Cancer Management Chapter 20: Melanoma and other skin cancers

March 9, 2010

Skin cancer is the single most common form of cancer, accounting for more than 75% of all cancer diagnoses. More than 1 million cases of squamous cell and basal cell carcinomas are diagnosed annually, with a lifetime risk of more than one in five. The vast majority of skin cancers can be cured with surgery alone. Resection is the mainstay of therapy, even for skin cancer involving regional lymph nodes or, in some cases, more distant metastatic sites.

Skin cancer is the single most common form of cancer, accounting for more than 75% of all cancer diagnoses. More than 1 million cases of squamous cell and basal cell carcinomas are diagnosed annually, with a lifetime risk of more than one in five. The vast majority of skin cancers can be cured with surgery alone. Resection is the mainstay of therapy, even for skin cancer involving regional lymph nodes or, in some cases, more distant metastatic sites.

Sun exposure is the predominant risk factor for squamous cell and basal cell skin cancer, and is the only known environmental risk factor for melanoma. Melanoma accounts for 4% of all skin cancer diagnoses (60,000 new cases of invasive melanoma per year) but contributes to 75% of deaths from skin cancer. Melanoma was diagnosed in 62,480 Americans in 2008 and accounted for 8,420 deaths (13.5% of patients diagnosed). The lifetime risk of melanoma for Caucasians is 1 in 39 for men and 1 in 58 for women in the United States. The 5-year survival was 92% of patients diagnosed with melanoma in the United States between 1996 and 2004, an increase of 10% compared with the years 1975 through 1977. This is almost certainly due to an increase in early detection.

The superficial nature of melanoma and other skin cancers supports campaigns to raise public awareness and healthcare provider expertise in detecting skin cancers at the earliest possible stage. Screening initiatives would be less likely to benefit the 10% of melanomas that are nonpigmented or arise from the choroid of the eyes or mucosal surfaces.

Epidemiology

Age The relationship between the incidence of melanoma and age is somewhat unique in comparison to other common cancers. There is not an exponential increase in risk with age but rather a more even distribution across age groups. The median age of diagnosis of melanoma is 53 years, almost 10 years younger than the median age of diagnosis of most common cancers. Forty-two percent of cases present in people younger than age 55, contributing to the third highest number of years of life lost across all cancers. In contrast, the incidence of squamous cell and basal cell carcinomas increases exponentially with age.

Gender Men are more likely than women to develop melanoma (67% higher incidence), and their prognosis is worse (136% higher risk of death from melanoma). The risk of recurrence in the setting of resected primary melanoma is higher, and their expected survival in the setting of regionally advanced disease is significantly shorter.

Location Basal cell cancers arise exclusively from cutaneous sites and are closely related to sites of skin that receive the most sun exposure, such as the scalp, face, neck, and arms. Squamous cell cancers can arise from numerous sites in the body, but the term squamous cell skin cancer is reserved for cutaneous sites. The vast majority of melanomas arise from cutaneous sites, but most cases of melanoma arise from intermittently sun-exposed skin. A small percentage of melanomas arise on acral surfaces of the hands and feet, which tend to be diagnosed at a later stage. Melanoma can arise from melanocytes adjacent to the retina or within mucosal surfaces in the oropharynx, sinuses, rectum, or vulva. These lesions typically present with greater local invasion and risk of distant spread than cutaneous melanoma.

Geography The rates of melanoma and other skin cancers are highest where fair-skinned Caucasians migrated to lower latitudes, with annual sun exposure that is substantially higher than their historically native climates. Australia, New Zealand, South Africa, and Israel bear a disproportionate burden of skin cancer. In Australia, melanoma is the third most common cancer. In the United States, Hawaii and the desert Southwest have the highest rates of skin cancer of all kinds and melanoma.

Race Caucasians are by far the most susceptible race for melanomas, as well as squamous cell and basal cell cancers. Hispanics have a lower incidence but represent the group at next highest risk. Asians and African-Americans have the lowest rates of skin cancer. For those populations, cutaneous melanomas arising from sun-exposed sites are uncommon but not unseen. All racial groups are equally likely to develop melanoma on the acral surfaces of the hands and feet or mucosal surfaces. Therefore, melanomas arising from these sites represent nearly all cases of melanoma in these more darkly pigmented racial groups.

Survival Skin cancers that are confined to the skin at presentation and with adequate staging evaluation have a high rate of cure. The 10-year survival for patients with invasive melanomas that are 1 mm or less in thickness and that lack ulceration is 97%. Melanoma that is microscopically present in regional lymph nodes is associated with a 10-year survival of 50% to 60%; when macroscopic or clinically apparent lymph nodes are detected, the 10-year survival is only 30% to 40%. The presence of more distant metastatic disease is associated with only a 5% possibility of survival 10 years from initial recognition.

Etiology and risk factors

Genetic predisposition Although there are families in which melanoma can occur with high likelihood, an underlying genetic predisposition can only be found in 3% all cases. The pedigrees have been identified because of their high likelihood of a mutation carrier developing melanoma. Lower penetrance genotypes remain to be elucidated. Nonetheless, the identification of the genes responsible for familial melanoma has contributed greatly to the understanding of the molecular pathophysiology of melanoma.

The clinical observations that patients with multiple dysplastic nevi were at greater risk of developing melanoma and that many such patients came from families with multiple affected individuals provided the first insight into a melanoma progression model that might be accelerated based on inborn genetic abnormalities. Two highly related genes were discovered to harbor germline mutations in roughly 50% of melanoma pedigrees: CDKN2A and CDK4. CDKN2A enocodes two products via alternate splicing of messenger RNA: p16INK4A and p14ARF. Each of these tumor suppressor genes exerts an inhibitory effect on cell cycle progression.

Xeroderma pigmentosum is a rare inherited disorder in which DNA repair mechanisms are compromised, particularly in response to ultraviolet (UV) light. Mutations in XP genes A through G have been identified as the underlying molecular event. Squamous cell and basal cell carcinomas and melanoma are prevalent in this population and at a young age. The near-complete penetrance of melanoma in these patients emphasizes the critical balance between UV-induced DNA damage and repair in risk for skin cancer. As DNA damage repair is mediated by a complex network of sensor and effector proteins, variability in the function of this system almost certainly underlies the variability in risk among the fair-skinned population.

Genetic variability in melanocortin-1 receptor (MC1R) has been clearly implicated in pigmentation of skin and hair and, more recently, to melanoma predisposition. It has been known for decades that melanoma is more prevalent among fair-skinned individuals with red or blond hair. Furthermore, blond-haired individuals with an inability to tan are at substantially greater risk of developing melanoma than blond-haired individuals who tan readily. Polymorphisms, distinct from mutations, in MC1R appear to account for skin and hair color differences among Caucasians. It appears that individuals with melanocortin receptors that have a muted response to increased melanocortin expression following sun exposure suffer the greatest UV-induced genetic damage, leading to a greater risk of melanoma.

Exposure Even the inheritance of CDKN2A and CDK4 mutations is insufficient to lead to melanoma in all carriers. It is clear that multiple genetic changes are required to give rise to invasive disease. UV damage is the best-described modifiable risk factor for melanoma, as well as squamous cell and basal cell skin cancers. It is believed that the acquired or somatic genetic changes that give rise to melanoma occur as a consequence of UV-induced genetic damage.

Epidemiologic data relate the risk of melanoma most closely to a connection between cumulative sun exposure, severe sun burns, or sun exposure during childhood, depending on the study. The disagreement between studies likely stems from methodologic differences in obtaining a sun exposure history, a heterogeneous effect of sun exposure and risk depending on the underlying genetic composition of the study population, or both. It has been clarified that melanoma arising on intermittently sun-exposed skin (such as the trunk) has its peak incidence among younger individuals and declines severely with increasing age. On the other hand, melanoma arising from chronically sun-damaged skin (such as the face, neck, and upper extremities) has the highest incidence in older individuals. With the rise in popularity of indoor tanning salons, data indicate that those who use them are at higher risk of melanoma than those who are sun exposed. There is little dispute regarding the causal link between sun exposure or tanning salon use and risk of melanoma; however, there is disagreement regarding the constituents of light (UV-A or UV-B) that contribute most to genetic damage. Laboratory studies support a connection for both and suggest that prevention strategies must take the entire UV light spectrum into account.

Prevention With the incidence of melanoma still rising, it is clear that primary prevention efforts have not yet taken hold. The only approach firmly rooted in evidence is to minimize sun exposure. The use of sun-protective clothing appears to be the next best strategy. There are conflicting data regarding the protective effect of suncreens for melanoma, although there is no controversy regarding their ability to prevent squamous cell and basal cell carcinomas. Protection against UV-A has been a long-standing feature of widely available suncreens, whereas UV-B protection has more recently been engineered into all mainstream products. It is possible that the more widespread of these wide-spectrum sunscreens will provide more meaningful protective effects over the coming decades.

Immunosuppression There is incontrovertible evidence linking immunosuppression and squamous cell skin cancer. This increased risk applies to patients with acquired immunodeficiency syndrome as well as transplant recipients on chronic immunosuppressive medications. The risk of developing primary melanoma in the setting of immunosuppression is less well established, but there is some evidence that patients who have a history of melanoma are more likely to develop disease recurrence in the setting of immunosuppression.

Nevi Patients with numerous benign nevi (small, regularly shaped, and uniformly pigmented moles) are at increased risk of melanoma, as are patients with relatively few dysplastic nevi (large, irregularly shaped, and heterogeneously pigmented moles). Patients in either group may have a five-fold increased risk of developing melanoma compared with those with few benign nevi or without dysplastic nevi. However, it is critical to recognize that these preexisting moles represent a risk factor for melanoma, rather than precursor lesions, in most cases. The vast majority of dysplastic nevi do not give rise to melanoma. There seems to be little value in resecting every nevus that appears dysplastic on clinical grounds. Surveying the skin regularly for new or changing moles has been most widely adapted for educating patients and healthcare providers. Beyond the age of 25 to 30, new mole formation warrants examination by a provider who is comfortable making the diagnosis of melanoma.

Congenital nevi are relatively rare compared with acquired nevi. In general, they do not suggest a predisposition to melanoma. A small number of children are born with so-called giant congenital nevi, also referred to as bathing trunk nevi.

History of melanoma Perhaps the single greatest clinical risk factor for melanoma is a personal history of melanoma. In addition to having surveillance for local and distant recurrence related to their prior melanoma, these individuals require lifelong observation for the emergence of a new primary melanoma; their risk is 10 times greater than that of the general population.

Actinic keratoses represent the precursor lesion for the vast majority of squamous cell skin cancers. They manifest as raised, nonpigmented lesions with a plaque-like surface. Typically, they have an erythematous base and arise on heavily sun-exposed skin. These lesions are typically treated with either excision, cryotherapy, topical chemotherapy (5-fluorouracil [5-FU]), or immunotherapy (imiquimod). It is unclear what effect nonsurgical therapy has on the long-term risk of progression to squamous cell carcinoma, but the relatively indolent nature of most actinic keratoses suggests that nonsurgical therapy followed by close observation is reasonable, particularly for lesions arising on cosmetically sensitive areas such as the face.

Burns (unrelated to sun exposure) predispose to the formation of squamous cell carcinoma on permanently scarred areas. Although not particularly prevalent among those with a burn history, a high clinical suspicion for malignancy must be taken for the patient with a new nonpigmented lesion or an area of ulceration on a preexisting healed burn.

Signs and symptoms The diagnosis of melanoma is based on complete pathologic evaluation of the skin. A growing skin lesion is the easiest description that a patient can provide a physician to assist in determining which skin lesions should be biopsied. ABCDs is a simple learning method (see below for more information) also to assist patients. This system, which is based on asymmetry of the skin lesion, irregular edges, color appearance, and diameter, can be helpful to patients but may lead to some melanomas being missed.

Diagnosis

Early diagnosis of melanoma Beyond prevention, early detection is the next priority in limiting the life-threatening potential of melanoma. Although tumor invasion into the dermis (the definition of invasive melanoma) cannot be assessed clinically, there are hallmarks of melanoma that can be appreciated by visual inspection. The “ABCD” algorithm, promoted by the American Cancer Society, has not proved to be a particularly effective method of educating the general population, but healthcare providers are better suited to implementing this system in clinical practice. (See the color atlas following this chapter.) ‘A’ indicates gross asymmetry of a pigmented lesion. ‘B’ refers to irregular or indistinct borders. ‘C’ regards color, which, in the case of melanoma, refers to black, dark brown, or blue color for newly formed lesions. In the case of preexisting moles, the loss of pigmentation in a portion of the lesion can indicate regression, which is worrisome for melanoma. Additionally, a focus of darker pigmentation within a preexisting mole is grounds for concern. ‘D’, or diameter of 6 mm or greater, raises concern for melanoma, whereas smaller lesions are rarely indicative of invasive lesions. None of these single criteria can be considered as grounds for performing a biopsy, rather lesions that satisfy multiple criteria warrant either close observation or biopsy of the most abnormal portion. Not included in the ABCD system is the papular or nodular characteristic of most melanomas. If melanoma arises within a preexisting mole, the raised element of an otherwise flat or macular lesion should be considered suspicious for melanoma. Some primary melanomas do not produce melanin and therefore lack the classic appearance.

Atypical nevi represent a risk factor for melanoma in that individual lesions can occasionally progress to invasive melanoma. More commonly, the presence of dysplastic nevi suggests an individual at risk for melanoma formation at other sites. The clinical definition of atypical nevi has never been formally established but generally refers to the presence of one or two of the ABCD features. Cutaneous photography has been routinely incorporated in the follow-up of patients with multiple clinically dysplastic nevi in specialized pigmented lesion clinics. Full-body cutaneous photography (dermoscopy) is increasingly available to patients and provides an objective baseline from which to judge change of preexisting lesions or the appearance of new lesions. In general, excision of all atypical nevi is impractical and does not adequately address the risk that patients face of developing melanoma in sites where no precursor lesion is found.

Nodular melanoma represents the most difficult subset of melanomas to diagnose at an early stage. These lesions are more rapidly proliferative than typical melanomas and are generally not pigmented. Furthermore, they are generally raised early in their development and have symmetric and well-demarcated boundaries. Thus, they do not meet the ABCD criteria. Nodular melanomas account for approximately 10% of melanoma cases but account for a disproportionate percentage of fatalities. The most useful clinical rule to apply in the assessment of lesions that have these features is that the de novo appearance of such lesions over a short time (months) warrants consideration of biopsy. As the features of nodular melanoma are common to some benign skin lesions, as well as basal cell carcinoma, the yield of biopsies for such lesions may be relatively low. Nonetheless, heightened awareness of this small but lethal subset of melanoma is needed.

Nonmelanoma skin cancers Classically, basal cell skin cancers are nodular and have a pearly, nonpigmented surface. Squamous cell carcinomas typically have a scaly, nonpigmented surface and can occasionally be ulcerated. Both types are generally slow to evolve. In most cases, squamous cell carcinomas are believed to arise from actinic keratoses in most cases. The differentiation between the two is based largely on size, with actinic keratoses being less than 1 cm in diameter. The appearance of nodularity or induration in preexisting keratinized lesions raises concern for tumor invasion into the dermis and warrants biopsy.

Skin examination of the entire body should be performed for any patient who presents with numerous benign-appearing nevi, any dysplastic nevi, or a history of even a single melanoma. The entire skin is at risk in these individuals and must be investigated for the appearance of new or changing lesions. Patients who fit into any of these risk groups should be instructed to perform monthly skin self-examination. Two mirrors are required to adequately examine the back if a partner is not available to assist in examining the back longitudinally.

Skin examination should be performed in a well-lit room with the patient completely disrobed. For patients with numerous moles, full-body cutaneous photography is extremely helpful to provide an objective baseline from which to judge change. For patients with only a few atypical moles, close-up photographs of those lesions may facilitate careful evaluation of those lesions while the remainder of the skin is surveyed for new lesions. A focused method for examining individual lesions, dermoscopy employs low-level magnification of the epidermis with tangential light applied to a liquid-skin interface. Examination of pigmented lesions with dermoscopy allows more precise visualization of patterns of pigmentation than is possible with the unaided eye. This method requires some degree of training and the availability of appropriate equipment. When regression of part or all of a preexisting mole is suspected, a Wood’s lamp can be helpful in bringing out the contrast in pigmentation between normally pigmented surrounding skin and an area where immunologic destruction of melanocytes has occurred.

Lymph node examination should be performed in all patients who are suspected of having invasive melanoma or large squamous cell, basal cell, or Merkel cell carcinomas. The closest lymph node basin is the most essential area to examine. The presence of a palpable lymph node mandates biopsy of the suspicious node, generally with fine needle aspiration (FNA). If the result of FNA is negative, a formal lymph node biopsy should be pursued for a clinically suspicious node. Whereas sentinel lymph node biopsy (SLNB) is performed routinely for patients with primary melanomas 1 mm or greater in depth, palpable lymphadenopathy with histologic evidence of melanoma, or nonmelanoma skin cancer, warrants proceeding directly to a regional lymph node dissection.

Biopsy techniques Excisional biopsy is recommended for any lesion that is suspected of being melanoma, squamous cell cancer, or basal cell carcinoma. This is especially true for lesions that can be entirely excised without concern for causing an unacceptable cosmetic result. Local anesthesia is generally all that is required. Narrow margins of 1 to 2 mm are sufficient when making an initial diagnosis of skin cancer. The biopsy should extend to the subcutaneous tissue to provide an adequate estimate of the depth of invasion, particularly for melanoma. Biopsy specimens should be placed in formalin and submitted for expert pathologic preparation including embedding in paraffin.

Incisional biopsy is an acceptable alterative for large lesions, especially when located on the face, neck, or distal extremities. In the case of possible melanoma, the most abnormal appearing area should be biopsied. As with excisional biopsies, the biopsy should include subcutaneous tissue to allow an estimation of thickness. A punch biopsy, of sufficient diameter to encompass the most abnormal appearing area, is the preferred method. Should the biopsy identify a melanocytic lesion with dysplasia or atypia, the entire lesion should be removed, when possible, with radial margins to 1 to 2 mm.

Techniques to avoid For lesions that are suspicious for melanoma, every attempt should be made to preserve the ability to assess involvement of margins and to perform immunohistochemistry of the primary tumor. The latter is critically important in borderline cases, where the diagnosis of melanoma is uncertain. Thus, shave biopsies and serial thin section techniques (such as Mohs surgery) are strictly contraindicated. Frozen biopsy assessment is inadequate for the diagnosis of melanoma and does not allow adequate material for review in difficult pathologic cases.

Pathology

Nonmelanoma skin cancers The two most common skin cancers are basal and squamous cell carcinomas. Both of these skin cancers arise predominately on sun-exposed areas and may be considered in the differential diagnosis with melanoma. The diagnosis of these skin cancers should be made after pathologic analysis of the skin specimen. Merkel cell carcinoma is a rare neuroendocrine tumor of the skin. Like melanoma, it has a propensity for regional nodal metastasis and for local, regional, and distant recurrence.

Staging and prognosis of melanoma

A great deal of information is available regarding factors that correlate with clinical outcome in

patients with melanoma. In patients with clinically localized disease, the most important prognostic factors are Breslow’s thickness, ulceration, and SLN status. Overall, however, 85% of melanoma patients present with clinically normal lymph nodes. In clinically node-negative patients, most investigators have found the microscopic degree of invasion of the melanoma, or microstaging, to be of critical importance in predicting outcome (Tables 1, 2, and 3).

Microstaging
Two methods have been described for microscopic staging of primary cutaneous melanomas.

Clark’s level Wallace Clark and associates devised a system to classify melanomas according to the level of invasion relative to histologically defined landmarks in the skin. Although Clark’s levels correlate with prognosis (lesions with deeper levels of invasion have a greater propensity for recurrence), the inherent problem with Clark’s system is that the thickness of the skin and hence the distance between the various landmark dermal layers varies greatly in different parts of the body. Furthermore, except for Clark’s level I (melanoma in situ), there is no scientific rationale for considering these landmarks to be biologic barriers to tumor growth. For example, there is no a priori reason to suspect that a lesion that reaches but does not invade the reticular dermis is inherently less

aggressive than a similar melanoma that penetrates the reticular dermis in an area where the skin is thinner. Despite this, Clark’s level remains of prognostic value for patients with melanoma less than or equal to 1 mm in depth and is incorporated into the staging system for these lesions (Table 1).

Breslow’s thickness An alternative microstaging method, described by Alexander Breslow, obviates some of the problems associated with Clark’s levels. In this method, the thickness of the primary tumor is measured from the top of the granular layer of the epidermis to the deepest contiguous tumor cell at the base of the lesion using a micrometer in the microscope eyepiece.

Ulceration The presence of ulceration in a primary melanoma has been recognized as one of the strongest negative predictive factors for long-term survival. Ulceration is defined as the lack of a complete epidermal layer overlying the melanocytic lesion. The presence of ulceration essentially upstages affected patients to the next highest T level. In other words, a patient with a 1.1- to 2-mm melanoma that is ulcerated will carry the same long-term prognosis as a patient with a

2.1- to 4-mm melanoma that does not have ulceration. The likelihood of finding ulceration is directly related to tumor depth: Patients with thin melanomas (≤ 1 mm) have a 6% rate of ulceration, whereas those with > 4-mm melanomas have a 63% incidence of ulceration. Along with tumor depth, ulceration is integral in determining a patient’s long-term prognosis and is an independent predictor of patient outcome.

It has been firmly established that there is an inverse correlation between Breslow’s tumor thickness and survival. More important, several studies have demonstrated that tumor thickness conveys more prognostic information than does Clark’s level of invasion. In addition, the measurement of tumor thickness is generally more reproducible and less subjective than is the determination of Clark’s level.

Regional lymph node involvement
In patients with intermediate-risk melanoma (1 to 4 mm in thickness), lymph node involvement is the strongest prognostic indicator in the staging of melanoma. Patients with nodal involvement at the time of their diagnosis have significantly decreased survival compared with those who do not, independent of the prognostic factors associated with the primary lesion. There is a direct relationship between the depth of invasion of the primary lesion and the potential for lymph node involvement. Among node-positive patients, the prognosis is more favorable in those with microscopic as opposed to macroscopic or clinically apparent disease. In addition, the increased number of nodes involved is associated with decreased survival in both microscopically detected and clinically apparent regional nodal metastasis. The worst prognosis occurs in patients with large, matted regional lymph nodes, whose outcome is similar to that of patients with stage IV disease.

Melanoma of unknown primary
Melanoma of unknown primary (MUP) occurs in up to 5% of patients who present to tertiary cancer referral centers.. Lymphadenopathy is the most common presenting feature, followed by identification of visceral metastases and cutaneous nodules. Patients should be evaluated by a dermatologist and should undergo a complete physical examination, including an anorectal and genital evaluation. A CT scan of the chest, abdomen, and pelvis and/or a whole-body PET scan are useful for staging patients with MUP. In the absence of other sites of disease, surgical resection of the metastatic lesion (or complete regional lymphadenectomy) is appropriate. In the absence of symptoms, an ophthalmologic consultation is probably not warranted.

Single dermal nodules with no identifiable primary lesion are typically treated in similar fashion to primary melanomas, with wide local excision and regional nodal evaluation SLNB if appropriate. The prognosis for patients who present with MUP is similar to that for somewhat more favorable patients with metastatic disease from a known primary.

Clinical and pathologic staging TNM staging system The melanoma staging committee of the American Joint Committee on Cancer (AJCC) revised the TNM staging system to reflect more accurately the impact of statistically significant prognostic factors that were validated on a multi-institution sample of over 17,000 melanoma patients. This new system is shown in Tables 1 and 2.

Some of the most significant changes include new cutoff points for T classification (T1 = ≤ 1.0 mm, T2 = 1.01–2.0 mm, T3 = 2.01–4.0 mm, T4 = > 4 mm), inclusion of ulceration as a factor in T stage (a: no ulceration, b: ulceration), inclusion of node size with node number for N stage, inclusion of node size (microscopic vs macroscopic) as a factor in N stage, and inclusion of lactate dehydrogenase (LDH) level as a factor in M stage. These improvements in staging based on known prognostic factors will allow for better stratification of patients in clinical trials.

For patients with stage IV disease (distant metastases), there are few long-term survivors, but significant differences in prognosis for disease limited to the skin, for subcutaneous and nodal sites vs visceral sites, and for patients with levels of LDH in the serum.

Other prognostic factors Age Overall, patients who are ≥ 65 years old have a survival rate that is decreased by 10% to 15% compared with their younger counterparts. This trend has been demonstrated in numerous studies. A different relationship between age and risk of lymph node involvement, however, is emerging. Data from the prospective Sunbelt Melanoma Trial and retrospective studies reveal that younger patients are significantly more likely to harbor nodal disease than older patients with similar lesions.

Gender Many studies have identified improved survival in women compared with men with melanoma when stratified by stage. The reasons for this are unclear.

Anatomic location There is a correlation between anatomic location and prognosis of primary melanoma. Those with lesions on the back, upper arms, neck, and scalp (BANS area) have a worse prognosis than those with lesions on the extremities. This may be due to the simple fact that these lesions cannot be evaluated easily by the patient because of their inopportune location rather than a manifestation of true differences in tumor biology. Men are more likely to develop truncal melanomas, whereas women are most likely to develop melanoma on their extremities.

Desmoplastic melanoma Desmoplastic melanomas represent a less common but clinically distinct spindle cell variant of melanoma with dense fibrosis and frequent neurotropism. Clinically, they are raised, firm nodules that are amelanotic in up to 40% of patients, commonly leading to a delay in diagnosis.They are usually deep lesions but have a more favorable prognosis than do conventional melanoma of a similar depth. The association of desmoplastic melanoma with local recurrence may be due to the fact that many of these lesions are misdiagnosed and undertreated, with regard to an appropriate radial margin of excision.

Additionally, although many of these lesions are deeply invasive at the time of diagnosis, desmoplastic melanomas are less likely to involve regional lymph node basins. Patients with “pure” desmoplastic melanoma are extremely unlikely to harbor regional nodal metastasis, and SLN biopsy is unlikely to provide useful information. Patients with “mixed” desmoplastic melanoma, containing a component of conventional melanoma, have a risk of regional nodal metastasis that is approximately half that of patients with purely conventional melanoma, and SLN mapping should be considered in these patients.

Despite the fact that these lesions are often relatively thick at presentation, patients with desmoplastic melanomas tend to have survival rates that are favorable compared with patients with conventional melanomas of a similar depth.

Angiolymphatic invasion Defined as invasion of tumor cells into the wall and/or lumen of vessels or lymphatics of the dermis or deeper structures, angiolymphatic invasion is uncommon in malignant melanoma. However, this finding is clearly associated with more aggressive tumors. Multiple large studies have shown worsened long-term survival and more frequent lymph node involvement in patients with angioinvasive melanomas. In fact, risk of lymphatic involvement increases as much as threefold, whereas 5-year survival is reduced by as much as 50%, in patients with vascular invasion compared with matched patients without vascular invasion.

Regression The finding of regression represents host immune response to invasive melanoma. Areas where invasive cells may have once existed are replaced by inflammatory reaction and fibrosis, which may make it impossible to determine the precise depth of the initial lesion histologically. There is continuing debate regarding the prognostic importance of regression; however, many clinicians believe that thin lesions that show signs of regression should be given higher consideration for surgical nodal staging, given the fact that the initial lesion may have originally been more deeply invasive. Although some studies have shown that regressed lesions have a higher propensity for lymph node metastases than nonregressed primary tumors of the same thickness, this finding has not been universally observed. Indeed, a recent retrospective single-institution study demonstrated no increased risk of SLN metastasis in patients with regressed melanoma, suggesting that SLNB should not be performed for lesions that would otherwise not be considered based on the presence of regression.

Tumor-infiltrating lymphocytes (TILs) Variable numbers of TILs are observed in melanoma. Tumors with a high number of TILs should carry an improved prognosis because of the active host response to tumor. In fact, many studies have shown this tendency. The absence of TILs in the primary melanoma has been associated with a higher risk of positive SLNs in a recent, large, single-institution study.

Treatment of melanomas

SURGICAL TREATMENT OF CUTANEOUS MELANOMA

Excision of primary lesion It was recognized over a century ago that tumor cells could extend within the skin for several centimeters beyond the visible borders of a melanoma, so that the risk of local recurrence relates to the width of normal skin excised around the primary tumor. Only much more recently was it realized that the thickness of the primary tumor influenced the likelihood of contiguous spread and that not all melanomas require the same excision margin. This realization prompted a number of randomized trials to determine the optimal excision margins for melanomas of different Breslow’s thicknesses.

Initially, a “one-size-fits-all” approach of taking a 5-cm margin around all cutaneous melanomas was adopted. With such wide margins, skin grafts were required after removal of melanomas on most parts of the body. Modern melanoma surgical care is based on level I medical evidence, and many large, prospective randomized trials have been conducted to determine the appropriate radial margin for cutaneous melanoma based on Breslow thickness.

A randomized trial conducted by the World Health Organization found that when a 1-cm margin of normal skin was taken around a melanoma < 1 mm thick, the local recurrence rate was exceedingly low (< 1%), and patient survival was just as good as if 3-cm margins were taken. For melanomas 1 to 2 mm in thickness, patient survival was the same for both margins of excision, but the local recurrence rate was higher with the 1-cm margin (3.3% after 10-year follow-up).

The Intergroup Melanoma Trial compared 2- vs 4-cm margins for all cutaneous melanomas between 1 and 4 mm in thickness of the trunk or proximal extremity. Most patients in this trial had melanomas less than or equal to 2 mm in thickness. In the intergroup trial, both local recurrence and survival were the same regardless of whether 2- or 4-cm margins were obtained. Skin grafts were less frequent and hospital stays shorter with the narrower margin.

In a recent trial conducted by the United Kingdom Melanoma Group, patients with higher-risk melanomas, those greater than or equal to 2 mm in depth, were randomized to radial margins of 1 cm vs 3 cm. Patients in the narrow-excision group had a higher risk of local or regional recurrence than those with a wider excision, and there was a trend toward decreased survival in the narrow-excision group, although this did not reach statistical significance. A more recent trial conducted in the United Kingdom addressed patients with primary melanoma greater than or equal to 2 mm in depth. Patients who underwent a 1-cm radial margin of excision had a higher risk of local/regional recurrence than those who underwent excision with 3-cm margins.

Current recommendations Based on these important studies, it is possible to make rational recommendations for excision margins for melanoma patients.

• Patients with melanoma less than or equal to 1 mm should undergo excision of skin and subcutaneous tissue for a radial margin of 1 cm.

• Patients with melanoma between 1 and 2 mm should undergo excision of skin and subcutaneous tissue for a radial margin of 1 to 2 cm.

• Patients with melanoma greater than 2 mm in depth should undergo a wide excision of skin and subcutaneous tissue of 2 to 3 cm.

• When the anatomic location of the primary tumor precludes excision of the margin (eg, on the face), at least 1 cm should be taken when feasible.

Management of regional lymph nodes

Clinically apparent lymphadenopathy Melanoma patients with clinically enlarged nodes and no evidence of distant disease (AJCC stage III) should undergo complete regional lymphadenectomy. The first step in evaluating palpable nodes is generally FNA. A negative or inadequate sample FNA may be repeated, with image guidance if necessary, or may lead directly to an open node biopsy, followed by complete lymphadenectomy in the event of a positive frozen section or touch-prep cytologic determination of metastasis.

Clinically normal nodes The surgical management of clinically normal nodes is determined by the characteristics of the primary lesion. A direct relationship between thickness of the primary lesion and nodal involvement has long been recognized. When the depth of the primary is unknown because of the biopsy technique or other factors, consideration should be given for SLN mapping.

SLNB is the primary screening tool for regional nodal staging. All potentially involved basins should initially be examined as part of a thorough history and physical examination.

Thin melanomas Patients with thin melanomas (< 1 mm Breslow depth) have a low risk of occult nodal involvement (< 5%). Some patients with melanomas 0.76 to 1 mm have a high enough risk of lymph node involvement to justify consideration of SLNB in addition to wide excision.

Historically, Clark’s level of IV or V was considered to be a significant prognostic indicator and continues to be included in the AJCC staging system. When ulceration is present, consideration should be given to SLNB. In addition, a high mitotic count should prompt consideration of regional nodal basin staging.

Intermediate-thickness melanomas Risk of nodal metastasis rises significantly with increasing depth of invasion. Patients with a 1-mm thick melanoma have approximately a 10% chance of nodal involvement, whereas those with a 4-mm melanoma have a 35% to 40% risk of nodal metastases. For these reasons, wide excision of the primary tumor is generally accompanied by SLNB for evaluation of the nodal basin involved. Patients who have evidence of metastasis in the SLN are offered completion lymphadenectomy as a standard approach. An alternative approach of wide excision alone with observation and serial physical examination of nodal basins has been advocated by some clinicians and has not been shown to reduce long-term survival if complete lymph node dissection is performed at the identification of clinically positive nodes. This approach is justified in a recently completed prospective randomized clinical trial of wide excision and SLN mapping compared with wide excision and nodal basin observation. There was no evidence that SLN mapping contributed to an overall survival benefit, but it was a powerful predictor of outcome. The use of SLNB vs observation continues to be an area of intense debate.

Regional lymphadenectomy The identification and management of nodal disease have evolved significantly over the past 20 years. Historically, complete lymph node dissections were performed on all patients with intermediate and thick melanomas with the belief that this would lead to a survival benefit. This was based on the observation that there was an approximately 20% survival advantage to patients found to have microscopic involvement of lymph nodes at “elective” lymph node dissection when compared with those patients who had clinically apparent nodal metastasis and underwent a “therapeutic” lymph node dissection. With elective node dissection, however, a significant percentage (80% to 85%) of patients underwent complete lymphadenectomy only to find that their nodal basin was free of disease. More importantly, a series of prospective randomized clinical trials failed to demonstrate a survival advantage for patients undergoing elective lymphadenectomy over wide excision only. For this reason, the standard approach prior to the use of SLN mapping was wide excision alone and clinical observation of the nodal basin.

By the current technique, which maximizes the success rate for sentinel node identification, technetium99m–labeled sulfur colloid is injected into the primary site preoperatively, and the potentially involved lymph basins are identified with a gamma camera. The patient is then taken to the operating room, where a hand-held gamma probe is used to identify the sentinel node(s). This technique, in combination with the use of vital blue dye, has led to a success rate exceeding 99% in most large current trials.

The intuitive hope that SLN mapping and biopsy followed by completion lymphadenectomy would result in improved survival was not borne out in a recently reported prospective randomized clinical trial. Even in the absence of a proven survival benefit, however, the staging advantages of SLNB are sufficiently compelling to justify its routine use in healthy patients with melanomas and a significant risk of nodal involvement. A large, prospective clinical trial is currently being conducted to determine the role of completion lymphadenectomy compared with close surveillance in SLN-positive patients.

Thick melanomas Deep melanomas harbor nodal metastases in up to 60% of patients; surgical evaluation of the nodal basin may be appropriate, as most studies suggest that SLN status retains prognostic status in patients with deep primary melanoma. Unless the information will direct further management, it is unlikely to be of benefit for these patients. In addition, a completion lymphadenectomy for those at high risk for local and regional recurrence may subject the patients to undue morbidity. Whether to perform a complete lymphadenectomy in patients at increased risk of recurrence is difficult, however, because it is much less likely to be of benefit in these patients than in those with intermediate-thickness melanoma. In addition, the lymphedema associated with complete node dissection may complicate treatment of subsequent recurrences in patients with deep melanoma of the extremity. Although elective or completion lymph node dissection (LND) has not proven beneficial in these patients, SLNB has been used extensively to guide diagnostic, prognostic, and therapeutic decision-making.

SURGICAL TREATMENT OF NONCUTANEOUS MELANOMA

Ocular melanomas generally do not have access to lymphatic channels, so the surgical principles outlined previously do not apply here. However, the unique propensity to metastasize hematogeneously, often to the liver after a long relapse-free interval, warrants further study of the biology associated with this primary site. Advances in understanding the biology of ocular melanomas may lead to adjuvant approaches different from therapies now under investigation for cutaneous primaries.

A diagnosis of ocular melanoma with no evidence of distant disease signifies that a decision must be made as to whether or not the eye can be spared. Some small melanomas situated peripherally in the retina can be excised with minimal loss of vision, but most cannot. For larger lesions, treatment options are enucleation (total removal of the eye) or implanted radiotherapy with a radioactive gold plaque fitted to the back of the eyeball immediately behind the tumor. A multi-institution, randomized trial comparing implanted radiotherapy with enucleation for local disease control and overall survival was completed by the COMSG; it appears that both techniques provide similar outcomes for all sizes of tumors.

Melanomas of the anus and vulva pose challenges in the treatment of both the primary lesion and regional nodes. These “mucosal” melanomas are notoriously difficult to cure and are almost always associated with a poor outcome. The surgical management of anal melanoma is controversial, but most surgeons prefer a wide excision, when possible, over an abdominoperineal resection (APR). In the past, an APR was more commonly used for patients with less-advanced disease who were viewed as potentially curable. This almost certainly explains the reported association of APR with long-term survivors. More recently, APR is reserved for patients with bulky disease or recurrent disease that is not amenable to wide excision, which is favored for patients with more localized/potentially curable disease. Not surprisingly, more recent studies demonstrate no survival advantage to APR, which appears to be equivalent to wide excision.

Radical resection is also less commonly performed for patients with vulvar melanoma, who also have a high risk for relapse and death. Function-preserving resection and nodal basin staging in the setting of relatively early disease are appropriate, but, as in patients with anal primaries, preemptive nodal staging in patients with advanced primary tumors is unlikely to impart a benefit to those whose nodal basin can be followed clinically.

Nasal sinuses or nasopharyngeal melanomas Melanomas arising in the nasal or nasopharyngeal mucosa should be widely excised to include adjacent bony structures, if needed. As in patients with anal or vulvar melanoma, node dissection is reserved for patients who have proven nodal involvement. Radiotherapy should be considered for those patients whose primary tumor cannot be fully removed from this site with adequate margins or as adjuvant therapy in patients unlikely to be controlled with surgical resection alone.

ADJUVANT THERAPY FOR MELANOMA

Interferon-alpha-2b (IFN-α; Intron A) was approved as a standard therapy on the basis of an adjuvant therapy trial conducted in patients with deep primary or N1 disease (ECOG). Interferon-alpha-2b was administered by IV, 20 mU/m2 for 5 consecutive days every 7 days for 4 weeks, during the “induction” phase. For a subsequent 48 weeks, 10 mU/m2 was administered by subcutaneous injection on alternate days for a total of three doses every 7 days in the “maintenance” phase.

This “high-dose” regimen was compared with observation. A statistically significant improvement in overall survival was demonstrated with IFN, compared with the observation arm; relapse-free survival was also improved. Three-quarters of the IFN patients experienced severe toxicities, most commonly fatigue, asthenia, fever, depression, and elevated liver transaminase levels. A quality-of-life analysis found that the toxicity associated with this regimen was largely compensated for by the prevention of disease relapse.

Due to uncertainty regarding the optimal dose and schedule of IFN, a second trial was initiated comparing the high-dose regimen and a low-dose regimen (3 mU by subcutaneous injection 3 times weekly for 24 months) as well as observation. No significant improvement in overall survival for either the high-dose or low-dose arm was observed compared with observation; relapse-free survival was improved by 22% in the high-dose arm compared with observation. A pooled analysis of E1684 and E1690, with longer follow-up for both trials, revealed a continued, statistically significant impact on relapse-free survival but not on overall survival.

High-dose IFN has been consistently shown to improve relapse-free survival compared with either observation or ganglioside GM2/keyhole limpet hemocyanin vaccination but does not confer a an overall survival advantage. A recent prospective randomized trial conducted by the HCOG compared outcome in patients treated with high-dose IFN as an adjuvant for 1 month vs 1 year and found no difference in overall survival or disease-free survival. Due to the significant toxicity associated with 1 year of high-dose interferon and the lack of an overall survival advantage, the use of high-dose IFN as an adjuvant therapy has markedly diminished.

Intermediate-dose regimens with IFN have been evaluated in several trials; they have less consistently demonstrated a relapse-free survival advantage compared with observation and have never resulted in a survival advantage. A recent trial conducted by the EORTC demonstrated no survival advantage for patients undergoing 13 months or 25 months of intermediate-dose adjuvant IFN after resection of stage IIB (deep primary melanoma) or stage III nodal disease when compared with observation.

The EORTC reported the results of a prospective randomized trial of observation vs pegylated IFN-α for 5 years. Again, there was an advantage to treatment in terms of relapse-free survival but no overall survival advantage for patients with high-risk resected melanoma. Nearly one-third of the patients in the treated arm discontinued therapy due to toxicity.

Recently, the DCOG reported an improvement in overall survival for patients treated with 2 years of adjuvant low-dose IFN-α compared with those treated with IFN and dacarbazine or surgery alone in patients with resected regional nodal metastasis. These studies indicate that there appears to be no continued role for 1 year of high-dose IFN for patients with high-risk resected melanoma. In addition, two regimens of intermediate-dose IFN show no advantage over observation in these same patients. Whether patients with nodal disease benefit from 2 years of low-dose IFN given as adjuvant therapy after node dissection awaits confirmation in additional trials.

Adjuvant chemotherapy Single-agent chemotherapy or combination chemotherapy regimens have not been systematically evaluated for the adjuvant treatment of melanoma because of the low response rates seen in patients with advanced disease. The largest randomized trial compared an IV administered regimen of carmustine (BiCNU; 80 mg/m2) every 4 weeks, dactinomycin (Cosmegen, 10 µg/kg), and vincristine (1.0 mg/m2) every 2 weeks for 6 months with observation among patients with resected stage III or IV melanoma. A significant improvement in relapse-free survival, but not overall survival, was observed in this small study. A randomized trial comparing dacarbazine and observation failed to demonstrate an improvement in either relapse-free or overall survival.

There is currently no standard adjuvant therapy following resection of metastatic melanoma. Therefore, observation remains the standard of care for patients in this setting, and investigational vaccines and targeted agents are currently being evaluated for efficacy in these patients and remain a priority.

Neoadjuvant therapy for resectable stage IIIC or IV melanoma remains an investigational approach. Neoadjuvant biochemotherapy (cisplatin, dacarbazine, vinblastine, interleukin-2 [IL-2], and IFN) was evaluated in a small study of patients with stage III melanoma. In the 50 patients with measurable disease, the response rate was 26%. High-dose IFN, however, produced a 55% clinical response and 15% pathologic response when used in a neoadjuvant setting in a small group of patients with clinically apparent nodal metastasis. Currently available chemotherapy regimens for melanoma are not sufficiently active to support use in neoadjuvant therapy.

Radiation therapy In vitro data from the 1970s demonstrated radiation resistance among melanoma cell lines and thus have led to the reluctance to use radiation therapy in the treatment of melanoma. More modern data have shown that many melanomas are sensitive to radiation, but the disinclination toward the use of radiation therapy for melanoma continues.

Another unanswered question is the proper fraction size to use in melanoma. In a large series of patients studied by Overgaard and colleagues, they found that the response rate of metastatic melanoma lesions was dependent on the fraction size. The complete response rate was 57% when fractions greater than 4 Gy were used, compared with 24% for those of less than 4 Gy. RTOG 83-05 is the only randomized trial that compared hypofractionated (8 Gy for 4 fractions) and conventional (2.5 Gy for 20 fractions) schedules of radiation therapy. In this study, they noted no difference in the complete and partial responses. Though a wide spectrum of fraction sizes have been used, this does not preclude the use of radiation but emphasizes the need for more research to better understand and create an effective fractionation schedule.

Adjuvant radiation treatment Radiation treatment has been increasingly employed in the postoperative setting to improve locoregional tumor control in a select group of patients. In the past decade, there have been multiple reports regarding patient outcomes with the use of adjuvant radiation treatment. The 5-year regional control has ranged from 87% to 94%, and the 5-year overall survival has ranged from 36% to 46% with the addition of radiation to surgery. Single institutions have reported that local recurrence is reduced to 6% to 11% when postoperative radiotherapy is used for high-risk lesions, relative to 19% seen in the O’Brien study and 30% in previous studies. High risk is defined as having a thick primary, desmoplasia or neurotropism, a head and neck site, close or positive (unresectable) margins, and multiple previous recurrences. In these reports, they found that serious complication rates after radiation therapy were low, which is likely secondary to the superficial nature of the treated volumes.

At the Royal Prince Alfred Hospital, 143 patients with melanoma metastases to the parotid and/or cervical lymph nodes were treated with surgery alone or surgery and postoperative radiation therapy. The patients who received radiation therapy had more aggressive features such as at least two lymph nodes, and extracapsular extension; there was a trend toward higher regional control in the group that received surgery and postoperative radiation therapy (94%) than with surgery alone (91%, P = .065).

As previously mentioned, there are known risk factors for recurrence of disease in patients with regional nodal metastases after surgical resection. A retrospective review of patients treated at M. D. Anderson from the 1980s to early 2000s revealed 434 patients with regionally advanced melanoma (cervical, axillary, inguinal, and epitrochlear lymphadenopathy) were treated with surgery and radiation therapy. The majority of patients had involvement of the head and neck region, with the trunk second in prevalence. At a median follow-up of 4.2 years, they found 5-year disease-specific survival was 49%, 5-year disease-free survival was 42%, and 5-year distant metastases-free survival was 44%. A multivariate analysis showed two factors to be associated with the actuarial rate of disease-specific survival and distant metastases-free survival: an increasing number of lymph nodes involved and primary ulceration. They also had a 5-year regional control rate of 89%.

The recommendation made from this study was that treatment is indicated for most patients with any of the following features: the presence of lymph node extracapsular extension, lymph node measuring at least 3 cm, at least four lymph nodes involved, or nodal recurrence after previous dissection of pathologically confirmed disease. The risk of long-term lymphedema is known to be high in patients with inguinal lymph node metastases; thus, it would be prudent to consider a higher threshold for treatment of these patients (eg, two or more of the above-mentioned factors). In a recent review on the use of radiation therapy in malignant melanoma, Stevens and McKay offered the following recommendations for postoperative radiation therapy after regional lymph node dissection:

• Patients with melanoma less than or equal to 1 mm should undergo excision of skin and subcutaneous tissue for a radial margin of 1 cm.

• Patients with melanoma between 1 and 2 mm should undergo excision of skin and subcutaneous tissue for a radial margin of 1 to 2 cm.

• Patients with melanoma greater than 2 mm in depth should undergo a wide excision of skin and subcutaneous tissue of 2 to 3 cm.

• When the anatomic location of the primary tumor precludes excision of the margin (eg, on the face), at least 1 cm should be taken when feasible.

Radiation therapy can provide some benefit and should be considered in the palliative treatment of melanoma metastatic to bone or other symptomatic sites.

Vaccines and other immunotherapy remain an investigational approach and are currently being tested among patients with resected stages II, III, and IV melanoma. The goal of this immunotherapy is to reverse immune tolerance of microscopic residual melanoma following surgery through the amplification of melanoma-specific cytolytic T cells. This approach is distinct from conventional immunization against pathogens to which the host is naive, where long-lasting, antibody-mediated immunity is sought by induction of memory B cells.

Melanoma peptide vaccines are currently being intensely evaluated in clinical trials. Peptides have been selected from two classes for clinical development: melanocyte-specific proteins and so-called cancer-testis antigens. The latter are proteins that have a restricted spatiotemporal expression and are not detectable in adult tissues outside the immunologic sanctuary site of the testes. The most widely studied peptides are derived from gp-100, tyrosinase, melan-A (all melanocyte-specific), and NY-ESO-1 and MAGE proteins (cancer-testis antigens). Preliminary evidence suggests that this approach may successfully engender specific and potent immunologic responses to the protein fragments being administered. Protection against disease recurrence in patients at high risk following surgery has not yet been demonstrated. Notably, no randomized data suggesting a detrimental effect from peptide vaccination in melanoma have been presented. Indeed, a recent prospective randomized trial reported a survival advantage for patients treated with a peptide vaccine. Patients with metastatic melanoma were randomized to receive a multivalent peptide vaccine and high-dose IL-2 vs high-dose IL-2 alone. This approach may prove promising in the adjuvant setting.

CTLA-4 blockade Ipilimumab (MDX-010) is a promising new agent for the treatment of patients with metastatic melanoma. Ipilimumab is a monoclonal antibody against CTLA-4 (cytotoxic T-lymphocyte–associated molecule-4). It has been used in patients with metastatic melanoma and renal cell carcinoma refractory to other immunotherapies. The overall response rate is 14%, with a significant percent of patients having a durable complete response.

The efficacy of ipilimumab is due to its ability to block the normal activity of CTLA-4, a molecule that tempers the immune system’s response to antigen. Although the exact antitumor activity of anti–CTLA-4 in unclear, it is thought that by inhibiting the activity of CTLA-4, the threshold for a full immune response to antigen may be lowered, and an immune response against tumor antigens is facilitated. Ongoing trials of ipilimumab as a therapeutic agent and as an adjuvant therapy will provide further guidance regarding the role this new drug will play in the treatment of patients with melanoma.

TREATMENT OF ADVANCED MELANOMA

Surgery The surgical treatment of patients with isolated or limited distant metastasis is controversial, as there are only a few systemic agents with demonstrated tumor responses, and only a few patients may be eligible for surgery. A number of retrospective studies have shown the validity of resecting pulmonary metastases in patients with melanoma, and more recent studies have shown similar survival rates for patients undergoing resection of limited number and sites of disease (usually up to four) for metastases to distant lymph nodes, skin and subcutaneous tissue, and lungs. Patients with isolated metastases in the liver, adrenal, brain, and gastrointestinal tract may undergo resection, but the survival of these patients may be less favorable.

Surgical resection may also be an important component of palliation. Bleeding or obstruction from small intestine metastases may be managed by resection. Although palliative resection can provide short-term effective relief of patient symptoms, numerous studies have shown that incomplete resection does nothing to enhance the length of survival.

Single-agent chemotherapy Melanoma is regarded as a relatively chemotherapy-refractory tumor. Durable objective responses have been observed in a small minority of patients with metastatic melanoma treated with single-agent chemotherapy, providing perhaps the most compelling reason to offer these agents routinely.

In a phase III trial comparing temozolomide (Temodar) and dacarbazine as first-line therapy, 3% of

patients on either arm experienced complete responses. The highest response rates have been observed with alkylating agents, platinum analog, and mictrotubule interactive drugs (Table 4). The reference standard to which novel agents and regimens are compared is dacarbazine. In the majority of trials, dacarbazine was administered IV at daily doses of 200 mg/m2 for 5 days every 3 or 4 weeks; however, 1,000 mg/m2 once every 3 or 4 weeks has been given in recent trials. The most common toxicities are myelosuppression and nausea. The severity of myelosuppression rarely requires the use of growth factor support, and the advent of potent antiemetics in recent years has significantly improved the tolerability of this agent. In the largest phase III trial that included a single-agent dacarbazine arm, the objective response rate was 3.5%, and the median progression-free survival was 1.5 months. Overall survival for dacarbazine-treated patients is 6 to 9 months and does not clearly differ from the natural history of metastatic melanoma. Fotemustine, a nitrosourea with modestly superior activity compared with dacarbazine, is available for clinical use in Europe but not in the United States.

Combination chemotherapy regimens As with other types of cancer, combinations of agents with some measurable single-agent activity in melanoma have been empirically developed. Clinical trials combining chemotherapy have produced promising response rates in single-arm, single-institution studies but have never demonstrated an improvement in overall survival compared with single-agent chemotherapy in multicenter, randomized trials. Given the increased toxicity associated with such regimens, the absence of a survival advantage limits their consideration as standard therapies. These treatments include cisplatin, vinblastine, and dacarbazine (CVD) and the Dartmouth regimen (cisplatin, carmustine, dacarbazine, and tamoxifen). Inpatient treatment is standard for these regimens, and the severity of myelosuppression often requires growth factor support not needed for single-agent chemotherapies. Given that longevity is not impacted for these patients who have such a short life expectancy, the quality-of-life detriment associated with combination therapy cannot be justified.

Tamoxifen Based on anecdotes of tumor regression associated with single-agent tamoxifen and laboratory evidence suggesting synergy with chemotherapy, clinical trials were undertaken combining tamoxifen with individual chemotherapy agents or combination regimens. As with other combination chemotherapy trials, single-arm trials of regimens containing tamoxifen appeared promising. However, in randomized trials where the same chemotherapy regimen was administered to all patients, and tamoxifen given only to the experimental arm, no evidence of clinical benefit was observed with the addition of tamoxifen.

Biologic therapies IL-2 was approved by the FDA as a treatment for metastatic melanoma on the basis of durable complete and partial remissions associated with the “high-dose” regimen. This regimen requires patients to be in excellent overall health to tolerate the physiologic stress of 5 days of inpatient therapy. The standard dose is 200,000 U/m2 repeated every 8 hours, for a maximum of 14 doses. This treatment is followed by a treatment break of 10 to 14 days and readmission for another course of therapy. Patients who demonstrate some degree of tumor regression are offered additional courses of therapy. Typical toxicities include fever, malaise, hypotension complicated by renal dysfunction, elevated levels of liver transaminases, and mood alterations. Clinical expertise among physicians and nurses is required to safely and effectively administer this therapy. As a consequence, the use of high-dose IL-2 has largely been restricted to high-volume referral centers.

IFN-α has demonstrated single-agent activity in metastatic melanoma that is comparable to chemotherapy, including occasional durable responses. A randomized trial comparing IFN-α with dacarbazine in metastatic disease has never been conducted. Given the significant toxicity associated with chronic administration of IFN, its use in the metastatic setting is limited.

Biochemotherapy regimens The distinct mechanism of action of biologic agents and evidence of single-agent activity have led investigators to combine these agents with chemotherapy for the treatment of metastatic melanoma. Given the toxicity associated with high-dose IL-2, it is not possible to safely administer chemotherapy concurrently. However, lower-dose IL-2 regimens can be safely coadministered with combination chemotherapy. Likewise, the high-dose IFN regimen that is the current standard therapy for adjuvant treatment of stages II and III melanoma cannot be easily combined with chemotherapy, whereas modified schedules of IFN can.

The first biochemotherapy regimens tested in phase II trials and then subsequently in large, randomized phase III trials combined IFN-α with dacarbazine. Despite early evidence suggesting a substantially higher response rate associated with the combination, definitive phase III trials failed to identify a survival advantage to this regimen over with dacarbazine alone.

More intensive regimens combining chemotherapy with biologic agents, requiring inpatient administration, have more recently been evaluated. A regimen containing cisplatin, vinblastine, dacarbazine, with concurrent IFN and IL-2, was evaluated in several phase II trials, a small, single-institution phase III trial, and a multicenter phase III trial in comparison to multiagent chemotherapy without the biologic agents. Although progression-free survival appeared modestly superior to that of the chemotherapy backbone alone, no improvement in overall survival could be demonstrated. Likewise, a trial in which the contribution of IL-2 to a regimen of cisplatin, dacarbazine, and IFN-α was isolated failed to demonstrate a significant improvement in outcome.

Targeted therapy The discovery of somatic genetic mutations in melanoma, underlying aberrant signal transduction in melanoma cells, has provided leads for the development of molecularly targeted therapy. The MAP kinase pathway, which is activated in the vast majority of melanomas due to mutations in BRAF, NRAS, and c-kit, has been the focus of most clinical investigations of signal transduction (kinase) inhibitors. Angiogenesis, necessary for the progression of all solid tumors, has proved to be a valuable therapeutic target in the treatment of other metastatic cancers and is currently being investigated in randomized trials in melanoma.

Combinations of targeted therapy with chemotherapy are being extensively studied. Sorafenib (Nexavar) has been approved by the FDA for treatment of renal cell carcinoma and hepatocellular carcinoma as a single agent, and is currently being evaluated in combination with chemotherapy in phase III trials in melanoma. This agent offers the potential to inhibit both the MAP kinase pathway and tumor angiogenesis. Bevacizumab (Avastin), proved to enhance the efficacy of chemotherapy in colon, non–small-cell lung, and breast cancers, is being evaluated in combination with chemotherapy in melanoma in a randomized trial.

Palliative radiation therapy Unfortunately, many patients with malignant melanoma will present with CNS metastases. These patients can be symptomatic, with complaints of headache, seizures, motor loss, or impaired mentation; these symptoms can be relieved with the use of whole-brain radiation therapy (WBRT). WBRT has been used with multiple different fractionation schemes, although none has been found to be superior to standard fractionation (300 cGy x10). These patients have a median survival of 2.2 to 4.9 months with the addition of WBRT, with local tumor control rates in the brain in the 70% to 80% range.

Stereotactic radiosurgery (SRS) involves the use of a large, single fraction of radiation to a limited target volume and rapid dose falloff outside the target. A standard linear accelerator is used to deliver one dose, usually between 15 and 24 Gy. Most centers that perform SRS have an upper size limit for the metastatic lesion (approximately 4 cm). Proponents of this treatment have shown less cognitive decline due to the limited area of radiation. There has been a small prospective series of patients with one to three brain metastases from melanoma, renal cell carcinoma, and sarcoma treated with SRS alone; they were found to have a 32% intracranial recurrence rate outside the SRS volume, suggesting there is a need for WBRT.

RTOG 95-08 was a randomized trial that looked at the addition of SRS to WBRT in patients with one to three brain metastases (including those with metastatic melanoma). Both treatment modalities produced an improvement in Karnofsky Performance Status (KPS) at 6-month follow-up, with no difference in mental status or neurologic death. Patients with only one brain metastasis who underwent WBRT and SRS had a survival advantage (median survival time, 6.5 vs 4.9 months; P = .0393).

Patients with recurrent disease or metastasis at multiple sites may require palliative radiation therapy. Common examples include skin metastases that have bleeding and fungation, painful bony metastases or lymph nodes metastases that push against soft tissue, bony metastases, neurologic compromise, and spinal cord compression from vertebral metastases. These patients may experience improvement in quality of life with a short course of radiation treatment.

Future directions
In addition to the previously mentioned trials in active research, there are multiple studies being performed at different institutions around the United States, some of which are assessing the use of radiosensitizers. The University of Utah is performing a phase II study looking at the response rate of measurable melanoma lesions with the use of cisplatin, a known radiomodifier, with radiation treatment. The M. D. Anderson Cancer Center is performing a phase I trial looking at the toxicity of using docetaxel (Taxotere)/temozolomide/cisplatin in patients with melanoma. There is still much to be understood about radiation treatment for patients with melanoma and how best to use it to its maximal efficacy in this subset of patients.

Initial treatment of nonmelanoma skin cancer

Surgery
Most nonmelanoma skin cancers can be conservatively excised with much narrower margins than are required for cutaneous melanomas. Excision margins of 0.5 to 1.0 cm are adequate for most nonrecurrent basal cell and squamous cell cancers and yield local recurrence rates under 5%, provided that histologically negative margins are achieved. For most tumors in most anatomic sites, these excision margins can be achieved using standard surgical techniques with local anesthesia and primary closure.

Radiation therapy
Radiation therapy is effective for the treatment of squamous cell and basal cell skin cancers, and is associated with a similar likelihood of cure as surgery. However, this practice is only considered standard in areas of the body where surgery would be disfiguring or debilitating. In instances where negative margins cannot be obtained without an unacceptable cosmetic result, adjuvant radiation therapy to treat microscopic residual disease appears effective. This practice is also typically extended to Merkel cell skin cancers and sarcomas of the skin. The use of radiation therapy to the surgical bed when negative margins have been obtained is more controversial and is not well supported by data.

Topical and intralesional therapyTopical therapy is a consideration for patients with large areas of skin affected by numerous or recurrent squamous cell or basal skin cancers. Immunocompromised patients and occasional patients with extremes of cumulative lifetime sun exposure are candidates for topical 5-FU or imiquimod. These therapies are generally not offered to patients who are otherwise good candidates for surgical resection. 5-FU is an antimetabolite chemotherapy that is effective in eradicating both actinic keratoses and squamous cell skin cancers. Following treatment, patients typically develop significant erythema in treated skin lasting for several weeks until the treated skin has adequately regenerated its epithelium. Imiquimod stimulates the activity of antigen-presenting cells in the skin when applied topically, which engenders an immune response in treated tumors. Complete regression of treated tumors is typically achieved. However, patients who are offered this therapy are typically at high risk for local recurrence and appearance of new lesions elsewhere and thus require lifelong surveillance and, potentially, repeated treatment.

Intralesional therapy for unresectable or recurrent basal cell skin cancers has been investigated. The only readily available agent that has been employed for this purpose is IFN-α. Comparative trials have not been conducted with intralesional vs topical therapies.

Management of recurrent nonmelanoma skin cancer Local recurrence complicates approximately 5% of cases of nonmelanoma skin cancer; if possible, surgical resection should be performed. In areas of skin where surgery would result in an unacceptable cosmetic result, radiation therapy can be offered. The vast majority of recurrent nonmelanoma skin cancers can be cured with re-excision or radiation therapy.

Regional lymph node involvement Among nonmelanoma skin cancers, Merkel cell carcinoma is the most likely to be associated with regional lymph node involvement. The prognosis of patients with lymph node involvement appears to be significantly worse than that of patients with localized disease. Thus, SLNB is justified as a staging procedure in patients with primary Merkel cell carcinoma. Squamous cell skin cancer is less likely to involve regional lymph nodes (5% of all cases), and SLNB is generally reserved for patients with primary tumors that are large, deeply invasive, or poorly differentiated. Lymph nodes should be evaluated with FNA to determine the need for lymph node dissection. Basal cell carcinoma infrequently spreads to regional lymph nodes.

Distant metastases occur infrequently in squamous cell carcinoma (2% of all cases) and generally are limited to patients who are immunocompromised. Fewer than 1 in 1,000 cases of basal cell carcinoma develop metastatic disease. Therefore, clinical trials in these patients have not been conducted. Combination chemotherapy regimens employed for squamous cell carcinoma from other sites are typically employed for patients with metastatic squamous cell carcinoma of the skin, although responses appear to be infrequent and usually short-lived.

 

References:

SUGGESTED READING

on skin cancers

Swetter SM, Geller AC, Kirkwood JM: Melanoma in the older person. Oncology 18:1187–1196, 2004.

Tsao H, Atkins MB, Sober AJ: Management of cutaneous melanoma. N Engl J Med 351:998–1012, 2004.

On Biology and epidemiology of melanoma

Chao C, McMasters KM: Relationship between age and other prognostic factors in melanoma. Am J Oncol Rev 3:446–454, 2004.

Curtin JA, Fridlyand J, Kgeshita T, et al: Distinct sets of genetic alterations in melanoma. N Engl J Med 353:2135–2147, 2005.

Demierre MF, Merlino G: Chemoprevention of melanoma. Curr Oncol Rep 6:406–413, 2004.

Kashani-Sabet M: Melanoma genomics. Curr Oncol Rep 6:401–405, 2004.

Tucker MA, Fraser MC, Goldstein AM, et al: A natural history of melanomas and dysplastic nevi: An atlas of lesions in melanoma-prone families. Cancer 94:3192–3209, 2002.

ON Surgical Treatment and imaging

Ballo MT, Ang KK: Radiotherapy for cutaneous malignant melanoma: Rationale and indications. Oncology 18:99–108, 2004.

Bonnen MD, Ballo MT, Myers JN, et al: Elective radiotherapy provides regional control for patients with cutaneous melanoma of the head and neck. Cancer 100:383–389, 2004.

DuBay D, Cimmino V, Lowe L, et al: Low recurrence rate after surgery for dermatofibrosarcoma protuberans: A multidisciplinary approach from a single institution. Cancer 100:1008–1016, 2004.

Hawkins WG, Busam KJ, Ben-Porat L, et al: Desmoplastic melanoma: A pathologically and clinically distinct form of cutaneous melanoma. Ann Surg Oncol 12:207–213, 2005.

Kammula US, Ghossein R, Bhattacharya S, et al: Serial follow-up and the prognostic significance of reverse transcriptase-polymerase chain reaction-Staged sentinel lymph nodes from melanoma patients. J Clin Oncol 22:3989–3996, 2004.

McMasters KM, Reintgen DS, Ross MI, et al: Sentinel lymph node biopsy for melanoma: Controversy despite widespread agreement. J Clin Oncol 19:2851–2855, 2001.

Mijnhout GS, Hoekstra OS, van Lingen A, et al: How morphometric analysis of metastatic load predicts the (un)usefulness of PET scanning: The case of lymph node staging in melanoma. J Clin Pathol 56:283–286, 2003.

Morton DL, Thompson JF, Cochran AJ, et al: Sentinel-node biopsy or nodal observation in melanoma. N Engl J Med 355:1307–1317, 2006.

Pawlik TM, Sondak VK: Malignant melanoma: Current state of primary and adjuvant treatment. Crit Rev Oncol Hematol 45:245–264, 2003.

Statius Muller MG, van Leeuwen PA, de Lange-de Klerk ES, et al: The sentinel lymph node status is an important factor for predicting clinical outcome in patients with stage I or II cutaneous melanoma. Cancer 91:2401–2408, 2001.

Thomas JM, Newton-Bishop J, A’Hern R, et al: Excision margins in high-risk malignant melanoma. N Engl J Med 350:757–766, 2004.

ON Surgical Treatment and imaging

Ballo MT, Ang KK: Radiotherapy for cutaneous malignant melanoma: Rationale and indications. Oncology 18:99–108, 2004.

Bonnen MD, Ballo MT, Myers JN, et al: Elective radiotherapy provides regional control for patients with cutaneous melanoma of the head and neck. Cancer 100:383–389, 2004.

DuBay D, Cimmino V, Lowe L, et al: Low recurrence rate after surgery for dermatofibrosarcoma protuberans: A multidisciplinary approach from a single institution. Cancer 100:1008–1016, 2004.

Hawkins WG, Busam KJ, Ben-Porat L, et al: Desmoplastic melanoma: A pathologically and clinically distinct form of cutaneous melanoma. Ann Surg Oncol 12:207–213, 2005.

Kammula US, Ghossein R, Bhattacharya S, et al: Serial follow-up and the prognostic significance of reverse transcriptase-polymerase chain reaction-Staged sentinel lymph nodes from melanoma patients. J Clin Oncol 22:3989–3996, 2004.

McMasters KM, Reintgen DS, Ross MI, et al: Sentinel lymph node biopsy for melanoma: Controversy despite widespread agreement. J Clin Oncol 19:2851–2855, 2001.

Mijnhout GS, Hoekstra OS, van Lingen A, et al: How morphometric analysis of metastatic load predicts the (un)usefulness of PET scanning: The case of lymph node staging in melanoma. J Clin Pathol 56:283–286, 2003.

Morton DL, Thompson JF, Cochran AJ, et al: Sentinel-node biopsy or nodal observation in melanoma. N Engl J Med 355:1307–1317, 2006.

Pawlik TM, Sondak VK: Malignant melanoma: Current state of primary and adjuvant treatment. Crit Rev Oncol Hematol 45:245–264, 2003.

Statius Muller MG, van Leeuwen PA, de Lange-de Klerk ES, et al: The sentinel lymph node status is an important factor for predicting clinical outcome in patients with stage I or II cutaneous melanoma. Cancer 91: