Introduction

Melanoma arises through malignant transformation of melanocytes and is considered a neural crest neoplasia. It is the most aggressive and lethal form of all skin cancers. Although it represents approximately 5% of all cutaneous malignancies, it is responsible for the vast majority of skin cancer-related deaths [1].

Embryologically, melanocytes are derived from the neural crest and colonized in the skin, the eyes, and several tissues during the development. Benign proliferation is commonly observed and results in melanocytic nevi. With the presence of diverse mutations, aberrations, translocations, and deletion, however, malignant transformation can take place, giving rise to various types of melanomas [2]. Although early stages can be cured with surgical excision, a proportion of patients still develop metastatic disease. As novel treatments are being rapidly developed, a tremendous improvement in the survival of metastatic patients has been achieved. For example, the average 1-year survival for stage IV melanoma patients are reported to be 74.5% with BRAF/MEK inhibitors and 71.9% for anti-PD-1 blockade alone or combined with anti-CTLA-4 [3•]. However, the treatment of advanced melanoma stages still remains challenging.

As the incidence of melanoma steadily increases in both sexes, further improvement in primary prevention and early detection strategies is crucial [4]. Still, the epidemiologic, genomic, and anatomic profiles of the disease significantly differ across the world and mostly depend on a constellation of environmental and (epi) genetic factors.

In the current review, we aim to outline the major epidemiologic, genetic, and anatomic differences of the various melanoma types across the globe and attempt to provide a background which is relevant to better understand the disease and its treatment.

Melanoma Incidence

Cutaneous melanoma is the 12th most common cancer worldwide with an estimated age-standardized incidence rate of 3.0 per 100,000 [5]. The incidence of the disease varies among populations, with highest rates reported in Australia and New Zealand (40.3 per 100,000 and 30.5 per 100,000, respectively) followed by North America and Northern and Western Europe. Lowest incidences are documented in South-Central and South-Eastern Asia (below 0.5 per 100,000) (Table 1). This variation is mainly attributed to heterogeneity of risk factors, such as ultraviolet (UV) exposure and genetically determined phenotypic characteristics. Recent epidemiological data show an annual overall rising incidence of melanoma with 232,000 new cases (1.6% of all cancers) in 2012 [5] and 351,880 cases (95% CI 281,633–445,036) in 2015 with an age-standardized rate of five cases per 100,000 persons (95% CI 4–7) [6]. The overall number of patients being diagnosed with melanoma is estimated to continue to increase in the decades ahead, mostly due to the lengthening of human lifespan and the aging of population [7].

Table 1 Estimated age-standardized rate (ASR) (per 100,000) of melanoma incidence and mortality by sex and region, 2012 (http://globocan.iarc.fr)
Full size table

In Australia, melanoma is the fourth most common cancer and accounts for 9.7% of all new cancer cases between 2009 and 2013 [8]. Between 1982 and 2016, the incidence rate of melanoma has increased from 27 cases per 100,000 in 1982 to 49 per 100,000 in 2016 for all age groups [8]. Among Australians aged 15–34, melanoma is the most common cancer (excluding non-melanoma skin cancers originating from keratinocytes). However for this age group, the incidence rate has dropped from 13 new cases per 100,000 in 1982 to 9.4 per 100,000 in 2016 [8]. With 72 cases per 100,000 per annum on average (2010–2014), the incidence of invasive melanoma in Queensland remains the highest recorded worldwide. A recent study analyzing incidence and mortality data in Queensland between 1995 and 2014, showed a stable or falling incidence of both thin (≤ 1 mm) and invasive (> 1 mm) melanoma in people under 60 years. This might serve as a good marker of effective prevention campaigns, reducing sunburns in early age, one of the strongest risk factors of melanoma. However, in the population over 60 years, the trend in incidence is still increasing [9]. Melanoma mortality rates remain stable or declining in both sexes under 40, while in males over 60, the mortality seems to increase with a rate of 2% per annum. These results suggest that long-running melanoma prevention campaigns and early detection strategies can contribute to reduce the incidence and mortality rates of melanoma in susceptible populations around the world. Yet, it also reflects that prevention has to be started in young age [10].

The incidence of melanoma is increasing in the US and in European countries as well and is anticipated to continue to rise. In the US, melanoma of the skin represents 5.2% of all new cancer cases with 22.3 per 100,000 numbers of new cases for 2010–2014 [11]. Notably, the incidence and mortality of the disease across the ethnic disparities seems to be heterogeneous. In the US in fact, melanoma is more common in Caucasians compared to African-Americans (AAs), with incidence rates of 34.4 per 100,000 men and 20.9 per 100,000 women compared to 1.1 per 100,000 men and 1.0 per 100,000 women, respectively [11]. Recent data have shown that ethnic minorities remain at greater risk for melanoma-associated mortality, probably due to the clinical heterogeneity and histological variation of the disease [12]. Whereas the most common form of melanoma among Caucasians is the superficial spreading melanoma (SSM); melanomas in the AA population occur more often on non-sun-exposed skin, such as the palms and the soles. Of these, acral lentiginous melanoma (ALM) is the most common form [12]. Although ALM accounts only 2–3% of all melanoma cases worldwide, they account for up to 70% of all melanomas in darker skin types [13]. The atypical localization and appearance seem to contribute to delayed diagnosis and lead therefore to outcomes.

In the European Union, melanoma was the seventh most common malignancy in 2012, accounting for 3% of all new cancers [14]. For the same year, the estimated age-standardized rates were 8.6 per 100,000 for males and 8.9 per 100,000 for females (Table 1) [15]. Between the European countries there are large incidence differences, varying from 19.2 in Switzerland to 2.2 per 100,000 in Greece [16]. Central and Eastern European countries have the lowest reported rates in Europe. This rate spectrum may be explained through differences in socioeconomic status, linked to tendency for recreational sun exposure, cultural, as well as phenotypic dissimilarities. Rate differences accruing between neighboring countries however, imply diagnostic delays and differences in reporting or registration [16]. In 2012, Switzerland registered the highest melanoma incidence in Europe, followed by the Scandinavian countries and the Netherlands. Specific to the Scandinavian population, the high incidence of melanoma may be attributed to a high-risk phenotype (fair skin, hair, and eye color) in combination to a tanning culture, high levels of UV exposure during (intermittent) holidays and indoor tanning [16].

Several studies suggest a significant gender difference in melanoma incidence and survival across the globe. In 2017, melanoma incidence rates were about 60% higher in men than in women, while death rates were more than double [4]. Higher melanoma rates have been mostly observed in elderly and male populations, whereas the female sex seems to represent an independent risk factor for early onset melanoma for women < 45 years old [17]. Gender disparities, regardless of other known predictive factors such as tumor thickness, ulceration, histologic subtype, location, and age, are thought to represent an independent prognostic factor in melanoma incidence and survival [18, 19]. According to recent data, the rising melanoma trends mostly affect the older age groups, whereas in the youngest age groups (24–44 years), the incidence seems to stabilize [20]. However, melanoma still affects mostly younger patients, with a median age diagnosis of 57–64 worldwide. The anatomical location of melanoma also varies according to gender. Males tend to present with worse clinical and histological characteristics at primary diagnosis; melanomas in male are more often located on the head, neck, and trunk, commonly ulcerated and have a higher Breslow thickness. On the contrary, the survival benefit among female patients seems to be attributed to those treated with early stage I–II of melanoma, but the survival advantage decreases in patients with a higher metastatic tumor burden [21]. The lower mortality rates in females seem to depend on both biological and behavioral differences regarding primary (sun behavior, UV protection) and secondary (skin checks) prevention [22].

Melanoma Mortality

Paradoxically, although the incidence of melanoma continues to rise, the mortality rates seem to be rising less rapidly. Melanoma mortality accounts for 0.7% of all cancer deaths, with global mortality rates ranging from 0.1 in women of South-Central Asia to 6 per 100,000 in men of Australia/New Zealand [5]. Most deaths occur in more developed regions and mostly among males, with a sex ratio of 1:3 [5] (Table 1). In the US, the number of deaths was reported to be 2.7 per 100,000 men and women per year (time period 2010–2014), with higher mortality rates among the middle-aged and elderly population (24.1% among people aged 75–84 with advanced melanoma) [11]. Across the European continent, there are important differences concerning the mortality, with highest rates of 3.2 in Norway and lowest of 1.0 and 0.9 in Romania and Greece, respectively. Countries from Central and Eastern Europe reported higher rates of advanced tumors and lower survival as compared to Romania and Greece [15]. This discordance between incidence and mortality of Central and Eastern European countries in comparison with their neighboring countries might be contributed to an extended secondary prevention through regular skin checks in Northern and Southern European countries. Indeed, the development of early detection strategies is thought to underlie a shift toward the increasing reporting of thin melanomas around the globe [23,24,25]. Recent epidemiological data confirm higher survival rates of thinner (≤ 1 mm) compared to thicker melanomas [26••]. However, an analysis of melanoma incidence and mortality in Queensland, Australia, showed that an increasing proportion of melanoma deaths occurred among patients who were initially diagnosed with thin melanomas (≤ 1 mm), compared to a melanoma thickness of > 4 mm [27]. This thin melanoma paradox has been also reported in a US study, investigating the melanoma thickness trends for 1988–2006 [28] and indicates a widespread of early detection strategies. It is possible that specific histological characteristics or other factors, such as genetic mutations, could explain this phenomenon; however, further clinical and molecular features need to be identified.

Melanoma Risk Factors and Anatomic Differences

Melanoma arises through multiple different causal pathways and reflects a dynamic interdependence between environmental factors and genetic alterations. Several factors have been identified that significantly influence the incidence and the clinical and oncogenic characteristics of cutaneous melanoma. These factors mainly comprise increased UV exposure, tanning bed use, family and personal history of melanoma, and certain phenotypical characteristics, such as fair skin and hair color.

Exposure to UV radiation is the major known environmental risk factor for melanoma development. More than 70% of cutaneous melanomas are thought to be caused by UV radiation exposure [29], while the association of melanoma risk and intermittent sun exposure has been also verified in a number of large analytical epidemiological studies [30]. The effect of UV spectrum differs between the UV components, subdivided by wavelength into UVA (320–400 nm), UVB (280–320 nm), and UVC (100–280 nm). UVC is blocked by ozone and does not reach surface of the earth; thus, UVB and UVA are responsible mainly for the UV-induced skin damage. Due to different electrophysical properties, UVB causes direct DNA damage, whereas UVA causes indirect DNA damage through generation of reactive oxygen species [31].

Observations based on the incidence of the disease and its association with the UV radiation show an increased incidence on population exposed to natural or artificial sunlight over long periods of time, as well as on sporadically sun-exposed skin in sites which are more susceptible to sunburns [30]. Since UVA (320–400 nm) is the primary source used in indoor tanning beds, indoor tanning is associated with increased risk of melanoma, with recent data providing strong evidence of higher melanoma risk with younger age at initiation and longer duration of usage of indoor tanning systems [10, 32]. Australia has already undertaken comprehensive skin cancer awareness campaigns to reduce the burden of skin cancer and has enacted a nationwide ban on tanning beds [9]. Indoor tanning legislation is constantly evolving, with youth access restriction now in place in many countries of Europa and in several states in the US [33].

Epidemiological data support two major pathways in the pathogenesis of cutaneous melanoma: one by cumulative sun exposure to the site of the future melanoma in sun sensitive people and other by early sun exposure and nevus proneness, promoted by host factors, intermittent sun exposure, or both. Approximately 25–33% of cutaneous melanomas derive from a benign, melanocytic nevus, whereas in patients with numerous nevi, this number may be as high as 50% [34, 35]. The divergent pathway model for the development of melanoma on sun-exposed skin identified differences in the sun exposure and the anatomic distribution of melanoma [36]. Transformation of nevi to melanoma most commonly occurs in non-chronically sun-damaged skin. Nevus-prone patients who have increased number of melanocytic nevi tend to develop melanomas at a younger age and on axial locations. On the other side, nevus resistant patients with fewer nevi tend to develop de novo melanomas on habitually sun-exposed skin, and at older ages [37]. Chronically sun-damaged (CSD) melanomas have higher mutation burden, later age of onset, and mostly occur on the head and neck areas [38•]. Non-CSD melanomas present earlier in life have lower mutation burden and appear on anatomical sites with intermediate levels of sun exposure, such as the trunk.

In contrast to cutaneous superficial spreading melanoma, the occurrence of nodular melanoma (NM) and mucosal melanoma (MM) seems to be independent of UV exposure. Specifically, in the case of NM, the influence of UV is controversially discussed in the literature. Some studies reported a higher prevalence of NM on sun-exposed skin such as the lower limbs, head, and neck. However, NM can also affect non-chronically sun-exposed body areas such as the trunk in fair- but also dark-skinned patients [39, 40].

Apart from environmental risk factors, phenotypic and genetic characteristics also have been consistently associated with an increased risk of melanoma development. A recent observational study investigating the clinical features associated with individuals at higher melanoma risk showed that a positive family history of melanoma is associated with higher risk for melanoma development in younger ages and on non-sun-exposed areas [41]. Notably, individuals with large or giant congenital melanocytic nevi (CMNs) at birth are at higher risk of melanoma development, which increases according to the size of CMN and is highest in those nevi traditionally designated as garment nevi [42, 43]. Also, personal history of a prior melanoma is a strong predictor for the development of a subsequent melanoma, with approximately tenfold increased risk [44]. However, subsequent melanomas appear to be thinner than the first melanomas, which can be explained through regular clinical examinations during the follow-up. Additionally, melanoma seems to appear more commonly in immunosuppressed patients, including patients with prior organ transplantation, hematologic malignancies, or human immunodeficiency virus infection, as well as patients taking immunosuppressive medication [45]. Direct drug-associated effects on melanoma carcinogenesis as well as drug-dependent impacts on a preferred environment promoting melanoma carcinogenesis are discussed in the literature as possible etiologic explanations [45].

Mutation Overview

Many phenotypic factors are genetically determined. The presence of certain pathogenic mutations determines distinct subsets of melanoma, which can be defined on the basis of their molecular phenotype. These findings have been already translated into recent therapeutic developments, and may lead to further therapeutic modalities and improve prevention strategies.

Using whole exome sequencing (WES) and whole genome sequencing (WGS), skin cancers have been identified as the most mutated cancers in human [46, 47] (10–110 mutations/Mb [48, 49••, 50]). Identification of mutated genes, implementation of this knowledge and creation of targeted therapies in melanoma have continued to serve as one of the highlight stories in translational medicine.

Among the discovered genetic aberrations, no universal mutation for all cutaneous melanomas has been identified. However, oncogene or tumor suppressor gene mutations almost always led to constitutive activation through mitogen-activated protein kinase (MAPK) pathway [51, 52]. Together with mutations in phosphoinositol 3-kinase (PI3K) pathway and in Wnt signaling pathway, they affect the majority of melanoma samples [53•, 54] (Fig. 1). The two most common mutations, BRAF and NRAS, are oncogenes within the MAPK pathway, whereas NF1 is a tumor suppressor gene, which contributes to regulation of the activation of RAS [55]. According to the frequency of mutations, TCGA Network proposed four genomically defined subdroups: BRAF-, RAS-, NF1-mutated, and triple wild-type (wt) [49••].

Fig. 1
figure 1

Mitogen-activated protein kinase (MAPK) and phosphoinositide 3-kinase (PI3K) pathways play a key role in the pathogenesis of melanoma [51,52,53•]

Full size image

BRAF-, RAS-, and NF1-Mutated Melanoma

Described over a decade ago, BRAF mutation remains the most relevant in terms of its implications on treatment decision making [56]. Affecting 40–60% of cutaneous melanomas, these are typically point mutations that lead to one amino acid substitution resulting in a constantly activated state [49••, 57]. The substitution from valine to glutamic acid (V600E) contributes for 74–86% of BRAF mutations, followed by V600K (10–30%) and less common substitutions of V600M, V600D, or V600R [49••, 58]. Although first approved for BRAF V600E-mutated melanomas, BRAF inhibitors proved to also be effective in tumors harboring other BRAF V600 mutations [59, 60]. Interestingly, over 80% of benign melanocytic nevi harbor mutations in BRAF gene [61]. However, in contrast to melanoma, MAPK pathway activation is brief in these lesions, and is followed by a growth arrest phase [62]. In the analysis of melanomas evolving from the preexisting nevi, the benign lesions harbored BRAF mutation exclusively, whereas additional mutations could be identified in each step of the invasion [63]. If compared to triple wt tumors, BRAF-mutated melanomas show much higher UV signature (30% vs 90%), which might represent an additional trigger for the melanoma-genesis [57, 58].

Of all melanomas, 28–30% harbor an activating mutation within the oncogene RAS family, most commonly NRAS, followed by KRAS and HRAS. NRAS mutations also have been identified in 18% of benign nevi [61]. Yet, as recently reported by Shain and colleagues [63], NRAS mutation was only present in the intermediate lesions. This mutation is routinely assessed in clinical practice, and, although no medication is approved for this genetic subtype of melanoma, MEK inhibitors lead to responses some in preclinical models [64] as well as clinical trials [65,66,67].

The most commonly mutated tumor suppressor gene NF1 is mutated in 13–17% of melanomas [49••, 68] and it is reported to carry loss of function mutations only in malignant tumors [63]. So far, no final clinical translation has been made for this mutation.

Triple Wild-Type Melanoma

Triple wt melanomas not only fails to show any of the most common mutations but these also harbor less UV signature, as compared to BRAF-, NRAS-, or NF1-mutated tumors [49••] (30% vs over 90%). Yet, these melanomas have more copy number alterations and complex structure changes. The triple wt tumors also show less loss of function mutations in tumor suppressor genes CDKN2A and TP53 [53•, 57, 69], but more of those seen in non-cutaneous melanomas, such as KIT, GNAQ, GNA1, and CTNNB1 [49••]. As KIT-mutated melanomas can be targeted with KIT inhibitors (imatinib and nilotinib), testing for KIT mutation is reasonable in BRAF wild-type melanomas [70, 71].

UV Signature and Epigenetic Regulation

UV light has long been reported to contribute highly to the development of melanoma, and the highest mutational load is observed in primary melanomas of ultraviolet (UV)-exposed non-glabrous skin (111 per Mb), whereas the lowest load is reported in patients with primary melanomas of glabrous skin and without the history of chronic sun exposure (3–14 per Mb) [50]. In 2015, the Cancer Genome Atlas (TCGA) Network published the data from the largest patient integrative analysis of cutaneous melanomas and observed mutations, representing the UV signature in 76% of primary and 84% of metastatic melanomas, independent of the site of primary tumor [49••]. This is, in part, an explanation for the impressive response to immunotherapies seen in cutaneous melanoma [72••, 73•, 74•], as high tumor mutational burden has been reported to be associated with higher response rates in anti-PD-1 therapy in a wide range of malignancies [75].

Epigenetic regulation causes changes in gene expression without affecting the DNA sequence alterations [76], typically through histone modification, chromatin remodeling complexes, and DNA methylation or demethylation [77]. A third of the 40 most commonly mutated genes in melanoma were found in the genes of epigenetic regulation, with up to 90% of melanoma samples showing a mutation in at least one gene associated with epigenetic regulation (e.g., genes responsible for histone modification and/or chromatin remodeling) [57].

A recent publication on high response rate to immunotherapy with anti-PD-1 antibody in desmoplastic melanomas (DM) hypothesize that high mutational burden, known to be typical for DM and also for other cutaneous melanomas [48, 49••, 50], is strongly contributing to the efficacy of the therapy [78]. Since immunotherapy is being more actively implemented in early adjuvant setting [79•], the mutational burden might be used as a biomarker for response to treatment.

Mutations in Different Subtypes of Melanoma

The genetic diversity of the various melanoma subtypes can be translated into optimal choices in the pharmacological treatment. SSM and NM often carry genomic alterations of clinical relevance, such as BRAF mutations. On the contrary, only 10–20% of melanomas arising in mucosal or acral locations harbor a MAPK pathway mutation [80]. Mutations or genomic amplification of KIT are predominantly found in acral and mucosal melanoma, with incidences of 10–20% and 20–30%, respectively; whereas, they are rarely identified in other subtypes of melanoma [13, 81]. In contrast, NRAS mutations occur with a fairly consistent prevalence rate of 15–20% in all subtypes of cutaneous and mucosal melanoma [82].

Pharmacological Treatment Approaches

Kinase inhibitors targeting the BRAF and MEK oncoproteins as well as checkpoint inhibitors have become indispensable treatment agents in the treatment of metastatic patients [3•, 72••]. Still, the mechanisms of resistance as well as potential biomarkers of response to these agents are currently elucidated, giving rise to innovative treatments and treatment combinations. Genomic alterations in the BRAF oncogene is associated with response to BRAF and MEK inhibitors. Also, MEK inhibitors have been evaluated to show some significant therapeutic response in NRAS-mutant subtypes [67]. The addition of a third inhibitor is currently being investigated in clinical trials, aiming to overcome resistance mechanisms and improve the patient outcomes. Candidates for third agents are cyclin-dependent kinase inhibitors (CDK), receptor tyrosine kinase (RTK) inhibitors, or an inhibitor targeting the phosphoinositide 3-kinase of the PI3K-pathway (Fig. 1) (NCT02159066) [83, 84]. Moreover, further molecular targets of the MAPK pathway are being tested in ongoing phase I trials, including ERK and CRAF/BRAF inhibitors (NCT02711345 and NCT02607813). As BRAF inhibition has also been shown to be associated with an increased expression of programmed death-ligand 1 (PD-L1) [85], combination therapy of BRAF/MEK inhibitors and anti-PD1 is currently examined in ongoing clinical studies (NCT02130466, NCT02967692, and NCT02908672). Checkpoint inhibitors targeting the interaction between the immune system and the tumor cell are particularly interesting as they are mutation independent and therefore seem to possibly target all melanomas. Yet, not all melanomas respond to immunotherapy, and further research has to elucidate the mechanisms of resistance here. As ALM and MM rarely carry BRAF mutations, studies were recently performed, investigating the effect of KIT inhibitors on KIT-mutated melanoma (NCT00424515 and NCT00788775). A clinical trial in patients with KIT-mutated or amplified ALM and MM showed treatment responses of 23.3% with imatinib [86].

Conclusion

Melanoma of the skin shows clear differences in incidence, mortality, genomic profile, and anatomic presentation dependent on the patient’s country of residence, ethnicity, and socioeconomic status. As the incidence of melanoma is increasing in both industrial and non-industrial countries, the development of new treatment options and strategies is crucial. Although the implementation of innovative therapies has dramatically improved patient outcomes during the recent years, the high costs are beyond affordable possibilities of health care systems in many non-developed countries. This leads to important treatment and subsequently overall survival disparities. Improved understanding of marked ethnical and genetic differences in melanoma causality and presentation may improve therapeutic developments and prevention strategies.