Advertisement

Histologic and Phenotypic Factors and MC1R Status Associated with BRAFV600E, BRAFV600K, and NRAS Mutations in a Community-Based Sample of 414 Cutaneous Melanomas

Open ArchivePublished:January 22, 2016DOI:https://doi.org/10.1016/j.jid.2015.12.035
      Cutaneous melanomas arise through causal pathways involving interplay between exposure to UV radiation and host factors, resulting in characteristic patterns of driver mutations in BRAF, NRAS, and other genes. To gain clearer insights into the factors contributing to somatic mutation genotypes in melanoma, we collected clinical and epidemiologic data, performed skin examinations, and collected saliva and tumor samples from a community-based series of 414 patients aged 18 to 79, newly diagnosed with cutaneous melanoma. We assessed constitutional DNA for nine common polymorphisms in melanocortin-1 receptor gene (MC1R). Tumor DNA was assessed for somatic mutations in 25 different genes. We observed mutually exclusive mutations in BRAFV600E (26%), BRAFV600K (8%), BRAFother (5%), and NRAS (9%). Compared to patients with BRAF wild-type melanomas, those with BRAFV600E mutants were significantly younger, had more nevi but fewer actinic keratoses, were more likely to report a family history of melanoma, and had tumors that were more likely to harbor neval remnants. BRAFV600K mutations were also associated with high nevus counts. Both BRAFV600K and NRAS mutants were associated with older age but not with high sun exposure. We also found no association between MC1R status and any somatic mutations in this community sample of cutaneous melanomas, contrary to earlier reports.

      Abbreviations:

      CI (confidence interval), MC1R (melanocortin-1 receptor), NRHC (non-red hair color), OR (odds ratio), RHC (red hair color)

      Introduction

      Melanoma is a potentially lethal cancer arising from the pigment cells, melanocytes. Although UV radiation from sunlight is the principal environmental cause for these cancers, there is increasing evidence that the effect of UV radiation on melanocytes is not the same for all people (
      • Whiteman D.C.
      • Pavan W.J.
      • Bastian B.C.
      The melanomas: a synthesis of epidemiological, clinical, histopathological, genetic, and biological aspects, supporting distinct subtypes, causal pathways, and cells of origin.
      ). Epidemiologic observations originally led to the concept that melanomas may arise through one of several pathways under a divergent pathway model for melanoma (
      • Whiteman D.C.
      • Stickley M.
      • Watt P.
      • Hughes M.C.
      • Davis M.B.
      • Green A.C.
      Anatomic site, sun exposure, and risk of cutaneous melanoma.
      ). This model suggested at least two different causal pathways to melanoma development, one pertaining to host susceptibility and nevus prevalence and the other associated with chronic sun exposure. Subsequent investigations strongly suggested that the molecular profile of tumors for several oncogenes, including BRAF and NRAS, reflected these causal pathways (
      • Curtin J.A.
      • Fridlyand J.
      • Kageshita T.
      • Patel H.N.
      • Busam K.J.
      • Kutzner H.
      • et al.
      Distinct sets of genetic alterations in melanoma.
      ,
      • Landi M.T.
      • Bauer J.
      • Pfeiffer R.M.
      • Elder D.E.
      • Hulley B.
      • Minghetti P.
      • et al.
      MC1R germline variants confer risk for BRAF-mutant melanoma.
      ,
      • Thomas N.E.
      • Edmiston S.N.
      • Alexander A.
      • Millikan R.C.
      • Groben P.A.
      • Hao H.
      • et al.
      Number of nevi and early-life ambient UV exposure are associated with BRAF-mutant melanoma.
      ,
      • Whiteman D.C.
      • Stickley M.
      • Watt P.
      • Hughes M.C.
      • Davis M.B.
      • Green A.C.
      Anatomic site, sun exposure, and risk of cutaneous melanoma.
      ). Several studies have now shown that melanomas arising on the trunk tend to occur in younger individuals and are associated with adjacent melanocytic nevi and BRAF mutations, and these appear biologically distinct from melanomas arising on chronically sun-exposed sites, such as the head and neck, which tend to occur in older individuals carrying other mutation profiles, including NRAS mutations (
      • Curtin J.A.
      • Fridlyand J.
      • Kageshita T.
      • Patel H.N.
      • Busam K.J.
      • Kutzner H.
      • et al.
      Distinct sets of genetic alterations in melanoma.
      ,
      • Hacker E.
      • Hayward N.K.
      • Dumenil T.
      • James M.R.
      • Whiteman D.C.
      The association between MC1R genotype and BRAF mutation status in cutaneous melanoma: findings from an Australian population.
      ,
      • Hacker E.
      • Nagore E.
      • Cerroni L.
      • Woods S.L.
      • Hayward N.K.
      • Chapman B.
      • et al.
      NRAS and BRAF mutations in cutaneous melanoma and the association with MC1R genotype: findings from Spanish and Austrian populations.
      ;
      • Lee J.H.
      • Choi J.W.
      • Kim Y.S.
      Frequencies of BRAF and NRAS mutations are different in histological types and sites of origin of cutaneous melanoma: a meta-analysis.
      ,
      • Maldonado J.L.
      • Fridlyand J.
      • Patel H.
      • Jain A.N.
      • Busam K.
      • Kageshita T.
      • et al.
      Determinants of BRAF mutations in primary melanomas.
      ,
      • Thomas N.E.
      • Edmiston S.N.
      • Alexander A.
      • Millikan R.C.
      • Groben P.A.
      • Hao H.
      • et al.
      Number of nevi and early-life ambient UV exposure are associated with BRAF-mutant melanoma.
      ,
      • van Elsas A.
      • Zerp S.F.
      • van der Flier S.
      • Kruse K.M.
      • Aarnoudse C.
      • Hayward N.K.
      • et al.
      Relevance of ultraviolet-induced N-ras oncogene point mutations in development of primary human cutaneous melanoma.
      ,
      • Whiteman D.C.
      • Watt P.
      • Purdie D.M.
      • Hughes M.C.
      • Hayward N.K.
      • Green A.C.
      Melanocytic nevi, solar keratoses, and divergent pathways to cutaneous melanoma.
      ,
      • Whiteman D.C.
      • Stickley M.
      • Watt P.
      • Hughes M.C.
      • Davis M.B.
      • Green A.C.
      Anatomic site, sun exposure, and risk of cutaneous melanoma.
      ). More recent data have emerged suggesting that different genotypes exist within BRAF mutant melanoma, and that melanomas harboring BRAFV600K mutations are associated with older age, male sex, higher levels of sun exposure, and poorer prognosis than BRAF V600E melanomas (
      • Jewell R.
      • Chambers P.
      • Harland M.
      • Laye J.
      • Conway C.
      • Mitra A.
      • et al.
      Clinicopathologic features of V600E and V600K melanoma—letter.
      ,
      • Mar V.J.
      • Liu W.
      • Devitt B.
      • Wong S.Q.
      • Dobrovic A.
      • McArthur G.A.
      • et al.
      The role of BRAF mutations in primary melanoma growth rate and survival.
      ,
      • Menzies A.M.
      • Haydu L.E.
      • Visintin L.
      • Carlino M.S.
      • Howle J.R.
      • Thompson J.F.
      • et al.
      Distinguishing clinicopathologic features of patients with V600E and V600K BRAF-mutant metastatic melanoma.
      ). Thus, there appear to be marked differences in the associations between sun exposure, melanocyte susceptibility, and host characteristics with a suite of melanoma mutations, strongly suggestive of different causal pathways to melanoma development.
      The melanocortin-1 receptor (MC1R) gene is a key determinant of human pigmentation with specific variants linked to red hair and melanoma risk (
      • Palmer J.S.
      • Duffy D.L.
      • Box N.F.
      • Aitken J.F.
      • O'Gorman L.E.
      • Green A.C.
      • et al.
      Melanocortin-1 receptor polymorphisms and risk of melanoma: is the association explained solely by pigmentation phenotype?.
      ,
      • Sturm R.A.
      • Duffy D.L.
      • Box N.F.
      • Newton R.A.
      • Shepherd A.G.
      • Chen W.
      • et al.
      Genetic association and cellular function of MC1R variant alleles in human pigmentation.
      ). An interaction between germline MC1R variants and somatic BRAF mutations was reported in tumors from US and Italian populations (
      • Fargnoli M.C.
      • Pike K.
      • Pfeiffer R.M.
      • Tsang S.
      • Rozenblum E.
      • Munroe D.J.
      • et al.
      MC1R variants increase risk of melanomas harboring BRAF mutations.
      ;
      • Landi M.T.
      • Bauer J.
      • Pfeiffer R.M.
      • Elder D.E.
      • Hulley B.
      • Minghetti P.
      • et al.
      MC1R germline variants confer risk for BRAF-mutant melanoma.
      ), suggesting that people carrying germline MC1R variants had a greater risk of developing a melanoma harboring a BRAF mutation in skin not damaged by sunlight. Analyses of Spanish and Austrian samples found no association between germline MC1R variants and somatic BRAF mutations across all tumor samples, but they did observe a modest trend between germline MC1R status and somatic BRAF mutations in melanomas of the trunk (odds ratio [OR] 1.8 [0.8–4.1], P = 0.1) but an inverse association between MC1R and BRAF for melanomas of the head and neck (OR 0.3 [0.1–0.8], P = 0.02) (
      • Hacker E.
      • Nagore E.
      • Cerroni L.
      • Woods S.L.
      • Hayward N.K.
      • Chapman B.
      • et al.
      NRAS and BRAF mutations in cutaneous melanoma and the association with MC1R genotype: findings from Spanish and Austrian populations.
      ). However, the association between germline MC1R variants and somatic BRAF mutations has not been replicated in other populations, including studies from the United States (
      • Thomas N.E.
      • Kanetsky P.A.
      • Edmiston S.N.
      • Alexander A.
      • Begg C.B.
      • Groben P.A.
      • et al.
      Relationship between germline MC1R variants and BRAF-mutant melanoma in a North Carolina population-based study.
      ), Australia (
      • Hacker E.
      • Hayward N.K.
      • Dumenil T.
      • James M.R.
      • Whiteman D.C.
      The association between MC1R genotype and BRAF mutation status in cutaneous melanoma: findings from an Australian population.
      ), and Germany (
      • Scherer D.
      • Rachakonda P.S.
      • Angelini S.
      • Mehnert F.
      • Sucker A.
      • Egberts F.
      • et al.
      Association between the germline MC1R variants and somatic BRAF/NRAS mutations in melanoma tumors.
      ). Indeed, Scherer and colleagues (2010) observed significantly lower frequencies of somatic BRAF mutations in carriers of MC1R variants. These conflicting findings across different populations underscore the complexity of gene-environment interactions for melanoma.
      Given the emergence of novel therapies targeting somatic mutations in melanoma, coupled with the desire to develop evidence-based primary prevention programs, there is a need to catalog the frequency of mutations in large samples of melanoma patients and to understand the mechanisms through which they arise. Here, we present the findings of an investigation into the epidemiologic, histologic, and genotypic associations with melanoma mutations, comprising a large, community-based sample of 414 primary cutaneous invasive melanoma patients arising in a high-incidence population exposed to very high levels of ambient UV radiation.

      Results

      Subject characteristics

      A total of 766 patients with primary invasive melanoma (32% female, mean age 58 years) were recruited for the parent epidemiologic study (
      • Kvaskoff M.
      • Pandeya N.
      • Green A.C.
      • Perry S.
      • Baxter C.
      • Davis M.B.
      • et al.
      Site-specific determinants of cutaneous melanoma: a case-case comparison of patients with tumors arising on the head or trunk.
      ). The majority of melanomas were classified histologically as superficial spreading melanoma (72%), with the remainder classified as lentigo maligna melanoma (13%), nodular (5%), or unclassified (10%). Tumors were generally thin; 65% were Clark level II, and 82% had Breslow thickness ≤1 mm. The analyses presented here were restricted to 414 patients for whom sufficient material was remaining for somatic mutation analysis (see Supplementary Figure S1 online). There were no significant differences between those genotyped (n = 414) and those who were not (n = 352) in terms of sex (71% vs. 65% males) or melanoma thickness distribution (84% vs. 81% ≤1 mm), but participants not genotyped were slightly older (56.3 years vs. 59.8 years, P = 0.04) and were more likely to have melanomas of the head and neck (9.2% vs. 17.6%, P < 0.001) and of the lentigo maligna subtype (15.7% vs. 24.7%, P = 0.005).

      Mutation frequencies

      Mutations were identified using the MelaCarta multiplex assay (Agena Bioscience, San Diego, CA). Mutually exclusive BRAF mutant and NRAS mutant tumors occurred at frequencies of 38.7% (V600E 67%, V600K 31%, other 12%) and 9.2% (Q61H 5%, Q61K 37%, Q61L 24%, Q61R 34%), respectively (Table 1 and Supplementary Table S1 online). Further statistical analysis was performed for the BRAF and NRAS mutant samples because of the low frequencies of mutations in other genes.
      Table 1Spectrum and frequency of mutations in primary cutaneous melanoma samples
      GeneNo.Frequency (%)
      BRAF16038.7
       V600E10766.9
       V600K3320.6
       Other2012.5
      CDK451.2
      CTNNB110.2
      EPHB6102.4
      ERBB461.5
      GNA1120.5
      GNAQ10.2
      KIT41.0
      KRAS102.4
      MEK30.7
      MET10.2
      NRAS389.2
       Q61H25.3
       Q61K1436.8
       Q61L923.7
       Q61R1334.2
      PDGFRA20.5
      PIK3CA61.5
      PTK2B30.7
      JAK210.2
      ABL120.5
      No mutations were observed in AKT3, CXCR4, EPHA10, NEK10, ROR2, EGFR, IDH1, and ATK1.

      Clinical and pathologic characteristics of lesions

      Overall, patients older than 70 years were significantly less likely to have BRAFV600E mutant melanomas than BRAF wild-type melanomas [OR 0.08, 95% confidence interval (CI) 0.03–0.19] but were more likely to have melanomas harboring BRAFV600K or NRAS mutations (Table 2 and Supplementary Table S2 online). BRAFV600E mutations were significantly more frequent in melanomas from women (P = 0.01), whereas NRAS mutations were more common in melanomas from men (P = 0.01). We observed that BRAFV600E (P = 0.01) and BRAFV600K (P = 0.047) mutant melanomas were more likely to harbor somatic mutations in other genes on the melanoma panel than NRAS mutant melanomas (Table 2 and Supplementary Table S2). Although numbers were small, melanomas carrying somatic mutations in KRAS or EPHB6 were more likely to carry additional mutations than melanomas without these mutations (see Supplementary Table S3 online). Although the prevalence of BRAF and NRAS mutations differed somewhat by histologic subtype and anatomic site, the differences were not statistically significant. We found no statistical evidence that the risks of BRAFV600E or BRAFV600K mutations differed by Clark level or tumor thickness; however, somewhat against expectation, we found melanomas with NRAS mutations were significantly less likely to have marked dermal elastosis (OR 0.26, 95% CI 0.07–0.95, P = 0.03). Melanomas with BRAFV600E mutations were significantly more likely to have contiguous neval remnants than wild-type melanomas (OR 1.94, 95% CI 1.14–3.31, P = 0.02), but melanomas carrying other BRAF or NRAS mutations were not significantly associated with this feature.
      Table 2Association between clinical and pathologic characteristics with BRAF/NRAS mutation status in cutaneous melanoma
      CharacteristicAge- and sex-adjusted odds ratio (95% confidence interval)
      BRAF V600E
      Comparison group were samples wild type for BRAF.
      (n = 107)
      BRAF V600K
      Comparison group were samples wild type for BRAF.
      (n = 33)
      Other BRAF mutation
      Comparison group were samples wild type for BRAF.
      (n = 20)
      Any NRAS mutation
      Comparison group were samples wild type for NRAS.
      (n = 38)
      Age (years)
       <501.00 (reference)1.00 (reference)1.00 (reference)1.00 (reference)
       50–590.28 (0.15–0.51)0.94 (0.27–3.28)0.66 (0.19–2.28)5.81 (1.59–21.22)
       60–690.14 (0.08–0.28)0.90 (0.28–2.92)0.75 (0.25–2.30)6.28 (1.78–22.18)
       ≥700.08 (0.03–0.19)2.24 (0.75–6.66)0.13 (0.02–1.15)3.75 (0.94–14.96)
      Age (continuous)0.93 (0.91–0.95)1.02 (0.99–1.05)0.98 (0.95–1.02)1.03 (1.00–1.05)
      Sex
       Female1.00 (reference)1.00 (reference)1.00 (reference)1.00 (reference)
       Male0.53 (0.33–0.86)1.68 (0.66–4.24)2.11 (0.60–7.44)3.89 (1.35–11.21)
      No. of other somatic mutations
       11.00 (reference)1.00 (reference)1.00 (reference)1.00 (reference)
       >15.94 (1.46–24.21)4.71 (1.02–21.74)1.82 (0.27–12.52)0.56 (0.15–2.07)
      Histologic type of melanoma
      Superficial spreading melanoma1.00 (reference)1.00 (reference)1.00 (reference)1.00 (reference)
      Nodular melanoma0.70 (0.21–2.36)0.38 (0.05–3.03)0.68 (0.08–5.72)2.08 (0.64–6.79)
      Lentigo maligna melanoma0.40 (0.12–1.41)0.67 (0.19–2.40)0.83 (0.27–2.54)
       Not stated (n = 41)
      Clark level
       21.00 (reference)1.00 (reference)1.00 (reference)1.00 (reference)
       ≥30.71 (0.42–1.23)0.70 (0.32–1.53)0.51 (0.18–1.47)3.01 (1.49–6.09)
      Tumor thickness (mm)
       ≤1.01.00 (reference)1.00 (reference)1.00 (reference)1.00 (reference)
       >1.00.74 (0.36–1.53)0.71 (0.26–1.97)0.48 (0.10–2.18)0.67 (0.41–2.67)
      Anatomic site of melanoma
       Trunk1.00 (reference)1.00 (reference)1.00 (reference)1.00 (reference)
       Head or neck1.00 (0.53–1.90)1.64 (0.72–3.72)0.44 (0.10–1.98)0.61 (0.24–1.54)
      Dermal elastosis
       Nil or mild1.00 (reference)1.00 (reference)1.00 (reference)1.00 (reference)
       Moderate0.84 (0.41–1.73)2.49 (0.90–6.97)0.30 (0.06–1.44)1.09 (0.47–2.52)
       Marked0.95 (0.44–2.05)1.53 (0.50–4.67)0.37 (0.08–1.82)0.26 (0.07–0.95)
       Missing (n = 50)
      Contiguous neval remnants
       No1.00 (reference)1.00 (reference)1.00 (reference)1.00 (reference)
       Yes1.94 (1.14–3.31)1.20 (0.54–2.65)1.47 (0.54–3.94)1.52 (0.74–3.10)
       Not stated (n = 9)
      1 Comparison group were samples wild type for BRAF.
      2 Comparison group were samples wild type for NRAS.

      Phenotypic and environmental factors associated with BRAF and NRAS mutations

      We observed strong positive associations between increasing nevus count and risk of BRAFV600E (p-trend = 0.03) or BRAFV600K (p-trend = 0.02) mutations (Table 3 and Supplementary Table S4 online), but no associations with NRAS mutations. In contrast, there were inverse associations between the numbers of excised skin cancers and BRAFV600E mutational status (p-trend = 0.04). The numbers of skin cancers were also inversely associated with BRAFV600K mutations, although the trend was of marginal significance (p-trend = 0.06). The measure of cumulative sun exposure (summed from a matrix capturing recreational and occupational sun exposure for all career episodes since leaving high school) showed an unusual pattern of association with BRAFV600E and BRAFV600K mutations. Although not statistically significant, relative risks of BRAF mutant melanoma were higher for patients with intermediate categories of cumulative sun exposure than for those with the highest levels of sun exposure. No consistent associations between markers of cumulative sun exposure and risks of NRAS mutant melanoma were observed, although it was notable that risk estimates were less than unity for all categories of solar keratosis counts and for having three or more skin cancers excised (Table 3). There were no consistent associations between hair or eye color and risks of any type of BRAF mutations or NRAS mutations. However, patients with BRAFV600K mutant melanomas were significantly more likely to have blue/gray eye color (OR 2.38, 95% CI 1.00–5.65, P = 0.049). In addition, patients with BRAFV600E mutant melanomas were significantly less likely (p-trend = 0.02) to report having any extent of facial freckling as a teenager compared with patients with BRAF wild-type melanomas; associations between freckling and other mutation types were not significant. A family history of melanoma was associated with BRAFV600E mutation status (OR 1.85, 95% CI 1.06–3.21, P = 0.03). Other characteristics were assessed for associations with BRAF/NRAS mutational status, but in the main these were unremarkable (see Supplementary Table S5 online).
      Table 3Association between phenotypic and environmental factors with BRAF/NRAS mutation status in cutaneous melanoma
      CharacteristicAge- and sex-adjusted odds ratio (95% confidence interval)
      BRAF V600E
      Comparison group were samples wild type for BRAF.
      (n = 107)
      BRAF V600K
      Comparison group were samples wild type for BRAF.
      (n = 33)
      Other BRAF mutation
      Comparison group were samples wild type for BRAF.
      (n = 20)
      Any NRAS mutation
      Comparison group were samples wild-type for NRAS.
      (n = 38)
      Total nevus count (quartiles)
       0–291.00 (reference)1.00 (reference)1.00 (reference)1.00 (reference)
       30–591.80 (0.72–4.48)2.25 (0.73–6.96)0.45 (.10–2.05)0.54 (0.20–1.47)
       60–1192.26 (0.91–5.59)2.36 (0.67–8.34)1.04 (0.29–3.76)0.73 (0.27–1.97)
       120+2.90 (1.16–7.29)5.03 (1.38–18.38)0.74 (0.18–3.15)0.93 (0.33–2.62)
      Total no. of solar keratoses
       01.00 (reference)1.00 (reference)1.00 (reference)1.00 (reference)
       1–40.64 (0.33–1.26)1.00 (0.31–3.25)1.61 (0.40–6.69)0.44 (0.17–1.13)
       5–90.52 (0.19–1.40)0.43 (0.08–2.44)0.51 (0.05–5.22)0.51 (0.16–1.63)
       10+0.45 (0.18–1.12)1.43 (0.43–4.80)2.57 (0.57–11.58)0.45 (0.17–1.21)
      Skin cancers excised
       01.00 (reference)1.00 (reference)1.00 (reference)1.00 (reference)
       1–20.68 (0.34–1.35)0.65 (0.37–1.15)1.01 (0.32–3.17)1.33 (0.57–3.14)
       3+0.48 (0.24–0.97)0.45 (0.26–0.79)0.35 (0.09–1.35)0.76 (0.33–1.79)
      Cumulative sun exposure (adult years)
       <1.61.00 (reference)1.00 (reference)1.00 (reference)1.00 (reference)
       1.6–2.81.98 (0.93–4.20)2.23 (0.63–7.85)1.59 (0.36–6.96)0.98 (0.31–3.11)
       2.8–4.52.17 (0.95–4.97)2.01 (0.55–7.27)1.89 (0.45–7.97)0.70 (0.21–2.35)
       >4.51.74 (0.67–4.48)1.04 (0.25–4.28)1.10 (0.21–5.80)1.42 (0.45–4.44)
      Hair color as a teenager
       Black/dark brown1.00 (reference)1.00 (reference)1.00 (reference)1.00 (reference)
       Light brown0.95 (0.50–1.80)1.88 (0.81–4.35)1.23 (0.31–4.82)1.03 (0.47–2.27)
       Red/auburn0.90 (0.40–2.03)3.46 (0.89–13.48)0.68 (0.21–2.22)
       Blond1.69 (0.85–3.37)1.23 (0.39–3.92)2.44 (0.61–9.83)0.65 (0.22–1.93)
      Eye color
       Brown1.00 (reference)1.00 (reference)1.00 (reference)1.00 (reference)
       Blue/gray1.33 (0.70–2.54)2.38 (1.00–5.65)1.05 (0.28–3.98)0.75 (0.27–2.09)
       Green/hazel0.74 (0.37–1.48)0.55 (0.15–1.94)0.66 (0.18–2.45)1.24 (0.52–2.96)
      Freckling as a teenager
       None1.00 (reference)1.00 (reference)1.00 (reference)1.00 (reference)
       Few0.50 (0.27–0.93)0.69 (0.29–1.65)1.49 (0.47–4.75)0.94 (0.42–2.06)
       Some0.47 (0.22–1.03)0.79 (0.26–2.45)1.06 (0.23–4.88)1.83 (0.72–4.65)
       Many0.34 (0.12–0.97)0.83 (0.21–3.31)2.56 (0.49–12.41)
      Family history of melanoma
       No1.00 (reference)1.00 (reference)1.00 (reference)1.00 (reference)
       Yes1.85 (1.06–3.21)1.24 (0.51–3.03)0.56 (0.17–1.84)0.44 (0.16–1.23)
      1 Comparison group were samples wild type for BRAF.
      2 Comparison group were samples wild-type for NRAS.

      Somatic mutations and MC1R variants

      In this series, 84% of melanoma patients carried one of the nine common MC1R variants, with 53% carrying red hair color (RHC) variants and 31% carrying non-red hair color (NRHC) variants (see Supplementary Tables S6 and S7 online). As expected, MC1R status was associated with red hair color (P < 0.001) and skin type determinants such as susceptibility to burn (P < 0.001) and propensity to tan (P < 0.001) and was inversely associated with nevus counts (P = 0.02; Table 4). There was no association between germline MC1R variants and somatic BRAF or NRAS mutations in melanoma samples overall (Table 5 and Supplementary Table S8 online). In site-specific analyses (trunk melanoma; head and neck melanoma), we found no evidence that the risks of BRAF mutations were associated with MC1R variants, regardless of the type of variant. We repeated the analyses by excluding patients with lentigo maligna melanoma subtype, but this made no material difference to our conclusions (see Supplementary Table S9 online). We also observed no association between the number of MC1R polymorphisms and either BRAF (P = 0.38) or NRAS (P = 0.83) mutation status.
      Table 4Association between MC1R and phenotypic characteristics
      Any MC1R variant (n = 637)
      MC1R status missing for 9 samples cohort = 766.
      RHC (n = 399)NRHC (n = 238)Wild type (n = 120)
      Hair color as a teenager
       Black/dark brown194 (30.5)99 (24.8)95 (39.9)54 (45.0)
       Light brown209 (32.8)120 (30.1)89 (37.4)40 (30.3)
       Red/auburn97 (15.2)91 (22.8)6 (2.5)4 (3.3)
       Blond136 (21.4)88 (22.1)48 (20.2)22 (18.3)
       Missing (n = 1)
      P-value0.001<0.0010.76Reference
      Eye color
       Brown411 (64.5)255 (63.9)156 (65.6)68 (56.7)
       Blue/grey136 (21.4)82 (20.6)54 (22.7)28 (23.3)
       Green/hazel90 (14.1)62 (15.5)28 (11.8)24 (20.0)
      P-value0.180.330.09Reference
      Freckling as a teenager
       None201 (31.6)98 (24.6)103 (43.3)65 (54.2)
       A few256 (40.2)165 (41.5)91 (38.2)37 (30.8)
       Some118 (18.5)83 (20.8)35 (14.7)12 (10.0)
       Many62 (9.7)53 (13.3)9 (3.8)6 (5.0)
      P-value<0.001<0.0010.18Reference
      Total nevus count (quartiles)
       0–29152 (23.9)105 (26.3)47 (19.8)14 (11.7)
       30–59169 (26.5)114 (28.6)55 (23.1)33 (27.5)
       60–119145 (22.8)81 (20.3)64 (26.9)34 (28.3)
       120+171 (26.8)99 (24.8)72 (30.3)39 (32.5)
      P-value0.020.0040.28Reference
      Propensity to burn
       Never/Rarely46 (7.2)19 (4.8)27 (11.3)20 (16.7)
       Sometimes174 (27.3)102 (25.6)72 (30.3)49 (40.8)
       Mostly187 (29.4)128 (32.1)59 (24.8)29 (24.2)
       Always230 (36.1)150 (37.6)80 (33.6)22 (18.3)
      P-value<0.001<0.0010.01Reference
      Propensity to tan
       Never74 (11.6)63 (15.8)11 (4.6)5 (4.2)
       Lightly199 (31.2)137 (34.4)62 (26.1)18 (15.0)
       Moderately283 (44.4)166 (41.6)117 (49.2)61 (50.8)
       Deeply81 (12.7)33 (8.3)48 (20.2)36 (30.0)
      P-value<0.001<0.0010.05Reference
      Abbreviations: NRHC, non-red hair color (MC1R variants: V60L, V92M, R163Q, I155T); RHC, red hair color (MC1R variants: R142H, D84E, R151C, R160W, D294H).
      Values are given as no. (%).
      All P-values are from multivariable logistic regression models adjusted for age and sex.
      1 MC1R status missing for 9 samples cohort = 766.
      Table 5Association between MC1R variants and somatic BRAF and NRAS mutations in cutaneous melanoma
      MC1RAge- and sex-adjusted odds ratio (95% confidence interval)Any NRAS mutation (n = 38)
      BRAF V600E (n = 107)BRAF V600K (n = 33)Any BRAF mutation (n = 160)
      All melanomas
       WT/WT1.00 (reference)1.00 (reference)1.00 (reference)1.00 (reference)
       Any variant1.13 (0.58–2.20)0.80 (0.32–2.02)1.32 (0.75–2.32)1.13 (0.47–2.74)
       RHC variant1.08 (0.53–2.20)0.49 (0.17–1.43)1.17 (0.64–2.14)1.01 (0.39–2.59)
       NRHC variant1.23 (0.58–2.59)1.30 (0.48–3.55)1.53 (0.82–2.86)1.37 (0.56–3.40)
      Trunk melanomas
       WT/WT1.00 (reference)1.00 (reference)1.00 (reference)NA
       Any variant1.38 (0.65–2.92)0.68 (0.25–1.88)1.46 (0.77–2.77)NA
       RHC variant1.20 (0.53–2.70)0.42 (0.13–1.41)1.24 (0.63–2.47)NA
       NRHC variant1.67 (0.72–3.86)1.14 (0.38–3.39)1.80 (0.89–3.65)NA
      Head and neck melanomas
       WT/WT1.00 (reference)1.00 (reference)1.00 (reference)NA
       Any variant0.40 (0.08–1.94)1.88 (0.18–19.42)0.87 (0.24–3.13)NA
       RHC variant0.51 (0.09–2.85)0.94 (0.08–11.08)0.87 (0.22–3.42)NA
       NRHC variant0.16 (0.01–2.13)5.26 (0.48–58.0)0.61 (0.10–3.63)NA
      Abbreviations: NA, not applicable; NRHC, non-red hair color (MC1R variants: V60L, V92M, R163Q, I155T); RHC, red hair color (MC1R variants: R142H, D84E, R151C, R160W, D294H); WT, wild type.

      Discussion

      We assessed the frequency of somatic mutations in 25 putative melanoma genes in a large community-based series of 414 primary cutaneous invasive melanomas. Mutation prevalences were <2.5% for all genes except BRAF and NRAS, which occurred mutually exclusively at frequencies of 39% and 9%, respectively. Previous community-based series of primary cutaneous melanomas have reported BRAF mutations at frequencies ranging from 32% to 39% in Australian populations, 20–22% in Spanish, Austrian, and German populations, 44% and 64% in Italian populations, and 43% in the United States (
      • Fargnoli M.C.
      • Pike K.
      • Pfeiffer R.M.
      • Tsang S.
      • Rozenblum E.
      • Munroe D.J.
      • et al.
      MC1R variants increase risk of melanomas harboring BRAF mutations.
      ;
      • Hacker E.
      • Hayward N.K.
      • Dumenil T.
      • James M.R.
      • Whiteman D.C.
      The association between MC1R genotype and BRAF mutation status in cutaneous melanoma: findings from an Australian population.
      ,
      • Hacker E.
      • Nagore E.
      • Cerroni L.
      • Woods S.L.
      • Hayward N.K.
      • Chapman B.
      • et al.
      NRAS and BRAF mutations in cutaneous melanoma and the association with MC1R genotype: findings from Spanish and Austrian populations.
      ;
      • Scherer D.
      • Rachakonda P.S.
      • Angelini S.
      • Mehnert F.
      • Sucker A.
      • Egberts F.
      • et al.
      Association between the germline MC1R variants and somatic BRAF/NRAS mutations in melanoma tumors.
      ,
      • Thomas N.E.
      • Edmiston S.N.
      • Alexander A.
      • Millikan R.C.
      • Groben P.A.
      • Hao H.
      • et al.
      Number of nevi and early-life ambient UV exposure are associated with BRAF-mutant melanoma.
      ). In all prior series, BRAFV600E mutations were at least three- to fourfold more common than BRAFV600K mutations, as we found. Importantly, we found that BRAF mutant melanomas were significantly more likely than wild-type BRAF melanomas to carry mutations in other genes on the MelaCarta panel (Agena Bioscience), whereas NRAS mutant melanomas were not.
      As expected, we found that the somatic mutation status of melanomas was correlated with a number of clinical and phenotypic features. BRAFV600E mutant melanomas were more likely than wild-type BRAF melanomas among women, younger patients, and those with high nevus counts, contiguous neval remnants adjacent to the tumor, and a family history of melanoma. These findings are in accordance with previous studies examining BRAF status and characteristics of patients with cutaneous melanoma (
      • Fargnoli M.C.
      • Pike K.
      • Pfeiffer R.M.
      • Tsang S.
      • Rozenblum E.
      • Munroe D.J.
      • et al.
      MC1R variants increase risk of melanomas harboring BRAF mutations.
      ;
      • Thomas N.E.
      • Edmiston S.N.
      • Alexander A.
      • Millikan R.C.
      • Groben P.A.
      • Hao H.
      • et al.
      Number of nevi and early-life ambient UV exposure are associated with BRAF-mutant melanoma.
      ). In addition, patients with BRAFV600E tumors were less likely than those with wild-type BRAF tumors to have phenotypic features indicative of high cumulative sun exposure, such as high numbers of actinic keratoses or a history of prior skin cancer excisions.
      Earlier reports have suggested that melanomas carrying BRAFV600K mutations have been exposed to higher levels of cumulative sun exposure than other melanomas, but we found no evidence supporting that conclusion (
      • Jewell R.
      • Chambers P.
      • Harland M.
      • Laye J.
      • Conway C.
      • Mitra A.
      • et al.
      Clinicopathologic features of V600E and V600K melanoma—letter.
      ,
      • Mar V.J.
      • Liu W.
      • Devitt B.
      • Wong S.Q.
      • Dobrovic A.
      • McArthur G.A.
      • et al.
      The role of BRAF mutations in primary melanoma growth rate and survival.
      ,
      • Menzies A.M.
      • Haydu L.E.
      • Visintin L.
      • Carlino M.S.
      • Howle J.R.
      • Thompson J.F.
      • et al.
      Distinguishing clinicopathologic features of patients with V600E and V600K BRAF-mutant metastatic melanoma.
      ). In our large series, patients with BRAFV600K melanomas were significantly less likely than patients with wild-type BRAF melanomas to report prior history of non-melanoma skin cancer and were not significantly different in terms of self-reported lifetime sun exposure, numbers of actinic keratoses, or dermal elastosis adjacent to the melanoma. Even though the series reported here is the largest and most comprehensively annotated to date, the number of cases with BRAFV600K mutations was still modest (n = 33), so our study suffers from a lack of statistical power to explore these associations fully. Pooling data from comparable studies to increase the sample size would permit more definitive assessments of the role of cumulative sun exposure in the development of BRAFV600K melanomas. We note with interest, however, that BRAFV600K melanomas were even more strongly associated with total nevus count than BRAFV600E melanomas, providing strong evidence that these tumors arise through a nevus-prone pathway.
      A synergistic relationship between germline MC1R variants and somatic BRAF mutations was suggested by Landi and colleagues (2006), in which people with MC1R variant genotypes carried a significantly increased risk of developing BRAF mutant melanoma in skin not damaged by sunlight. Analyses of Spanish and Austrian samples found a modest trend between germline MC1R status and somatic BRAF mutations in melanomas from trunk sites with an inverse association between MC1R and BRAF for melanomas of the head and neck (
      • Hacker E.
      • Nagore E.
      • Cerroni L.
      • Woods S.L.
      • Hayward N.K.
      • Chapman B.
      • et al.
      NRAS and BRAF mutations in cutaneous melanoma and the association with MC1R genotype: findings from Spanish and Austrian populations.
      ). Other studies conducted in North Carolina (
      • Thomas N.E.
      • Kanetsky P.A.
      • Edmiston S.N.
      • Alexander A.
      • Begg C.B.
      • Groben P.A.
      • et al.
      Relationship between germline MC1R variants and BRAF-mutant melanoma in a North Carolina population-based study.
      ), Australia (
      • Hacker E.
      • Hayward N.K.
      • Dumenil T.
      • James M.R.
      • Whiteman D.C.
      The association between MC1R genotype and BRAF mutation status in cutaneous melanoma: findings from an Australian population.
      ), and Germany (
      • Scherer D.
      • Rachakonda P.S.
      • Angelini S.
      • Mehnert F.
      • Sucker A.
      • Egberts F.
      • et al.
      Association between the germline MC1R variants and somatic BRAF/NRAS mutations in melanoma tumors.
      ) have not observed associations between MC1R status and increased risk of somatic BRAF mutations. This latest investigation, comprising a community-based sample of 414 patients with cutaneous melanoma of predominantly Northern European and Anglo-Celtic ancestry exposed to high levels of ambient UV radiation, also found no association between germline MC1R variants and somatic BRAF mutations. These conflicting findings across different populations highlight the complexity of gene-environment interactions in the development of melanoma. The model proposed by
      • Thomas N.E.
      • Kanetsky P.A.
      • Begg C.B.
      • Conway K.
      • Berwick M.
      Melanoma molecular subtypes: unifying and paradoxical results.
      to explain this discordance illustrated opposing effects of MC1R status and highlighted a role for pigmentation in photoprotection and generation of oxidative stress. The allele frequencies of seven common nonsynonymous MC1R variants (V60L, D84E, V92M, R151C, R160W, R163Q, D294H) differ significantly between Northern European (France, Netherlands, Britain/Ireland) and Southern European populations (Italy and Greece) (
      • Gerstenblith M.R.
      • Goldstein A.M.
      • Fargnoli M.C.
      • Peris K.
      • Landi M.T.
      Comprehensive evaluation of allele frequency differences of MC1R variants across populations.
      ). We also observed this difference in our cancer cohorts, with >60% of the Australian, US, and German cohorts carrying one of the MC1R RHC variants, whereas the Spanish, Austrian, and Italian cohorts carrying MC1R RHC variants only accounted for <45% (Supplementary Table S6). The downstream effects of MC1R on cellular function appears to vary depending on the polymorphisms, thus it is possible that the discordance between studies could relate to the variation in MC1R allele frequencies in the different populations as well as the differences in environmental conditions and patterns of UV radiation exposure.
      Melanoma risk is intricately associated with pigmentation characteristics, and genomewide association studies have revealed a number of genetic variants involved in pigmentation, including MC1R, ASIP, OCA2, SLC45A2, TYRP1, and TYR (
      • Bishop D.T.
      • Demenais F.
      • Iles M.M.
      • Harland M.
      • Taylor J.C.
      • Corda E.
      • et al.
      Genome-wide association study identifies three loci associated with melanoma risk.
      ,
      • Duffy D.L.
      • Zhao Z.Z.
      • Sturm R.A.
      • Hayward N.K.
      • Martin N.G.
      • Montgomery G.W.
      Multiple pigmentation gene polymorphisms account for a substantial proportion of risk of cutaneous malignant melanoma.
      ). The discordant results across studies examining solely MC1R status as a determinant for developing somatic BRAF mutant melanoma may also be due to the confounding role of other pigmentation genes. It must also be noted that given the relatively small sample size of all studies examining the association of MC1R variants and BRAF mutant melanoma, we cannot rule out the possibility that the differences in results are attributable to chance alone. To expand this work, our future focus needs to be on modeling the complex regulation of pigmentation as a factor of genetic interactions and through larger studies or meta-analyses.
      Strengths of our study include the population-based sampling frame and the detailed epidemiologic data (including physician counts of nevi and actinic keratoses, blinded to genotype status) accompanying the tumor specimens. The call rate for somatic mutations was high using the MelaCarta platform (Agena Bioscience), with mutation status determined for 98% of samples genotyped. Although we did not fully sequence the entire MC1R gene, the variants genotyped in this study comprise >95% of the nonsynonymous changes observed (
      • Kanetsky P.A.
      • Rebbeck T.R.
      • Hummer A.J.
      • Panossian S.
      • Armstrong B.K.
      • Kricker A.
      • et al.
      Population-based study of natural variation in the melanocortin-1 receptor gene and melanoma.
      ). We do not believe that further sequencing to identify rare MC1R variants could materially alter our null findings. A potential weakness was the relatively limited number of samples for analysis because of insufficient tumor material remaining for mutation analysis after sections had been cut for diagnostic purposes. This is to be expected from a community-based study conducted in Queensland, Australia, where the majority of patients present with thin melanomas (<1 mm). To assess possible selection bias, we compared the prevalence of phenotypic (including skin type, hair and eye color, freckling density, and counts of nevi and actinic keratosis) and histologic (contiguous neval remnants, thickness, anatomic site) characteristics as well as the age and sex among those participants with and those without tumor blocks available for analysis. We found that the participants whose tumors were not genotyped were slightly older, were more likely to have melanomas of the head and neck, and were more likely to be of the lentigo maligna subtype than those who were genotyped, but in other respects they were not significantly different. Given these features, it is possible that our sample had a higher prevalence of BRAF mutations than melanomas arising in the general population, although there is no reason to conclude that the associations between BRAF mutation statues and phenotype or other factors would differ.
      In conclusion, these data from a large, well-characterized, community-based sample of cutaneous melanomas provide robust estimates of the somatic mutation frequencies of putative melanoma genes. The study confirmed that BRAF mutant melanomas differ from wild-type melanomas with regard to associations with sun exposure, nevus propensity, and host characteristics, with largely similar patterns of association for BRAFV600E and BRAFV600K melanomas. There was no evidence that MC1R status conferred particular risks of mutations in BRAF, NRAS, or other genes. Taken together, these findings highlight the diversity of mutation profiles in melanoma and the heterogeneity of pathways through which these cancers arise.

      Materials and Methods

      Subjects

      We compared the prevalence of BRAF and NRAS mutations in formalin-fixed, paraffin-embedded melanoma specimens from 414 patients ascertained from southern Queensland (latitude 27 degrees south), Australia. Detailed descriptions of subject selection and data collection for this study have been described previously (
      • Kvaskoff M.
      • Pandeya N.
      • Green A.C.
      • Perry S.
      • Baxter C.
      • Davis M.B.
      • et al.
      Site-specific determinants of cutaneous melanoma: a case-case comparison of patients with tumors arising on the head or trunk.
      ,
      • Kvaskoff M.
      • Pandeya N.
      • Green A.C.
      • Perry S.
      • Baxter C.
      • Davis M.B.
      • et al.
      Solar elastosis and cutaneous melanoma: a site-specific analysis.
      ). Briefly, eligible patients were residents of greater Brisbane, Australia, who were diagnosed between April 1, 2007, and September 30, 2010, with a histologically confirmed primary invasive cutaneous melanoma arising on the head, neck, or trunk. Those with metastatic melanoma or a previous diagnosis of melanoma were not eligible. No acral lentiginous melanoma, spitzoid, or nevoid lesions were included in this study. Of 1,456 eligible patients for the initial epidemiologic study, 808 (55%) completed questionnaires, 766 (53%) provided written informed consent to obtain specimens of archived melanoma tissue, and 414 (28%) patients had sufficient tissue remaining for mutation analysis (Supplementary Figure S1). The age, sex, site, and histology subtype distribution of the 414 patients who were genotyped differed from the 352 patients who were not, as described earlier.
      Approval to perform this study was given by the Human Research Ethics Committee of the QIMR Berghofer Medical Research Institute. The study adhered to the Declaration of Helsinki, and all participants gave written informed consent to take part in the study.

      Histologic assessment

      At the time of histologic diagnosis, collaborating dermatopathologists assessed the extent of solar elastosis in the skin adjacent to the melanoma using a scale of four categories (nil, mild, moderate, and marked) as previously described (
      • Kvaskoff M.
      • Pandeya N.
      • Green A.C.
      • Perry S.
      • Baxter C.
      • Davis M.B.
      • et al.
      Site-specific determinants of cutaneous melanoma: a case-case comparison of patients with tumors arising on the head or trunk.
      ,
      • Kvaskoff M.
      • Pandeya N.
      • Green A.C.
      • Perry S.
      • Baxter C.
      • Davis M.B.
      • et al.
      Solar elastosis and cutaneous melanoma: a site-specific analysis.
      ). In addition, they assessed each tumor’s histologic type, tumor thickness, and presence of neval remnants adjacent to the tumor. The anatomic site of each melanoma was abstracted from the pathology report and was confirmed directly with the patient.

      DNA isolation

      Hematoxylin and eosin-stained sections of each patient’s melanoma were assessed for areas of normal and tumor tissue, and the percentage of tumor cells was recorded. Formalin-fixed, paraffin-embedded tissue sections were dissected to select areas in which melanoma cells dominated over stromal cells. Punch biopsies (2 mm) were taken from each tumor block, deparaffinized in xylene, and washed twice in absolute ethanol. DNA was isolated using Qiagen GeneRead DNA FFPE Kit (Qiagen, Hilden, Germany), with additional proteinase K digestion at 56°C for 3 hours. DNA quantification was determined by spectrophotometry Qubit (Life Technologies, Carlsbad, CA). Saliva samples were also collected, and DNA was extracted for MC1R genotyping from saliva samples using Oragene saliva kits (DNA Genotek, Ottawa, Ontario, Canada) following the manufacturer’s instructions.

      Genotyping

      Genotyping was performed on the mass spectrometric genotyping platform using an optimized multiplex assay of 25 common mutations found in melanomas (MelaCarta Panel, Agena Bioscience), which includes AKT3, BRAF, CDK4, CXCR4, CTNNB1, EPHA10, EPHB6, ERBB4, GNA1, GNAQ, KIT, KRAS, MEK, MET, NEK10, NRAS, PDGFRA, PIK3CA, PTK2B, ROR2, EGFR, IDH1, JAK2, ATK1, and ABL1. An optimized multiplex assay of all nine common variants of MC1R (I155T, R142H, D84E, R160W, D294H, V92M, R163Q, V60L, R151C) were used as previously described (
      • Duffy D.L.
      • Box N.F.
      • Chen W.
      • Palmer J.S.
      • Montgomery G.W.
      • James M.R.
      • et al.
      Interactive effects of MC1R and OCA2 on melanoma risk phenotypes.
      ). Participants with none of the MC1R variants listed were classified as wild type (WT) for these analyses. People carrying one or more of the RHC alleles (R142H, D84E, R160W, D294H, R151C) were classified as RHC variants, and people carrying one or more of the NRHC alleles (I155T, V92M, R163Q, V60L) were classified as NRHC variants (Supplementary Tables S6 and S7). People carrying both RHC and NRHC alleles were classified as RHC variants (Supplementary Table S7).

      Phenotypic characteristics and sun exposure history

      Relevant exposure data (including sun exposure history and skin sensitivity) were collected from study participants through a self-completed, structured questionnaire as described previously (
      • Kvaskoff M.
      • Pandeya N.
      • Green A.C.
      • Perry S.
      • Baxter C.
      • Davis M.B.
      • et al.
      Solar elastosis and cutaneous melanoma: a site-specific analysis.
      ). After completing the questionnaire, each participant was examined by the same dermatologist, who recorded hair and eye color and counted the number of melanocytic nevi (defined as brown to black pigmented macules or papules of any size that are darker than the surrounding skin). Using a standard international protocol (

      English DR, MacLennan R, Rivers J, Kelly J, Armstrong BK. Epidemiological studies of melanocytic nevi: protocol for identifying and recording nevi. IARC internal report no90/002. Lyon, France: International Agency for Research on Cancer; 1990.

      ), nevi were counted on the back, neck, face, and upper limbs (left and right) using a transparent plastic stencil. The numbers of actinic keratoses (defined as superficial, rough scaly areas with erythematous background and ill-defined margins) were counted on the dorsum of hands and forearms, and on the face.

      Statistical analysis

      We performed simple cross-tabulations and calculated Pearson χ2 and/or Fisher exact test (for cells with expected count <5) as a measure of statistical association. We used multivariable logistic regression to calculate ORs and 95% CIs as measures of association between patient/tumor characteristics and mutation status. We included terms for age stratum (<40, 40–49, 50–59, 60–69, 70+ years) and sex to control for possible confounding introduced by the study design. P < 0.05 was considered statistically significant, and all such tests were two sided. We tested for trend by including each category as an ordinal variable in the multivariable model, with category values taken as the midpoint of the range. All analyses were performed using the SAS 9.4 statistical software package (SAS Institute, Cary, NC).

      Conflict of Interest

      The authors state no conflict of interest.

      Acknowledgments

      The authors would like to thank all participants for their time and generosity, as well as Sullivan and Nicolaides Pathology, Queensland Medical Laboratory and IQ Pathology for their support and assistance in performing this study. This work was funded by a Project Grant (#442960) and a Program Grant (#552429). AG, NH, and DW are supported by fellowships from the National Health and Medical Research Council (NHMRC) of Australia.

      Supplementary Material

      References

        • Bishop D.T.
        • Demenais F.
        • Iles M.M.
        • Harland M.
        • Taylor J.C.
        • Corda E.
        • et al.
        Genome-wide association study identifies three loci associated with melanoma risk.
        Nat Genet. 2009; 41: 920-925
        • Curtin J.A.
        • Fridlyand J.
        • Kageshita T.
        • Patel H.N.
        • Busam K.J.
        • Kutzner H.
        • et al.
        Distinct sets of genetic alterations in melanoma.
        N Engl J Med. 2005; 353: 2135-2147
        • Duffy D.L.
        • Box N.F.
        • Chen W.
        • Palmer J.S.
        • Montgomery G.W.
        • James M.R.
        • et al.
        Interactive effects of MC1R and OCA2 on melanoma risk phenotypes.
        Hum Mol Genet. 2004; 13: 447-461
        • Duffy D.L.
        • Zhao Z.Z.
        • Sturm R.A.
        • Hayward N.K.
        • Martin N.G.
        • Montgomery G.W.
        Multiple pigmentation gene polymorphisms account for a substantial proportion of risk of cutaneous malignant melanoma.
        J Invest Dermatol. 2010; 130: 520-528
      1. English DR, MacLennan R, Rivers J, Kelly J, Armstrong BK. Epidemiological studies of melanocytic nevi: protocol for identifying and recording nevi. IARC internal report no90/002. Lyon, France: International Agency for Research on Cancer; 1990.

        • Fargnoli M.C.
        • Pike K.
        • Pfeiffer R.M.
        • Tsang S.
        • Rozenblum E.
        • Munroe D.J.
        • et al.
        MC1R variants increase risk of melanomas harboring BRAF mutations.
        J Invest Dermatol. 2008; 128: 2485-2490
        • Gerstenblith M.R.
        • Goldstein A.M.
        • Fargnoli M.C.
        • Peris K.
        • Landi M.T.
        Comprehensive evaluation of allele frequency differences of MC1R variants across populations.
        Hum Mutat. 2007; 28: 495-505
        • Hacker E.
        • Hayward N.K.
        • Dumenil T.
        • James M.R.
        • Whiteman D.C.
        The association between MC1R genotype and BRAF mutation status in cutaneous melanoma: findings from an Australian population.
        J Invest Dermatol. 2010; 130: 241-248
        • Hacker E.
        • Nagore E.
        • Cerroni L.
        • Woods S.L.
        • Hayward N.K.
        • Chapman B.
        • et al.
        NRAS and BRAF mutations in cutaneous melanoma and the association with MC1R genotype: findings from Spanish and Austrian populations.
        J Invest Dermatol. 2013; 133: 1027-1033
        • Jewell R.
        • Chambers P.
        • Harland M.
        • Laye J.
        • Conway C.
        • Mitra A.
        • et al.
        Clinicopathologic features of V600E and V600K melanoma—letter.
        Clin Cancer Res. 2012; 18: 6792
        • Kanetsky P.A.
        • Rebbeck T.R.
        • Hummer A.J.
        • Panossian S.
        • Armstrong B.K.
        • Kricker A.
        • et al.
        Population-based study of natural variation in the melanocortin-1 receptor gene and melanoma.
        Cancer Res. 2006; 66: 9330-9337
        • Kvaskoff M.
        • Pandeya N.
        • Green A.C.
        • Perry S.
        • Baxter C.
        • Davis M.B.
        • et al.
        Site-specific determinants of cutaneous melanoma: a case-case comparison of patients with tumors arising on the head or trunk.
        Cancer Epidemiol Biomarkers Prev. 2013; 22: 2222-2231
        • Kvaskoff M.
        • Pandeya N.
        • Green A.C.
        • Perry S.
        • Baxter C.
        • Davis M.B.
        • et al.
        Solar elastosis and cutaneous melanoma: a site-specific analysis.
        Int J Cancer. 2015; 136: 2900-2911
        • Landi M.T.
        • Bauer J.
        • Pfeiffer R.M.
        • Elder D.E.
        • Hulley B.
        • Minghetti P.
        • et al.
        MC1R germline variants confer risk for BRAF-mutant melanoma.
        Science. 2006; 313: 521-522
        • Lee J.H.
        • Choi J.W.
        • Kim Y.S.
        Frequencies of BRAF and NRAS mutations are different in histological types and sites of origin of cutaneous melanoma: a meta-analysis.
        Br J Dermatol. 2011; 164: 776-784
        • Maldonado J.L.
        • Fridlyand J.
        • Patel H.
        • Jain A.N.
        • Busam K.
        • Kageshita T.
        • et al.
        Determinants of BRAF mutations in primary melanomas.
        J Natl Cancer Inst. 2003; 95: 1878-1890
        • Mar V.J.
        • Liu W.
        • Devitt B.
        • Wong S.Q.
        • Dobrovic A.
        • McArthur G.A.
        • et al.
        The role of BRAF mutations in primary melanoma growth rate and survival.
        Br J Dermatol. 2015; 173: 76-82
        • Menzies A.M.
        • Haydu L.E.
        • Visintin L.
        • Carlino M.S.
        • Howle J.R.
        • Thompson J.F.
        • et al.
        Distinguishing clinicopathologic features of patients with V600E and V600K BRAF-mutant metastatic melanoma.
        Clin Cancer Res. 2012; 18: 3242-3249
        • Palmer J.S.
        • Duffy D.L.
        • Box N.F.
        • Aitken J.F.
        • O'Gorman L.E.
        • Green A.C.
        • et al.
        Melanocortin-1 receptor polymorphisms and risk of melanoma: is the association explained solely by pigmentation phenotype?.
        Am J Hum Genet. 2000; 66: 176-186
        • Scherer D.
        • Rachakonda P.S.
        • Angelini S.
        • Mehnert F.
        • Sucker A.
        • Egberts F.
        • et al.
        Association between the germline MC1R variants and somatic BRAF/NRAS mutations in melanoma tumors.
        J Invest Dermatol. 2010; 130: 2844-2848
        • Sturm R.A.
        • Duffy D.L.
        • Box N.F.
        • Newton R.A.
        • Shepherd A.G.
        • Chen W.
        • et al.
        Genetic association and cellular function of MC1R variant alleles in human pigmentation.
        Ann NY Acad Sci. 2003; 994: 348-358
        • Thomas N.E.
        • Edmiston S.N.
        • Alexander A.
        • Millikan R.C.
        • Groben P.A.
        • Hao H.
        • et al.
        Number of nevi and early-life ambient UV exposure are associated with BRAF-mutant melanoma.
        Cancer Epidemiology Biomarkers Prev. 2007; 16: 991-997
        • Thomas N.E.
        • Kanetsky P.A.
        • Begg C.B.
        • Conway K.
        • Berwick M.
        Melanoma molecular subtypes: unifying and paradoxical results.
        J Invest Dermatol. 2010; 130: 12-14
        • Thomas N.E.
        • Kanetsky P.A.
        • Edmiston S.N.
        • Alexander A.
        • Begg C.B.
        • Groben P.A.
        • et al.
        Relationship between germline MC1R variants and BRAF-mutant melanoma in a North Carolina population-based study.
        J Invest Dermatol. 2010; 130: 1463-1465
        • van Elsas A.
        • Zerp S.F.
        • van der Flier S.
        • Kruse K.M.
        • Aarnoudse C.
        • Hayward N.K.
        • et al.
        Relevance of ultraviolet-induced N-ras oncogene point mutations in development of primary human cutaneous melanoma.
        Am J Pathol. 1996; 149: 883-893
        • Whiteman D.C.
        • Pavan W.J.
        • Bastian B.C.
        The melanomas: a synthesis of epidemiological, clinical, histopathological, genetic, and biological aspects, supporting distinct subtypes, causal pathways, and cells of origin.
        Pigment Cell Melanoma Res. 2011; 24: 879-897
        • Whiteman D.C.
        • Stickley M.
        • Watt P.
        • Hughes M.C.
        • Davis M.B.
        • Green A.C.
        Anatomic site, sun exposure, and risk of cutaneous melanoma.
        J Clin Oncol. 2006; 24: 3172-3177
        • Whiteman D.C.
        • Watt P.
        • Purdie D.M.
        • Hughes M.C.
        • Hayward N.K.
        • Green A.C.
        Melanocytic nevi, solar keratoses, and divergent pathways to cutaneous melanoma.
        J Natl Cancer Inst. 2003; 95: 806-812