slightly worse males prognoses ; however, some studies report equal gender distributions and prognoses. Similar to adults, clinical characteristics concerning for melanoma in children include rapid size increase, bleeding, color change, itch, lymph node enlargement, subcutaneous mass, pain, and distant metastases. The most common primary tumor sites are the extremities, followed by the trunk. Most melanomas in children arise from large or giant congenital nevi. Histologic features of melanomas in children and adults are identical, including asymmetry, atypical melanocytes, mitoses, pagetoid spread, irregular epidermal nest formation, and failure of maturation with progressive depth. The diagnosis is difficult when the lesions demonstrate spitzoid morphology because most such lesions represent Spitz nevi. When spitzoid lesions in children exhibit poor maturation and contain deep mitoses near their bases, the diagnosis of melanoma should be strongly considered. In questionable cases, err on the cautious side by performing excision with appropriate margins, and if necessary, also perform SLN excision. The 10-year cause-specific survival rate in children with melanoma is 89.4%.
Special problems in the histologic diagnosis of cutaneous melanoma
There are several problems clinicians and pathologists must be aware of that can create difficulty in the diagnosis of melanoma other than those described previously. Some of these represent mimics of melanoma whereas others involve certain settings in which melanoma may develop or be altered by other factors.
Dyplastic Nevi
DN are associated with an increased risk of melanoma (up to a 10-fold increased risk) when present in an individual with a personal or family history of melanoma. Melanoma risk increases as the number of DN increases. DN have been reported in up to 34% to 56% of melanoma cases although these findings are somewhat controversial because the criteria used to distinguish between DN and malignant melanoma (MM) as well as DN and other forms of nevi are somewhat subjective. Several studies have investigated the ability of pathologists to distinguish melanoma from DN. Concordance rates in these studies have ranged from 49% to 76%, which illustrates the diagnostic challenge melanocytic lesions present to pathologists.
Criteria for the histologic diagnosis of DN include both architectural and cytologic features. Architectural features include poor circumscription, lentiginous melanocytic hyperplasia, subepidermal fibroplasia, and variable degrees of associated dermal lymphohistiocytic infiltrate. Cytologic features include the presence of random cytologic atypia and small dermal melanocytes with at least some maturation with dermal descent. In the past, some investigators have proposed classifying the cytology of DN as mild, moderate, or severe. Due to difficulty in distinguishing between mild and moderate dysplasia, however, a simplified classification scheme using low-grade and high-grade dysplasia was proposed. Nevertheless, none of the proposed classification schemes have proved to provide guidelines for the consistent diagnosis or classification of DN.
Nevi on Special Sites
Certain body sites show a tendency to express melanocytic nevi in a way that may closely simulate melanoma. These sites include the breast, scalp, ear, umbilicus, perineal area, genital area, and flexural sites, such as the axilla, neck, popliteal, and antecubital fossa ( Fig. 7 ). Clinicians and pathologists must recognize the potential for mimicry at these sites to avoid overdiagnosing melanoma. Histologic features that may cause benign nevi to simulate melanoma include large size of individual melanocytes, large nests of melanocytes, and slight pagetoid spread of melanocytes in the epidermis. An especially confusing situation arises in nevi that develop on genital areas affected by lichen sclerosus et atrophicus. These nevi, like melanoma, may demonstrate prominent pagetoid spread and atypia but in actuality are benign.
Congenital Nevi Biopsied in Neonates
Congenital nevi biopsied shortly after birth and rarely up to 2 to 3 years may simulate melanoma by exhibiting pagetoid spread of large melanocytes with abundant cytoplasm throughout the epidermis. This is believed to result from migration of evolving nevus cells (melanoblasts) through the epidermis. As congenital nevi mature, the single melanocytes in the epidermis congregate in nests that then become positioned at the DE junction and in the dermis. In addition to prominent intraepidermal melanoctyic proliferation, there exists concurrent expression of nests of typical-appearing melanocytes in the dermis situated between and among collagen bundles and around adnexal structures. The presence of these benign nests is characteristic of a congenital nevus. Clinical correlation, including the location of the lesion, size of the lesion (large congenital melanoctyic nevi carry a 2.5%–5% risk of developing into melanoma), and presence of evolution in the lesion, remains valuable in making an accurate diagnosis.
Congenital Melanoblastic Proliferation
Occasionally seen in congenital nevi are large nodular aggregates of small hyperchromatic melanocytes. Pleomorphism and mitoses may occur in these lesions simulating melanoma. Distinguishing histologic features indicating a benign nature include symmetry, good circumscription, and evidence of maturation. Furthermore, this lesion is almost always present at birth, matures with age, and may undergo complete regression.
Persistent (Recurrent) Nevi
Nevi that have been previously biopsied or damaged by trauma, such as shaving, may develop histologic features that simulate melanoma. In these situations, melanocytes in the epidermis may be large and can coalesce with involvement at all levels of the epidermis, including the granular and cornified layers. In addition, scattered mitotic figures may be present. Characteristic of these lesions is confinement of atypical changes in the area of epidermis directly above the scar tissue in the dermis ( Fig. 8 ). If the melanocytic proliferation extends beyond the margin of the scar, diagnosis of melanoma should be considered. Clinical correlation, as well as review of any available histologic sections from prior biopsies at that site, may be helpful. If doubt exists, conservative re-excision would be a reasonable action.
Single Cell Melanocytic Proliferations
In some instances, a benign proliferation of single melanocytes in the epidermis may occur and simulate melanoma in situ, which begins in a similar fashion. This may occur in many different settings, including in the presence of an underlying congenital nevus on sun-damaged skin, on the eyelid, in the epidermis overlying fibrous papules, within solar lentigines, and on mucocutaneous sites, such as the nail unit. Benign features include a lower density of melanocytes compared with melanoma in situ as well as a lack of nests. The presence of an obvious underlying congenital nevus in the dermis beneath the epidermal proliferation allows the diagnosis to be made with certainty. Clinical correlation may be valuable because most of these lesions do not demonstrate visible pigment. This is particularly important when dealing with excision margins of melanoma.
Regressed Melanoma
After regression, melanomas exhibit a characteristic pattern of dermal fibrosis in the upper dermis along with telangiectasias, lymphocytes, and abundant melanophages. The fibromucinous stroma differs from fibroplasias seen in scars. In regression, fibrosis occurs broadly and mirrors the position of the original melanoma that was present previously. Residual scattered single melanocytes may remain in the epidermis along with a few nests of atypical melanocytes in the fibrotic dermis.
Metastatic Melanoma
Cutaneous metastases of melanoma are typically characterized by atypical melanocytes in the deep dermis or subcutis arranged in either nodules or diffusely in cords and strands. Occasionally, involvement of the superficial dermis and epidermis occurs. Generally, these lesions are as deep as they are broad and contain atypical melanocytes with mitoses throughout. These lesions can be small, however, which further complicates diagnosis.
A distinct form of metastatic melanoma expresses atypical but delicate pigmented spindle or dendritic melanocytes in the dermis. These lesions may simulate blue nevi; however, this variant of metastatic melanoma usually develops rapidly and demonstrates at least some atypia with occasional mitoses, allowing the diagnosis to be made.
Artifact from Monsel Solution
Dermatologists commonly use Monsel solution (ferrous subsulfate) to control bleeding after a biopsy. This solution contains iron that, when applied to the skin, is engulfed by histiocytes. The iron-containing histiocytes appear brown and may simulate atypical melanocytes or melanoma. Careful inspection reveals a refractile quality of the pigment similar to the appearance of hemosiderin. In addition, the cells themselves in this instance are not significantly pleomorphic or atypical.
Techniques to improve accuracy of histopathologic diagnosis of melanoma
Special Stains and Immunohistochemical Stains
As discussed previously, melanoma may present with a wide spectrum of histologic features that mimic epithelial, hematologic, mesenchymal, and neural tumors. For poorly differentiated tumors or amelanotic melanomas, chemical and immunohistochemical stains are often necessary for correct diagnosis. A commonly used silver stain, which stains melanin, is the Fontana-Masson stain. Immunohistochemical stains directed against S-100 protein remain the most sensitive marker for melanocytic lesions (sensitivity for melanoma is 97%–100%). The name S-100 is derived from the observation that this 21-kDa calcium-binding protein is soluble in 100% saturated ammonium sulfate solution. Unfortunately, S-100 stain lacks specificity, because the protein is also expressed in nerve sheath cells, myoepithelial cells, adiopocytes, chondrocytes, Langerhans cells, and all tumors derived from these cells.
Most spindle or desmoplastic lesions only express S-100 positivity, which gives the other available stains little value in these cases. Particularly challenging is differentiating DM, which can appear scar-like, from actual scar tissue, because both lesions stain positive for S-100. Chorny and Barr found the spindle cell component from previously biopsied nonmelanocytic neoplasms showed S-100 positivity in 90% of cases. In addition, they found 100% S-100 positivity in re-excision scars from previously biopsied nevomelanocytic lesions. They concluded the presence of S-100 positive spindle cells in scars may create a diagnostic pitfall, particularly in the evaluation of re-excision specimens of DM. To further exacerbate this potential pitfall, the vast majority of desmoplastic lesions only express S-100, rendering the remaining arsenal of melanoma stains useless.
A more specific, albeit less sensitive, marker for melanoma, is HMB-45. This marker stains the cytoplasmic premelanosomal glycoprotein, gp100. Reported sensitivities for melanoma range from 69% to 93%; sensitivity for primary melanomas is greater than metastatic melanoma. HMB-45 is specific for melanocytic differentiation and was reported to be absent 100% of the time in nonmelanocytic tumors in several studies. HMB-45 is also expressed, however, in tumors other than melanoma, including PEComas, sweat gland tumors, meningeal melanocytomas, CCS of the tendons, and aponeuroses as well as certain ovarian steroid cell, breast, and renal cell carcinoma tumors. Fortunately, most of these tumors are histologically distinct from melanomas.
A cocktail that stains for HMB45, Mart-1, and tyrosinase exhibits a high sensitivity for all forms of melanoma and is a complementary marker to polyclonal S-100 protein. Additional stains that are available clinically include MITF, NKI/C3, CD10 antigen, and vimentin. MITF has good sensitivity (81%–100%) and specificity (88%–100%) for melanoma but is less specific for spindle cell lesions and stains tumors of many other lineages. NKI/C3 is not widely used due to poor specificity and high cost. CD10 antigen, a neutral endopeptidase expressed by a variety of mesenchymal tumors, is significantly up-regulated during the process of metastasis in melanoma and may help distinguish between primary versus metastatic melanoma. Vimentin exhibits good sensitivity (96%) and if negative can be used to rule out melanoma. Newer markers for melanoctyic differentiation that are being investigated but are not yet widely used clinically, include MUM-1, melanocortin 1, SM5-1, PNL2, and TRP-1/TRP-2.
Cytogenetics
Understanding of the genetics of melanoma has greatly improved in recent years. Melanoma is recognized as a heterogenous tumor characterized, in part, by DNA mutations that lead to either a loss of function of tumor suppressor genes or activation of oncogenes. Different subtypes of melanoma express characteristic mutational and karyotypic profiles. Two important and well studied pathways are implicated in the pathogenesis of melanoma: the mitogen-activated protein kinase (MAPK) (or RAS-[B]RAF-MEK-ERK) pathway, which contributes to the regulation of cell growth, and the phosphatidylinositol-3-kinase-PTEN (PI3K/AKT/PTEN/mTOR) pathway, which is involved in the regulation of cell death. In melanoma, the MAPK and PTEN/AKT pathways may become activated aberrantly by one of several activating oncogenic mutations. Known mutations involve BRAF, NRAS, and HRAS.
The activating BRAF V600E mutation was found in 60% of melanomas in one study. These melanomas, however, are the type that arises in association with intense intermittent sun exposure. Some melanoctyic nevi, however, contain activating BRAF mutations that play a role in melanocytic growth arrest: stimulating oncogene-induced senescence. In these lesions, the progression to melanoma might occur after the loss of cell cycle regulators, such as p16.
Many studies have looked at genetic mutations in specific variants of melanoma, particularly those variants where distinction between benign and malignant is challenging. A few studies looking at BRAF and NRAS have shown these mutations are common in spitzoid melanomas but absent in Spitz nevi. Other studies, however, have disputed this claim. HRAS mutations seem limited to Spitz nevi, so detection of this mutation favors the diagnosis of Spitz nevus over spitzoid melanoma. Congenital nevi tend to have NRAS mutations, not BRAF mutations. Blue nevi are characterized by GNAQ mutations. In the hereditary melanoma syndrome, germline mutations in the cyclin-dependent kinase inhibitor 2A (CDKN2A) encoding p16 tumor suppressor proteins is frequently mutated. Further studies are needed to assess the potential diagnostic value of these mutations.
Array Comparative Genomic Hybridization
Array comparative genomic hybridization (CGH) is a powerful technique that can assess the entire genome for genetic aberrations. A landmark study by Bastian and colleagues in 2003 used CGH to evaluate genetic aberrations in melanocytic lesions. They found multiple genetic aberrations in 96% of melanomas whereas nevi only rarely exhibited mutations. The only nevi to express mutations were Spitz nevi, where a specific single gain on the short arm of chromosome 11 was found. This mutation seen in Spitz nevi was not appreciated in any of the melanomas studied. Subsequent studies evaluating array CGH in congenital nevi showed no aberrations whereas melanomas arising in congenital nevi showed aberrations typical of de novo cutaneous melanoma. Additionally, proliferative nodules seen in congenital pattern nevi have genetic aberrations but these differ from those seen in melanoma. Ultimately, the use of array CGH will likely be limited to research because of its high cost, time-consuming nature, need for special equipment/expertise, and the requirement of a significant amount of relatively pure (primarily tumor) tissue.
Fluorsescence in Situ Hybridation
Fluorescence in situ hybridatization (FISH) technique has emerged as the preferred molecular technique to interrogate chromosomal abnormalities because of its low cost compared with CGH, efficacy with use of archival tissue, and increasing availability in pathology laboratories. This technique is limited in the sense it can only investigate a maximum of 4 genetic aberrations per experiment.
Currently, a 4-probe melanoma FISH test targeting chromosomes 6 (6p25, 6q23, and centromere) and 11 (11q13) is commercially available. A validation study of this test revealed a sensitivity of 87% and specificity of 95%. Subsequent studies have found similar results. The majority of these investigations, however, have only studied the performance of the 4-probe melanoma FISH test in unequivocal cases. This test would not be ordered in cases where the diagnosis of a melanocytic lesion is certain. Thus, if FISH proves to have diagnostic value, it will be in ambiguous cases. Gerami and colleagues evaluated the performance of the 4-probe FISH test in 27 cases of ambiguous melanocytic lesions. After 5+ years of follow-up, 6 cases developed macroscopic metastatic disease whereas 21 cases remained disease-free. Of the 6 patients who went on to develop metastatic disease, all 6 of them had a positive FISH test. Six of the 21 disease-free patients, however, were FISH positive as well. A limitation of this investigation is that follow-up can only prove malignancy of the lesion. Follow-up cannot differentiate benign lesions from malignant lesions that were cured with removal. Nevertheless, mounting numbers of cases with long-term event-free follow-up and negative FISH tests increases confidence in the ability of this test to identify benign lesions.
The efficacy of FISH has been explored in several other settings. One study found FISH can be used to help distinguish benign intranodal nevi from metastatic melanoma. In a study of lentiginous junctional melanomas of the elderly, 84% of cases tested positive using FISH. FISH might also be helpful in distinguishing between epithelioid blue nevi and blue nevus-like metastatic melanomas based on a small study that showed excellent sensitivity (90%) and specificity (100%). Similarly, another small study used FISH to distinguish nevoid melanomas from mitotically active nevi. They found the 4-probe melanoma FISH test performed perfectly in this setting (100% sensitive and specific).
Several problems with FISH exist in addition to a paucity of data evaluating FISH in equivocal cases. For one, the currently widely available 4-probe test is not sufficiently sensitive or specific to accurately diagnose melanoma by itself. The results of this test must be interpreted in conjunction with histologic and clinical information. Also, FISH requires 20 to 30 well-visualized cells to get an accurate count of fluorescent signals, so small melanoctyic lesions may not have sufficient cells for FISH analysis. Additionally, FISH relies on mathematic algorithms to define positive and negative results. If rigid standards of technique and assessment are not followed, diagnostic discrepancy and uncertainty increase and limit the diagnostic value of FISH. One study investigated intraobserver variation in FISH and found that 3 pathologists, evaluating the same data, failed to reach agreement in 32% of cases. This study calls into question the reproducibility of FISH tests. For FISH tests to be of any diagnostic use, it is important that pathologists adhere to stringent standards at every step of the testing process.
Pathologists face a difficult situation when deciding when to order a FISH test. Gerami and Zembowicz advocate using this test as an adjunctive test to routine pathologic examination and clinical correlation in borderline cases. Other investigators advocate using FISH in challenging cases where no consensus diagnosis is reached, particularly in those cases where clinical consequences are substantial. Further studies are needed to clarify the diagnostic value of FISH.
Proliferation Markers
Many malignancies exhibit identifiable proliferation markers that can aid in both diagnosis and prognosis. Ki-67, a nuclear antigen present in all active phases of cell cycle proliferation, is the most widely used proliferation marker in pathology. This antigen is not expressed during nonproliferative stages of the cell cycle (G0 and early G1). Antibodies to Ki-67 can be used to label melanocytes in an effort to correlate the rate of melanocyte proliferation with histologic malignancy.
Smolle and colleagues were the first group to investigate the proliferative potential of melanoctyic lesions by staining for Ki-67. They examined 25 melanocytic skin tumors using antibodies to Ki-67 on fresh frozen tissue. Using the parameter of growth fraction (percentage of positively stained nuclei to total nuclei), they found statistically significant differences between nevi, primary melanomas, and MMs. Similarly, Boni and colleagues successfully used MIB1, another antibody to Ki-67, to stain Ki-67 on formalin-fixed, paraffin-embedded specimens. This discovery permitted Ki-67 staining of archival tissues. Since the work of Smolle and colleagues, several studies looking at Ki-67 as a proliferation marker in melanoma have found similar results. Benign nevi typically express Ki-67 positivity in less than 5% of tumor cells, although some reports have shown up to 15% positivity in Spitz nevi and DN. Conversely, melanoma lesions stain positive for Ki-67 in 13% to 30% of cells. Spitz neoplasms also typically demonstrate increased Ki-67 staining. There are conflicting data concerning whether Ki67 is an independent risk factor for adverse outcome in melanoma.
Another marker of cellular proliferation is proliferating cell nuclear antigen (PCNA). There is increased expression of this marker in MM compared with benign nevi; however, Spitz nevi also show increased expression of PNCA. Like Ki-67, the prognostic implication of PCNA remains subject to debate.
Cyclins (A, B, D1, D3, and E), another class of proliferation marker, bind and activate cyclin-dependent kinases causing the cell to progress through the cell cycle. Cyclin A, rarely expressed in benign nevi, is expressed in 42% to 99% of melanomas. Florenes and colleagues reported an inverse correlation between cyclin A staining and disease-free survival in superficial melanoma but not in nodular melanoma. Similarly, cyclin B is rarely expressed in benign nevi but is expressed in approximately 50% of melanomas. The prognostic implication of cyclin B has not been adequately studied. Cyclins D1 and D3 are rarely expressed in benign nevi and commonly expressed in melanoma. No prognostic significance was seen in cyclin D1, whereas increased cyclin D3 staining was associated with early relapse and decreased survival in superficial melanomas only.
Other markers of proliferation under investigation include p16, p21, p27, p53, HDM2, and GADD proteins. HDM2 and GADD proteins possess promise as prognostic markers in melanoma; however to this point, none of the proliferation markers has been definitely shown to have independent prognostic value.
Related posts:
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
