Epidemiology of Melanoma
Incidence
The incidence of cutaneous malignant melanoma (CMM) has been steadily rising in the United States and in other Western countries for decades, and CMM is now considered an epidemic cancer.1 Compared to all other cancers, the incidence of CMM among whites in the United States increased the most during the period 1975-2001, from 8.7 per 100,000 to 22.6 per 100,000, an increase of more than 150%.2 In 2007, there is estimated to be 59,940 new cases of melanoma and 8,110 deaths from melanoma.3 Over the past 2 decades, epidemiologic studies of melanoma identified major environmental, host, and genetic risk factors. These factors play a complex role individually and in combination in determining melanoma risk.
Environmental Risk Factors
The major environmental risk factor for melanoma is exposure to ultraviolet radiation, both the longer UVA radiation (320-400nm) and the shorter wavelength UVB radiation (290-320nm). A complex relationship between sun exposure and response to sun exposure is suggested by epidemiologic data, with intense intermittent sun exposure at any age being more important for risk than total lifetime exposure.4 Measures of sun exposure can be imprecise when based on recall of lifetime exposure; other studies use different tools to improve the accuracy of this measure. One large case-control study used residential history and estimates of midrange UVB flux at more than 30 locations throughout the United States as a measure of cumulative sun exposure, and demonstrated that a 10% increase in average annual UVB flux was associated with a 19% increase in melanoma risk among men and a 16% increase among women.5 In this study, risk of melanoma rose with increasing time spent outdoors, even among those individuals who tanned well.
Host Risk Factors
Major host factors associated with melanoma include increased numbers of banal and dysplastic nevi, fair hair color (red or blond), light eye color, many freckles, and an inability to tan.4 Having a previous melanoma or nonmelanoma skin cancer also increases the risk for melanoma.4 Patients affected by certain genetic conditions have an increased risk of melanoma; these include xeroderma pigmentosum, retinoblastoma, Li-Fraumeni syndrome, and Werner syndrome.6
Familial Melanoma
About 5% to 12% of CMMs develop in individuals with at least one affected first-degree relative,7 suggesting that family history of CMM is a major risk factor. Familial melanoma refers to this clustering, although the term is defined variably across studies and geographic areas. In familial melanoma compared with nonfamilial melanoma, age at diagnosis is typically earlier, lesions are generally thinner, and there is a higher frequency of multiple primary melanomas (MPMs), but the lesions are histologically similar and the clinical course is not significantly different.7 A recent study using the Utah Cancer Registry and Utah Population Database to generate a familiality score showed no difference in prognostic and survival statistics between melanoma cases with high and low familial risk.8 By studying families with several members affected with melanoma, mutations in melanoma susceptibility genes CDKN2A (OMIM*600160)9,10 and CDK4 (OMIM*123829)11 were identified. Familial melanoma likely occurs by chance as well as by shared genetic and nongenetic factors, such as hair color, type of nevi, mutations in melanoma susceptibility genes, and patterns of sun behaviors.
Genetic Risk Factors: High-Risk Melanoma Susceptibility Genes
Two high-risk melanoma susceptibility genes are currently known: CDKN2A, located on chromosome 9p21 and CDK4, located on chromosome 12q14.12 Both genes exhibit autosomal dominant inheritance patterns. CDKN2A is a complex tumor suppressor gene that encodes proteins p16 and p14ARF, products that help maintain cell cycle control. The p16 protein regulates G1-phase exit through inhibition of the CDK4-mediated phosphorylation of the retinoblastoma protein (Rb), and p14ARF induces cell cycle arrest or apoptosis through the p53 pathway.12 CDK4 plays a role in the retinoblastoma pathway as well, although it functions as an oncogene.
CDKN2A is the major high-risk melanoma susceptibility gene yet identified. To date, cosegregating germline mutations in CDK4 have been detected in only 6 melanoma-prone families worldwide.13,14 Examination of 2 families with CDK4 mutations compared with 17 families with CDKN2A mutations demonstrated similar distributions of clinical characteristics, including age at diagnosis, number of melanoma tumors, and number of nevi.15 The rarity of CDK4 mutation-positive families has precluded further investigation.
Given that CDKN2A and CDK4 mutations are present in only a subset of familial melanoma kindreds, it is likely that other high-risk melanoma susceptibility genes exist. Previous family studies using linkage analysis identified regions on chromosome 1p22 and 1p36, suggesting the possibility of additional melanoma susceptibility genes in these locations.12,16 Also, studies of the chromosome 9p21 region suggest the possibility of other melanoma susceptibility loci there.12 However, no other high-risk melanoma susceptibility genes have yet been found.
Genetic Testing for Melanoma Susceptibility Genes
Recommendations Regarding Clinical Genetic Testing
The American Society of Clinical Oncology and the Department of Health and Human Services Secretary's Advisory Committee on Genetic Testing made recommendations regarding genetic testing for cancer susceptibility genes that are helpful guides for determining the benefit of widespread genetic testing.17-19 These recommendations as applied to melanoma susceptibility genes were considered by several groups including the International Melanoma Genetics Consortium, formed in 1997 and currently composed of research groups from Australia, Europe, North America, and the Middle East.6,20 Potential benefits of genetic testing for melanoma susceptibility genes include improved motivation for sun protection and skin surveillance, a lowered threshold for biopsying suspicious lesions, and early detection of primary melanoma.6
Risks and benefits of genetic testing for CDKN2A mutations have not been extensively studied, although one European group recently examined reasons why those not participating in genetic testing for CDKN2A research studies were reluctant to do so.21 This study showed that not participating was based on rational and emotional motivation; rationally motivated individuals had more accurate risk knowledge and lower anxiety scores than emotionally motivated individuals, leading the authors to suggest that informed decision making regarding genetic testing should be improved. It might be difficult to directly translate findings from a population with government-sponsored healthcare to a population with largely private insurance-sponsored care; therefore, additional surveys of participants in different healthcare settings might further improve our understanding of the motivation for and the risks and benefits of testing for CDKN2A mutations.
Clinical genetic testing for melanoma susceptibility genes is presently a controversial issue.22-25 Currently, genetic testing for CDKN2A mutations outside of defined research protocols is not recommended by most research groups for several reasons, including the low likelihood of detecting mutations, the difficulty of interpreting test results, and because knowledge of mutation status does not affect medical management. These issues are described in more detail below. Given the extreme rarity of CDK4 mutations, we restrict our discussion of genetic testing for melanoma to CDKN2A mutations.
CDKN2A Mutation Prevalence is Low
Several population and clinic-based studies demonstrated that even among high-risk groups, the frequency of detecting mutations in known melanoma susceptibility genes is low. A large population-based study examined 482 Australian families, which were classified into high, intermediate, and low-risk groups based on number of cases of melanoma in the family beyond that expected, with the number of relatives, their ages, sex, and birth cohorts taken into account.26 Of note, CDKN2A mutations were found only in the high-risk families, and the frequency of detecting mutations in this group was only 10.3%. The frequency of detecting CDKN2A mutations in melanoma cases from the general Australian population was estimated to be 0.2%.
The observation that individuals with CDKN2A mutations tend to have an earlier age at CMM diagnosis than the general population was the motivation for studies in which young melanoma patients were tested for CDKN2A mutations. These series, which included a population-based study of children with melanoma in Australia,27 a population-based study of adolescents with melanoma in Australia,28 a clinic-based study of adults under age 40 with melanoma in the United States,29 and a clinic-based study of adults under the age of 31 with melanoma in Spain,30 all demonstrated a low frequency of CDKN2A mutations ranging from 1.4% to 5%, with many of the mutation-positive individuals having a family history of melanoma.12
The presence of MPM is associated with increased detection of CDKN2A mutations as demonstrated in Table 1. In patients with MPM and no family history of the disease, approximately 10% have germline mutations in CDKN2A.6,12 A recent international population-based study, however, reported that only 1.8% of individuals with MPM without a family history had CDKN2A mutations.31 The reason for the variation in estimates is unclear, although it might reflect the different means of ascertainment used in these studies or, possibly, the different levels of confirmation of melanomas obtained from relatives. Among MPM patients, the frequency of detecting a CDKN2A mutation is higher in those with a family history of melanoma.13,31,32
About 20% to 40% of melanoma-prone families with 3 or more affected members from North America, Europe, and Australia have germline mutations in the CDKN2A gene,6 as demonstrated in Figure. As the number of melanoma cases in a family increases, the frequency of mutations increases. The frequency of detectable mutations, shown in Table 1, is less than 5% for families with only 2 affected members; it increases to greater than 50% for families with more than 6 affected members.6,7
Using data from 116 familial melanoma patients from Boston, a regression model for estimating CDKN2A carrier probability was developed recently based on a patients' age, number of primary melanomas, and number of affected family members, and was validated in an external data set of 143 familial melanoma patients from Toronto.33 Although this model is a potential tool for genetic risk assessment, it has not yet been examined in non-North American study populations. As demonstrated in studies of CDKN2A mutation prevalence in the general population and familial melanoma kindreds, widespread screening for CDKN2A mutations is likely to have a very low yield.34
Interpretation of Results of CDKN2A Mutation Testing Is Complex: Penetrance Estimates Vary With Geography and Population Tested
For widespread genetic testing to be considered, the interpretation of results should be adequate17,18; adequacy of interpretation includes both the validity of a test result and the ability to interpret risk conferred by the presence of a mutation. The observation that many CDKN2A mutation carriers, including some individuals homozygous for CDKN2A mutations, do not develop melanoma prompted studies evaluating the penetrance of CDKN2A mutations. The International Melanoma Genetics Consortium performed the largest family-based evaluation of penetrance in CDKN2A mutation carriers and reported geographic variation of penetrance estimates. In this study of 80 families with documented CDKN2A mutations and multiple family members with CMM from Europe, Australia, and the United States, penetrance estimates differed according to the local population incidence rate of melanoma, from 13% in Europe, to 50% in the United States and 32% in Australia by age 50 years, and from 58% in Europe, to 76% in the United States and 91% in Australia by age 80 years.35 These data suggest that similar factors affect both population incidence of melanoma and penetrance of CDKN2A mutations.
Comparing family-based studies with population-based studies can further our understanding of CDKN2A mutation penetrance. Estimates of lifetime risk of melanoma among relatives of CDKN2A mutation carriers who had melanoma were obtained from a large population-based study of 3,550 individuals with first or subsequent melanoma from Italy, Australia, Canada, and the United States.31 In this study, lifetime risk estimates were much lower than those from family studies, with 14% by age 50 years and 28% by age 80 years, as shown in Table 2. These risk estimates likely represent an upper bound for non-family-based estimates because relatives were identified by a proband already having at least 1 melanoma. The higher penetrance estimates obtained from family studies compared with studies of the general population are not unexpected and implicate the involvement of other genetic, environmental, and/or host factors in modifying melanoma risk in familial melanoma. For example, certain variants of MC1R (OMIM*155555), a pigmentation gene, dysplastic nevi, and total number of nevi influence CDKN2A penetrance in familial melanoma.36-40 Due to variations in penetrance estimates and our lack of knowledge of why these variations exist, the ability to interpret risks conferred by CDKN2A is not presently adequate.
Interpretation of Results of CDKN2A Mutation Testing Is Complex: Risk Estimates of Nonmelanoma Cancers Are Currently Unknown
In addition to melanoma, other cancers have been associated with CDKN2A mutations, including pancreatic, breast, and neural system tumors. According to several studies, there is an increased risk for pancreatic cancer in melanoma-prone families with CDKN2A mutations.34,41 However, it is not possible at this time to accurately predict the genotype or phenotype that predisposes an individual with a CDKN2A mutation to develop pancreatic cancer. A subset of families with CDKN2A mutations, predominantly in Sweden, was found to have an increased risk of breast cancer.42 Neural system tumors may also associated with CDKN2A mutations that affect p14ARF. These studies, however, are based on very small numbers of patients and families.41
Nonmelanoma cancers can contribute to morbidity and mortality within some families with melanoma. Risk estimates for these cancers are largely unknown, contributing to the difficulty of interpreting a positive test result and counseling those with CDKN2A mutations. A Dutch study estimated the cumulative risk of pancreatic cancer to be 17% by age 75 years in mutation carriers with a specific type of CDKN2A mutation known as p16-Leiden.43 Further studies in different populations with a wide range of mutations and good environmental risk factor assessment are needed to improve knowledge of risk estimates of nonmelanoma cancers in CDKN2A mutation carriers.
CDKN2A Mutation Status Does Not Currently Influence Medical Management
Currently, individuals diagnosed with melanoma or at high risk for melanoma should be followed closely with serial skin examinations. Knowledge of CDKN2A mutation status does not affect clinical management of individuals from familial melanoma pedigrees at this time. Within familial melanoma pedigrees, the absence of detectable CDKN2A mutations does not reduce a family member's risk of developing melanoma to that of the general population, and therefore, would not change clinical management of these individuals. More than 60% of melanoma pedigrees test negative for CDKN2A and CDK4 mutations,6 suggesting the involvement of low penetrance genes as well as currently undiscovered genes in melanoma development or types of mutations in known genes not identified by current tests. For example, certain MC1R variants are considered low-penetrance genes themselves in addition to their effect on CDKN2A penetrance.12 In some populations, individuals within CDKN2A mutation-positive families who test negative for the mutation (considered true negatives) still have an elevated risk of developing melanoma above the general population, and therefore, should be followed closely. In these families, approximately 9% of individuals who develop melanoma do not have the family's mutation.20 Individuals from families with multiple cases of melanoma who test negative may develop a false sense of security about their risk, forgo precautions, and decline screening interventions, all of which could be harmful.6 Therefore, all patients with a positive family history of melanoma, irrespective of mutation status, should be treated as high-risk and enrolled in surveillance and education programs.
There are currently no evidence-based guidelines concerning nonmelanoma cancer screening in CDKN2A mutation-positive individuals. Regarding pancreatic cancer, there are currently no adequate tests available for population-based screening.44 A few small studies have examined the feasibility of screening high-risk individuals, defined by family history of pancreatic cancer, using invasive imaging studies including endoscopic ultrasound and endoscopic retrograde cholangiopancreatography.45,46 These screening methods also have associated risks and potential morbidity. A recent review of familial pancreatic cancer and its treatment reported that even the best screening efforts will not detect all patients with invasive pancreatic cancer, a disease with treatments that do not usually result in long-term cures except for small, lymph node-negative, margin-negative cancers treated with surgical resection.47 Although an algorithm for pancreatic cancer surveillance in CDKN2A mutation-positive patients has been suggested,48 definitive screening guidelines for enrolling CDKN2A mutation-positive individuals are not well established. With improved understanding of the risk of nonmelanoma cancers conferred by CDKN2A mutations and improved screening methods for these cancers, medical management of a mutation carrier might be affected in the future.
CDKN2A Mutation Testing in the Context of Other Familial Cancer Syndromes
Genetic testing for CDKN2A mutations is similar in some respects to testing for other familial cancer syndromes, such as Hereditary Breast/Ovarian Cancer Syndrome (HBOC), but there are important differences. Similar to CDKN2A mutations and melanoma, mutations in BRCA1 (OMIM+113705) and BRCA2 (OMIM+600185) account for a proportion, but not all, of HBOC. However, genetic testing for mutations in BRCA1 and BRCA2 can dramatically change the clinical care of the individual being tested. The National Comprehensive Cancer Network Clinical Practice Guidelines in Oncology include recommendations for early screening with mammography and/or magnetic resonance imaging, prophylactic surgical interventions, and even possibly chemoprevention for BRCA1/BRCA2 carriers.49 Within HBOC families that carry mutations in BRCA1/BRCA2, individuals who test negative for the mutation (true negatives) have a risk of breast and ovarian cancer similar to women from the general population. Thus, these women do not have to pursue the risk-reducing interventions recommended for mutation carriers.49,50 In contrast, individuals from CDKN2A mutation-positive families who test negative still have a risk for melanoma elevated above that of the general population and should not abandon close surveillance of their skin.
Another difference between HBOC and familial melanoma is the main intervention used to reduce cancer risk at this time. It is important to identify BRCA1/BRCA2 carriers within HBOC families so that true negatives are not subject to invasive and life-altering procedures such as bilateral mastectomy and oophorectomy unnecessarily. The main interventions in CDKN2A mutation carriers are education to change sun-related behaviors as well as frequent skin examinations and skin biopsies, which are relatively minor procedures and generally well-tolerated. Finally, patients diagnosed with breast cancer and identified as BRCA1 or BRCA2 mutation carriers might have different treatment courses compared with patients with sporadic breast cancer, such as bilateral mastectomy rather than local resection, or perhaps different systemic therapy. Currently, patients diagnosed with melanoma are treated similarly regardless of CDKN2A mutation status. Therefore, CDKN2A mutation status does not affect clinical management for familial melanoma in the same way that identifying BRCA1 and BRCA2 mutations does for individuals with HBOC.
Management of Individuals at High Risk for Melanoma
All individuals considered at high risk for melanoma should be managed similarly regardless of CDKN2A mutation status. Individuals at increased risk for melanoma should be identified. Family history information should be obtained from all individuals with melanoma or dysplastic nevi, and first-degree relatives should be screened because of the increased risk of CMM in melanoma-prone families with or without CDKN2A mutations.7 These individuals should be educated about sun protection, avoidance of intense sun exposure, and other preventive measures, such as learning how to detect dysplastic nevi characteristics and melanoma warning signs.51
Management recommendations also include examination of the entire skin surface by a skilled healthcare provider every 6 months from 10 years, or earlier for suspicious nevi, until nevi are stable and annually thereafter.6 Frequency of skilled examinations should increase after a recent melanoma diagnosis and during times of active nevus changes, including puberty and pregnancy. Additionally, individuals at high risk should perform monthly total body self-examinations with assistance from a partner or family member, particularly for the scalp and back. Parents need to examine their children's skin. Photography of lesions can be a useful tool for providers and patients to compare lesions over time. Indications for biopsy and removal of a pigmented lesion are the same among individuals with or without a family history of melanoma.6 Prophylactic removal of nevi is not recommended as melanomas can occur on previously normal skin.52 Following high-risk patients clinically with frequent examinations can lead to early detection of thin melanomas, which, when excised with adequate margins, have a high survival rate.6,7
Oncology nurses may be involved in assessing cancer risk in patients, educating patients about genetic testing, and using genetic technologies to improve cancer prevention and treatment strategies.53 Identifying patients at high risk for melanoma due to, for example, multiple cases of melanoma in a family, early onset of melanoma in a family, an excess of nonmelanoma cancers in a melanoma-prone family, or the presence of multiple primary melanomas in an individual, is important. Furthermore, oncology nurses should be aware that 8% of melanomas occur in patients with previous malignancies, particularly melanoma, prostate cancer, breast cancer, and non-Hodgkin lymphoma.54 Although widespread clinical genetic testing for CDKN2A mutations is not recommended, oncology nurses and other providers performing risk assessment counseling have an important role in referring such patients to research studies as well as to dermatologists or other healthcare providers with expertise in melanoma and pigmented lesions. Relevant resources are the Web sites of the International Melanoma Genetics Consortium (http://www.genomel.org) and the National Cancer Institute PDQ Cancer Genetics Services Directory (http://www.cancer.gov/search/geneticsservices).
Conclusion
Genetics, host factors, and environmental exposures contribute to the development of melanoma. Mutations in known melanoma susceptibility genes CDKN2A and CDK4 play an important role in melanoma development especially in the context of familial melanoma, although this role is not completely delineated at this time. Due to the low likelihood of identifying mutations even in high-risk individuals, the limitations on adequate interpretation of a positive test result, and the lack of change in medical management resulting from a test result, predictive genetic testing for CDKN2A mutations outside of research settings is not currently recommended by most research groups. In the future, enhanced knowledge of the genetic, host, and environmental risk factors for melanoma and their interactions might provide a role for clinical genetic testing for melanoma susceptibility genes that could contribute to an improvement in melanoma prevention, detection, and treatment. Until that time, individuals identified at high risk of developing melanoma for any reason should enter surveillance and education programs for melanoma prevention and early detection.
ACKNOWLEDGMENTS
The authors thank Barbara Rogers for her assistance with the tables and figure and June Peters for her review of the manuscript.
References