Julie R. Lange,Leslie A. Fecher,
William H. Sharfman,
Rhoda M. Alani,
Radha Mikkilineni,
Suzanne L. Topalian,
Charles M. Balch
SUMMARY OF KEY POINTS
|
Incidence
Etiology and Epidemiology
Pathology
Clinical Findings
Differential Diagnosis and Staging
Primary Therapy and Salvage Therapy: Surgical
Complications
Primary Therapy and Salvage Therapy: Systemic
Prognosis
|
INTRODUCTION
In the early part of the 20th century, melanoma was considered a rare disease, and often was recognized only at an advanced stage. Today, melanoma is one of the most common malignancies, and it usually is recognized at an early stage, when survival rates are high, often with surgery as the only necessary treatment. Melanoma can occur at all ages, and, therefore, can be associated with potential loss of many years of productive life. For early melanoma, improved understanding of the natural history of melanoma and a number of well-run studies of surgical treatment of melanoma have resulted in standard surgical procedures that are much less radical than those commonly performed just a few decades ago. Advanced melanoma is difficult to treat and presents opportunities for research to improve treatment strategies. Interest in the biology of melanoma continues to increase; ongoing research in the biology of melanoma and its potential application to the medical treatment of melanoma may lead to improved management strategies. This chapter reviews the current treatment of melanoma.
EPIDEMIOLOGY
According to the American Cancer Society, approximately 59,940 new cases of melanoma and 8110 deaths attributable to melanoma were expected to occur in the United States in 2007. It is the sixth most frequently diagnosed malignancy in both men and women, occurring slightly more commonly in men than in women.[1] The incidence of melanoma has risen dramatically over the last several decades, and it is now estimated that the lifetime risk of melanoma in the United States is approximately 1 in 49 for men and 1 in 73 for women.[1] Melanoma can occur in children and teenagers: the Centers for Disease Control and Prevention estimates that in 2002 there were 475 cases of melanoma in persons younger than 20 years of age.[2] The incidence of melanoma in persons under 20 years of age rose 2.9 % per year from 1973 to 2001.[3] Mortality from melanoma is significant. The overall mortality attributable to melanoma has increased because of the great increase in the incidence of the disease; however, case mortality rates have improved, and today 92% of patients are alive 5 years after their diagnosis.[1] Because melanoma often affects young and middle-aged adults, the potential for years of life lost is great. For most of the 20th century, the incidence of melanoma in populations of European origin rose faster than any other cancer except lung cancer. Current reports indicate that the incidence of melanoma is stabilizing or slightly decreasing, particularly among younger adults.[4]
CLINICAL RISK FACTORS
Most known risk factors for melanoma fall into one of two categories: susceptibility of the host and exposure to ultraviolet irradiation. Persons who are most susceptible to melanoma are fair-skinned, with a tendency to sunburn. They may have more than 20 benign moles, atypical moles, or congenital moles. A family history of melanoma also is associated with increased risk.
Skin Type
Fair-skinned persons with a tendency to burn are at higher risk than darker-skinned persons. Red or blond hair and blue or green eyes are associated with increased risk. Freckling is common.[5]
Common and Atypical Nevi
Common benign nevi, if numerous, are associated with increased risk. [6] [7] Atypical nevi are flat macules greater than 5 mm, with variable pigmentation, asymmetrical outlines, and indistinct borders. Atypical nevi are found in 2% to 7% of the white population, but in approximately 40% of patients with melanoma. The presence of atypical nevi implies a greatly increased risk of melanoma, particularly when combined with a family history of melanoma. There is a rare, autosomal dominant syndrome of atypical nevi with variable penetrance. These nevi may occur on non–sun-exposed areas. The lifetime risk of melanoma in some affected families is nearly 100%, with many melanomas occurring de novo and apparently not in pre-existing atypical nevi.[8]
Giant Congenital Nevi
Patients with large nevi present at birth or in early childhood are at increased risk for melanoma. Giant congenital melanocytic nevi often are elevated, and usually are large, prominent, irregular, pigmented lesions. They may follow a dermatome distribution and may cover large areas. They occur in fewer than 1 in 20,000 births. Malignant changes may occur in the deeper areas of the dermis and may, therefore, be difficult to detect.[9] Excision early in life when possible with limited morbidity is advised.
Personal History of Melanoma
Patients with a personal history of melanoma have an approximate 5% lifetime risk of a second melanoma. Persons who have had melanoma need lifelong skin screening.
Exposure
Ultraviolet (UV) radiation exposure, to both UVA and UVB, is strongly associated with subsequent development of melanoma. The exposure history most strongly associated with subsequent melanoma development is intermittent intense exposure, particularly a history of blistering sunburns in childhood. Exposure to UV radiation is of particular interest because it is a modifiable risk factor. However, such exposure is not necessary for the development of melanoma—areas with no sun exposure, such as the soles of the feet and the anal area, can be sites of primary melanoma. The development of melanoma is, like many cancers, a result of a complex interaction between patient susceptibility and carcinogen exposure.
Risk Management
Identification of persons with no prior diagnosis of melanoma who are at increased risk of melanoma can be difficult, and the role of routine screening is poorly defined. A case-control study evaluated the relative risk of melanoma in non-Hispanic whites, using data including common risk factors such as complexion, exposure, and number of moles. The authors developed a statistical tool that can be used to estimate an individual's risk of melanoma. This tool may prove useful in counseling at-risk individuals for reduced UV exposure and regular screening.[10] Further research is needed in risk assessment, risk management and in disease prevention.
CLINICAL PRESENTATION AND BIOPSY
The most important clinical feature of a cutaneous melanoma is change in the color, size, perimeter, or contour of a mole or pigmented skin lesion.[11] Sometimes patients report an itching sensation around a mole or an unusual sensation in a mole. Cutaneous melanoma can arise anywhere on the body. The most common site in women is the lower extremity; the most common site in men is the back. Clinical features typical of melanoma include variegated color, irregular borders, or a history of a change in such a skin lesion. A simple mnemonic to remember concerning features of a pigmented lesion is “ABCDE”: Asymmetry, Border irregularity, Color change/variegation, Diameter change, and Evolution (change over time) of the lesion.[12] Some melanomas do not have typical features; some are nonpigmented and may resemble other dermatologic entities such as basal cell carcinoma, squamous cell carcinoma, dermatofibroma, or seborrheic keratosis.
Melanoma appears in five clinical growth patterns: superficial spreading melanoma (SSM); nodular melanoma (NM); lentigo maligna melanoma (LMM); acral lentiginous melanoma (ALM); and desmoplastic melanoma (DM). The clinical features, anatomic distribution, ethnic distribution, and etiology are distinctive. SSM, the most common type of cutaneous melanoma among the white population, is largely responsible for the increased incidence of melanoma over the last few decades and may arise from a pre-existing nevus over a period of months to a few years. The average age at diagnosis is 51 years, which is one to two decades earlier than that for LMM or ALM. SSM has a predominant radial growth phase, both clinically and histologically. These lesions are more common on the trunk in men and on the legs in women, and typically are larger than common benign moles, with asymmetry, notched or irregular borders, and multiple colors (brown, black, pink, and gray).
NM is the second most common type of melanoma. It is similar to SSM in terms of age at diagnosis and anatomic distribution, but does not have a precursor radial growth phase. The lesions usually are shiny, smooth nodules, often with a single color, usually black, dark brown, or bluish. Nodular melanomas usually are thicker and thus are diagnosed at a more advanced stage than SSM. The survival rates and prognosis for SSM and NM are virtually the same when matched for thickness and ulceration. Both SSM and NM appear to be associated with acute exposure to UVB irradiation in fair-skinned persons who tend to sunburn rather than suntan. SSM is more common in individuals living in areas with greater UVB exposure.
LMM occurs on chronically sun-exposed skin, especially the face and neck, in older persons. More than 75% of such patients are older than 60 years of age, and they typically have a history of a slowly growing mole that has been present for a decade or more. Histologically, LMM lesions have a predominant radial growth phase, and there is associated solar elastosis of the surrounding skin as a result of chronic sun damage. These lesions probably are less aggressive in their metastatic behavior compared with other growth patterns, although this is controversial.
ALM is relatively uncommon. In contrast to SSM, NM, and LMM, which occur almost exclusively in fair-skinned persons, ALM can occur in any ethnic group or in persons with any degree of skin pigmentation. Lesions occur on the palms, soles of the feet, and nail beds. They tend to present as more locally advanced lesions. Even when accounting for their greater tumor thickness at presentation, they tend to be more aggressive in their behavior.
DM is uncommon. Lesions often are nondescript papules, plaques, or nodules, and may be nonpigmented; the appearance of DM may be more suggestive of basal cell carcinoma or a verruca. DM often occurs in the head and neck area and can be associated with neurotropism. The biologic behavior of DM is similar to that of a soft tissue sarcoma, in that lesions can invade across fascia and along peripheral nerves. DM has a higher rate of local recurrence than the other common forms of melanoma.
Biopsy Technique
Proper biopsy of a suspicious lesion is critical for accurate staging.[13] When melanoma is in the differential diagnosis, a full-thickness biopsy is done to interpret the maximum tumor thickness and the presence or absence of ulceration accurately. Excisional biopsy with a narrow margin of normal-appearing skin is preferred for small lesions and can be performed on most small lesions. The biopsy scar should be oriented to be compatible with a subsequent wide local excision should the lesion prove to be melanoma. On the extremities a longitudinal or oblique incision is preferred. On the trunk or the head and neck the biopsy should be oriented parallel to the skin lines. Punch biopsy is appropriate for large lesions or those at a vital anatomic site where one would want to know the diagnosis before removing the entire lesion. Punch biopsy should be performed at the most raised area. Final determination of the tumor thickness cannot be made until the entire lesion has been excised and examined. A thin shave biopsy of suspicious lesions should be avoided, because it may compromise histologic interpretation and proper measurement of thickness.
PATHOLOGY
Melanoma Histopathology
The histopathologic diagnosis of melanoma often is straightforward and relies heavily on the presence of specific alterations in growth pattern relative to that of benign nevi. Melanocytes are identified easily and differentiated from keratinocytes by their propensity for “retraction artifact” on well-prepared, routine hematoxylin and eosin (H & E)-stained sections. In normal skin they usually are located in the basal layer of the epidermis, where the ratio of melanocytes to keratinocytes is approximately 1 : 9. Benign nevi are characterized by the presence of nevus cells that are arranged in clusters or nests. Their nuclei are small, without hyperchromasia or mitoses. Nevic melanocytes are arranged uniformly and predominantly as nests and occasional single units at the rete tips of the dermo-epidermal junction and within the dermis. Nevic cells “mature” or diminish in size as they descend into the dermis. Significant alterations in these basic growth patterns are cause for concern, and when they occur, a diagnosis of melanoma must be considered.
In contrast to nevi, and with few exceptions, melanomas are asymmetric and poorly circumscribed. Melanocytes, as nests and single units, may be present at all levels of the epidermis, a pattern known asPagetoid spread. Dermal melanocytes do not mature evenly, and are characterized by enlarged, hyperchromatic nuclei. In situ melanoma is defined as an irregular proliferation of atypical melanocytes confined to the epidermis. Invasive melanoma is defined as the presence of melanoma cells, at a minimum, within the papillary dermis. The concept of radial and vertical growth phases of melanoma initially was introduced by Wallace Clark and colleagues.[14] Radial growth phase refers to the progressive intraepidermal proliferation of melanocytes that occurs in superficially invasive melanoma in the papillary dermis. It precedes the vertical growth phase of melanoma in all subtypes except nodular melanoma, which lacks prominent radial growth. Approximately one third of melanomas arise in association with a pre-existing nevus. Distinguishing between melanoma and a benign process can be challenging, particularly when attempting to distinguish a Spitz nevus from a Spitzoid melanoma or a nevus from a nevoid melanoma.[15] In these instances, the histologic differences may be subtle, and interpretation by an experienced dermatopathologist is necessary.
PROGNOSIS AND MICROSTAGING OF MELANOMA
Multivariate statistical analysis has revealed several features of primary melanoma that have prognostic significance and should be incorporated into pathology reports of invasive melanomas.[16] At a minimum, melanoma pathology reports should include site, Breslow thickness, presence or absence of histologic ulceration, Clark's level, mitotic rate, presence or absence of vascular or neural invasion, presence and degree of regression, degree of lymphocytic infiltrate, presence or absence of a pre-existing nevus, and subtype classification. Greater Breslow thickness, presence of histologic ulceration, elevated mitotic rate, microscopic satellite deposits, and angiolymphatic invasion are associated with a poorer prognosis.[17] Clark level, histologic subtype, and the presence of a coexisting nevus are of limited value in determining prognosis.[16] Histologic regression is defined as the absence of melanoma in a focal area that, instead, consists of dermal fibroplasia mixed with blood vessels, lymphocytes, and melanophages; melanoma is present in the immediate adjacent areas. Regression of the radial growth phase of melanoma is associated with poorer prognosis.[18] Regression of the vertical growth phase, which is less rigorously defined as the presence of tumor-infiltrating lymphocytes (TILs), in contrast, is associated with a better prognosis.[19]
Breslow Thickness
Alexander Breslow determined that melanoma thickness correlates with prognosis. A micrometer is used to measure the distance from the top of the granular layer to the deepest level of invasion, excluding adnexal extension. A Breslow thickness of 0.75 mm or less is associated with excellent prognosis. Breslow thickness is the strongest and most reproducible prognostic feature of the primary tumor.[20]
Ulceration
Histologic ulceration is defined as the loss of continuity of the epithelium over the tumor surface and has prognostic significance. Increasing breadth of the ulceration, specifically of 3 mm or greater or defined as involving the majority of the epithelium, has a greater chance of correlating with true ulceration. Ulceration is included in the AJCC staging system.[21]
Histologic Subtypes of Melanoma
Lentigo maligna melanoma (LMM) arises predominantly on sun-exposed areas of the head, neck, and extremities. These lesions may arise de novo or in association with a pre-existing nevus. These melanomas are characterized histologically by the confluence of enlarged hyperchromatic or atypical melanocytes arranged as single units at the dermo-epidermal junction. Prominent adnexal extension of confluent melanocytes often is present. Pagetoid spread is not always evident. Adjacent atypical keratinocytes and prominent dermal elastosis often are present, attesting to the long-standing nature of sun exposure at these sites. Complete excision of LMM sometimes is difficult, because the extent of the pathologic abnormality is not always clinically apparent. The invasive component of LMM may consist of spindled or epithelioid melanocytes.
Superficial Spreading Melanoma
Superficial spreading melanoma (SSM) is characterized by a marked intraepidermal proliferation of atypical melanocytes which are present at all levels of the epidermis. If the tumor is invasive, atypical melanocytes are found in the dermis and are arranged as irregular nests, fascicles or lobules that fail to mature. The melanocytes may appear spindled, epithelioid, or even nevus-like. Melanin pigmentation may be irregular. The dermal component may have an associated lymphohistiocytic infiltrate, with or without prominent melanophages.
Nodular Melanoma
Nodular melanoma, by definition, lacks a radial growth phase and shows extensive dermal invasion. The presence of a pre-existing nevus or an epidermal connection allows for distinction from metastatic melanoma in most cases.
Acral Lentiginous Melanoma
Acral lentiginous melanoma occurs on the hands and feet, including subungual locations. The most common location of this subtype is the sole of the foot. This subtype of melanoma can occur in all races, and is the most common subtype found in darkly pigmented races. The poorer prognosis of this subtype has been attributed to the likelihood of higher stage at diagnosis.
Desmoplastic or Spindle-cell Melanoma
Desmoplastic or spindle-cell melanoma is characterized by a fibroblastic or scar-like proliferation of spindled melanocytes. Desmoplastic melanoma tends to have a surrounding sclerotic stromal response associated with blood vessels and characteristic lymphoid aggregates scattered throughout the dermis. They may occur alone or in association with an overlying lentigo maligna melanoma. This subtype of melanoma ranges from a paucicellular, fibrosing variant, with a prominent stromal response, to a high-grade cellular and spindled or sarcomatous subtype exhibiting only focal desmoplasia. Most desmoplastic melanomas stain strongly and diffusely with S100, but are negative or only focally positive for other melanoma markers such as HMB45. They must be differentiated from other malignant spindle cell tumors of muscle, epithelial, neural, or fibroblastic origin, typically by the use of a panel of immunohistochemical markers, as discussed earlier. In some series, the paucicellular variant of desmoplastic melanoma has been shown to have improved survival relative to other, more common types of melanoma.[22]
PATHOGENESIS
Melanoma arises from transformation of melanocytes, which are of neural crest origin. Most melanocytes reside in the basal layer of the epidermis or within benign common nevi. Melanocytes synthesize melanin using the enzyme tyrosinase, and thus, under normal conditions, help to protect against UV damage. Melanoma induction by UV radiation is a multistep process, involving both UVB and UVA. Melanoma has been conceptualized as growing first in a radial growth phase with little risk of metastatic behavior. This phase is followed by the vertical growth phase with the capacity for metastasis. Different clinical and histologic features are associated with the different phases. Tumor progression likely is the consequence of multiple genetic events.
Melanoma can arise in a pre-existing nevus or de novo. Most melanomas probably do not arise from pre-existing benign nevi. Only about 20% to 30% of melanomas are pathologically associated with melanocytic nevi, and epidemiologic studies strongly suggest that the risk of an individual nevus undergoing malignant transformation is low.[23] Thus, although the presence of large numbers of nevi is associated with increased risk of melanoma, the actual precursor lesion is still in question. It is likely that both the stepwise evolution of melanocytic nevi to melanoma and the de novo onset of melanoma from malignant conversion of epidermal melanocytes are mechanisms of tumorigenesis.
BIOLOGY
Over the last decade, much progress has been made in defining the genetic events that influence melanoma development. Most recently, systematic approaches to identifying cancer-associated genetic defects have been undertaken, abetted, in large part, by the sequencing of the human genome. Such studies have brought new molecular pathways to light in the area of melanoma investigation. Greater knowledge of the molecular events that govern melanoma development is expected to result in the design of targeted therapeutic strategies that may improve patient outcomes.
RAS and RAF and the MAP Kinase Pathway
Although mutations of RAS genes are uncommon in human melanomas, mouse models of melanoma suggest that activation of this pathway in conjunction with inactivation of the p16/INK4a tumor suppressor pathway is important for tumor development.[24] Recent studies have supported the importance of this pathway in melanoma development, because a genome-wide screen of alterations in RAS or its downstream effectors identified activating mutations of the serine/threonine kinase BRAF in 59% of melanoma cell lines and six of nine primary melanomas.[25] Most BRAF mutations identified (80%) were accounted for by a single amino acid substitution (V600E, formerly identified as V599E) that rendered the kinase constitutively active. Subsequent studies showed a similarly high incidence of activating BRAF mutations in benign nevi,[26] suggesting that activation of BRAF kinase may be an initiating event for melanocyte proliferation, but is unlikely to be an important mediator of malignant conversion to melanoma. Additional studies have not identified BRAF kinase germline mutations in large-scale evaluations of patients with familial melanomas, [27] [28] [29] but it has been suggested thatBRAF mutations in nevi may serve as markers of melanoma susceptibility in an individual.[26]
Interestingly, studies of mucosal, uveal, and other noncutaneous melanomas identified considerably fewer activating mutations of BRAF versus intermittently sun-exposed cutaneous melanomas, suggesting that the high-frequency BRAF mutation is targeted specifically to cutaneous melanocytic lesions that are intermittently sun exposed, rather than those that are chronically exposed or unexposed. [30] [31] [32]
Although evaluation of primary melanocytic lesions confirmed the high rate of activating BRAF kinase mutations in both benign and malignant lesions, [33] [34] it was specifically noted that early, radial-growth-phase melanomas showed the lowest incidence of BRAF mutation (10%),[34] suggesting that BRAF kinase activation may not be necessary for malignant conversion of melanocytes. Because most large radial growth phase melanomas would be of the “lentigo maligna” type, such chronic sun-exposed melanomas could, alternatively, be initiated through non-BRAF-associated genetic events. Indeed, a recent study of genome-wide alterations in DNA copy number and BRAF and N-RAS mutational status in primary human melanomas demonstrated correlation between genomic changes and unique melanoma subtypes, stratified according to site and UV exposure.[35] Array-based comparative genomic hybridization, DNA sequencing, and immunohistochemical analyses were used to determine DNA copy number and BRAF/N-RAS mutational status, where primary melanomas were classified into four groups: mucosal, acral, and cutaneous melanomas with and without chronic sun-damaged (CSD) skin (defined by presence of solar elastosis). Most cutaneous melanomas on non-CSD skin (intermittent sun-exposed) possessed mutations in BRAF or N-RAS (59% and 22%, respectively). Other common findings included alterations in DNA copy numbers of CDK4, CCND1 (cyclin D1 gene), CDKN2A (p16INK4a), and chromosome 10 (site of PTEN).[35]
Current data on BRAF kinase mutations in benign and malignant melanocytic lesions suggest that these mutations are not sufficient to induce malignant conversion of melanocytes to melanoma. However, it may be necessary for the mutations of BRAF or other genes within the MAP kinase pathway to be activated for melanoma to develop. Thus, inhibition of BRAF kinase may be a useful therapeutic intervention when used in conjunction with other therapeutic modalities. Because therapeutic targeting of tyrosine kinases has resulted in effective treatment for a variety of malignancies, there has been much excitement over the development of BRAF kinase inhibitors as treatments for melanoma. Sorafenib, originally thought to be a raf kinase inhibitor, but clearly a multitargeted tyrosine kinase inhibitor (TKI), has been investigated in a variety of solid tumors, including melanoma.
Cell-Cycle Regulatory Proteins in the Development of Melanoma
p16/INK4A and the Retinoblastoma Pathway
Cell-cycle regulatory proteins are required for the precise regulation of cell growth and division and, therefore, are critical targets in the malignant conversion of all cells. Frequent deletions in the 9p21 locus were found in familial primary melanomas and melanoma cell lines, and linkage studies eventually led to the identification of the p16/INK4a gene as a candidate tumor-suppressor gene for familial melanoma. [36] [37] It later was discovered that this same genetic locus (referred to as CDNK2A) also encodes a second tumor suppressor gene, p14/p19ARF. The p16/INK4a gene encodes an inhibitor of the cyclin-dependent kinases, CDK4 and CDK6, and leads to cell-cycle arrest at the G1 phase of the cell cycle. Because CDK4/CDK6 phosphorylates the retinoblastoma protein pRb and inactivates its tumor suppressor function, this was considered a major mechanism of p16/INK4a tumor suppression in melanomas. Inherited mutations of the gene encoding the cell-cycle regulatory protein p16/INK4a or its associated cyclin-dependent kinase, CDK4,[38] predispose patients to melanoma. Defects in the p16/INK4a gene or CDK4 play a role in the development of a relatively small percentage of sporadic melanomas, and only about 20% of familial melanoma cases harbor p16/INK4a mutations, suggesting that most familial melanoma cases are associated with other genetic defects. Expression of p16/INK4a is silenced in sporadic melanomas via epigenetic inactivation through promoter methylation[39]; however, this form of gene silencing seems to be activated in a limited number of primary tumors. Of note, in thin sporadic melanomas, loss of p16 expression is associated with disease progression, despite the low incidence of loss of heterozygosity at the p16/INK4a locus, p16/INK4a intragenic mutations, and p16/INK4a promoter methylation (<10% in lesions thinner than 4 mm). [40] [41] This finding suggests that alternative mechanisms exist to allow for decreased expression of p16/INK4a in these early lesions. Other mechanisms of p16/INK4a inactivation in melanoma are under investigation, including transcriptional repression of the p16/INK4a promoter itself.
An area of interest is the potential role of Id helix-loop-helix transcription factors in melanoma initiation. In general, high Id expression levels are found in proliferative, undifferentiated cells, a feature that is characteristic of tumor cells. Id genes have been identified as potential proto-oncogenes because overexpression of Id proteins in primary cells promotes cellular immortalization. [42] [43] Id gene expression also is elevated in various tumor cell lines as well as a broad spectrum of primary human tumors.[44] In situ evaluations have shown a correlation between tumor invasiveness, aggressiveness, and progression and Id expression. Recently, Id1 was shown to be a repressor of the familial melanoma gene p16/INK4a, [45] [46] and early studies of primary melanocytic lesions showed that Id1 expression correlated with decreased p16/INK4a expression in early melanomas that were confined to a radial growth phase. In addition, later stages of melanoma that did not express Id1 had sustained genetic mutations that inactivated the p16/INK4a gene. The data suggest that Id1 transcriptional repression of p16/Ink4a may represent one of the earliest mechanisms of dysregulation of p16/INK4a expression, resulting in melanoma initiation, and that Id1 expression may be a useful marker for malignant disease in melanocytic lesions of questionable malignant potential.[47] Large-scale studies are underway to determine the utility of Id1 expression in melanocytic lesions as both a marker of malignant disease and an independent predictor of clinical outcome.
p53 and Melanoma
As with the retinoblastoma pathway, the p53 tumor suppressor pathway often is inactivated in cancers; however, alterations of p53 itself rarely are seen in melanomas (0–25%).[48] p53 is a transcription factor that functions to maintain the genome during cellular stress, including the DNA damage induced by UV radiation. Downstream functions of p53 include repair of DNA damage, growth arrest through inhibition of the cell cycle, senescence, and apoptosis.[49]
Given the importance of such functions to the prevention of cancer, it is not surprising that the regulation of p53 functions in the cell is complex. HDM2, a negative regulator of p53, promotes p53 loss through ubiquitin-mediated proteasomal degradation and is overexpressed in early melanomas.[50] As mentioned earlier, the p16/INK4a gene has been shown to reside at a genetic locus that encodes a second tumor suppressor gene, p14/ARF. This gene functions through the p53 tumor suppressor pathway by binding to MDM-2 and decreasing its ubiquitin ligase functions. Melanoma-associated CDKN2A mutations have been documented to occur within p16/INK4a alone, p14/ARF alone, or both genes simultaneously, confirming the independent tumor suppressor functions of each.[51] PAF-1, a downstream effector of p53 involved in induction of apoptosis, has been shown to be inactivated in metastatic melanomas,[52] suggesting that p53 dysfunction in melanoma is related to targeting of p53-associated molecules.
Apoptotic Pathways and the Development of Melanomas
The process of apoptosis, or programmed cell death, is critical to cellular responses to stress and is a major pathway of cell death induced by radiation therapy and traditional chemotherapy. Melanoma cells can be resistant to therapies that have shown efficacy in other tumor types; this is related, in part, to their ability to evade normal apoptotic signals.[53] Over the last decade, the particular signaling cascades regulating apoptotic pathways have been delineated. The two major apoptotic pathways have been designated the extrinsic pathway, which is induced on activation of cell-membrane-associated death receptors by their associated ligands, and the intrinsic pathway, which is dependent on mitochondrial membrane permeability in response to cellular stress signals. Both pathways result in activation of caspases that are critical effectors of apoptosis. Melanomas evade both intrinsic and extrinsic apoptotic pathways, critical determinants of tumor response to traditional cytotoxic therapies. The cytochrome c-associated factor Apaf-1 is downregulated in advanced melanomas and influences the death response of melanoma cells to cytotoxic agents.[52] In addition, Fas and TNF-related apoptosis-inducing ligand (TRAIL) death receptors are downregulated by a variety of mechanisms, leading to impaired activation of the extrinsic apoptotic pathway.[53] Therapeutic strategies aimed at circumventing impaired apoptotic pathways in melanoma are likely to require multiple interventions to ensure the continued activation of effective death pathways, given the variety of resistance mechanisms present in melanoma cells.
CLINICAL EVALUATION AND IMAGING
On initial presentation, patients with newly diagnosed thin or intermediate-thickness melanoma rarely show evidence of metastatic disease. Most melanoma metastases cause symptoms or can be discovered on physical examination. Appropriate evaluation of patients with newly diagnosed primary cutaneous melanoma includes a thorough history, a skin examination to search for other primary skin cancers, and examination of the regional node basins. For patients with a primary melanoma 1 mm or thicker, chest x-ray and liver enzymes, including lactate dehydrogenase (LDH) are reasonable screening tools. In patients with findings on history and physical examination that raise concern or abnormal findings on chest x-ray or liver function tests, further imaging studies are warranted.
Liver function tests, including LDH, are commonly used screening tools for patients with newly diagnosed invasive melanoma, although there is no universal agreement about the utility of these tests for screening. Extensive routine radiographic evaluation of asymptomatic patients with American Joint Committee on Cancer (AJCC) stage I, II, or IIIA melanomas with routine computed tomography (CT) or positron emission tomography (PET) rarely shows distant metastases. For most asymptomatic patients, a standard chest x-ray provides adequate basic radiologic screening. Patients who have findings that raise concern on screening history and physical examination or on chest x-ray or liver enzyme tests should undergo further evaluation.
For patients with suspected intra-abdominal or hepatic metastases based on abnormal findings on physical examination or abnormal liver chemistry findings, contrast-enhanced spiral CT scan of the abdomen should be obtained. Magnetic resonance imaging (MRI) also can be used to detect melanoma deposits by a high signal on T1-weighted images. Other than MRI, a bone scan is probably the most sensitive test to detect skeletal metastatic disease, but a careful history and directed radiographs are necessary to ensure that areas of uptake do not represent old trauma or inflammation.
PET and CT scans obtained with 2-[18fluorine]-fluoro-2-deoxy-D-glucose(18FDG) have gained acceptance as a tool for detecting metastatic melanoma and for following the results of systemic or surgical therapy in stage IIIB, IIIC, or IV melanoma. [54] [55] [56] PET scanning is based on metabolic changes that could detect early metastatic disease in high-risk patients and has a reported sensitivity of 78% to 100% in detecting metastatic melanoma. [57] [58] False-positive findings have been observed in association with inflammatory responses and second primary or metastatic tumors. Wagner and associates[59]reported that the sensitivity of PET with FDG (FDG-PET) for detection of metastatic melanoma in lymph nodes depends on sufficient tumor volume. FDG-PET begins to detect metastatic tumor in lymph nodes reliably at volumes greater than approximately 80 mm3, but sensitivity decreases rapidly below this point. Fused PET-CT scans also allow correlation of metabolic findings with morphologic findings and have come into frequent use in the evaluation and treatment planning of patients with stage III and IV melanoma.
Prognostic Features
In a multifactorial analysis of 13,581 patients with localized melanoma (either clinically or pathologically), the two most powerful and independent characteristics of the primary melanoma, among all of the prognostic variables analyzed, were tumor thickness and ulceration.[16] No other feature of the melanoma or the patient with localized melanoma had the predictive capability of these two factors. Other factors that were statistically significant prognostic factors were patient age, site of the primary melanoma, level of invasion, and sex. Subsequent studies have demonstrated that mitotic rate is an independent predictor of melanoma survival outcome and risk of sentinel node metastasis. [60] [61] [62] [63] [64]
Thickness
In virtually all studies analyzing the prognosis of patients with stages I and II melanoma using a Cox regression analysis, melanoma thickness is the strongest predictor of outcome. Increasing melanoma thickness correlates with increasing risk of local recurrence, regional metastasis, and distant metastasis, and with poorer melanoma-specific survival. [16] [64] [65] Melanoma thickness is a continuous variable for which there are no naturally occurring breakpoints that delineate different biologic risks for melanoma-specific mortality.
Ulceration
Melanomas with histologic ulceration are more biologically aggressive, and are associated with a substantially increased risk of metastasis. In virtually every Cox regression analysis of prognostic factors that includes ulceration, melanoma ulceration portends a significantly worse prognosis and a higher risk of metastatic disease compared with nonulcerated melanomas of equivalent thickness. [16] [66]Melanoma ulceration correlates with increased tumor thickness, mitotic rate and increasing age, further suggesting that this factor is associated with increased risk of metastatic behavior.
Mitotic Rate
Tumor mitotic rate is a reflection of the proliferation rate of the primary melanoma. Although the pathologic definition has not yet been standardized, mitotic rate is an independent factor in predicting both incidence of sentinel node metastases and survival. [60] [61] [62] [63] [64] [67] Mitotic rate also has been associated with ulceration, further supporting its recognition as an indicator of poor prognosis. [62] [63] [68]
Site of Primary Melanoma
The anatomic site of the primary melanoma correlates significantly with survival, with trunk and head and neck sites having a poorer prognosis than extremity sites. The biological reasons for this are unclear; nevertheless, several large studies have shown this to be an independent prognostic indicator of survival. [69] [70]
Age
Although older patients have thicker melanomas and a higher incidence of ulcerated melanomas, age is an independent adverse prognostic factor, even after multivariate adjustment for other factors.[71]There is a consistent and incremental decline in both 5- and 10-year survival rates with each decade increase in age. Many studies have shown that older patients have a lower survival rate, especially those older than 60 years of age. [66] [72]
Anatomic Site
The anatomic site of primary melanoma is directly associated with prognosis, with extremity lesions having a better prognosis than truncal or head and neck lesions. [69] [70]
Level of Invasion
The level of dermal invasion has been viewed as a valuable prognostic factor for decades. [73] [74] There is less reproducibility in this determination than for thickness, and when all of the prognostic factors are analyzed in a multifactorial fashion, the level becomes much less important than thickness and ulceration for melanomas thicker than 1 mm. However, among patients with thin (<1 mm) melanoma, the level of invasion was more predictive of survival outcome than was tumor ulceration.[16]
Sex
Men in general have a worse prognosis than do women of similar melanoma presentation and stage, a finding that has been consistent across many studies [69] [75]
STAGING CLASSIFICATION
The current staging system went into effect in January 2003.[76] The AJCC Melanoma Database consisted of a total of 30,450 patients with melanoma. Of this group, 17,600 patients (58%) had information available for all of the factors required for the proposed TNM classification and stage grouping. Of the 17,600 patients included in this analysis, 12,837 (73%) had at least 5 years of follow-up information, 8633 (49%) had at least 10 years of follow-up, and 2485 (14%) had at least 20 years of follow-up. The data were merged from prospective databases of patients who did not receive any adjuvant systemic therapy and all of whom had quality control measures in place for data entry, pathology, and surgery. The current melanoma TNM categories are listed in Table 73-1 , and the stage groupings are shown inTable 73-2 . The 15-year survival curves for patients with stage I to IV melanoma are shown in Figure 73-1 .[21] The distinction between clinical and pathologic staging is worth noting.
Table 73-1 -- American Joint Committee on Cancer TNM Definitions
|
PRIMARY TUMOR (T) |
|
Tx: unable to assess (e.g., shave biopsy or regressed melanoma) |
|
T0: no evidence of primary tumor |
|
Tis: melanoma in situ |
|
T1: 0.1-1.0 mm thick |
|
T1 a: without ulceration and level ll/lll |
|
T1 b: with ulceration or level IV/V |
|
T2: 1.01-2.0 mm thick |
|
T2a: without ulceration |
|
T2b: with ulceration |
|
T3: 2.01-4.0 mm thick |
|
T3a: without ulceration |
|
T3b: with ulceration |
|
T4: >4.0 mm thick |
|
T4a: without ulceration |
|
T4b: with ulceration |
|
REGIONAL LYMPH NODES (N) |
|
Nx: Cannot be assessed |
|
NO: No regional lymph node metastasis |
|
N1: metastasis in one node |
|
Nla: micrometastasis (clinically occult) |
|
Nib: macrometastasis (clinically apparent) |
|
N2: metastasis in two to three nodes or in-transit metastases without nodal metastases |
|
N2a: micrometastases |
|
N2b: macrometastases |
|
N2c: in-transit met(s)/satellite(s) without metastasis in nodes |
|
N3: metastasis in four or more regional nodes, matted notes, or in-transit metastasis or satellite(s) with metastasis in node(s) |
|
DISTANT METASTASES (M) |
|
Mx: cannot be assessed |
|
M0: no distant metastasis |
|
M1: Distant metastasis |
|
M1a: skin, subcutaneous, distant lymph node metastases |
|
M1b: lung metastases |
|
M1c: all other visceral metastases, or distant metastases at any site with elevated serum LDH |
From Greene FL, Page DL, Fleming ID, et al (eds): AJCC Cancer Staging Manual, 6th ed. New York: Springer-Verlag, 2002.
Table 73-2 -- American Joint Committee on Cancer Pathologic Stage Grouping
|
Stage 0: |
TisNOMO |
|
Stage IA: |
T1aNOMO |
|
Stage IB: |
T1bNOMO |
|
T2aN0M0 |
|
|
Stage IIA: |
T2bN0M0 |
|
T3aN0M0 |
|
|
Stage IIB: |
T3bN0M0 |
|
T4aN0M0 |
|
|
Stage IIC: |
T4bN0M0 |
|
Stage IIIA: |
T1-4aN1aMO |
|
T1-4aN2aM0 |
|
|
Stage IIIB: |
T1-4bN1aMO |
|
T1-4bN2aM0 |
|
|
T1-4aN1bMO |
|
|
T1-4aN2bM0 |
|
|
T1-4a/bN2cM0 |
|
|
Stage IIIC: |
T1-4bN1bMO |
|
T1-4bN2bM0 |
|
|
AnyTN3M0 |
|
|
Stage IV: |
anyT, any N, M1 |
From Greene FL, Page DL, Fleming ID, et al (eds): AJCC Cancer Staging Manual, 6th ed. New York: Springer-Verlag, 2002.
|
Figure 73-1 Fifteen-year survival curves for stages I, II, III, and IV melanoma. For each curve, N = the number of patients in the AJCC melanoma database used to calculate rates. Curve differences are all highly significant (P < 0.0001). (From Balch CM, Buzaid AC, Soong SJ, et al: Final version of the American Joint Committee on Cancer staging system for cutaneous melanoma. J Clin Oncol 19:3635, 2001.) |
Clinical Staging
Patients with clinical stage I and II disease have invasive melanoma with no evidence of metastases at either regional or distant sites, based on clinical, radiologic, or laboratory evaluation. Patients with clinical stage III melanoma have clinical or radiologic evidence of regional metastases in the regional lymph nodes or evidence of intralymphatic satellite or in-transit metastases. The clinical or radiologic assessment of the regional lymph nodes is inherently difficult, especially with respect to assessing the presence and number of metastatic nodes. There are, therefore, no subgroup definitions of clinically staged patients with nodal or intralymphatic regional metastases. Patients with clinical stage IV melanoma have metastases at a distant site and are not subgrouped.
Pathologic Staging
In contrast to clinical staging, greater accuracy is possible in defining distinctive prognostic subgroups for appropriate patients for whom pathologic information is available about the regional lymph nodes (after sentinel or complete lymphadenectomy). Pathologic stage I and II melanoma has no evidence of regional or distant metastases, based on absence of nodal metastases after pathologic examination of the regional lymph nodes and absence of distant metastases based on routine clinical and radiologic examination.
Pathologic stage III melanoma has pathologic evidence of regional metastases, either in the regional lymph nodes or in intralymphatic sites. Quantitative classification of pathologic nodal status requires that pathologists perform a careful examination of the surgically resected nodal basin and report on the number of lymph nodes examined and the number of nodes involved with metastases, and that they determine whether the nodes are microscopically or macroscopically involved. Pathologic stage IV melanoma has histologic documentation of metastases at one or more distant sites.
Clinical versus Pathologic Staging
The ability to stage patients more accurately with sentinel node evaluation has had a profound effect on staging, treatment planning, and the conduct of clinical trials of patients with melanoma. [77] [78] The widespread use of sentinel lymphadenectomy has led to considerable stage migration of patients previously staged as “node-negative” when their nodal metastases were undetected.[79] These patients have shown the extraordinary heterogeneity of metastatic risk for stage III melanoma. [16] [66] Survival differences between patients with clinically and pathologically staged disease were statistically significant among all T substages, except T4b.
TNM Criteria
The primary criteria for the T classification are tumor thickness (measured in millimeters) and the presence of histologic ulceration. The T category thresholds of melanoma thickness are defined in even integers (i.e., 1.0, 2.0, and 4.0 mm), because they represent a statistical “best fit” and are most compatible with current thresholds in clinical decision-making. [80] [81] A clinically convenient and widely used threshold of 1.0 mm or less is used for T1 melanomas. T2 melanomas are 1.01 to 2.0 mm thick, T3 melanomas are 2.01 to 4.0 mm thick, and T4 melanomas are more than 4.0 mm thick. Melanoma ulceration heralds a high risk for metastases; survival rates for patients with an ulcerated melanoma are remarkably similar to those for patients with nonulcerated melanomas of the next highest T category.[21] [73] The level of invasion is an independent prognostic feature of “thin” (T1) melanomas, but not of thicker lesions. [16] [73] Therefore, the level of invasion is incorporated into the stage grouping definitions of T1 melanomas only. In the cohort of T1 melanomas, the assignment of T1a is restricted to patients whose lesions meet three criteria: (1) lesion thickness of 1.0 mm or less; (2) absence of ulceration; and (3) depth of invasion limited to level II or III. T1b melanomas are defined as those with a thickness of 1.0 mm or less and with the more aggressive features of level IV or V, or with ulceration. All T2, T3, and T4 melanomas are defined according to the thickness and ulceration criteria as described previously, but not according to the level of invasion.
The definitions used for clinical and pathologic staging of stage III disease are more complicated than those used for the other stages because of the need to accommodate the marked diversity in the natural history of pathologic stage III melanoma. This is demonstrated by fivefold differences in 5-year survival rates for defined substages, ranging from 69% for patients with a nonulcerated melanoma (regardless of thickness) and a single clinically occult nodal metastasis to a low of 13% for patients with an ulcerated melanoma of any thickness and four or more clinically apparent nodal metastases documented by therapeutic lymphadenectomy ( Table 73-3 ).[16]
Table 73-3 -- Five-Year Survival in Node-Positive Melanoma Patients
|
|
NONULCERATED PRIMARY |
ULCERATED PRIMARY |
|
No. Positive Nodes and Tumor Burden |
Percent ± SE (No.) |
Percent ± SE (No.) |
|
Microscopic involvement |
||
|
1 |
69 ±3.7 (252) |
52 ±4.1 (217) |
|
2-3 |
63 ±5.6 (130) |
50 ±5.7 (111) |
|
>4 |
27 ± 9.3 (57) |
37 ± 8.8 (46) |
|
Macroscopic involvement |
||
|
1 |
59 ±4.7 (122) |
29 ± 5.0 (98) |
|
2-3 |
46 ± 5.5 (93) |
25 ±4.4 (109) |
|
>4 |
27 ±4.6 (109) |
13 ±3.5 (104) |
From Balch CM, Soong SJ, Gershenwald JE, et al: Prognostic factors analysis of 17,600 melanoma patients: validation of the American Joint Committee on Cancer melanoma staging system. JClin Oncol 2001;19:3622.
|
SE, standard error. |
The stage groupings for pathologic stage III melanoma use four criteria to assign patients with regional metastases to one of three groups, designated as stages IIIA, IIIB, and IIIC ( Fig. 73-2 ). Patients with pathologic stage IIIA disease have three or fewer microscopic (clinically occult) nodal metastases arising from a nonulcerated melanoma (T1–4aN1aM0 and T1–4aN2aM0). Three subgroups of patients with pathologic stage IIIB disease have equivalent survival rates: (1) those with three or fewer microscopic (clinically occult) nodes arising from an ulcerated primary melanoma (T1–4bN1aM0 and T1–4bN2aM0); (2) those with three or fewer gross metastatic nodes and a nonulcerated primary (T1–4aN1bM0 and T1–4aN2bM0); and (3) those with satellite or in-transit metastases, but no evidence of nodal or distant metastases (T1–4 a/bN2cM0). Patients with stage IIIC disease include (1) those with four or more microscopic metastatic nodes and an ulcerated primary melanoma (T1–4bN2aM0); (2) those with three or more grossly involved nodes and a nonulcerated primary lesion (T1–4aN2bM0 and T1–4aN3M0); and (3) those with any combination of satellite or in-transit metastases and nodal metastases[16] [21] (see Fig. 73-2 ).
|
Figure 73-2 Five-year survival rates in the AJCC melanoma database displaying different stage groupings for stage III melanoma. (From Balch CM, Buzaid AC, Soong SJ, et al: Final version of the American Joint Committee on Cancer staging system for cutaneous melanoma. J Clin Oncol 19:3635, 2001.) |
In patients with distant metastases, the site (or sites) of metastasis and elevated serum levels of LDH are used to classify the M categories into three groups: M1a, M1b, and M1c. The 1-year survival rates range from 40% to 60%. [16] [21] The factors that are most predictive of poor survival are the site and number of metastases and elevated serum LDH levels. [82] [83] [84] Patients with distant metastases to the skin, subcutaneous tissue, or distant lymph nodes are categorized as M1a. Patients with metastasis to the lung are categorized as M1b, and have an “intermediate” prognosis when comparing 1-year survival rates.[85] Patients with metastases to all other visceral sites have a relatively worse prognosis, and are designated as M1c. When the serum LDH level is elevated above the upper limits of normal at the time of staging, patients are classified as M1c, regardless of the site of distant metastasis.
PROGNOSIS
Patients with Primary Melanoma
In a multifactorial analysis of 13,581 patients with localized melanoma (either clinically or pathologically), the two most powerful and independent characteristics of the primary melanoma were tumor thickness and ulceration. Other statistically significant prognostic factors were patient age, patient sex, site of the primary melanoma, and level of invasion.[16]
Patients with Regional Metastases
There are four major determinants of outcome for patients with pathologic stage III melanoma: (1) the number of metastatic lymph nodes; (2) the tumor burden, either microscopic (clinically occult and detected pathologically by sentinel node biopsy or elective lymphadenectomy) or macroscopic (clinically apparent by physical or radiologic examination and verified pathologically); (3) ulceration of the primary melanoma; and (4) satellite or in-transit metastases. [86] [87] Table 73-2 shows the stage groupings for stage III melanoma. Figure 73-2 shows the survival rates for these patients.
Patients with Advanced Disease
The site of metastasis, the number of metastatic sites, and elevated serum LDH levels were the factors that were most predictive of poor survival in all studies analyzing prognosis in patients with distant metastases using a Cox regression analysis. [88] [89] Patients with distant metastasis to the skin, subcutaneous tissue, or distant lymph nodes have a relatively better prognosis compared with patients with metastasis to another anatomic site. [88] [90] Lung metastases are associated with better prognosis compared with other visceral sites.[16] The number of metastases at distant sites is as an important prognostic factor. [82] [83] There is significant variability in the use of diagnostic tests to comprehensively search for distant metastases. Until the indications for testing and the types of tests used are better standardized, the number of metastases cannot be used reliably for staging purposes. Elevated serum LDH is among the independent factors most predictive of decreased survival, even when the site of metastasis and the number of metastases are considered. [88] [89]
TREATMENT OF THE PRIMARY
Locoregional Disease
Margins
Every primary melanoma requires wide local excision of the surrounding skin to decrease the risk of local and satellite recurrences. In most cases, wide local excision can be done as an elliptical excision with primary closure. Excision is carried down to a plane between the superficial and the deep (muscular) fascia. Primary closure often can be facilitated by a length-to-width ratio of the ellipse of approximately 3 : 1 and by undermining the flaps to close the defect with an advancement flap, if needed. In areas where the skin edges cannot be approximated primarily, a split-thickness skin graft or a rotational flap can be used to cover the defect. The skin graft donor site preferentially should be chosen outside the area of potential in-transit metastasis. Today, surgical excision margins are less radical than in most of the 20th century; the previously routine margin of 5 cm is no longer used. Primary invasive melanoma should be excised with a 1- to 2-cm minimum radial margin of normal-appearing skin around the biopsy site, with the margin determined by the anatomic site and the thickness of the melanoma.
Five randomized clinical trials have been conducted that address surgical margins around a primary melanoma. [91] [92] [93] [94] [95] Three trials involved patients with “early melanomas,” those less than 2.0 mm in thickness and with a low risk for local recurrences. The WHO Melanoma Program, the first to conduct a prospective clinical trial of patients treated in Europe and the United States, randomized patients with melanomas smaller than 2 mm to either a 1- or 3-cm margin of excision. [92] [96] This seminal clinical trial showed that melanomas less than 2 mm in thickness could be safely excised with a 1-cm margin. The Swedish and European (French) randomized trials were initiated at the same time but were published later. [91] [93] These two trials were more conservative; both studies randomized melanomas smaller than 2 mm to either a 5- or a 2-cm radial margin of skin excision. Both studies concluded that the narrower 2-cm margin was safe.
Two clinical trials were conducted for thicker melanomas: the Intergroup Melanoma Trial conducted in the United States on melanomas 1.0 to 4.0 mm in thickness, comparing 2- vs. 4-cm radial margins, and the British study of melanomas larger than 2.0 mm in thickness, comparing 1- vs. 3-cm margins of excision. The Intergroup Melanoma Trial showed that the narrower 2-cm surgical margin was safe with regard to local and regional recurrences, and overall survival.[94] The British study concluded that the 1-cm margin might not be as safe in this cohort of higher-risk patients, since the disease-free survival rates were lower in the narrower excision margin cohort, although there was no difference in overall survival.[95] This study, therefore, leads to some circumspection about adopting a 1-cm surgical margin in all patients, because a narrower margin appeared to be associated with some increased risk for the subsequent development of local and regional metastases.
Difficult Sites
EAR.
Primary melanoma of the ear usually can be treated with wedge excision. Primary closure usually is possible, with acceptable cosmetic results. Complete amputation is used only for extensive involvement of the ear or for a local recurrence that is not amenable to wedge excision.
FINGERS AND TOES.
Digital melanoma often requires amputation at the midphalanx proximal to the melanoma and follows the same margin recommendations as other sites. The great toe is the most common site of melanoma of the digits and generally requires amputation at the midproximal phalanx. Sufficient skin and soft tissue often can be saved on the plantar surface to allow soft tissue coverage of the stump. As long as the metatarsophalangeal joint is preserved, this amputation causes essentially no functional impairment. Primary melanoma of the distal second through fifth toes usually requires amputation at the midproximal phalanx. Melanoma of the distal finger or nail bed typically requires amputation one phalanx proximal to the melanoma. Melanoma of the proximal finger or web space may be managed with soft tissue excision, with preservation of underlying tendon and bone and coverage with a full-thickness skin graft. For optimal preservation of function, patients with melanoma of the hand should be treated by a hand surgeon.
SOLE OF THE FOOT.
Primary melanoma on the sole of the foot requires wide local excision down to the plantar fascia, with either skin graft or soft tissue coverage. When primary melanoma occurs on the instep, a split-thickness skin graft may be sufficient. On weight-bearing surfaces, the tendons should be preserved, and a myocutaneous free flap can be used to provide optimal coverage.
Management of the Clinically Negative Regional Node Basin
Most melanoma experts recommend lymphatic mapping and sentinel lymphadenectomy as a staging procedure for patients with clinical stage I or II melanoma if their primary tumor is at least 1 mm thick, or, if thinner, when the melanoma is ulcerated or is level IV or V (T1bN0M0). The morbidity of the procedure is low and the staging information gained is valuable. Patients found to be node positive are then classified as having pathological stage III melanoma and usually should undergo a therapeutic lymphadenectomy of that nodal basin and should be considered for systemic adjuvant therapy.
Accurate detection of the sentinel node depends on mapping the lymphatic drainage from the skin directly next to the melanoma in a coordinated effort by the surgeon and the nuclear medicine specialist. Preoperative lymphoscintigraphy is vital to the success of sentinel node biopsy. With truncal or head and neck primary lesions, drainage to more than one node basin often occurs, and it is important to retrieve the sentinel node(s) from each node basin in which a sentinel node is identified. [78] [97] Preoperative lymphoscintigraphy facilitates the identification of sentinel nodes that lie outside traditional node basins or at unexpected sites. Sentinel nodes occasionally have been identified at popliteal and epitrochlear sites and in the triangular intramuscular space in the back, at the supraclavicular fossa, and at internal mammary and paravertebral sites.
The sentinel node should be placed in formalin for permanent fixation. Frozen section analysis is discouraged; it is preferable to save the lymph node intact for permanent processing only, so that multiple sections can be taken with immunohistochemical stains (e.g., HMB-45, S-100, Melan-A) to look for microscopic evidence of disease. Immunohistochemical staining increases the node-positive rate by 10% to 12% over staining with H & E alone. [98] [99] The risk of histologic node positivity is related to the thickness of the primary melanoma [78] [100]: In patients with lesions less than 0.76 mm thick, the chance of finding a positive sentinel node is minimal; for melanomas 0.76 to 1 mm thick, the chance of finding a metastatic sentinel node is 5% to 6%; for melanomas 1.1 to 1.5 mm thick, the chance is 7% to 8%; for melanomas 1.5 to 4.0 mm thick, the chance is 18% to 19%; and for melanomas 4 mm or thicker the chance is 29% to 34%.
Sentinel lymphadenectomy, initially described by Morton and colleagues, is a highly accurate, minimally invasive method of identifying those primary melanoma patients who have clinically occult nodal metastases. In a recent report of the results of a randomized prospective study of 1269 patients with intermediate-thickness melanomas (1.2–3.5 mm), Morton and colleagues [101] [102] confirmed the prognostic significance of the sentinel node and documented improved disease-free survival among the patients having a sentinel node biopsy. This report conclusively showed that nodal metastases could be detected 16 months earlier (median) than with the nodal observation approach in the 20% of patients who had nodal metastases. In the SLN biopsy arm, SLN status was the most significant predictor of survival in a multifactorial analysis; these results con- firm other analyses examining the prognostic significance of SLN biopsy. [78] [103]
With regard to regional disease control, the nodal observation group had multiple follow-up examinations before nodal metastases became clinically detectable; most of those who suffered a regional recurrence did require a radical lymphadenectomy, as it was usually a solitary site of metastasis. In the group with regional recurrence, there were a larger number of metastatic lymph nodes (3.3 metastatic lymph nodes vs. 1.4 in the SLN group). The implications for this higher number of nodal metastases are important. First, the subsequent regional recurrence rates increase significantly as the number of metastatic nodes in a nodal basin increases, and reaches about a 20% regional failure rate after standard lymphadenectomy in patients with four or more metastatic nodes (compared with <5% in patients with a single metastatic node). In many melanoma centers, patients with multiple, grossly detectable nodal metastases would be considered for adjuvant radiation therapy to the nodal basin, and would be more likely to receive adjuvant high-dose interferon. In contrast, patients with one or two nodal micrometastases would not be considered for adjuvant radiation therapy and would be candidates for other systemic treatments, especially on clinical trials, such as melanoma vaccine trials, in lieu of high-dose interferon.
With regard to survival rates, the two randomized groups overall were comparable; however, among the 16% of patients in each randomized cohort who had nodal metastases, survival rates were higher in the arm with early surgical intervention as directed by the SLN biopsy. The dilutional effect on overall survival rates by the 84% of patients who, in retrospect, never had nodal metastases, will make it difficult ever to demonstrate a survival advantage of the group. It would also be difficult—perhaps even impossible—to design a randomized trial differently to address this issue, because the presence or absence of nodal metastases is known only retroactively, after either the SLN procedure or close clinical examination of the regional nodal basin by palpation or ultrasound.
The use of SLN biopsy has changed the current management of melanoma dramatically, and it offers many significant benefits. First, it provides accurate prognostic information. Second, patients with a positive sentinel node biopsy can undergo a completion lymphadenectomy when the nodal metastatic burden is low. Third, patients with nodal micrometastases can be considered for adjuvant therapy or clinical trials if deemed appropriate. And finally, future trials of systemic therapy can use the results of sentinel node biopsy to assure more homogeneous groups of patients entering trials. The concept of the sentinel node is now well-established and should be used in all melanoma patients where the staging information will be useful for staging and treatment planning.
Management of Regional Nodal Metastases
Complete lymphadenectomy currently is standard treatment for patients with identified regional nodal metastases ( Fig. 73-3 ). The goals of surgery include staging, regional control of disease, and, possibly, improved survival. Most patients who require lymphadenectomy are those who have histologically positive sentinel nodes. In an era when most node-positive patients have micrometastases found by sentinel node biopsy, the value of routine dissection for patients with microscopically positive nodes has been questioned, and clinical trials are ongoing to determine the magnitude of any benefit in this population.
|
Figure 73-3 Surgical management of clinical stage I and II melanoma. Sentinel nodes should be evaluated with multiple levels of immunohistochemical staining. All patients with positive nodes or melanoma >4 mm should be strongly considered for interferon or a clinical trial of systemic adjuvant therapy. Some patients with negative nodes and melanoma >1.5 mm may qualify for clinical trials and should be considered for them if interested. SN, sentinel node; SNBx, lymphatic mapping and sentinel node biopsy; WLE, wide local excision. |
Patients presenting with clinically suspicious nodes should be evaluated with fine-needle aspiration, if possible, with excisional biopsy done only if the results of fine-needle aspiration are indeterminate. Patients with biopsy-proven bulky nodal disease should be evaluated with baseline spiral CT scans, a complete blood count, and measurement of liver enzymes, including LDH, to rule out identifiable distant disease before proceeding with nodal surgery. As a general principle, lymphadenectomy should be anatomic. The nodal contents are excised in a single block of tissue within their surrounding fatty tissue, preserving motor nerves and muscle whenever possible.
Axillary Dissection
Therapeutic axillary dissection for melanoma should include complete dissection of levels I, II, and III. The long thoracic nerve and the thoracodorsal neurovascular bundle are left intact unless they are directly invaded by tumor. A closed-suction drain is placed. Patients should have no appreciable loss of range of motion or motor function. Complete dissection carries approximately a 10% risk of lymphedema in the upper extremity.
Inguinal and Iliac Dissection
For patients with positive inguinal or femoral nodes, anatomically complete inguinofemoral dissection is performed. The boundaries of the dissection extend superiorly to approximately 6 cm above the inguinal ligament, medially to the pubic tubercle and the midbelly of the adductor longus, laterally to the anterior superior iliac spine and the lateral border of the sartorius, and inferiorly to the apex of the femoral triangle. In patients not known to have iliac nodal metastases in preoperative imaging, the femoral canal is opened and Cloquet's node (the lowest node in the iliac chain) is removed. If Cloquet's node is negative, iliac and obturator dissection usually is not done, and the femoral canal is closed. The sartorius muscle is taken down from its insertion to the anterior superior iliac spine, rotated over the femoral vessels, and tacked in place to the edge of the inguinal ligament and the fascia of the adductor longus. A closed-suction drain is placed in the inguinofemoral area. Inguinal dissection wounds have an infection rate of at least 10% to 15%. The risk of symptomatic lymphedema of the lower extremity is approximately 20%. Routine measures to reduce the risk of lymphedema include a program of wearing a fitted compression garment at 20 to 30 mm Hg during the day for the first 6 months postoperatively and leg elevation when possible.
Absolute indications for iliac and obturator dissection include the finding of a positive Cloquet's node intraoperatively and the finding of enlarged, suspicious iliac or obturator nodes on preoperative imaging. Iliac and obturator dissection may also be considered if four or more lymph nodes positive for disease are found at the inguinofemoral level or if the patient has bulky disease at the inguinofemoral level. The boundaries of the dissection are from the inguinal ligament inferiorly to the bifurcation of the common iliac vessels superiorly and to the obturator vessels medially.
Cervical Dissection
The extent of cervical lymphadenectomy depends on the location of the positive node and the presence of direct invasion into the structures of the neck. When a positive sentinel node is found, a functional neck dissection preserving the internal jugular vein, spinal accessory nerve, and sternocleidomastoid is appropriate. These structures should be sacrificed only if they are directly invaded by tumor.
Radiation Therapy
The risk of regional recurrence sometimes can be lowered by judicious use of radiation therapy after lymphadenectomy in selected patients at particularly high risk, such as those with four or more positive nodes or bulky nodal disease with extracapsular extension. Although there are no prospective randomized studies showing superior outcome with the addition of postoperative radiation, a few retrospective studies strongly suggest that radiation may improve regional control in patients at particularly high risk for treatment failure. [104] [105] [106] [107]
Management of Locoregional Recurrence
Local Recurrence
Local recurrence rates after appropriate wide local excision are low. With long-term follow-up of intermediate-thickness lesions in the Intergroup Melanoma Trial, 2.1% to 2.6% of patients overall had a local recurrence.[66] Patients at particularly high risk for local recurrence include those with thick primary lesions (>4 mm) and those with histologic ulceration, desmoplastic features, or a high mitotic rate. Local recurrence can be a sign that the patient either has or soon will have systemic disease. When local recurrence is detected, the minimal systemic workup that should be done includes careful physical examination, chest x-ray, and measurement of liver enzyme levels, including LDH. If any abnormalities are found by this examination, CT scans, PET scan, or other investigations may be indicated. Local recurrences are treated with wide local excision with 1-cm margins whenever possible ( Fig. 73-4 ). Radiation therapy or systemic IL-2 therapy for unresectable or confluent lesions also is a consideration in selected patients.
|
Figure 73-4 Surgical management of stage III and IV melanoma. All stage III and IV patients should be strongly considered for systemic therapy. Radiation therapy may be considered after lymphadenectomy for patients with >4 positive nodes or extranodal extension. |
In-Transit Disease
In-transit metastases of melanoma appear as identifiable tumor nodules in the subcutaneous or cutaneous tissues between a primary site and its nearest draining node basin. Approximately 2% to 4% ofpatients with melanoma eventually have in-transit disease after excision is performed for localized primary melanoma.[108] As with local recurrence, in-transit disease can be a harbinger of impending systemic disease; therefore, patients with in-transit disease should undergo a staging evaluation. If limited in-transit lesions are present and they are amenable to excision, wide local excision with negative margins is the treatment of choice.
Nodal Recurrence
Patients with recurrence in a previously undissected node basin should undergo evaluation and complete node dissection, as described previously. Recurrence in a previously dissected basin should be evaluated with a staging workup, including CT scans and measurement of liver enzymes, including LDH, and should undergo surgical excision of the area of recurrence. Adjuvant radiation therapy can be considered as well.
Isolated Limb Perfusion or Infusion
The indications for isolated limb perfusion or infusion are confined to extensive in-transit lesions or local recurrences involving an extremity. [109] [110] Limb perfusion with hyperthermia and melphalan has never been shown to be associated with improved survival, but it has a role in securing local and regional control for patients with unresectable local recurrence or a large volume of in-transit disease. [110] [111] Still under study is the addition of tumor necrosis factor, which in some studies shows a suggestion of improved disease response, but in others appears to offer no benefit. [112] [113] However, the risk of regional toxicity is significant; isolated limb perfusion should be performed only in centers with experience with the technique, preferably in the setting of a clinical trial. A technique of isolated limb infusion using a low-flow infusion of melphalan and actinomycin without oxygenation via percutaneous catheters has been reported that is technically simpler and has reported responses similar to those of melphalan limb perfusion with hyperthermia. [109] [114] Overall response rates were reported at 85% (41% complete response) with median response duration of 16 months.[114]
SYSTEMIC THERAPY AND SPECIAL TOPICS
Systemic Adjuvant Therapy
The recently revised AJCC staging system for melanoma allows us to predict an individual's chance of survival more accurately and to identify the patients most likely to benefit from adjuvant therapy, e.g., those with relapse rates of 50% or greater.[76] Trials of adjuvant therapies for high-risk patients have been largely unsuccessful. In 1996, based on a randomized trial showing a survival benefit of 1 year, high-dose interferon alfa-2b was approved by the U.S. Food and Drug Administration (FDA) as adjuvant therapy for patients with resected stage IIB (>4 mm primary) or stage III (regional lymph node involvement) disease.[115] Subsequent randomized trials of high-dose interferon alfa-2b have shown a consistent disease-free survival benefit, but no significant overall survival benefit.[116] The use of high-dose interferon alfa-2b remains controversial. A variety of therapies, including vaccines, biochemotherapy, and biologic agents, continue to be evaluated in clinical trials for high-risk patients in the adjuvant setting, but to date none have improved overall survival rates.
Early Adjuvant Therapy Trials
A variety of chemotherapeutic agents, including dacarbazine (DTIC)[117] and carmustine (BCNU),[118] showed no benefit in the adjuvant setting. The biologic agents bacillus Calmette-Guérin[119] and interferon-gamma[120] did not improve survival in high-risk postoperative patients. A small, prospective, randomized study showed a survival benefit for patients who took megestrol acetate compared with control subjects.[121] One of four randomized studies of levamisole hydrochloride showed a survival benefit for patients in the treatment arm.[122]
High-Dose Interferon Alfa-2b
Three large published randomized trials conducted under the auspices of the Eastern Cooperative Oncology Group (ECOG) evaluated 1 year of treatment with high-dose interferon alfa-2b in high-risk patients. All defined high-risk patients as those with primary melanomas at least 4 mm thick or with regional lymph node involvement. The high-dose interferon alfa-2b dose schedule was the same in all three studies: 20 million units/m2 intravenously (IV), Monday through Friday for 4 weeks, followed by 10 million units/m2 subcutaneously (SC), three times a week for the remaining 11 months. In ECOG protocol 1684, the first trial, 287 patients were randomized to receive high-dose interferon alfa-2b or no treatment postoperatively. In 1996, a statistically significant disease-free survival (DFS) benefit (1.72 years vs. 0.98 years) and an overall survival (OS) benefit (3.82 years vs. 2.78 years) were seen in the interferon alfa-2b arm, at a median follow-up of 6.9 years.[115] The FDA then approved high-dose interferon alfa-2b for high-risk patients, as defined in this study. While at a median follow-up of 12.6 years, the DFS benefit persisted, the OS benefit did not, with both arms now having an equal number of deaths.[123]
The follow-up study, ECOG protocol 1690, involved a three-way randomization with patients receiving high-dose interferon alfa-2b, low-dose interferon alfa-2b (3 million units/day SC three times a week for 2 years), or no therapy. The study enrolled 642 patients; statistical analysis in this study focused on hazard ratios. At a median follow-up of 4.3 years, high-dose interferon alfa-2b led to a reduction in the risk of recurrence compared with no treatment (hazard ratio = 1.28). However, there was no difference in OS at 5 years.[116] The relapse-free survival (RFS) benefit, with no OS benefit, was confirmed at median follow-up of 6.6 years as well.[124] Of note, two confounding issues were that the patients in the observation arm had a surprisingly good survival rate, median survival of 6 years, even after relapse and that some of these patients eventually received high-dose interferon alfa-2b therapy.
In ECOG protocol 1694, high-dose interferon alfa-2b was compared with a GM2-ganglioside vaccine in 880 high-risk patients (high risk as defined previously). The Data Safety Monitoring Committee stopped this study early, at a median follow-up of 16 months, because a statistically significant DFS benefit (hazard ratio = 1.47) and an OS benefit (hazard ratio = 1.52) were noted for the high-dose interferon alfa-2b arm. The estimated DFS rate at 2 years for the patients receiving high-dose interferon alfa-2b was superior (62% vs. 49%), as was OS (78% vs. 73%).[125] Because there was no observation arm, it is impossible to know whether the vaccine was inferior or equivalent to no therapy.
A pooled analysis of ECOG trials upheld the finding of prolonged RFS, but not OS, in patients treated with high-dose interferon alfa-2b vs. observation on two-sided univariate log-rank analysis of E1684 and E1690 pooled data. Multivariate models adjusted for negative prognostic factors also confirmed a significant RFS benefit in pooled populations.[124] A meta-analysis of adjuvant interferon randomized trials also attempted to clarify the varied results.[126] RFS and OS were evaluated, and a subgroup analysis was performed according to interferon alfa-2b dose. Again, a significant RFS improvement was seen in those treated with high-dose interferon alfa-2b (three studies) compared with control (hazard ratio = 0.83, CI: 0.77–0.90). However, the benefit in OS was less clear as the confidence interval crossed 1.0 (hazard ratio = 0.93, CI: 0.85–1.02). Odds reduction in RFS appeared to correlate directly with increasing dose of interferon alfa-2b; namely, the greatest reduction in odds for recurrence was in the high dose groups (hazard ratio = 0.74). Yet again, no consistent benefit in OS was seen.[126]
High-dose interferon alfa-2b is associated with a number of potential toxicities, which necessitate dose reduction in approximately 75% of patients. Side effects include fatigue, anorexia, flu-like symptoms, depression, liver abnormalities, and cytopenia. Patients require frequent blood draws: weekly for 4 weeks, and then monthly.[115] The treatment is expensive, and some insurers do not reimburse for home SC administration. Many patients are not candidates for high-dose interferon alfa-2b because of age or comorbid conditions.
The results of studies of high-dose interferon alfa-2b continue to inspire much controversy. [123] [127] Since the FDA approved high-dose interferon alfa-2b in 1996, the treatment regimen has not been widely embraced by U.S. medical oncologists and has not been accepted at all by European practitioners.[123] There is a consistent DFS benefit in multiple studies that cannot be denied. However, the OS benefit faded in ECOG protocol 1684 and was never seen in ECOG protocol 1690. In ECOG protocol 1694, at a median follow-up of 2 years there was a survival benefit, but no subsequent data have been published. If there is a survival benefit to high-dose interferon alfa-2b therapy, it is likely small.
Other Interferon Alfa-2b Schedules
A number of other interferon alfa-2b schedules have been evaluated in the adjuvant setting. High-dose interferon alfa-2b given for 3 months was ineffective.[128] Low-dose interferon alfa-2b schedules of 3 million units/day SC three times a week for 3 years for patients with stage III disease,[129] or for 18 months for patients with primary melanomas more than1.5 mm thick,[130] did not show an OS benefit compared with observation. The European Organization for Research and Treatment of Cancer (EORTC) studied intermediate-dose schedules of interferon alfa-2b, 10 million units SC five times a week for 4 weeks, then three times a week for 12 or 24 months in a randomized controlled trial of 1388 patients with stage IIB or III melanoma (protocol 18952).[131] The measured endpoints were distant-metastasis-free interval (DMFI), distant metastasis-free survival (DMFS), and OS. After a median follow-up of 4.5 years, neither group evidenced a statistically significant improvement in DMFI or OS. There was a trend toward significance in DMFI and OS in patients treated with 2 years of interferon alfa-2b vs. observation; among those patients, those with stage IIb melanoma showed the greatest effect. These findings were thought to underscore the impact of tumor burden and duration of treatment on treatment effect. Another EORTC protocol (18991) that involves an extended course of pegylated interferon alfa-2b for 5 years versus observation in 1256 stage III melanoma patients reported that subjects with micrometastasis in regional nodes had improved relapse-free survival, but not improved overall survival, when treated with 5 years of pegylated interferon alfa-2b.[131]
Vaccines
Currently, increasing numbers of melanoma vaccines are in preclinical development and in phase I, II, and III clinical trials in the adjuvant setting. Despite the scientific promise of melanoma vaccines, to date none has proven beneficial in the adjuvant setting when tested in a randomized clinical trial. A notable disappointment in the field of melanoma research came in 2005 when the phase III study of CancerVax (CancerVax Corporation) in resected stage III melanoma, as well as the phase III study in resected stage IV melanoma, were stopped early by the respective external data safety monitoring committees. After interim analyses, the likelihood of a survival benefit in the Canvaxin-treated patients was sufficiently low to warrant trial closure. CancerVax was an allogeneic whole-cell vaccine composed of three highly antigenic, irradiated melanoma cell lines. In a phase II trial, CancerVax appeared to improve survival compared with historical controls.[132] M-Vax (AVAX Technologies, Philadelphia, PA), an autologous whole-cell melanoma vaccine modified by the hapten dinitrophenol and then irradiated, also improved survival in patients with stage III disease compared with historical control subjects and demonstrated that development of a delayed type-hypersensitivity response (DTH) correlated with improved DFS and OS. [133] [134] A phase III study of M-vax in metastatic melanoma is ongoing.
In addition to the Canvaxin trial just mentioned, a limited number of other phase III adjuvant vaccine trials have been completed. A vaccinia melanoma oncolysate vaccine did not improve outcome in patients with stage III melanoma.[135] In a large randomized trial, a vaccinia viral lysate of melanoma showed no benefit in patients with melanomas more than 4 mm thick or with regional lymph node involvement.[136] Melacine (Corixa Corp., Seattle, WA), a cell-lysate vaccine, was compared with observation by the Southwest Oncology Group (SWOG) in patients with melanomas 1.5 to 4 mm thick without lymph node involvement. No DFS benefit was seen.[137] However, in a retrospective analysis, vaccinated patients who were positive for HLA-A2 or C3 had a DFS rate of 77%, compared with 64% for those who received observation only.[138] Unfortunately, no further studies are planned. In a study by Bystryn and colleagues,[139] polyvalent shed-antigen vaccine showed a survival benefit in a very small phase III trial that involved 38 patients with stage III disease. In ECOG protocol 1694, discussed earlier, the GM2-ganglioside vaccine was inferior to high-dose interferon alfa-2b at a relatively short median follow-up.[125]
Granulocyte-Macrophage Colony-Stimulating Factor
Granulocyte-macrophage colony-stimulating factor (GM-CSF) is a colony-stimulating factor approved by the FDA for treatment of bone marrow transplant graft delay or failure and for speeding neutrophil recovery after induction chemotherapy in elderly patients with acute myelogenous leukemia. It also promotes melanoma antigen presentation through activation of macrophages, monocytes, and dendritic cells. GM-CSF was studied in 48 high-risk patients with melanoma (stage III with >4 involved lymph nodes or stage IV) after surgical resection with no evidence of disease. The outcome was much better in the treated patients compared with historical controls, with a median survival of 37.5 months vs. 12.2 months.[140] Based on these results, ECOG designed a double-blinded, randomized, six-arm phase III study of various combinations of a peptide vaccine, GM-CSF, and placebos of both (protocol 4697) in resected high-risk stage III and stage IV patients, which recently closed to accrual.
Adjuvant Biochemotherapy
With the initial success of biochemotherapy in advanced disease,[141] it seemed logical to design a trial for the high-risk adjuvant setting. Intergroup trial S0008 is comparing three cycles of biochemotherapy with 1 year of high-dose interferon alfa-2b administration in very-high-risk patients with stage III disease (i.e., ulcerated primary lesion with regional lymph node involvement, grossly involved or clinically palpable nodes, matted nodes, two or more involved nodes, or satellite lesions). The study continues to accrue patients, although the disappointing recently released data on biochemotherapy in patients with stage IV disease casts doubt on the utility of this intensive, toxic approach in stage III disease.[142]
Neoadjuvant Therapy
A newer strategy involves neoadjuvant therapy for regionally advanced melanoma. A dramatic response recently was reported in patients with stage IIIB-C melanoma who were treated with standard high-dose interferon alfa-2b: IV infusion daily for 4 weeks prior to curative surgical resection, followed by SC injections three times a week for 11 months.[143] Of the 20 enrolled patients, 11 demonstrated an objective clinical response and 3 evidenced a complete pathologic response. Fifty percent of patients were free of recurrent disease at median follow-up of 18.5 months.[143] Biochemotherapy in the neoadjuvant setting also has been evaluated in two phase II studies. [144] [145] Both studies showed promising results that urge further evaluation in randomized phase III setting.
MANAGEMENT OF ADVANCED DISEASE
Diagnosis and Evaluation
When metastatic disease is suspected, the diagnosis should be pathologically confirmed whenever possible. Often this can be accomplished with a minimally invasive procedure, such as excisional biopsy, fine-needle aspiration, or core biopsy. Routine staining of the pathology slides plus immunohistochemical staining with S100, HMB-45, and anti-MART-1/Melan-A, should confirm the diagnosis of melanoma and differentiate it from other malignancies. In the rare circumstance in which it is necessary to differentiate clear cell sarcoma from melanoma, the presence of the t(12;22)(13;q13) translocation can be used to exclude the diagnosis of melanoma. Patients should be fully staged with appropriate imaging studies before therapy is initiated. All patients should undergo an MRI scan of the brain, and whole-body imaging with either spiral CT scans of the chest, abdomen, and pelvis or combined PET/CT. Patients with bony symptoms should have a bone scan, plain x-rays, or MRI, depending on the clinical assessment.
Role of Surgery in Advanced Melanoma
Because highly effective chemotherapy is not available, surgery can be an appropriate treatment for isolated metastases. Surgical excision of isolated metastatic melanoma can provide quick and effective palliation and, in some cases, long-term survival. [146] [147] Surgical candidates should be selected carefully. Surgery should be used only in cases of accessible lesions and when the risk of perioperative morbidity is acceptable. Isolated visceral metastases, especially to the brain and lung, are amenable to surgical therapy. The same is true for both symptomatic and asymptomatic gastrointestinal metastases, as well as for lesions in the skin, subcutaneous tissues, or distant lymph nodes.
Chemotherapy
DTIC is an alkylating agent that is converted to its active metabolite, 5-(3-methyl-1-triazeno) imidazole-4-carboxamide (MTIC), in the liver. As a single agent, it has been studied extensively in metastatic melanoma. In early studies by the Central Oncology Group, DTIC had an overall response rate of 20%, with 5% of patients achieving complete responses.[148] Most of the patients responding to this treatment had nodal or cutaneous metastases. Subsequent randomized studies showed objective response rates of 5% to 20%. [148] [149] With modern antiemetic agents, DTIC is very well tolerated. Although multiple dosage schedules have been studied, 800 to 1000 mg/m2 IV over 1 hour every 3 to 4 weeks is more convenient and at least as effective as any other schedule.[149] Unfortunately, DTIC has not been proven to provide a survival benefit compared with supportive care or other treatments. Temozolomide, also an alkylating agent, is a pro-drug of MTIC. In contrast to DTIC, temozolomide is orally available and penetrates the blood–brain barrier. It currently is FDA-approved for the treatment of primary brain tumors, but not for melanoma. In a phase II trial, temozolomide at a dose of 150 to 200 mg/m2 orally on days 1 to 5 in a 28-day cycle had similar activity to DTIC, with a complete response rate of 5% and an overall response rate of 21%, with some responses in the central nervous system (CNS).[150] In a head-to-head comparison, the survival rates were similar—6.4 months for DTIC and 7.7 months for temozolomide.[151] Both agents are well tolerated. The oral administration route for temozolomide is attractive to patients, but without FDA approval, it is not always adequately reimbursed by insurance. Temozolomide also is used on an extended schedule, 75 mg/m2 orally every day for 6 weeks followed by a 2-week break. This is the preferred dosing when concurrent external beam radiation to the CNS, or elsewhere, is indicated, since temozolomide is radiosensitizing. [152] [153] Studies currently are in progress to determine whether one dosing regimen is better than the other. A phase I study showed greater drug exposure with an extended dosing regimen,[154] whereas selective CD4+ lymphopenia is more common with the extended regimen.[155] Consideration should be given to prophylaxis against Pneumocystis jiroveci (formerly Pneumocystis carinii) pneumonia in these patients, especially those who are receiving concomitant radiation.
A long list of chemotherapeutic agents has shown low levels of activity in metastatic melanoma, including the platinum compounds cisplatin[156] and carboplatin,[157] BCNU,[158] vindesine,[159]paclitaxel,[160] docetaxel,[161] and vinorelbine.[162] None has been shown to be superior to DTIC as a single agent, in either efficacy or toxicity profile. Two-drug combinations have not yet yielded superior results to single-agent DTIC. A small, randomized trial comparing DTIC with DTIC plus tamoxifen showed almost a doubling of the response rate with the addition of tamoxifen.[163] This result, however, was not confirmed in a follow-up study.[164] A three-drug combination of cisplatin, vinblastine, and DTIC showed encouraging early results in a phase III trial comparing the combination with single-agent DTIC, but unpublished follow-up data showed no advantage to the combination.
Two combination chemotherapy regimens—bleomycin, vincristine, lomustine, and DTIC (BOLD) and a combination of three alkylating agents (DTIC, BCNU, and cisplatin) with tamoxifen, known as the Dartmouth regimen—showed great promise in the treatment of metastatic disease and became quite popular among clinicians during the 1980s and early 1990s. The BOLD regimen demonstrated an objective response rate of greater than 40% in a phase II trial.[165] However, in a phase III study, the response rate dropped to the single digits and the regimen fell out of favor.[166] In several single-institution studies of the Dartmouth regimen, very high response rates of greater than 50% were consistently reported.[167] In one single-institution study, when tamoxifen was removed from the regimen to minimize the risk of deep vein thrombosis, the response rate dropped precipitously.[168] In a randomized trial performed by the National Cancer Institute of Canada in which the Dartmouth regimen was compared with the Dartmouth regimen minus tamoxifen, no survival advantage was shown to occur with the addition of tamoxifen.[169] However, there were more objective responses in the tamoxifen group, particularly in patients who would now be categorized as stage M1a. In an Intergroup study led by investigators at Memorial Sloan-Kettering Cancer Center, the Dartmouth regimen was compared with single-agent DTIC. In this large randomized study, no survival benefit was seen with the Dartmouth regimen. Of note, again there were more objective responses seen in the combination chemotherapy arm.[149] Further, in a randomized trial comparing the Dartmouth regimen with Melacine (Corixa Corp.), a melanoma cell-lysate vaccine, alone, survival rates were equally poor in both arms, but toxicity was significantly less with the vaccine.[170]
Immunotherapy
Since the early 1980s, two biologic therapy agents, interferon-alfa (2a and 2b) and interleukin-2 (IL-2) have been studied extensively in metastatic melanoma. In multiple studies of single-agent interferon alfa in metastatic disease, the objective response rate was approximately 15%. With single-agent interferon alfa, there are very few complete responses, and a survival benefit has never been demonstrated. Most responses are in patients with soft tissue or lymph node involvement (stage M1a). [171] [172] Higher, more toxic doses appear more active, but low, nontoxic doses (1 to 3 million units/m2 SC) appear inactive in the metastatic setting.[172] There is no randomized clinical study showing a survival benefit when interferon-alfa is added to any single agent, such as DTIC[164] or IL-2,[173] or to any combination therapy regimen, including the BOLD[174] and Dartmouth regimens.[175]
IL-2, originally known as T-cell-derived growth factor, was first reported by Rosenberg and associates[176] to have significant response rates in metastatic melanoma when given at high doses. While it does activate several types of immune effector cells, including lymphokine-activated killer cells, natural killer cells, B and T lymphocytes, and macrophages, its mechanism of action remains unclear. Its principal anti-melanoma activity is believed to be mediated by the activation of melanoma-specific cytotoxic T cells. High-dose IL-2 has been extensively studied in metastatic melanoma. Two large series have been published using the high-dose schedule of 600,000 to 720,000 units/kg IV every 8 hours for 14 doses originally described by Rosenberg and associates.[176] The National Cancer Institute reported an overall response rate of 18% and a complete response rate of 5%, with some of the responses being quite durable.[177] Similar results were reported by Atkins and associates.[178] The striking finding in both studies was that of a few, very durable responses, some maintained for longer than 120 months.
Adoptive immunotherapy with lymphokine-activated killer cells does not appear to improve the response rate seen with high-dose IL-2 alone.[179] The response rate with tumor-infiltrating lymphocytes and IL-2 was reported to be 40%, but this has not been confirmed in a phase III trial.[180]
High-dose IL-2 can cause substantial toxicity. The patient must be hospitalized for administration of the drug and subsequent close monitoring. Major side effects include fluid retention, renal failure, myocardial ischemia, and neurologic changes. Patients with underlying cardiac, renal, or pulmonary disease are not candidates for high-dose IL-2 therapy.[181] Patients with any risk factors for cardiac disease should have a normal cardiac stress thallium study before receiving high-dose IL-2. Many low-dose, alternative IL-2 regimens have been evaluated, including low-dose bolus administration, continuous infusion, and SC administration. Although these regimens are associated with less toxicity, none has shown a significant objective response rate and none has provided long-term remission. Polyethylene glycol modified-IL-2[182] and liposomal IL-2[183] have not shown any substantial clinical benefit.
Based on the in vitro synergy between IL-2 and interferon alfa-2b in metastatic melanoma, several clinical trials have been conducted combining the two agents. Activity was seen only with doses of IL-2 that required inpatient administration.[173] The hope that combined therapy with relatively nontoxic, low-dose, outpatient IL-2 and interferon would yield significant clinical activity in metastatic melanoma has not been realized.
Biochemotherapy
In the early 1990s, several investigators began to study intensive regimens that included combination chemotherapy and the biologic agents interferon alfa-2b and IL-2. Single-institution studies showed significant objective response rates. Richards and colleagues[184] combined BCNU, cisplatin, and DTIC with high-dose IL-2 and interferon in a sequential fashion and obtained an overall response rate of 55% and a complete response rate of 14%. Legha and colleagues[141] combined the cisplatin, vinblastine, and DTIC (CVD) chemotherapy regimen with continuous-infusion IL-2 and SC interferon alfa-2b, first sequentially and then in a combined regimen, with all of the drugs given over a 5-day period and the cycle repeated every 3 weeks. The combined regimen consisted of cisplatin, 20 mg/m2 IV on days 1 through 4; vinblastine, 1.5 mg/m2 IV on days 1 through 4; DTIC, 800 mg/m2 IV on day 1 only; IL-2, 9 million units/m2 daily by continuous infusion on days 1 through 4; and interferon alfa-2b, 5 million units/m2 SC on days 1 through 5, and on days 7, 9, 11, and 13; and was repeated every 21 days. The authors reported an overall response rate of 64% and a complete response rate of 21%. The combined regimen seemed as active as the more protracted sequential regimen.
Two large single-institution randomized phase III studies have been reported comparing combination chemotherapy with biochemotherapy. The National Cancer Institute Surgery Branch compared a regimen of cisplatin, DTIC, and tamoxifen with the same regimen followed immediately by high-dose IL-2 and interferon alfa-2b. The overall response rate was higher in the biochemotherapy arm, but median survival was superior in the chemotherapy arm (15.8 vs. 10.7 months).[185] At M.D. Anderson Cancer Center, CVD was compared with CVD plus IL-2 and interferon alfa-2b given sequentially. The overall response rate and median survival rate were superior in the biochemotherapy arm (48% vs. 25% and 11.8 months vs. 9.5 months, respectively).[186]
The encouraging results reported in these studies led to a large randomized Intergroup study led by ECOG (protocol 3695).[142] This study compared CVD with the combined CVD, IL-2, and interferon alfa-2b regimen of Legha,[141] with several modifications, including reducing the dose of vinblastine by 25%, using prophylactic G-CSF, and limiting the number of chemotherapy cycles to four. This modified Legha biochemotherapy regimen previously was studied in a phase II trial and yielded an objective response rate of 48% and a complete response rate of 20% in 40 patients.[187] In ECOG 3695, 416 patients were enrolled between 1998 and 2002 and no previous chemotherapy or IL-2 was permitted. Sixty percent of the patients had received high-dose interferon alfa-2b prior to the development of metastatic disease. The overall response rate was 17.1% in the biochemotherapy arm vs. 11.4% in the chemotherapy arm, and complete response rates were 3% vs. 1.4%, respectively. OS was equally poor in both arms of the study, 8.7 months for biochemotherapy vs. 8.4 months for chemotherapy.[142] The results of this large, well-done, randomized trial were very disappointing. It may be that the reduced vinblastine dose and the lack of familiarity by physicians and nurses with the complex biochemotherapy regimen contributed to the low response rates in the cooperative group setting. However, the results show that biochemotherapy, as given in this trial, leads to very few durable responses and should not be considered a standard therapy. It is possible that when this study is published with a more detailed analysis, subsets of patients may be identified who are more likely to benefit from biochemotherapy, such as those who have an excellent performance status, low-volume disease, or no previous treatment with interferon alfa-2b. It seems unlikely that any variation of biochemotherapy currently being evaluated will prove superior to the regimen studied in ECOG 3695.
O'Day and colleagues[188] explored the use of maintenance biotherapy (IL-2 and GM-CSF) for patients who achieved stable disease or partial remission with biochemotherapy. One hundred thirty-three patients were enrolled and treated on a 1-year program consisting of IL-2, 1 million units/m2 SC Monday through Friday; GM-CSF, 125 mg/m2 SC, 2 weeks on and 2 weeks off; and seven 2-day IV infusions of decrescendo IL-2. The reported 12- and 24-month survival was 57% and 23%, respectively, and 12% of patients were disease-free at a median follow-up of 30 months. Although the results are intriguing, in light of the low response rates to biochemotherapy reported in the ECOG 3695 trial, it seems that only a very small number of patients would benefit from this approach.
New Therapies
New treatment options are under development for patients who have advanced locoregional or widely disseminated melanoma. From a clinical and basic research perspective, melanoma occupies the crossroads of molecular biology and immunology. As an externally visible tumor, it offers a unique opportunity to investigate lesions at the earliest stages of carcinogenesis for molecular events or signatures portending progression, invasion, and dissemination. Knowledge of the genetic basis for aggressive melanoma behavior has led to the design of molecularly targeted therapies. In addition, melanoma is among the most immunogenic of all human cancers, and as such has been the prototype for defining cancer-specific antigens and developing anti-cancer immunotherapies.
IMMUNOTHERAPY
Cancer vaccines are a form of active immunotherapy, the effects of which depend on target-specific activation of the patient's immune system. Vaccines directed against melanoma-associated or melanoma-specific proteins (antigens) have proved capable of enhancing antitumor immune responses in patients that can be detected in vitro, and yet have had limited clinical success in the setting of advanced metastatic disease. Most melanoma vaccine trials for advanced melanoma have been nonrandomized phase I/II studies. These have included inoculation with whole melanoma cells or gene-modified cells, heat shock proteins, naked DNA, recombinant viral vectors, recombinant proteins, synthetic peptides, and dendritic cells pulsed with peptides or cell lysates.[189] Peptide vaccines have been studied most intensively, due to ease of manufacturing at relatively low cost, low potential for toxicity, and well-developed laboratory techniques for immunomonitoring. The targeted antigens have included commonly expressed cancer-testis antigens (e.g., MAGE, NY-ESO-1) and melanoma differentiation antigens (e.g., tyrosinase, gp100, MART-1/Melan-A). Clinical trials have explored the best way to administer peptide vaccines, whether as a single peptide, with multiple peptides binding to HLA class I alone or to both class I and class II (eliciting both cytotoxic and helper T cell responses); with amino acid substitutions that augment their immunogenicity; or in combination with other biologic agents. However, the results of melanoma vaccine trials to date, and cancer vaccine trials overall, have been generally disappointing in the setting of advanced disease, with objective response rates of less than 5%.[190] It is possible that melanoma vaccines will be more efficacious in the adjuvant setting, against microscopic disease burdens.
The immunosuppressive in vivo milieu of the tumor microenvironment is now thought to play a critical role in determining the outcome of interactions between the immune system and cancer. Recent studies indicate that immune responses are tightly regula-ted by positive and negative signals, through receptor-ligand interactions on the cell surface. Specific interactions between costimulatory or coinhibitory receptors on resting and activated T lymphocytes (CD28 family of molecules), and their ligands on tumor cells or professional antigen presenting cells (B7 family), trigger biochemical signals leading to cascades of transcription and expression of downstream genes in the cell nucleus. These are among the earliest events regulating the initiation, differentiation, functional maturation and termination of innate and adaptive immune responses. CTLA-4, an inhibitory member of the CD28 family of molecules which binds to B7.1 and B7.2 on antigen presenting cells, has been recently targeted in phase I and II clinical trials via infusions of blocking antibodies to treat patients with advanced stage III/IV melanoma. [191] [192] Objective response rates (complete + partial responses) of 13%, including durable complete responses, have been observed, but a significant rate of serious autoimmune complications (approximately double the response rate) has limited the use of this agent.[193] Current investigations aim to identify molecular markers that might make it possible to select patients most likely to benefit from this therapy and least likely to develop autoimmune responses. However, available clinical data indicate a significant correlation between autoimmunity and tumor regression, [192] [193] consistent with the mechanism of action of anti-CTLA-4. Blocking antibodies directed against other CD28 and B7 family members are under clinical development, and based on information from preclinical models they may provide a more favorable therapeutic ratio than anti-CTLA-4. These include antibodies blocking PD-1, a coinhibitory receptor on T cells, and B7-H1, a ligand for PD-1 that is expressed on most melanomas.[194] Because preliminary data have revealed PD-1 expression on highly activated melanoma-specific T cells stimulated by cancer vaccines and other means, the combination of PD-1:B7-H1 blockade with immunization is envisioned for future clinical trials.
Adoptive immunotherapy, a form of passive immunotherapy involving the transfer of ex vivo-expanded tumor-specific T lymphocytes into immune-replete or depleted patients, has been under study for the past two decades as a treatment for advanced metastatic melanoma.[195] Preclinical models have indicated the potential advantages of removing tumor-specific lymphocytes from the immunosuppressive in vivo milieu, and manipulating them in vitro to express a “favorable” phenotype of rapid proliferation and antitumor reactivity (cytokine secretion, cytolysis, highly avid T cell receptors) prior to transfer back into the autologous cancer-bearing host. For patients with at least one surgically resectable metastatic lesion, tumor infiltrating lymphocyte (TIL) therapy seems to offer the highest probability of objective clinical response. Adoptive TIL transfer in the context of lymphodepleting chemotherapy and high-dose IL-2 has yielded an objective response rate of 51%, including heavily pretreated patients who have not responded previously to high-dose IL-2 therapy.[196] As expected, serious toxicities related to chemotherapy-induced cytopenias and IL-2 administration, as well as some significant autoimmune events, were encountered. For patients without resectable tumors, or whose tumors fail to yield reactive TIL, alternative T cell transfer options have been explored in the clinic. These include melanoma peptide-specific T cell clones and cell lines derived from peripheral blood, [197] [198] or peripheral blood lymphocytes genetically engineered to express melanoma peptide-specific T cell receptors.[199] These approaches have yielded relatively low objective response rates (0 to 15%), possibly due to the lack of T cell help, failure of transferred T cells to traffic to tumor sites, expression of PD-1 or other coinhibitory receptors on T cells, or emergence of antigen-negative tumor variants. Currently, clinical development of adoptive T cell transfer still is restricted to a limited number of medical centers due to the complex and intensive nature of this treatment.
Anti-Angiogenic Agents
Melanoma is a highly vascular tumor, and because new blood vessel formation is thought to play an important role in its pathobiology, the clinical use of anti-angiogenic agents is being actively investigated in advanced metastatic disease. Thalidomide, which has anti-angiogenic as well as immunomodulatory properties, is ineffective as a single agent against metastatic melanoma.[200] However, recent clinical trials have investigated the activity of thalidomide in combination with extended dosing of temozolomide. In a phase II trial which excluded patients with brain metastases or prior chemotherapy, an objective response rate of 32% was observed among 38 patients with advanced stage IIIC or IV disease.[201] A markedly lower response rate of 12% was observed in a phase II study using the same dosing regimen, which included patients with brain metastases and prior chemotherapy.[202] Of note, a phase II trial of the same combination regimen in patients with brain metastases, with or without extracranial metastases, was discontinued prematurely because of a high rate of serious or lethal adverse events (31%), particularly thromboembolic events, in the absence of objective tumor regressions.[203] Newer anti-angiogenic agents evaluated in early-phase melanoma clinical trials include the thalidomide analog lenalidomide (CC-5013, Revimid)[204]; the anti-avb3 integrin antibody MEDI-522[205]; and anti-vascular endothelial growth factor (anti-VEGF, bevacizumab, Avastin).[206] Owing to the mechanism of action of this class of antineoplastic agents, it is possible that disease stabilization, rather than objective tumor regression, will prove to be the most appropriate indicator of efficacy.[207]
Targeted Therapies
Recent scientific advances have increased our understanding of the molecular events promoting melanoma carcinogenesis and maintaining the cancer cell phenotype. Most commonly, these involveaberrations in the mitogen-activated protein kinase (MAPK) pathway supporting cell proliferation, the phosphatidylinositol 3′ kinase (PI3K) pro-survival pathway, and/or the melanocyte-stimulating hormone (MSH)/microphthalmia-associated transcription factor (MITF) melanocyte lineage-specific survival pathway. [35] [208] Knowledge of these genetic events has permitted the rational targeting of critical molecules supporting melanoma growth and survival.[209] In the largest randomized trial conducted to date in patients with advanced stage III/IV melanoma, Bcl-2 antisense oligonucleotide (oblimersen, Genasense [Genta Incorporated]) was combined with DTIC in an attempt to enhance the chemosensitivity of melanoma cells by blocking pro-survival mechanisms. Studies have shown variable overexpression of Bcl-2, which inhibits the intrinsic apoptosis pathway, in melanomas. Compared to patients treated with DTIC alone, patients receiving the combination regimen showed small but statistically significant increases in objective response rate, durable response, and progression-free survival. Increased benefits were observed in a subset of patients with normal serum LDH, suggesting that such patients should be selected for future trials with oblimersen.[210]
BRAF is a member of the MAPK signaling pathway, which transduces extracellular signals via cell surface receptor tyrosine kinases to promote cell activation and proliferation. A somatic Val600Glu missense mutation in the kinase domain of BRAF is associated with more than 50% of melanomas, as well as with smaller percentages of some other cancers, and causes constitutive activation of the MAPK pathway contributing to melanoma progression.[25] The multikinase inhibitor sorafenib (Nexavar [Onyx Pharmaceuticals]), which targets mutant and wild-type BRAF as well as c-Kit, VEGFR-2, VEGFR-3, and some other proliferation and angiogenesis receptors, recently was shown to be ineffective against advanced melanoma as a single agent.[211] In addition, in 2006 the corporate drug sponsor announced failure to demonstrate improvement in PFS, the primary endpoint, in its international randomized phase III trial comparing sorafenib to placebo in combination with carboplatin and paclitaxel in chemotherapy-refractory patients with advanced melanoma.[212] An ECOG trial that is of similar design but targets chemotherapy-naïve patients and has OS as its endpoint, is currently in progress.
Several clinical trials exploring the efficacy of imatinib (Gleevec) in advanced melanoma have been conducted, despite varied data regarding the expression of c-Kit tyrosine kinase receptor in melanoma.[213] [214] The results of two phase II trials using high-dose imatinib as a single agent have shown no evidence of clinical efficacy. [215] [216] However, a recent report has demonstrated the association of genetic abnormalities in c-Kit (mutations, amplifications) with distinct clinical subtypes of melanoma, specifically acral melanoma, mucosal melanoma, and melanomas arising from chronically sun-exposed skin.[32] Kit abnormalities were not observed in melanomas arising from skin without chronic sun damage. Of note, this pattern of genetic abnormality is distinct from that observed in melanomas with mutated NRAS/BRAF, underscoring the fact that clear genetic subtypes of melanoma exist. Thus, rational selection of patients for treatment with imatinib as well as other molecularly targeted agents in future clinical trials will depend on the identification of subsets of melanoma patients harboring the relevant genetic alterations.
Ending Treatment
Patients with a poor performance status, comorbid conditions, multiple brain metastases, or advanced age are unlikely to benefit from intensive systemic therapy. These patients also may experience more side effects than healthier patients from the currently available therapies. Providing comfort measures only may be a reasonable option in some patients with metastatic melanoma. Likewise, if first-line therapy for metastatic melanoma has been unsuccessful, the patient should undergo a thoughtful assessment of his or her overall status before additional antitumor therapy is given. In any case, supportive care and comfort measures are essential aspects of oncologic practice.
Special Clinical Situations in Stage IV Disease
Solitary Metastasis
Occasionally, patients have a solitary metastatic lesion. There have been several reports of prolonged survival after surgical resection of solitary brain, lung, and liver metastases, regardless of whether patients also received postoperative adjuvant therapy.[146] A number of clinical trials are underway for patients with stage IV melanoma and no evidence of disease, and patient enrollment should be strongly considered. There is no established role for adjuvant therapy in this setting, including interferon alfa-2b, outside of a clinical trial.
Brain Metastasis
Patients with brain metastases generally do poorly. Standard therapy is whole-brain irradiation, which offers some palliation. Patients with a solitary brain lesion and no disease elsewhere, or responding or slowly growing systemic disease, should be considered strongly for neurosurgical resection, stereotactic radiosurgery, or gamma-knife therapy.[217] For patients with multiple brain metastases and a good performance status, consideration should be given to stereotactic radiosurgery or gamma-knife surgery, after whole-brain radiation. Those patients with measurable brain metastases also should be considered for temozolomide therapy, in addition to local measures, because this is the only systemic treatment that penetrates the blood–brain barrier. In the appropriate setting, these approaches can provide significant local control and palliation.
Unusual Problems
Unknown Primary Site
Patients can present with stage III or IV disease without a history of previously diagnosed melanoma. All of these patients should have a thorough skin examination, including the anal region. An eye examination is appropriate if the metastatic pattern is consistent with ocular melanoma. Unless there are specific intestinal or gynecologic signs or symptoms to suggest a mucosal primary, invasive tests, such as colonoscopy or upper endoscopy, are not recommended. Often a primary site is not found. Sometimes patients provide a history of an unusual skin lesion arising and disappearing without biopsy or treatment, or a history of a skin lesion that was cauterized or frozen. Patients who appear to have only regional lymph node involvement should undergo a potentially curative regional lymph node dissection, like any other patient with regional metastases. These patients should then be considered for systemic adjuvant therapy. Patients with isolated or disseminated metastatic disease should be treated according to the same paradigms as those with known primaries. Periodic follow-up with a dermatologist still is recommended.
Uveal Melanoma
Uveal, or ocular, melanoma is rare. Although it can be morphologically similar to cutaneous melanoma, clinically it behaves differently. Primary ocular melanoma most often is treated with iodine 125 plaque brachytherapy or enucleation.[218] Ocular melanoma metastasizes via a hematogenous route, often spreading primarily to the liver. Sometimes the metastatic process moves relatively slowly, with several years passing before recurrence with metastatic disease. It appears to respond less often to chemotherapy and biologic therapy, although this was not confirmed in a review of the SWOG experience.[219] Some biologic correlates may explain these differences: human leukocyte antigen expression appears downregulated on ocular melanoma cells,[220] the BRAF mutation is not present,[221] and distinct chromosomal abnormalities, such as monosomy 3, are present.[222] Although initial results using the BOLD plus interferon alfa-2b regimen in metastatic ocular melanoma were promising, these results were not confirmed in a large trial.[223] Some patients with primarily liver metastases appear to benefit from chemoembolization[224] or liver perfusion.[225] Many clinical trials for metastatic melanoma exclude patients with ocular melanoma.
Mucosal Melanoma
Primary mucosal melanoma also is rare. Patients often have advanced disease at the time of initial diagnosis because primary sites in locations such as the gastrointestinal tract or sinuses make early detection difficult. This, too, appears to be a distinct disease, although morphologically similar to cutaneous melanoma. [32] [35] Interestingly, a retrospective analysis showed a surprisingly good response rate of 44% for biochemotherapy in 18 patients with metastatic anorectal melanoma.[226] Currently, treatment options for metastatic disease are the same as those for cutaneous metastatic melanoma.[227]
TREATMENT COMPLICATIONS
Serious postsurgical complications are uncommon. The risks of bleeding and infection after most surgeries for melanoma are small. After sentinel node biopsy, some patients have a transient lymphocele at the site of node excision. The risk of lymphocele can be minimized by tying off the lymphatics during the sentinel node resection. If a lymphocele is large or painful, a simple office aspiration should provide adequate management. Small, asymptomatic lymphoceles can be observed and usually resolve on their own. Lymphedema has been reported in patients after sentinel node biopsy of the axilla and the inguinal area, but the incidence is low.[228]
Complete lymphadenectomy for regionally metastatic melanoma carries a risk of seroma, sensory loss, and lymphedema. A few patients have a seroma or prolonged drain output. Lymphedema is the most feared common complication of lymph node dissection and can occur after axillary or inguinal lymphadenectomy. The risk of lymphedema in this population in general is low, and most cases are mild to moderate and controllable with diligent care.
FOLLOW-UP AND SURVEILLANCE PLANS
Most melanoma recurrences are recognized first by the patient him- or herself or on a routine physical examination. A study from the Sydney Melanoma Unit reported that 73% of all first melanoma recurrences were detected by the patients themselves.[229] Mooney and colleagues[230] reported the results of a surveillance program using physical examination, blood tests, and chest x-rays for 1004 patients with stage I or II cutaneous melanoma. Physical examination detected 72% of recurrences, constitutional symptoms indicated 17% of recurrences, and chest x-ray showed 11% of recurrences. Among the 373 patients followed in a surveillance program at the Yale Melanoma Unit, of the 78 patients who had recurrences, 76% were diagnosed by a complete history and physical examination alone.[231] Routine laboratory or radiologic studies have never been shown to be beneficial for follow-up of patients with early stage melanoma.
For patients with stage III or IV disease there is no consensus about utility of, or the optimal frequency of, routine surveillance imaging. Occasionally, false-positive lesion findings on CT scan can lead to anxiety, further diagnostic tests, or even biopsy of lesions that may be unrelated to the melanoma. CT scans are useful for evaluating suspected pulmonary, mediastinal or intra-abdominal metastases, and often are of greatest use in the evaluation of patients with symptoms that raise concern or new findings on physical examination.
In patients with known distant metastases, scans are most useful when the presence of additional metastases would alter the treatment plan or when better definition of lesions is required for treatment planning or patient entry into a research protocol. FDG-PET has a sensitivity of 78% to 100% in detecting metastatic melanoma. False-positive findings may be seen in patients with inflammatory processes, such as sarcoid, and in those with second primary tumors. [57] [58] [232] PET also can be used to determine the need for further diagnostic procedures, such as radiologically guided needle biopsy of suspicious, accessible lesions.
In a patient with symptoms, an abnormal finding on physical examination or laboratory tests, or an abnormal x-ray, the definitive diagnosis of metastatic melanoma can be made only by a biopsy. Excisional or needle biopsy is relatively easy to perform when the suspected metastasis is easily accessible. However, in the right clinical setting, radiologic studies are sufficient for a clinical diagnosis, especially if the metastases involve more than one site and the abnormality was not present on previous studies.
ISSUES FOR THE FUTURE
Today, most melanomas are found at an early stage and thus are very treatable. In fact, surgery often is the only necessary treatment. However, some melanomas are at an advanced stage when they are detected, and even some early melanomas eventually will recur. There is no routinely reliable and successful way to treat patients with advanced disease. Today the most important ways to improve melanoma-related survival are prevention and early detection.
Although much progress has been made in identifying the molecular defects that are common to malignant melanocytic lesions, few of these discoveries have thus far been translated to effective therapies. It is expected that further research in the molecular genetics and the immunology of melanoma will result in an explosion of information about the events governing melanoma onset and progression. Given the limited pathologic materials available for study, multi-institutional collaboration will be required to substantially improve our understanding of the pathogenesis of melanoma. Ultimately, these discoveries will provide the framework for the design of targeted therapies for future melanoma patients.
Copyright © 2008 Elsevier Inc. All rights reserved. - www.mdconsult.com
Abeloff: Abeloff's Clinical Oncology, 4th ed.
Copyright © 2008 Churchill Livingstone, An Imprint of Elsevier
REFERENCES