More than two million Americans were diagnosed with skin cancer in 2014, making it the most common malignancy in the United States and accounting for considerable morbidity. Most of the skin cancers are basal or squamous cell in origin with approximately 80% of these representing basal cell carcinomas (BCCs). Of these nonmelanoma skin cancers, most are curable but still approximately 2,000 people die each year from squamous cell carcinoma (SCC) or BCC.¹ Melanoma accounted for approximately 76,100 cases and was responsible for an estimated 9,710 deaths in 2014, which far surpasses the number of deaths due to all other skin malignancies combined.² Melanoma continues to increase in incidence at a higher rate than any other cancer in the United States, except for non-small-cell lung cancer in women. Approximately 5,120 cases of nonepithelial skin cancer cases with 3,270 deaths were diagnosed in 2014.1 These less common tumors of the skin include Merkel cell cancer, Kaposi sarcoma (see Chapter 25), and mycosis fungoides (MF).

Melanoma arises from pigment-producing melanocytes that migrate to the skin and eye from the neural crest during embryologic development. Approximately 5% of melanoma occurs in noncutaneous sites such as the eye and mucous membranes of the oropharynx, sinuses, vagina, and anus.³ Patients can present with regional lymph node involvement or distant metastatic disease without any primary being identified. This occurs in approximately 5% of patients. Melanoma occurs more commonly in men than women and has a peak age at incidence of approximately 50 years. Owing to the young age of many melanoma patients, this disease takes a striking toll in terms of the average number of years of life lost per patient in the United States. The incidence of the disease has increased in the United States to the point where melanoma is now the sixth most common cancer in men or women. The substantial increase in incidence is presumably due to increased exposure to sunlight (primarily ultraviolet B radiation), with the greatest risk of melanoma felt
to be in those who have intermittent intense sun exposure, particularly in fair-skinned, light-haired individuals with red and blonde hair, and blue or green eyes. The cultural emphasis on sun-tanned skin as an indicator of physical health and beauty has played a major role in this increase. Depletion of the ozone layer may contribute as well. Sunny parts of the United States have the highest incidence of the disease, especially California, Florida, Arizona, and Texas, which include three of the four most populous states in the United States. One particular melanoma subtype, lentigo maligna melanoma, which often occurs on the face, may be more closely associated with long-term occupational sun exposure and is seen in farmers and other outdoor workers. Patient education in prevention, including use of sun-protective clothing, performing outdoor activities at times other than the brightest sunlit hours of the day, use of topical sunscreens, refraining from use of sun-tan parlors, use of skin self-examination, and avoiding sun-tanning (“tanned skin = damaged skin”), should be emphasized. Individuals with xeroderma pigmentosa, an autosomal-recessive disorder, typically incur multiple basal and squamous skin cancers and melanoma because their skin lacks the ability to repair damage induced by ultraviolet radiation.

Melanomas arise not only from sporadic or familial atypical nevi but also from other congenital and acquired nevi; however, approximately half of cutaneous melanomas arise without a clearcut precursor lesion. Individuals who have more than 20 benign nevi are at increased risk for melanoma. Approximately 10% of patients with melanoma have a family history of this cancer. Careful surveillance should be carried out in patients with these risk factors. Suspicious-appearing lesions or lesions that appear to have changed coloration, shape, height, or have bled should be excised. The familial atypical multiple mole melanoma syndrome is characterized by a young mean age at diagnosis (34 years) and multiple lesions. The most common germline mutation seen in familial melanoma occurs in the tumor suppressor gene CDKN2A. CDKN2A, PTEN, NRAS, and BRAF mutations have also been seen in nonfamilial melanoma.⁴,⁵

Clinical features, classically known as “ABCDE,” that raise suspicion for melanoma include:

Other characteristics of concern include history of recent growth, change in pigmentation, ulceration, itching, or bleeding. Any pigmented lesion that returns after excision should be reevaluated with biopsy. Nonpigmented skin lesions that behave like melanoma should be examined with immunohistochemical stains S-100 and HMB-45 as 1% to 2% of melanomas are amelanotic.³

There are four clinical types of primary cutaneous melanoma.³ Superficial spreading melanoma is the most common type, accounting for 70% of melanomas. It is commonly found on the trunks of men and lower extremities of women. Nodular melanoma comprises 10% to 15% of melanomas and has an early vertical growth phase. It is commonly found on the trunks of men. Those lesions associated with intermittent sun exposure are often (50% to 60%) BRAF mutated but C-KIT wild type. Lentigo malignant melanoma accounts for approximately 10% of cases. It is characterized by flat, large (1 to 5 cm) lesions located on the arms, hands, and face of the elderly (median age 70 years) in particular and is known for a relatively longer radial phase. Acral lentiginous melanoma is seen in approximately 3% to 5% of cases and occurs primarily on the palmar surfaces of the hands, plantar surfaces of the feet, and under nails on the digits. This melanoma subtype is most commonly seen in individuals with darker-pigmented skin and is felt not to be as closely related to sun exposure as the other subtypes. Mutations in exons 9 and 11 of the C-KIT gene are more commonly observed in acral lentiginous and mucosal melanomas than other subtypes, but still only occur in about 20% to 30% of these cases.⁴

In general, melanoma is felt to show two distinct growth phases: an initial radial phase during which the melanoma enlarges in a horizontal/superficial pattern above the basal lamina of the skin, followed eventually by a vertical growth phase characterized by invasion deeply with exposure to lymphatic vessels and the vasculature. It is during the vertical growth phase that metastases are felt to be most likely to occur (Table 15.1).

Melanoma has a proclivity for direct nodal spread presumably through the lymphatics, but a significant proportion of lesions exhibit hematogenous spread as well. Common sites of metastases include lung, liver, bone, subcutaneous areas, and, primarily in late stages, brain. However, melanoma can spread to virtually any site and can imitate virtually any solid malignancy in its pattern of spread. Following diagnosis, approximately 25% of patients will develop visceral metastases. An additional 15% may develop disease limited to lymph nodes. Patients who present with lymph nodal or metastatic involvement without any obvious primary site may have undergone spontaneous remission of the primary, a phenomenon that may be attributable to some degree of immune system involvement. Interestingly, those patients may have a better outcome than similarly staged patients with known primaries. Patients with “cancer of unknown primary” should have their biopsy material analyzed with the immunohistochemical stains S-100 and HMB-45 to consider the possibility of melanoma.

Ocular melanoma is the most common malignancy of the eye in adults. It may occur in any eye structure that contains melanocytes, although uveal tract sites predominate, followed by choroid, ciliary body, and iris in decreasing frequency. Standard therapy may consist of either enucleation (often utilizing a “no touch” technique) or brachytherapy with radioisotopes such as iodine-125. A recently published large randomized study of those two treatments revealed that for primary uveal tumors less than 5 mm in depth, the outcome for survival was identical. This tumor metastasizes most frequently to the liver and appears to be less sensitive to both biologic agents and chemotherapy than is cutaneous melanoma.

Melanoma is staged according to the American Joint Committee on Cancer staging system.⁶ All patients should have a careful history and physical examination with special attention to the skin including scalp, mucous membranes, and regional lymph nodes. Pathology report should include thickness of the primary lesion, presence of absence of ulceration, number of mitoses per millimeter squared, and presence of lymphatic invasion.²,⁷,⁸ Laboratory studies should include complete blood count, blood urea nitrogen, serum creatinine, liver panel, alkaline phosphatase, and serum lactate dehydrogenase at baseline. A baseline chest radiograph is obtained to evaluate for pulmonary lesions. A computed tomography (CT) scan can be considered if clinically warranted. Elevation of liver function tests warrants further imaging of the liver, most typically with CT. Unexplained bone pain should also
be evaluated with CT or magnetic resonance imaging. Primary lesions equal to or thicker than 1.0 mm are at higher risk of regional lymph node involvement; therefore, the use of sentinel node mapping is recommended for lesions between 0.76 and 1 mm and above. Important recent additions to the staging criteria include the concept of nodal disease burden, especially at the time of the sentinel node biopsy, and the presence or absence of mitoses in the primary lesion, which is a very significant negative prognostic factor.⁷,⁸

The standard treatment for skin lesions suspected of being melanoma is excisional biopsy rather than incisional or “shave” biopsies. A subsequent wide and deep excision is required to provide adequate tumor-free margins as melanoma has a known propensity for local recurrences. While there is some variation in recommendations, most experts would advocate a 1-cm tumor-free margin for melanomas less than 1 mm in thickness and 1- to 2-cm margins for deeper primary lesions if technically possible, following current National Comprehensive Cancer Network guidelines.⁹ Additionally, for primary lesions of at least 1 mm, sentinel node mapping is recommended. Lymph node “drainage areas” are assessed via a specific lymph node (sentinel node[s]—sometimes more than one) into which lymph-borne metastases generally first occur.¹⁰ The absence of tumor involvement in the lymph node is associated with a reduced risk of nodal spread and systemic relapse in general and eliminates the need for subsequent dissection of that nodal basin.¹¹,¹² In the recent Multi-Center Sentinel Lymph Node Trials 1, 1,347 patients with primary melanomas of 1.2 to 3.5 mm, felt to be at intermediate risk of recurrence, were randomly allocated to receive either observation or a sentinel node biopsy, with completion of lymphadenectomy if the sentinel node was positive and observation only if negative.¹³ A delayed lymph node dissection was performed in case of nodal recurrence in either group. Preliminary results from that trial suggest that there was no survival advantage for the performance of a sentinel lymph node biopsy in this risk group, although it reduced the relative risk of recurrence at any site by 26%, reduced the absolute chance of recurrence locoregionally from 15.6% to 3.4%, and confirmed that those with a positive sentinel node had a worse outcome than those with a negative sentinel node biopsy.

Eastern Cooperative Oncology Group (ECOG) protocol E1684 was a large randomized adjuvant trial of interferon (IFN)-α_2b in patients with deep primary lesions (>4 mm thick) or regional lymph node involvement that showed statistically significant improvement in overall survival in the treated group compared with the observation group.¹⁴ The regimen used was IFN 20 million IU/m²
intravenously (IV) 5 days/week for 4 weeks (as a “loading phase”) followed by 10 million IU/m² subcutaneously (SC) 3 days/week for 48 weeks as a maintenance phase. Toxicity was significant, but quality-of-life analysis demonstrated overall benefit. The follow-up study of observation versus the same IFN regimen versus a lower-dose regimen, ECOG 1690, also showed a significant disease-free survival advantage over the observation arm but not a benefit in overall survival for the high-dose regimen.¹⁵ The difference between these two studies may be that patients on the observation arm in the subsequent trial (1,690) may have been treated with immunotherapy (including IFN or interleukin [IL]-2) at the time of relapse. The final published randomized adjuvant study of high-dose IFN compared with a vaccine demonstrated a benefit in terms of relapse-free and overall survival for the IFN arm.¹⁶ A phase II study that evaluated a month of high-dose IFN versus the standard 1 year of drug for high-risk resected melanoma patients pioneered in ECOG 1684 suggested that the 1-month regimen was inferior.¹⁷ One month of high-dose IFN was also compared with chemotherapy for resected mucosal melanoma patients, with no advantage for the IFN cohort.¹⁸ Several recent meta-analyses of randomized trials of high-dose IFN have shown that while there is a statistically significant and consistent advantage in relapse-free survival, the benefit for overall survival is very modest, at 2% to 3%.¹⁹,²⁰ Nonetheless, given that patients with deep cutaneous primaries and/or lymph node involvement are at high risk for metastatic recurrence and that the majority of patients who suffer metastatic relapse will die of their disease, it is reasonable to treat such high-risk patients with either adjuvant high-dose IFN or to consider entrance into a clinical trial. While predictive biomarkers for the utility of adjuvant IFN are sought after and have not been clearly defined, the onset of vitiligo is often associated with a favorable outcome, as are other manifestations of autoimmunity, although their assessment is complicated by lead-time bias.²¹ This is because patients that appear to benefit from treatment will stay on therapy longer, increasing the likelihood of developing vitiligo.

Peginterferon, which prolongs the half-life of the drug, allowing it to be delivered weekly, has been tested in several adjuvant trials in resected melanoma. In patients whose lesions were detected at sentinel node biopsy, there was an advantage in disease-free survival for peginterferon. In long-term follow-up of an EORTC randomized trial of peginterferon compared with placebo in patients with stage III resected disease, patients with ulcerated primary lesions that had nodal disease proven only by a sentinel lymph node biopsy had clear prolongation of survival with peginterferon.²²,²³ Ipilimumab, a CTLA-4-blocking human antibody, has been tested in a randomized phase III trial compared with placebo in patients with resected stage III melanoma.

There was an increase in relapse-free survival for patients receiving up to 3 years of ipilimumab at 10 mg/kg compared with those receiving placebo, albeit at the cost of significantly increased side effects, chiefly so-called immune related adverse events.²⁴ Chemotherapy as a single adjuvant modality has not been shown to be more beneficial than observation alone, and high-dose IFN with chemotherapy confers no difference in relapse-free or overall survival compared with the single-agent arms.²⁵ Biochemotherapy was also compared with IFN in the adjuvant setting, though improvement was seen in relapse-free survival, no improvement was seen in overall survival plus there was increased toxicity in the biochemotherapy arm.²⁶

Given the current data, IFN or observation remains the standard adjuvant option in the United States. Ongoing studies of ipilimumab and nivolumab hope to change this standard in the future.

While melanoma is considered relatively resistant to chemotherapy, certain favorable prognostic factors do lend themselves to longer survivals with single-agent or multiagent chemotherapy, high-dose IL-2, or even newer approved agents like ipilimumab, pembrolizumab, or nivolumab. These include ECOG performance status 0 or 1; subcutaneous, lymph node, or pulmonary metastasis with normal lactate dehydrogenase (M1a or M1b disease); no prior chemotherapy; normal marrow, renal, and hepatic function; and absence of central nervous system (CNS) metastases. The biologic basis for these findings has not been fully elucidated. When reviewing potential therapy for stage IV melanoma patients, patient characteristics as well as the natural history of their metastatic disease must be considered. Surgical resection should be considered for one or even multiple sites of metastases that are resectable with a reasonable surgical procedure associated with modest morbidity. Failing that, for unresectable disease, an evaluation should be made whether the patient has indolent disease, suggesting that immunotherapy with frontline ipilimumab, or second-line pembrolizumab or nivolumab might be more appropriate, or more aggressive disease, indicating that other approaches are warranted, particularly a BRAF plus MEK combination if the patient’s tumor has a BRAF V600 mutation. While a clinical trial should generally be a first consideration for treatment-naïve or second-line patients with metastatic disease, high-dose IL-2 could also be considered in any line of therapy for the subset of patients that qualify for that rigorous therapy (Table 15.2).

This class of agents has been extensively tested in melanoma, a histology where they have met with some success.

a. Ipilimumab is a human immunoglobulin G1 (IgG1) antibody directed against the immune checkpoint protein cytotoxic T-lymphocyte antigen-4, expressed on activated T cells.²⁷ It has been extensively tested in stage IV melanoma and is an active drug, with objective response rates of 7% to 20%.²⁸ Many responses are sustained over time, and ipilimumab has been shown to induce novel patterns of response, with slow regression over 6 to 12 months, progression followed by regression, and new lesions appearing while other baseline lesions continue to regress. This drug has also demonstrated unique side effects that are directly related to its immune mechanism of action, called immune-related adverse events, consisting of rashes, colitis with diarrhea, hepatitis, pancreatitis, and hypophysitis leading to pituitary insufficiency. In relapsed/refractory melanoma at 3 mg/kg, ipilimumab plus a peptide vaccine had superior overall survival compared with the peptide vaccine alone, which is the first positive controlled randomized trial in stage IV melanoma ever conducted, and established a new standard of care.²⁹ A subsequent
randomized trial of ipilimumab with dacarbazine compared with dacarbazine plus placebo showed a statistically significant improvement in median and overall survival, further establishing ipilimumab as a front- or second-line standard of care drug in melanoma.³⁰,³¹ Recent work has suggested that CTLA-4 blockade with ipilimumab induced recognition of mutated “neo-antigens” expressed by melanomas and expanded the repertoire of antigen-specific T-cell recognition.³²

b. Pembrolizumab is an IgG4 humanized antibody directed against the checkpoint protein programmed death-1, or PD-1, expressed on T and B cells. In phase I and II trials, it showed a 40%+ response rate in patients with treatment-naïve melanoma, and a 21% to 25% response rate in patients who had failed prior ipilimumab. Grade 3 to 4 dose-limiting toxicity was 12% or less, and only 3% to 4% of patients had to stop treatment because of side effects.³³ In a recent randomized trial of two different doses of drug at 2 mg/kg or 10 mg/kg, no differences in response or progression-free survival were observed, and with a response rate of 24% and excellent duration of response the dose of 2 mg/kg given IV every 3 weeks was recently approved by the FDA for patients who failed ipilimumab, or if BRAF mutated, both ipilimumab and a BRAF inhibitor.³⁴ In a recent phase III randomized trial, progression-free and overall survival were significantly prolonged in ipilimumab-refractory patients who received pembrolizumab compared with chemotherapy. Ongoing trials will determine the role and potential benefit of pembrolizumb in treatment-naïve patients. Biomarker analyses of the trials above have shown that tumor PD-L1 expression and T-cell infiltrate are associated with a favorable outcome.³⁵

c. Nivolumab is an IgG4 human antibody directed against PD-1 that has very recently been approved by the FDA for patients with ipilimumab-refractory melanoma, or those who failed both a BRAF inhibitor and ipilimumab if their melanoma was BRAF V600 mutated. In initial trials it demonstrated activity in heavily pretreated, therapy-refractory melanoma patients at doses ranging from 0.1 to 10 mg/kg, with excellent duration of response and long median survival of 18 months.³⁶,³⁷ It induced a 32% response rate with long duration of response in that population, compared with 11% for chemotherapy, and a 9% rate of dose-limiting grade 3 to 4 toxicities, compared with 31% for chemotherapy.³⁸,³⁹ It adds another agent to the melanoma armamentarium for treatment-refractory disease. In a randomized trial of nivolumab compared with DTIC chemotherapy in previously untreated melanoma patients, the response rate was a robust 40%, with significantly better PFS and OS observed for the immunotherapy.⁴⁰