Head and Neck Mucosal Melanomas

Unlike its cutaneous counterpart, exposure to sunlight is not an etiologic factor for mucosal melanoma. Although irritants and carcinogenic compounds in the air, such as tobacco smoke, have been implicated in the development of this malignancy, the potential role of these compounds is unclear. In assessing the influence of cigarettes, Axell and Hedin demonstrated that chemical and physical stimulation caused hyperproduction of the melanocytes in the oral epithelium, resulting in oral pigmented lesions. ^[7]

Derived from the neural crest, melanocytes contain melanin pigment and are found in the basal layer of the epidermis, in the mucous membrane, and in the eyes. Melanocytes in the nasal cavity can be found in the respiratory epithelium, the nasal glands, the superficial and deep stroma of the nasal septum, and the middle and inferior turbinates. In oral mucosa, melanocytes are located along the tips and peripheries of the rete pegs. The function of melanocytes in mucosa is unclear. In physiologic states, the melanocytes in mucous membranes do not produce melanin and contain only nonmelanized melanosomes in their cytoplasm. However, they produce substantial amounts of melanin under pathologic conditions, such as Addison disease and neoplasm. In the skin, melanin aids in the absorption of ultraviolet light, acts as a scavenger for cytotoxic intermediates, and may play a role in nervous system development. ^[8]

A study Vos et al indicated that the CD8+ T-cell, cytotoxic cell, and IFN-γ signatures are stronger for head and neck mucosal melanomas than for mucosal melanomas of lower body sites. The investigators suggested that this may “reflect a more potent immune reaction induced by” a higher tumor mutational burden for head and neck mucosal melanoma than is found at other organ sites, as previous research has described. However, in the small cohort of patients in the Vos study, the “relatively more favorable immune profile” for head and neck mucosal melanomas did not manifest as a significantly better objective response rate (ORR) to immune checkpoint inhibitor (ICI) therapy (discussed below) for these lesions as compared with mucosal melanomas at lower sites (36% vs 25%, respectively). ^[9]

Mucosal melanoma of the head and neck region has a peak incidence in patients aged 60-80 years and is far rarer in children. A few authors have reported a slight male predominance, but among the few long-term survivors, no sex predilection is noted.

In consequence of their hidden location, head and neck mucosal melanomas are usually diagnosed in a locoregionally advanced clinical stage, with a rate of 5-48% of regional and 4-14% of distant dissemination. The nasal cavity is the most common location within the head and neck area. The exact origin of sinonasal tract disease is often difficult to ascertain because of anatomic limitations and advanced stage at presentation with involvement of multiple subsites.

Within the nasal cavity, disease of the anterior portion of the nasal septum accounts for most cases, followed by the middle turbinate, then the inferior turbinate. However, melanoma is virtually nonexistent in the superior turbinate, the olfactory region, or the ethmoid sinus. The palate and alveolar gingiva are the most common sites in the oral cavity. Other reported locations include the lower and upper labial mucosa, buccal mucosa, and tongue. Very rarely, mucosal melanoma can be found in the pharynx, larynx, or upper esophagus. The most common site in the larynx is the supraglottic region.

Prevalence of mucosal melanomas in the head and neck appears to vary by race. Mucosal melanoma, particularly in the oral cavity, is relatively common in Japan. In Ugandans, 10% of all melanomas are located in the oral or nasal cavities and are mucosal in origin. Those in the nasal cavity account for 2.6% of melanomas at all sites.

Most patients with sinonasal mucosal melanoma present with unilateral nasal obstruction, epistaxis, or a combination of the two. Discharge, epiphora, facial pain, and swelling are more common in advanced cases.

The most prevalent clinical presentation of tumors within the oral cavity is a painless mass. Commonly, ulceration and bleeding can be present. Determining whether a mucosal melanoma is a primary or metastatic lesion is often extremely difficult, because cutaneous melanoma may metastasize widely, including to the mucous membranes. Patients with a history of cutaneous or ocular melanoma or nevi that have regressed should be considered to have metastatic melanoma rather than primary mucosal melanoma. ^[10]

The most important features in defining a primary lesion from a metastatic lesion are site of involvement, presence or absence of pigment, overlying mucosal ulceration, extension along salivary gland ducts, and vascular and perineural invasion.

The differential diagnosis for intraoral melanotic lesions include mucosal melanoma, amalgam tattoos (focal argyrosis), melanotic macule, oral mucosal nevi, and melanoacanthoma.

Amalgam tattoos are blue-black discolorations of the oral mucosa secondary to the deposition of amalgam during restorative or surgical procedures, and they are the most common reason for mucosal pigmentation in the oral cavity. Amalgam tattoos are twice as common as melanotic macules and 10 times more common than oral nevi.

Melanotic macules are benign, pigmented lesions that do not appear to have malignant potential and are most commonly found on the vermilion border.

Oral nevi are relatively rare, with a prevalence of 0.1%. These lesions are more common in females (female-to-male ratio, 2:1) and Black individuals. Although definitive transformation to melanoma has not been documented, oral nevi may represent precursor lesions.

A melanoacanthoma is a lesion that is thought to develop through a reactive process and is not related to melanoma formation. However, it may mimic melanoma because of its rapid increase in size in a few weeks.

Pigmentation of the oral mucosa may also occur secondary to an inflammatory reaction to heavy metals or tobacco. Drugs commonly associated with mucosal pigmentation include antimalarials, estrogen, clofazimine, 5-fluorouracil, ketoconazole, busulfan, azidothymidine, doxorubicin, and minocycline.

For more information, see the Medscape Drugs & Diseases articles Pathology of Benign Melanocytic Nevi and Pathology of Dysplastic (Atypical, Clark) Melanocytic Nevi.

Gross appearance of the tumor is often variable, ranging from macular to ulcerated and nodular. Tumors arising in the nasopharynx are most commonly friable and polypoid in character. The clinical color of oral melanomas varies from black to gray to purple to red to white. Some lesions are uniform in color, whereas others exhibit marked variations. Grossly noticeable pigmentation occurs in approximately 75% of oral melanomas but in only 50% of sinonasal melanomas.

Mucosal melanosis has been reported to be frequently encountered adjacent to oral melanoma and may exist for a considerable period of time before diagnosis. In Japan, almost two thirds of oral melanomas are reported to be associated with melanosis; however, a preexisting pigmented lesion is not usually associated with mucosal melanomas in White patients. One suggestion is that melanosis represents the radial phase of the growth of the tumor and precedes the vertical component by years.

In contrast to cutaneous melanoma, the presentation of melanoma involving head and neck mucosal surfaces is typically at a more aggressive vertical growth phase, with invasion of the underlying submucosa. As a result of the advanced stage at discovery, most do not have an associated radial growth (superficial spreading) phase.

The presence of a pigmented lesion in the oral or nasal cavities should raise suspicion of mucosal melanoma, and a biopsy of the lesion should be promptly obtained. Diagnosis of pathologic disease is dependent on the identification of intracellular melanin. With a fresh tissue sample, dopa-positive melanocytes can be identified using the dopa reaction to demonstrate tyrosinase activity. Nonetheless, immunohistochemistry is often required to diagnose malignant melanoma, because only 50-70% of lesions in the oral cavity demonstrate melanin.

Microscopically, two histologic patterns can be described: an in situ pattern in which the neoplasm is limited to the epithelium and the epithelial-connective tissue interface, and an invasive pattern in which the neoplasm is found within the supporting connective tissue. ^[11] A combined pattern is usually typical of most advanced lesions. Several different cell types make up the tumor. Spindled, plasmacytoid, and epithelioid tumor cells arranged in a sheetlike, organoid/alveolar, neurotropic, or desmoplastic configuration may be observed. Melanin pigment is noted in almost 90% of lesions.

Only 50-70% of lesions in the oral cavity demonstrate melanin. Therefore, immunohistochemical analysis is often required to diagnose malignant melanoma. Melanomas react strongly with the alpha subunit of S-100, which is a calcium-binding protein found in neural tissues. However, this protein is present in a variety of normal and neoplastic cells. The frequency of S-100 immunoreactivity in mucosal melanoma varies from 86-100%.

A reactive antigen more specific to melanoma cells is HBM-45 (melanoma cytoplasmic antigen). ^[12] Melanoma also reacts with antivimentin and NK1/C-3 antibodies but not with antikeratin or antileukocytic antigen antibodies.

No universally accepted staging system for mucosal melanoma exists. Because of the absence of histologic landmarks analogous to the papillary and reticular dermis, the prognostic value of various levels of invasion, as established in the Clark classification for cutaneous melanoma, does not apply to mucosal melanoma. The following is the system that suffices for staging:

• Stage I – Localized disease

• Stage II – Metastases to regional lymphatics

• Stage III – Distant metastatic disease

Most patients (75-83%) present with stage I disease. Among patients with oral cavity lesions, however, a higher prevalence of stage II disease is noted.

The Memorial Sloan Kettering Cancer Center proposed a microstaging system for stage I disease based on the architectural anatomy of the mucosa, as follows ^[13] :

• Level I - In situ mucosal melanoma without invasion or with microinvasion

• Level II - Invasion up to the lamina propria

• Level III - Deep invasion into bone, cartilage, or skeletal muscle

According to the authors, these levels represent different microanatomic compartments separated by tissue barriers that are easily and reliably identifiable on light microscopy. They found this staging system to be a significant and independent predictor of survival in patients with localized, lymph node–negative, stage I disease. ^[6]

Because the appearance of melanoma on CT scanning is not specific (homogeneously enhancing mass), some authors suggest the use of MRI for its diagnosis. Melanin has paramagnetic properties that can affect signal and produce a characteristic intensity pattern on MRI. The appearance is hyperintense on T1-weighted sequences and hypointense on T2-weighted sequences. ^[14] This is presumably related to chelated metal ions or free radicals known to exist in melanin. Although the intensity pattern on MRI is specific for mucosal melanoma when visible, it is not found in all tumors, especially amelanotic melanoma.

The treatment of choice for mucosal melanoma, like cutaneous melanoma, is complete excision with adequate negative margins. Because of the presence of surrounding vital structures, positive surgical margins are common. Unfortunately, mucosal melanoma tends to spread radially and involve large areas of mucosa.

Controversies in the treatment of head and neck melanomas include lymphoscintigraphy with sentinel node biopsy, nodal dissection, margin size, the role of radiation therapy, and reconstruction.

Moreover, the consideration of extensive surgery should take into account operative morbidity and a high disease recurrence risk. ^{[15, 16]}

Therapeutic neck dissection is indicated for lymph node metastasis in the neck. Elective lymph node dissection in local disease only is not recommended because of the low frequency of subsequent development of lymph node metastasis. The issue of sentinel node biopsy, which is achieving growing acceptance in the treatment of cutaneous malignant melanoma, has not been studied in mucosal melanoma.

A retrospective database study by Abiri et al indicated that in patients with head and neck mucosal melanoma, the overall survival rate is greater in those who undergo surgery with immunotherapy (SI) than in those who are treated with surgery alone (hazard ratio [HR] = 0.672). Overall survival was also better in study patients who underwent SI or surgery with adjuvant radiotherapy and immunotherapy (SRI) than in persons who were treated with surgery with adjuvant radiotherapy but no immunotherapy (HRs = 0.425 and 0.594, respectively). ^[17]

The role of radiotherapy in the treatment of mucosal melanoma is not clearly defined, and malignant melanoma has traditionally been regarded as relatively insensitive to radiation, but some studies have suggested a positive benefit. ^{[3, 4, 5]} In vitro studies on cutaneous melanoma cell lines show that they may not be intrinsically radioresistant, and high dose-per-fraction treatment is necessary. Ample clinical and basic scientific evidence lends support to the theory that melanoma has a high capacity for sublethal damage repair, making it resistant to conventional fractionation schemes. Therefore, treatment is often more successful with higher doses.

Shibuya et al reported equal or better local control rates when using radiation as primary treatment modality. ^[18] In spite of its beneficial effect, radiotherapy is usually applied as an adjuvant modality reserved for positive surgical margins, local recurrence, or palliation. ^[19] Statistical analysis has not confirmed that surgery with additional radiotherapy improves the patient’s overall survival significantly, and, therefore, the role of postoperative radiotherapy is not settled.

In a review of 69 patients with mucosal melanoma, Temam et al found local control rates were 26% with surgery alone and 62% with postoperative radiation therapy, even though the individuals in the radiotherapy group had much more locally advanced tumors. ^[20]  In patients who underwent surgery alone, the median local disease-free survival period was 9 months, while a median survival period of 33 months was found in patients who underwent postoperative radiotherapy. For patients in the postoperative radiotherapy group, however, distant metastases developed faster than they did in patients in the surgery group. ^[20]

Owens et al did a comparable study at the University of Texas MD Anderson Cancer Center and showed that the addition of radiotherapy decreased the rate of local disease recurrence but did not significantly improve survival. ^[21]

Another study conducted by Temam et al found that postoperative radiotherapy improves local control. ^[22] This was particularly significant for small tumors for which the occurrence of metastases was related to local control and its corollary, survival.

Cutaneous malignant melanoma has been treated with neutron beam radiotherapy, with good results in controlling disease locally in patients with stage III melanoma. Whether these results can be extrapolated to the mucosal counterpart of the disease remains to be seen. ^[23]

Morris et al found that radiation-induced oral mucositis was a complication of boron neutron capture therapy, particularly with the use of p -boronophenylalanine. ^[24]

Chemotherapy/immunotherapy is usually used with an adjuvant or palliative intention. The most frequently used chemotherapy agents are dacarbazine, the platinum analogues, the nitrosoureas, and the microtubular toxins.

Immunotherapy is currently effective only in a small percentage of patients with malignant melanoma. ^[6] Increased response rates have been observed when IL-2 and IFN-α are used with cisplatin.

In a study of 44 patients with mucosal melanoma, including 18 (40.9%) with lesions of the head and neck, Moya-Plana et al found that in cases of unresectable and/or metastatic disease, immunotherapy with pembrolizumab, a programmed cell death protein 1 (anti-PD-1) inhibitor, had a better benefit-risk ratio than did immune treatment with ipilimumab, a cytotoxic T-lymphocyte-associated protein 4 (CTLA-4) inhibitor. (These agents are also known as immune checkpoint inhibitors [ICIs].) First-line immunotherapy with ipilimumab was associated with an objective response rate (ORR) of 8.2%, with one complete response, while first-line treatment with pembrolizumab was associated with an ORR of 35%, with four complete responses. The investigators also found that for first-line treatment with ipilimumab and pembrolizumab, median progression-free survival (PFS) was 3 months and 5 months, respectively. ^[25]

A study by Ohshima et al looked at 32 patients with advanced head and neck mucosal melanoma and found better overall survival rates in those individuals who were treated with ICIs (n = 14) than in those who were not (n = 18). The 1-year overall survival rate for the ICI patients was 85.7%, compared with 54.5% in the non-ICI group. For patients in whom ICI treatment led to a complete or partial response, the 1-year overall survival rate was 100%, compared with 71.4% for those who did not respond but instead had stable or progressive disease. ^[26]

However, a study of almost 1500 patients in the Netherlands with mucosal melanoma, diagnosed between 1990 and 2019 (including 34% with head and neck lesions), found that although survival did improve after immune and targeted therapies for the disease were introduced, these treatments appear to have brought about only a limited absolute survival benefit. Such benefit, for example, was seen in only a 2-month improvement in patients with regional or distant-spread disease. ^[27]

Consider adjuvant systemic therapy—including immunotherapy and targeted treatments (in the presence of an appropriate mutation)—subsequent to radical primary surgery. When unresectable and metastatic head and neck mucosal melanoma exists, immunotherapy with ICIs and targeted treatments (including BRAF or MEK inhibitors and KIT inhibitors) are among the palliative systemic therapies. ^[15]

Melanoma of the mucosal membrane appears to have a lower prevalence of regional lymph node metastases than melanoma of the skin does, both at presentation and at recurrence. Lesions of the oral cavity have a higher prevalence than those occurring in either the nasal or the pharyngeal cavities. Overall, 18% of patients have lymphatic metastases at presentation. The average distant metastatic rate at presentation is 10%.

Primary site recurrence occurs in approximately 40% of nasal cavity lesions, 25% of oral cavity lesions, and 32% of pharyngeal tumors. Overall primary site recurrence ranges from 55-66%, with a range of 16-35% for nodal recurrence. Most recurrences occur within the first 3 years. ^[28]

A study by Heppt et al of 444 individuals with mucosal melanoma, including of the head and neck, female genital tract, and anorectal region, found that 32.4% of patients had local relapses, with these occurring most commonly in cases of head and neck tumors. ^[29]

The mucosal melanoma of the oral cavity is very aggressive, and the absence of any standardized treatment protocol makes its prognosis unfortunate. The difficulties to obtain free surgical margins, the elevated tendency to invade in depth, and the early hematogenous metastasis have been referred as features that may explain its bad prognosis. As such, the prognosis for mucosal melanoma is generally quite poor, with a 3-year mortality rate of higher than 50%.

A retrospective study by Dréno et al indicated that owing to its high potential for metastasis, sinonasal mucosal melanoma has a poor prognosis. The report, which included 44 patients, 25% of whom had an initial tumor classification of T4, found the cumulative overall survival rates at 1 and 5 years to be 71.5% and 33%, respectively. ^[30]

A study by Abt et al using the Surveillance, Epidemiology, and End Results (SEER) Program database found the mean survival time for patients with paranasal sinus mucosal melanoma to be 16 months, versus 26 months for those with nasal cavity mucosal melanoma. ^[31]

Negative surgical margins and the size of the primary lesion in head and neck mucosal melanoma do not appear to be predictive of outcome. Five-year survival rates range from 15-30%, with a median survival time of 25 months. Gingival melanoma has a slightly greater 5-year survival rate (18%) than that of palatal melanoma (11%), with a longer median survival time (46 mo vs 22 mo, respectively).

Prognostic and predictive factors

According to a study at Memorial Sloan-Kettering Cancer Center, clinical stage at presentation, tumor thickness greater than 5 mm, vascular invasion on histologic studies, and development of distant failure are the only independent predictors of outcome of mucosal melanoma of the head and neck. ^[32]

A retrospective study by Song et al of 82 patients with oral mucosal melanoma indicated that cell type, ulceration, mitotic rate, pigmentation, necrosis, and vascular invasion are prognostic factors for overall survival, with multivariate analysis revealing cell type to be an independent risk factor (although not in cases of localized disease). In addition, cell type, mitotic rate, and the presence of tumor-infiltrating lymphocytes were reported to be risk factors for distant metastasis, with tumor-infiltrating lymphocytes cited as an independent risk factor. ^[33]

In the aforementioned study by Abiri and colleagues, the investigators reported that the survival prognosis in head and neck mucosal melanoma is negatively impacted by older age; paranasal sinus localization; T4, N1, or M1 classification; and positive surgical margins. ^[17]

The previously mentioned study by Dréno et al reported that clinical risk factors for poor prognosis in sinonasal mucosal melanoma include headache, facial pain, and trigeminal V2 nerve anesthesia, while nasal involvement alone, without sinus involvement, indicates a greater likelihood of survival. ^[30]

Histologic predictors of survival in patients with localized, lymph node–negative (stage I, N0M0) primary mucosal melanomas of the head and neck show that microstaging according to invasion into the following three tissue compartments are found to be a significant and independent predictor of poor survival in such patients:

• Level 1 – Melanoma in situ

• Level 2 – Invasion in the lamina propria only

• Level 3 – Invasion into deep tissue

The presence of sarcomatoid and pseudopapillary architecture and undifferentiated cells also appear to be associated with significantly poor disease-specific survival (DSS). Nodal involvement reduces median survival time to 18 months. Multiple local recurrences are the most common cause of treatment failure. Because of the natural history of this disease, however, 5-year survival rate data are of limited use, because the patient is constantly at risk of death from local recurrence and melanomatosis.

Regional vs local disease

Studies have shown that patients who present with regional disease have no worse survival than those who present with local disease only. The only clinical finding that appears to have definitive prognostic significance is the presence of distant metastasis at the time of diagnosis.

Temam et al reported that 68% patients developed distant metastasis soon after the end of the treatment even though local control was achieved in 44% of these patients. ^[22] Because wide surgical margins are often mutilating, this high rate of early distant metastasis underscores the need to detect lung, liver, and bone metastases before surgery is performed. Goerres et al showed that 18F-fluorodeoxyglucose (FDG) positron emission tomography (PET) scanning may be suitable for the staging and/or restaging of these patients. ^[34]

Pediatric versus adult disease

Mucosal melanomas in children display a much less aggressive character than do similar lesions in adult patients. However, cutaneous melanoma behaves very similarly in the pediatric and adult populations. The indolent behavior of mucosal melanomas in pediatric patients suggests that these lesions represent a different phenotype than that encountered in adult patients.

Primary melanoma arising in the mucous membranes is an unusual clinical entity. The best chance for a favorable outcome comes with early detection and excision. Currently, despite aggressive therapy, the prognosis for people with mucosal melanoma is extremely poor. Local treatment failure is a significant problem for most treated patients; however, patients with local treatment failure can be treated multiple times, resulting in prolongation of survival.

Distant metastases continue to be synonymous with rapid clinical deterioration and short survival time after detection. Surgery remains the mainstay of treatment, although adjuvant radiation therapy has had an increasing role in the treatment of mucosal melanoma. A clearer understanding of the biology of this disease process may yield more specific immunotherapy techniques.