Epidemiological and genetic factors underlying melanoma development in Italy

SUMMARY 

Among human cancers, melanoma remains one of the malignancies with an ever-growing incidence in white populations. Recent advances in biological and immunological therapeutic approaches as well as increased efforts for secondary prevention are contributing to improve the survival rates. It is likely that a significant fall in mortality rates for melanoma will be achieved by further increase of the early detection through a more accurate selection of the higher-risk individuals (i.e., carriers of predisposing genetic alterations). A similar scenario occurs in Italy. In the present review, we have considered data on incidence, survival and mortality rates of melanoma in Italian population, including evaluation of the main risk factors and genetic mutations underlying disease susceptibility.

Practice points.

Incidence & mortality of melanoma

• A gradient of incidence rates was observed for melanoma within the Italian population, with higher frequencies reported in North Italy (about 14 new cases per 100,000 inhabitants per year for both sexes) as compared with those found in middle (about 12/100,000 inhabitants per year) or south (less than 7/100,000 inhabitants per year) Italy. Conversely, the mortality rates appear more homogeneous across the country (about two and 1.5 deaths per 100,000 inhabitants for males and females, respectively).

Melanoma risk factors

• In Italian population, the presence of a light phototype, history of sunburns, high density of cutaneous nevi and/or familial recurrence represent the main risk factors for melanoma, similarly to what described in other white populations.

Familial melanoma

• As in other countries with low incidence of melanoma, families with two affected members represent the majority of familial melanoma patients in Italy. Consistently with worldwide data, about one-tenth of the total melanoma patients is recognized as a familial case in Italian population. Incidence of multiple primary melanoma has been estimated of about 6% (range: 1.2–8.2%) of Italian melanoma patients.

Genetic susceptibility

• CDKN2A is the most frequently mutated high-penetrance gene in both familial and sporadic melanomas within Italian population. The role of other low-penetrance genes (such as MC1R, ASIP, TYR and TYRP1) in melanoma susceptibility is yet to be defined and needs further investigations.

High-penetrance genes, CDKN2A & CDK4

• About 25% of familial melanoma cases and 4% of sporadic melanoma cases are carriers of germline mutations in the CDKN2A gene in Italy. Among analyzed melanoma patients, the CDK4 germline mutations occur in about 5% of familial cases and none of sporadic cases. Prevalence of CDKN2A germline mutations is much higher in melanoma patients from North Italy as compared with those from South Italy (intermediate frequencies were observed in cases from middle Italy).

Prevention of melanoma

• A light phototype and a familial recurrence of skin cancer are factors involved in increasing cutaneous density of both common and atypical nevi. Lack of a significant impact of educational programs on protection from sun exposure and prevention of sunburns is reported in the few existent studies.

The incidence of melanoma has been growing fast during last decades in white populations [1]. Environmental and genetic risk factors have been implicated into the development of the disease. An intermittent exposure to ultraviolet radiation in combination with endogenous factors like a light phototype, a large number of common nevi and/or the occurrence of atypical nevi – seems to play an important role in melanoma pathogenesis [2,3]. As compared with dark-skinned populations, fair-skinned individuals exposed to similar levels of incident sunlight present a significantly higher incidence of melanoma [2,4]. From the genetic point of view, a family history positive for recurrence of melanoma and association with other malignancy types (mainly, pancreatic carcinomas) represents a constitutive, crucial risk factor in a subset of cases. Approximately, one-tenth of melanoma patients present at least one additional affected family member, suggesting that genetic alterations may deeply contribute to melanoma susceptibility [5]. Although pathogenetic mechanisms are yet to be completely understood, melanocytic transformation has been demonstrated to occur by sequential accumulation of genetic alterations (inactivation of tumor suppressor genes and/or activation of oncogenes).

Starting from the epidemiologic characteristic of the disease, here we evaluated the impact of all known risk factors in determining the rates of melanoma onset in Italian population.

Incidence & mortality of melanoma

In the world, the melanoma is ranked as 16th most frequent cancer in men and 18th in women [1]. With approximately 102,000 new cases per year (26,000 deaths/year) in men and 98,000 new cases per year (21,000 deaths/year) in women, melanoma represents the 1.5% of the total cancers in men and 1.6% in women, worldwide in 2012 [1].

Over the past five decades, the incidence of melanoma has steadily increased among the white population. In the USA, the incidence of melanoma raised from six cases per 100,000 inhabitants per year at the beginning of 1970 – 21 cases per 100,000 inhabitants per year at the beginning of 2000, showing an increase in the incidence rate of more than three-times. A similar behavior has been encountered during the same period in the countries of Central Europe, with incidence rates that have increased from 3–4 to 10–15 cases per 100,000 inhabitants per year [6]. Trend of melanoma incidence among western countries seems to indicate that the rates will even continue to rise in the next two decades. By contrast, mortality curves are characterized by stabilized rates in most western countries, with majority of melanoma patients presenting with thin lesions (Breslow thickness <1.00 mm) at the time of diagnosis, due to a continuous improvement of the secondary prevention and early detection [6,7].

Considering the geographical distribution of melanoma, incidences have been found to vary according to the population's origin: Australia, the USA and Sweden present a disease prevalence higher than that reported in Europe (except Sweden) [1,6–9]. In European population, there is a gradient of melanoma incidence moving from northern countries (where incidence is higher) to southern countries [8,9].

In Italy, a standardized incidence rate of 11.0 cases per 100,000 males and 11.1 cases per 100,000 females has been recently reported by the Italian Association of Cancer Registries (AIRTUM) [10], which is actually covering the 51% of the geographical areas across the country and including about 30 million inhabitants. Overall, it has been estimated that almost 7000 new diagnoses of melanoma are registered every year in Italy.

Incidence rates were evaluated over a long period, from 1986–2009, in Italian population. As for other western populations, there was a continuous and significant increase in incidence rates in both men, from 6.2/100,000 in 1986 to 11.0/100,000 in 2009, and women, from 6.5/100,000 in 1986 to 11.1/100,000 in 2009 (Figure 1). Despite the cutaneous melanoma represents less than 3% of all cancers in 2009 (without taking into consideration all nonmelanoma skin tumors), its frequency was pretty higher in individuals aged <50 years, among whom the disease represents the 8–10% of the total cases of cancer in both sexes [10]. In this sense, analysis of the age at disease onset within the most recent period of time (2005–2009) revealed that melanoma became the second and third most prevalent malignancy before 50 years of age among men and women, respectively (Table 1). However, a notably constant increase in incidence occurred in men older than 55 years; considering the distribution of incidence rates in relation to age, the peak of incidence was indeed registered at 70–75 years for males and, to a much lower level, over 75 years for females [10].

Figure 1. . Distribution of standardized incidence and mortality rates for melanoma in Italy over the period 1986–2009.

(A) Males; (B) females.

ASR: Age-specific rate.

Table 1. . Prevalence of melanoma in Italy, according to the onset age.

Males (age at diagnosis)			Females (age at diagnosis)
0–49 years (% of total incident cases)	50–69 years (% of total incident cases)	≥70 years (% of total incident cases)	0–49 years (% of total incident cases)	50–69 years (% of total incident cases)	≥70 years (% of total incident cases)
Testis: 12	Prostate: 23	Prostate: 20	Breast: 41	Breast: 35	Breast: 21
Melanoma: 9	Lung: 15	Lung: 17	Thyroid: 14	Colon-rectum: 12	Colon-rectum: 17
NHL: 8	Colon-rectum: 14	Colon-rectum: 14	Melanoma: 7	Corpus uteri: 7	Lung: 7

Similarly to disease distribution in European populations, a gradient of melanoma incidence rates between the northern, central and southern parts of the country have been reported (Table 2 & Figure 2). In particular, melanoma presents the highest prevalence (standardized incidence rates of 14.0 and 14.2 new cases per 100,000 inhabitants per year in males and females, respectively) in North as compared with that observed in middle (about 12/100,000 inhabitants per year in both males and females) or south (less than 7/100,000 inhabitants per year in both males and females) Italy (Table 2 & Figure 2). Sardinia, whose population is genetically homogeneous and separated from that of the rest of the country due to strong genetic drift, presented the lowest incidence rates in Italy (Table 2).

Table 2. . Distribution of melanoma standardized incidence, mortality and survival rates in Italy .

Characteristic	Geographical origin
	North-east		North-west		Middle		South		Sardinia		Italy
	M	F	M	F	M	F	M	F	M	F	M	F
Incidence (ASR per 100,000)	13.8	14.3	14.5	13.9	12.1	12.1	6.4	6.7	6.0	5.5	11.0	11.1
Mortality (ASR per 100,000)	2.3	1.5	3.2	2.1	2.3	1.8	2.0	1.3	1.4	1.1	2.3	1.5
5-year survival, % (95% CI)	82 (80–84)	89 (87–90)	84 (81–86)	90 (88–92)	81 (79–84)	88 (85–90)	73 (68–77)	82 (78–86)	80 (65–93)	84 (66–92)	82 (80–83)	89 (88–89)

ASR: Age-specific rates; F: Females; M: Males.

Figure 2. . Incidence and mortality for melanoma in Italy.

Crude incidence (A) and mortality (B) rates for melanoma in Italy over the period 2005–2009, according to the patients’ geographical origin.

In contrast to the differences in incidence, the mortality rates appear more homogeneous across the country. Crude overall mortality was 2.3/100,000 for males (2.6/100,000 in North, 2.3/100,000 in middle, 2.0/100,000 in south Italy) and 1.5/100,000 for females (1.7/100,000 in North, 1.8/100,000 in middle, 1.3/100,000 in south Italy) per year (Table 2). Considering the fraction of deaths per year as compared with new cases per year across the three Italian geographical areas, a slight increase in mortality rates has been however reported for melanoma in men originating from south Italy (Figure 2). Overall, the cumulative risk of melanoma occurrence or disease-related death before 85 years of age was 1.3% for males and 1.1% for females or 0.3% for males and 0.2% for females, respectively [10].

Finally, the relative survival at 5 years from diagnosis was found to be improved in both European and Italian populations. A recent study, based on data from 107 cancer registries in 29 European countries (EUROCARE-5) for the period 2000–2007, including more than 10 million patients with cancer diagnosed up to 2007 and followed up to 2008, showed that the 5-year relative survival for melanoma was 81% in men and 88% in women [11]. The same trend has been reported in Italy, with a relative survival after 5 years from diagnosis estimated of 82% in men and 89% in women (Table 2). Moreover, in European countries a gradient in reduction of the 5-year survival has been registered with the increase of the diagnosis age (89% among patients diagnosed at age before 45 years versus 72% among patients with disease onset at age of 75 years or more) [11]. In Italy, the same reduction in survival has been registered with the increasing age at diagnosis as well as according to geographical distribution; indeed, survivals reduced moving from northern to southern areas of the country, especially for male patients. The reasons underlying this latter finding have been recently investigated [12]; differences in survival could be imputed to either a minor provision and organization of skin mass screenings or a minor proneness to undergo cutaneous surveillance by people from southern Italian regions (socioeconomic status and lifestyle habit may contribute to support such a behavior). Overall, survival rates improved over the years in both Europe and Italy: from 82% and 79% in 1999–2001 to 86% and 84% in 2007–2009 for men and women, respectively [9,10]. Several factors impact on such an improvement of survival in patients with melanoma across the years: diffusion of effective screening programs, increase of early diagnoses (with higher percentages of initial disease stages at diagnosis), effectiveness of current therapeutic strategies [9]. However, some disparities in the melanoma survival across the Europe exist, reportedly due to differences in health care expenditures [13].

For the prevalence rates, data analysis showed that about 86,000 Italian patients (about 37,000 men and 49,000 women), who had been diagnosed with melanoma during their lifetime, were alive in 2009 [14].

Melanoma risk factors

Sun exposure has been identified in many epidemiological studies as the main environmental cause of melanoma, although this tight relationship is still debated [15,16]. In particular, it has been reported a direct positive association with the intermittent exposure to the sun for recreational purposes as compared with the continuous occupational sun-exposure [17–20]. The study by Gandini et al. [18] analyzed the results of 57 published papers (52 case–control and five cohort studies), calculating a relative risk (RR) for melanoma of 1.34 (95% CI: 1.02–1.77) in subjects exposed to ultraviolet (UV) radiation compared with nonexposed.

The UV radiations have a wave length between 100 and 400 nm and are classified into UVA (400–315 nm), UVB (315–280 nm) and UVC (280–100 nm) [21]. Previously, the UVB radiations were considered the only part of the solar spectrum acting as carcinogen; from 2009, the International Agency for Research on Cancer (IARC) however classified the full UV radiation spectrum as carcinogenic to humans [22].

Under classification of sun exposure in three types (intermittent [i.e., outdoor sports and activities, holidays in sunny places], continuous [occupational exposure] and total [coexistence of both intermittent and continuous exposures]), the risk, in comparison to nonexposed individuals, was significantly higher in subjects with an intermittent exposure (RR: 1.61; 95% CI: 1.31–1.99) as compared with subjects with a continuous exposure (RR: 0.98) [18]. Many epidemiological studies have also examined the association between artificial UV exposure and melanoma risk. Probability to develop such a disease has been estimated as higher in subjects who were exposed for the first time at age younger than 40 or 35 years (RR: 1.69; 95% CI: 1.32–2.19) or 1.75 (95% CI: 1.35–2.26), respectively [23,24]. Finally, the risk has been demonstrated to increase with intensity and duration of artificial UV exposure sessions [22]. In Italy, a decree (n. 110 on 12/05/2011) aiming at regulating the use of UV-emitting equipment for cosmetic purposes was promulgated by the Ministry of Economic Development, which explicits the prohibition on use of these devices for children under 18 years old, pregnant women, persons who suffer or have suffered from skin cancer and those who tan with difficulty or are prone to sunburn.

Individuals with a history of sunburn had a risk of about two-times higher (RR: 2.03; 95% CI: 1.73–2.37) compared with those with a negative history [18], especially when sunburns had occurred in childhood (RR: 1.9; 95% CI: 1.6–2.3) as compared with those in adolescence (RR: 1.6, 95% CI: 1.4–1.9) or in young adulthood (RR: 1.4; 95% CI: 1.3–1.6). The study by Chang et al. [25] analyzed the association between sun exposure and melanoma risk at different latitudes, through evaluation of 15 case–control studies (5700 melanoma cases and 7216 controls). The authors have shown that the intermittent exposure stratified by high latitude of subjects’ residence was a risk factor for melanoma of the trunk (RR: 1.7; 95% CI: 1.4–2.2) and limbs (RR: 1.4; 95% CI: 1.1–1.7), but not for that of the head and neck (RR: 1.1; 95% CI: 0.8–1.4). Conversely, the occupational sun exposure was associated with the risk of melanoma of the head and neck in areas of low latitude (RR: 1.7; 95% CI: 1.0–3.0) [25]. No significant differences were observed for the risk associated with sunburn in childhood for melanoma of the trunk or limbs and head and neck, according to the different latitudes [25].

Evaluation of the scientific literature suggested that, for white populations, the number of melanocytic nevi is a good indicator of risk for melanoma and atypical nevi may have an independent role [26–28]. In addition, the risk of melanoma is low in subjects with few common nevi, whereas it is higher in patients with multiple nevi and atypical nevi [27]. Again two meta-analyzes have provided clues in this sense [29,30]. The study by Gandini et al. [29] analyzed the results of 46 papers (38 case–control studies and eight cohort studies), confirming that the number of common nevi is an important risk factor, with an increase in risk associated with the presence of a total number of more than 100 nevi as compared with a limited number (<15) of nevi (RR: 6.89; 95% CI: 4.63–10.25) and the occurrence (≥5 vs 0) of atypical nevi (RR: 6.36; 95% CI: 3.80–10.33). The meta-analysis by Olsen et al. [30] evaluated 49 studies further confirmed the role of the presence of atypical nevi, demonstrating that for subjects with at least one atypical nevus the relative risk for melanoma was 3.63 (95% CI: 2.85–4.62).

In summary, the results from the literature show that there is a positive association between sun exposure and melanoma risk and that the risk is high for all ages. Moreover, the risk increases with the number of sunburns that occurred in the course of life. Finally, the number of common nevi, atypical nevi, and multiple nevi should be considered as important risk factors for melanoma.

Considering the phenotypic characteristics within white populations, an increased risk of melanoma has been demonstrated for patients with a high density of freckles (RR: 2.1; range: 1.7–6.9), clear/blue eyes (RR: 1.6; range: 1.0–4.5), red (RR: 3.6; range: 1.2–7.8) or blond (RR: 1.6; range: 1.1–4.9) hair and/or fair skin (RR: 2.0; range: 1.6–9.0) [31,32]. Such a wide variability in melanoma risk is mostly due to an inappropriate risk adjustment for all known confounding factors: population origin and residence latitude, duration of sun exposure, additional phenotypic characteristics and familiarity for melanoma [31,32]. Overall, there is reasonable scientific evidence that subjects with fair skin, eyes and hair – features characterizing the photo-type I/II – have an approximately double risk of melanoma than those with darker skin, eyes and hair encountering for the photo-type IV.

In Italian population, risk factors for melanoma are mostly overlapping with those described in western countries [33] and reported above. A prolonged recreational sun exposure – especially when associated to fair skin and blue eyes, the recurrent sunburns and/or an elevated number of total nevi – always acts as strong risk factors for melanoma, with an even higher prevalence of such features in familial cases as compared with sporadic ones [34–36]. All these factors are included in risk prediction models for melanoma [37,38].

Familial melanoma

Over the past 20 years, many epidemiological studies have assessed the outstanding role of family history in conferring a high risk of cutaneous melanoma [31,39–40] as well as of additional primary melanomas [41]. A meta-analysis by Gandini et al. [31], based on evaluation of 60 published papers (58 case–control studies and two cohort studies), confirmed that subjects with a positive family history of melanoma have a risk of about two-times higher (RR: 1.84; 95% CI: 1.41–2.14) of developing melanoma as compared with those without familiarity.

The familial melanoma, which accounts for about 10–12% of the total cases, is generally identified according to standardized criteria [42]: at least three affected members, or two affected members and presence of at least one of the following subcriteria: at least one affected member younger than 50 years at onset, or a case of pancreatic cancer in a first- or second-degree relative, or one affected member with multiple primary melanomas.

In 2009, a study based by Lechman et al. [43] proposed a revision of the criteria for identifying familial melanoma or, more in general, for selecting individuals with putative genetic susceptibility to melanoma. In particular, the variability of both the incidence of the disease and the penetrance of mutations in susceptibility genes should be associated with the geographic location and ethnic origin of melanoma patients. As previously evidenced, Italy is recognized as a country with a generally low incidence of melanoma (south Italy presents one of the lowest incidences in Europe; see above). Therefore, Italian patients with melanoma and at least one additional family member with invasive melanoma or pancreatic carcinoma among first- or second-degree relatives into the same family branch might be classified as putative familial cases to be addressed to genetic counseling and testing. As confirmation of this, in France (which is also classified as a country with a pretty low melanoma incidence [1]), families with two melanoma patients represent the majority of familial melanoma cases and are considered as eligible for genetic testing – especially in presence of a family member with early diagnosis of melanoma or multiple primary melanoma (MPM) [44]. For the same reason (origin in a geographical area with low incidence of melanoma), Italian patients with sporadic MPM (at least two synchronous or metachronous primary melanomas) should be usually recommended to undergo genetic testing

A positive family history for melanoma or atypical nevi has been implied into the development of MPM in the same subject, whose incidence has been reported ranging between 1.2 and 8.2% in Italian patients with a primary melanoma [45–48]. In particular, it has been estimated that MPM may occur in about 6% of Italian melanoma patients [48]. On the other side, 18–38% MPM cases reported a positive family history of melanoma [45] and 38–46% of them reported a history of atypical nevi [45,47]. The analysis of the time interval between the first and the second melanoma showed that 51–59% of patients with MPM had the diagnosis of the second primitive lesion within 12 months, even as consequence of an increased attention to skin screening [45,49]. Data from the literature suggest that, although less than 10% of patients with primary melanoma can develop additional primary lesions in their lifetime [50], all cases diagnosed with such a disease should be recommended not only to undergo dermatological examination, but also self-examine their skin regularly and protect themselves to sunlight or artificial UV radiation for life.

In summary, as stated by the GENOMEL consortium [51], criteria for offering a genetic testing in a geographical area with lower incidence of melanoma – like the Italian one – are the following ones: at least, two primary melanomas in an individual; or two cases of melanoma among first- or second-degree relatives (blood relatives); or one case of melanoma and one case of pancreatic cancer in first- or second-degree relatives (blood relatives).

For vast majority of melanoma families it is established a pattern of inheritance of susceptibility in an autosomal dominant fashion with incomplete penetrance and variable expression of the disease. Members of melanoma families display invasive melanomas but also lentigo maligna and in situ lesions [52]. The presence of melanoma and numerous atypical nevi in a family is diagnosed as the familial atypical mole melanoma syndrome, which sometimes also includes pancreatic cancer [53]. Pancreatic cancer, however, is also present in some melanoma kindreds who do not have atypical nevi [53].

Genetic susceptibility

From the genetic point of view, as with other cancers, it has been possible to document the involvement of different genes in the early stages of development and progression of melanoma. Some of these genes encode proteins that directly control the cell cycle and, if mutated at germline level, increase very significantly the susceptibility to melanoma (high-penetrance genes). The only two genes included into this group and recognized to confer the highest susceptibility to melanoma are the CDKN2A (also called multitumor-suppressor MTS1), located in the 9p21 chromosomal region, and the CDK4, located at 12q13.6. Germline mutations in CDKN2A are identified in up to 40% of melanoma kindreds, with percentages varying widely across countries, whereas CDK4 mutations are very rare [50,54–56]. The role of these two genes worldwide and in Italian population will be discussed below.

Another class of genes (low-penetrance genes), whose variations are much common in the population, may confer a limited risk of melanoma. These latter sequence variants may play different roles in cancer susceptibility, through modifications of the risk conferred by high-penetrance genes or functional interactions with other low-penetrance genes or interference with the biological effects by environmental factors.

Recent large scale genome-wide association studies have identified several common low penetrance genes contributing to the predisposition of different cutaneous disease or conditions: risk of melanoma (TYR and MTAP [57]), development of nevi (MTAP and PLA2G6 [58]), coexistence of increased nevus density and melanoma (IRF4 [59]) and occurrence of both basal cell carcinoma and melanoma (CDKN2B and KLF14 [60]).

Among others, MTAP gene, which is located closely to the CDKN2A locus at chromosome 9p21, seems to provide the main contribute in inducing development of nevi and melanomas [57,58]. Detection of simultaneous impairments of both MTAP and CDKN2A loci suggests the existence of interactions between these two genes [57–58,61]. Overall, a growing amount of additional genes has been associated to either cutaneous pigmentation (MC1R, TYR, TYRP1, ASIP, TPCN2, NCKX5 and OCA2), nevogenesis (TERT/CLMPT1L, CDK6/XRCC1, and PARP1), or melanomagenesis (MC1R, ATM, MX2, ARNT, CASP8, MITF and POT1) over the recent past years [61–65]. Some sequence polymorphisms in MC1R [64] and a rare germline variant in MITF [66] have been reported to increase the risk of both familial and sporadic melanoma, whereas rare missense mutations in POT1 may induce melanoma susceptibility in a limited subset of families [63]. For MC1R, multiple variants of this gene can modify melanoma risk and age at onset in CDKN2A-positive families; indeed, CDKN2A mutation carriers with MC1R variants had a statistically significant lower median age at diagnosis [67]. The E318K MITF variant contributes in predisposition to nodular melanoma and renal cell carcinoma and/or pancreatic cancer in French and Italian populations [68,69]. Although the MC1R, ASIP, TYR and TYRP1 genes seem to act as major determinants of hair and skin pigmentation, their role in melanoma development remains unclear [61,64].

High-penetrance genes: CDKN2A & CDK4

The major melanoma susceptibility genes are CDKN2A and CDK4. Although both genes present an high penetrance among families with melanoma recurrence, the frequency of their mutations is completely different; as many as two-fifths of melanoma kindreds carry germline mutations in CDKN2A, whereas CDK4 mutations have been identified in just few families worldwide [50,54–56]. In particular, a variable percentage – ranging between 20 and 40% of families with three or more affected, according to the geographical origin of the patients (see below) – may be positive for germline mutations in the CDKN2A gene, while a percentage of less than 5% may carry germline mutations in the exon 2 of the CDK4 gene [54,55].

The CDKN2A gene, representing the major gene involved in melanoma predisposition, encodes two proteins which are known to act as tumor suppressors:

• p16^INK4a (encoded by exons 1α, 2 and 3), which is involved in the regulation of the cell cycle inhibitor of cyclin-dependent kinases CDK4 and CDK6;

• p14^ARF (a product of an alternative splicing that includes exons 1β and 2), which is involved in the stabilization of apoptosis-regulator p53 protein.

Alterations in these interacting genes, which regulate cell proliferation and survival, have been demonstrated to allow melanoma cells to steadily progress through the cell cycle and grow uncontrollably [55,70].

The prevalence of unselected melanoma patients to carry a mutation in the CDKN2A tumor suppressor gene is between 0.2 and 2%, varying in relation to the incidence of the disease in each particular geographical area [50]. The frequency of CDKN2A mutations increases from 8–12% in patients with a sporadic MPM and 47% in patients with MPM and a positive family history [71].

Data from the GenoMEL consortium indicated that approximately 28% of families carrying CDKN2A mutation may have one or more pancreatic cancer; more in general, studies from different populations have estimated the lifetime risk of pancreatic cancer to be 17–25% [72,73]. Looking at specific geographical areas, a significant association of CDKN2A mutations with pancreatic cancer was demonstrated in Europe and North America but not seen in Australia [74]. In Italy, extensive studies have been conducted in North Italy only, estimating the lifetime risk to be lower than 10% [75,76].

The penetrance of CDKN2A mutations is thus influenced by geographical origin. In particular, mutation rates were reported to define a melanoma risk of 13% in Europe, 50% in the US and 32% in Australia by 50 years of age; the lifetime risk of melanoma has been estimated to be 58% in Europe, 76% in the USA and 91% in Australia [50,77]. Since it has been demonstrated that the incidence rates of melanoma may vary in different populations as a combination of constitutional factors and UV exposure levels [78], one could speculate that the same factors affecting melanoma incidence may also mediate CDKN2A mutation penetrance.

The first studies related to families from North Italy reported a rate of CDKN2A mutations of 30%, in families with two affected first-degree relatives, and 45% in those with three affected family members [79,80]. Pooling data from the mutation screenings in melanoma patients followed up at the main centers across the country [36,81–96], about one-fourth of familial cases (including at least two individuals affected by melanoma) were found to carry a mutation in the CDKN2A gene (Table 3). As for other patients’ collections from western countries, CDKN2A mutation frequency in Italy was reported to be mainly dependent on the total number of affected members, varying from less than 20% in families with two affected members to about 60% in those with four or more affected members (not shown) [88,89].

Table 3. . Distribution of CDKN2A germline mutation frequencies in Italy.

Type of patients	Geographical origin (n/total; %)			Total cases (n/total; %)
	North	Middle	South
Familial cases	132/399; 33.1	46/275; 16.7	14/91; 15.4	192/765; 25.1
Sporadic MPMs	13/84; 15.5	6/87; 6.9	3/72; 4.2	22/243; 9.1
Sporadic cases	34/703; 4.8	7/187; 3.7	14/811; 1.7	55/1701; 3.2
Total cases	179/1186; 15.1	59/549;10.7	31/974; 3.2	269/2709; 9.9

As previously stated, patients with a cutaneous melanoma present a higher incidence of additional melanomas (the second primary lesion mostly occurs within the first year from the diagnosis of the first melanoma; risk decreases progressively with time) [45,97]. The probability to find CDKN2A mutations increases with the occurrence of MPM, especially in patients with at least one affected family member (configuring the case of families with higher recurrence of the disease due to the presence of at least three familial melanomas) [50]. In Italy, stratification of the patients with familial melanoma according to the presence of MPM revealed an extreme variability of the likelihood to detect a CDKN2A germline mutation, whose frequency was ranging from less than 20% (presence of at least two affected family members only) to about 50% (more than two affected members and one MPM) or even nearly all cases (more than two affected members and two or more MPMs) (Figure 3).

Figure 3. . CDKN2A germline mutations in familial melanoma patients, according to number of MPM into the family.

Distributions of frequencies (A) and mutation rates (B) are presented.

Focusing on the sporadic melanoma cases, a limited fraction (77/1944; 3.9%) of the Italian patients included into the present evaluation presented a germline mutation in the CDKN2A gene (Table 3). Again, patients with synchronous or asynchronous multiple primary melanoma and absence of any recurrence of the disease in family – classified as sporadic MPMs – presented a higher frequency (22/243; 9.1%) of CDKN2A mutations as compared with patients with sporadic melanoma only (55/1701; 3.2%).

Overall, nearly one tenth (269/2709; 9.9%) of Italian melanoma patients were carriers of CDKN2A germline mutations (Table 3). Taking also into consideration some lack of sensitivity of the screening approaches previously used in the various studies (mostly, Sanger sequencing assays with a limit of detection of mutant alleles of about 20%, 5–10 times inferior of that achieved by other techniques such as pyrosequencing or real-time PCR [98]), which may reduce the total number of positive cases, we can affirm that Italy presents a low frequency of germline mutations in the CDKN2A gene among unselected melanoma patients. On the other side of the coin, it is worth underlining that some discrepancies in mutation prevalence may be due to less stringent selection criteria used in our population-based studies (i.e., vast majority of our melanoma families showed two affected members only or sporadic cases were pooled without stratifying them for age at diagnosis).

Considering the distribution of the frequencies of CDKN2A sequence variants within different Italian regions (according to the geographical origin of melanoma patients), a gradient of mutation prevalence was found (Table 3 & Figure 4). In contrast to a higher frequency of CDKN2A germline mutations observed in both familial (about 33%) and sporadic (about 6%) cases from North Italy, prevalence among the same subsets of patients originating from south Italy was markedly lower (about 15 and 2%, respectively), with intermediate values in cases from middle Italy (Table 3 & Figure 4). These observations strongly suggest that discrepancies in CDKN2A mutation frequency may be due to the presence of a different constitutional predisposition or ‘genetic background’ within Italian populations. Consistency between CDKN2A mutation frequencies and standardized incidence rates of melanoma (higher in North Italy as compared with those in south Italy; see above) represents a further clue that the same factors that affect the incidence of the disease may also underlay CDKN2A-based melanoma susceptibility.

Figure 4. . Distribution of CDKN2A mutation frequencies in melanoma patients from different Italian areas.

These data seem to also indicate that genetic factors predisposing to melanoma may be more heterogeneous in south than in North Italy. In other words, one could postulate that multiple low-penetrance genes may participate in determining the susceptibility to melanoma within southern Italian regions. As confirmation of this, melanoma patients from Sardinia – which presents a incidence of melanoma even slightly lower than that registered in populations from south Italy (roughly, four new cases per 100,000 inhabitants per year; Table 2) – were found to lack mutations in CDKN2A or any known major gene involved in melanoma susceptibility (CDKN2B and CDK4) [84,89].

The 292 mutations identified in the CDKN2A gene among familial and sporadic Italian melanoma patients were stratified according to the type of sequence variations and the correspondent effects on the gene products. The distribution of CDKN2A mutations showed a vast majority of missense mutations, including recurrent variants of possible or proven common origin (mostly, the G101W mutation, recognized as a variant with founder effects in some part of Italy (Figure 5) [80,88].

Figure 5. . Types of CDKN2A germline mutations in melanomas from Italian population.

Mutation designation (according to the gene product modification) and frequency (referred to a total of 292 mutations) are reported.

As expected, a very small proportion of melanoma cases of Italian origin presented mutations in the CDK4 gene. Two (4.7%) mutations, R24C and R24H, have been observed in a series of 427 familial cases across the country [56,79–81,83,85–86,90–91,96]. This frequency is consistent with that reported in other western populations [40,54–55]. No mutation has been described in 551 sporadic melanoma patients of Italian origin [86,91].

Despite limited data are available, other types of CDKN2A or CDK4 rearrangements – such as large deletions or duplications in the correspondent 9p21 and 12q13.6 loci – seem to be infrequently involved in the development of melanoma within the Italian population [81,88,99]. However, identification of melanoma families with evidence of linkage to chromosome 9p21 and without any deleterious mutation in the CDKN2A gene seems to indicate that other genetic factors with location at this genomic level may be involved in melanoma susceptibility (as previously mentioned, MTAP could be considered as a good candidate, but its role in melanomagenesis is yet to be defined).

Prevention of melanoma

Identification of individuals with a higher risk of melanoma (i.e., positive for family recurrence of the disease or predisposing genotypic/phenotypic characteristics) can be considered as the most feasible strategy to promote selective screening aimed at early diagnosis of such a malignancy. Indeed, mass screening, which is widely recognized as the best management of secondary prevention of tumors, is not feasible in melanoma due to the high expenses associated with the need for the dermatologists to examine the skin of nearly all adolescents and adults from white populations like that from Italy.

Our data showed a heterogeneous distribution of incidence within the Italian population, with the highest rates registered in Northern areas of the country. Although south Italy is usually exposed to higher total levels of UV radiation due to both its latitude (the closer the equator, the higher the UV levels) and the richness of sunny days during the year (UV radiation levels are highest under cloudless skies) [100], the prevalence of individuals with lower-risk phenotypic characteristics (i.e., photo-types III and IV) may probably justify the decreased trend of melanoma incidence in populations originating from this part of the country. Anyway, as repeatedly stated above, the role of each risk factor into the pathogenesis of melanoma in Italy has to be considered as similar to that described in white populations from other geographical areas throughout the world.

In the prevention of melanoma, it is difficult to quantify how much reduction of its incidence and mortality may be achieved through alteration of sun exposure habits in subjects at risk. In Italy, some studies relating to the variation of the number of nevi in relation to UV irradiation have evaluated how prevention programs, which mainly act on protection from sun exposure, may decrease the risk of melanoma [101,102]. In general, poor results have been reported in European population from evaluation of the education interventions in increasing the use of sunscreen and other measures of photo-protection (clothing, protected zones, etc.) among school-age children and young adults [103–105]. Consistently, Naldi et al. [106] observed no significant impact of educational programs on the reduction of sunburns among elementary school children. Recently, it has been indicated that low phototypes and a family history of skin cancer may act as the main features to determine a positive sunburn history in younger age as well as an increased skin density of both common and atypical nevi [107].

Conclusion

In Italy, a combination of genetic and constitutional factors (i.e., genomic backgrounds and phenotypic features) seem to contribute to the heterogeneous distribution of both melanoma incidence rates and susceptibility mutation frequencies across the country. The scenario depicted by our observations represents a further strong indicator that both epidemiologic and mutational analyses need to be investigated in every different geographical areas.

In a broader view, surveillance programs and early diagnosis acquire more and more the value of fundamental tools to decrease morbidity and mortality of melanoma. To improve the efficiency of such programs, it is becoming crucial the identification of high-risk cohorts through an increasing use of genetic and molecular tools toward a more accurate individual classification.

Future perspective

In very next future, selection of high-risk individuals could be fostered by the introduction in clinical practice of the emerging massively parallel sequencing or next-generation sequencing methodologies, in order to achieve a more comprehensive pattern of clinically-relevant predisposing mutations. Indeed, high throughput techniques, such as the whole exome sequencing, may permit to identify low-frequency variants with intermediate penetrance and, thus, provide multiple advantages for a more efficient genotyping, with the assessment of more accurate phenotypic associations. Cost decreasing of such technologies will allow first a wider application in medical diagnostics research (i.e., development of next-generation sequencing panels with disease-associated gene variants for identification of mutation carriers at a higher risk of melanoma – and then their validation in clinical trials, with the full respect of all ethical issues).