Defining the patient at high risk for melanoma


Ashfaq A. Marghoob, md
Memorial Sloan-Kettering Cancer Center
800 Veterans Memorial Highway, 2nd Floor Hauppauge, New York 11788
Tel: 631-863-5100


In this practical review, we aim to help clinicians identify patients who are at significant risk of developing malignant melanoma. Universal screening is challenging, thus it is important to effectively single out patients who have a high risk of developing the disease. We provide a summary of pertinent questions to review when taking the patient's history, point out the phenotypic features to note during skin examination, and suggest risk stratification as a means to plan initial and long-term surveillance strategy. We mention personal and family history of melanoma as prime risk factors for melanoma, yet the review also focuses on the patient who has no history of melanoma, either in himself or his family, and the proper ways to evaluate his likelihood of developing the disease.


Although melanoma is amenable to early detection, there has been no decline in the mortality rate of this disease.1 One American dies of metastatic melanoma every hour.2 Melanoma screening is not formally recommended by any of the major prevention task forces due to lack of prospective studies on the impact of screening on mortality.3 Accordingly, melanoma screening is not routinely performed in clinical practice, even by dermatologists. It is estimated that only 30% of dermatologists perform total skin examinations on their patients and about half of all dermatologists selectively screen only high-risk patients.4 Intuitively, melanoma is an ideal candidate for low-cost, low-morbidity, and highly effective screening. A large pilot study of melanoma screening conducted in Australia shows screening to be feasible and highly likely to be efficacious.5,6 Furthermore, mathematical modeling indicates that melanoma screening is cost-effective.7 Unlike primary prevention efforts (e.g. sun protection), which entail behavior modification and require decades for an observable impact on mortality, early detection has the potential to impact mortality more immediately.8 Because of a lack of highly effective systemic therapies for advanced melanoma, we posit that screening for melanoma promises the most rapid and dramatic decrease in mortality from the disease.

Three major reasons may explain why patients continue to die of advanced melanoma. First, patients oftentimes ignore suspicious-looking lesions and fail to seek medical advice in a timely manner to allow for early detection.9 Fortunately, public education efforts have dramatically improved melanoma self-detection. Further increases in public awareness will help detect these melanomas while they are still curable.10 Second, a subset of melanomas such as nodular melanoma can grow extremely rapidly at rates of up to 0.5 mm a month.11 Although some of these rapidly growing melanomas may be found fortuitously at an early stage during screening examinations, most are not amenable to early detection via periodic screenings. Finally, physicians do not routinely perform total body screening examinations on their patients.12 The latter should be performed on a completely disrobed patient, as patients who undergo complete cutaneous examinations are six-times more likely to have a melanoma detected than those undergoing partial examinations.13

Screening strategies include population-based screening (e.g. open clinics that screen all comers), opportunistic screening, as mentioned above, and targeted screening of high-risk patients.14 An individual is at high risk for melanoma if his or her risk of melanoma exceeds that of the general population. An example of opportunistic screening would be performing a total body examination on a patient presenting with an unrelated complaint such as a fungal infection. A survey of patients with melanoma found that, while 63% of them had visited their regular physician within the year prior to their diagnosis, only 20% overall reported having had their skin checked in that time.12 Targeted screening focuses on patients at high risk for melanoma such as those with a past diagnosis of melanoma or a family history of the disease. Presumably, the high-risk population will demonstrate a higher incidence of the disease than the general population. Although ideally everyone should be screened for this potentially curable cancer, the high incidence of melanoma coupled with the current shortage of dermatologists,15–17 mandates strategic allocation of resources in a most effective manner. Selective, targeted screening of individuals at high risk for the disease likely provides the most immediate method to help curtail the mortality rate of melanoma.18–20

How does one identify the high-risk patient for targeted screening? Melanoma is a multi-factorial disease thought to arise from an interaction between genetic susceptibility and environmental exposure (Fig. 1). A patient with a personal or family history of the disease is, from the outset, at increased risk for developing a melanoma. In the absence of a personal or family history of the disease, high-risk patients may declare themselves through phenotypic features that are markers of increased susceptibility to melanoma, such as numerous nevi, fair skin, or evidence of sun damage. In other cases, high-risk patients are first identified during the patient interview and medical history. Although intake forms, which include questionnaires to be filled out by the patient, are helpful in assessing risk, studies of patient self-assessment of their melanoma risk have yielded conflicting results, with patients at times underestimating their risk.21–24 Patients may underestimate their mole counts by failing to include flat or nonpigmented lesions.22,25 Thus, in risk assessment, the dermatologist should not rely solely on patient self-report, but rather take a proactive approach and aim to stratify a patient’s risk based on patient self-assessment (vis-à-vis the questionnaire), the interview, and the skin examination in tandem.

Figure 1.

 Overarching etiology of melanoma. Melanoma is a multifactorial disease with both genetic (left) and environmental (right) risk factors contributing to development of the disease

Melanoma risk factors elicited from patient history

Personal history of melanoma

A patient with a personal history of melanoma is clearly at greater risk for subsequent melanoma. It is commonly quoted that 1–8% of patients with prior history of melanoma will develop multiple primary melanomas (MPM).26 A prospective study of 4484 patients diagnosed with a primary melanoma, however, found the estimated cumulative five-year risk of a second primary tumor to be higher and reported the risk for the entire cohort to be 11%, with almost half of that risk occurring within the first year.27 For melanoma patients with a positive family history, the risk for a second primary melanoma increases to 19%.27 One-third of these melanomas are found concurrently with the diagnosis of the first melanoma, while the remainder are found at follow-up.28 Thus, screening frequency should be the highest during the first year following initial diagnosis but surveillance is indicated for life, as some second melanomas are diagnosed as long as 30 years later. Melanomas detected in patients with a past history of the disease are usually found at earlier, thinner stages,29 indirectly indicating that screening and increased awareness are effective strategies.

Patients with MPMs also tend to manifest many dysplastic nevi (DN),27 which can complicate their examination and augment their risk. The five-year risk of a second primary melanoma in a melanoma patient who has DN is 2–3 times the risk of a melanoma patient who does not have DN.27,30

In addition to the detection of new primary lesions, skin examinations in patients with a personal history of melanoma are geared to detect recurrent and metastatic disease. While prognosis relates primarily to tumor thickness,31 roughly 15% of melanoma deaths are attributable to metastases of thin melanomas.32 While metastases from thin lesions are rare, they contribute to mortality because the incidence of thin melanomas far outnumbers that of thick melanoma. The five-year survival rates for patients with Stage 1 melanoma are 96.9% Surveillance Epidemiology and End Results (SEER) or 92.1% American Joint Committee on Cancer (AJCC).32 However, thin melanomas may remain indolent and present with metastatic disease following an extended disease-free interval. Thus, the 10-year survival rates are lower (94.3% and 83.6%, according to SEER and the AJCC, respectively) than the 5 year survival rates.32 The clinical implication is that long-term monitoring of patients with a previous history of thin melanoma is important. Careful assessment for loco-regional and distant metastatic disease should be performed in any patient with a history of invasive melanoma.

Personal history of nonmelanoma skin cancer

Nonmelanoma skin cancer (NMSC), including basal cell carcinoma (BCC) and squamous cell carcinoma (SCC), like melanoma, is strongly related to one’s history of UV exposure, and thus the risk of melanoma in patients who have had BCC or SCC is increased. Better detection of melanomas in patients with NMSC may be due to the effect of increased surveillance. Estimates of the relative risk of melanoma in patients with a history of NMSC have depended on the site of the NMSCs and whether melanomas in situ were included in the analysis. One of the higher estimates was cited by Marghoob et al. who reported a 17-fold increased risk for melanoma in patients with BCC and/or SCC compared with age- and sex-matched controls; and 80% of the melanomas occurred on clothed sites, contrasting the tendency of BCCs to arise on the head and neck.33 A large meta-analysis found the risk to be lower, with a reported relative risk of 4.28 (95% CI: 2.80–6.55).34 Another cutaneous tumor reportedly sharing an association with melanoma is mycosis fungoides. A study of 250 patients with mycosis fungoides found a 15-fold increase in risk for melanoma, possibly because of the effects of immunosuppressive therapy.35

Family history of melanoma

Four decades ago, dermatologists began to appreciate that family history of melanoma is a strong risk factor for the disease.36 Nearly 8–12% of melanoma patients display a familial propensity for the disease.37Family history is a broad term encompassing various degrees of risk depending on the number of relatives and the degree of relatedness to the patient. Family history is often difficult to validate by history alone; patients often report that a relative had skin cancer, but they are not sure which type or do not appreciate the differences between melanoma and NMSC.38 Confirmation of family history by pathology report is desirable. Furthermore, a large majority of the relatives of melanoma patients are entirely unaware of the added risk to them39 and may therefore choose to leave out important information. Several clues should alert the dermatologist or general practitioner to a hereditary familial syndrome, including multiple cases in one side of the family, multiple primary cancers in one individual, and melanoma diagnosis at a young age (< 40 years).

Familial Atypical Multiple Moles and Melanoma (FAMMM) syndrome occurs when two or more first-degree relatives (i.e. parents, siblings, or children) have multiple DN and a history of melanoma.40,41 Germline mutations in CDKN2A as well as CDK4 and ARF, inherited in an autosomal dominant with incomplete penetrance, have been shown to co-segregate with the disease. Melanomas in these families develop at a younger age than usual.42 Although family history is one of the strongest risk factors for the eventual diagnosis of melanoma, it is not an independent prognostic factor, and melanomas in FAMMM patients are clinically similar to sporadic melanomas in terms of overall survival.43 As genetic testing for mutations in CDKN2A is not currently recommended for routine screening purposes, identification of a patient as a member of a high-risk FAMMM family may be determined by taking a meticulous history of melanoma in all first- and second-degree relatives. There are higher rates of CDKN2A positivity in individuals with three or more primary invasive melanomas and/or families with at least one invasive melanoma and two or more other diagnoses of invasive melanoma and/or pancreatic cancer among first- or second-degree relatives on the same side of the family.44 Of note, because of shared p16 susceptibility mutations, an association between familial pancreatic cancer and FAMMM syndrome has been recognized.45 Lynch et al. have reported that around 10% of pancreatic cancer families may also have the FAMMM syndrome,46 thus patients suspected to have FAMMM should undergo frequent skin examinations. Utilizing dermoscopy and photography can assist and facilitate in follow-up examinations.42

Even in the absence of a clear-cut case of hereditary melanoma, a positive family history of the disease remains a strong risk factor, and a patient with at least one first-degree relative with melanoma has a 1.7-fold risk of developing the disease,34 whereas the risk is 9-fold in the setting of two first-degree relatives.47 Begg et al. showed using SEER data that the cumulative risk for melanoma for a male first-degree relative of a melanoma patient increases by 0.8% above the risk of the general population at the age of 50 years, by 3.6% at 70 years, and by 6.9% at 80 years. In female relatives, the corresponding data were 1.3% at 50 years, 4.0% at 70 years, and 6.1% at 80 years.48 Patients from melanoma pedigrees are also at risk for MPMs. One study found that 21% of patients presenting with MPM had a family history of melanoma compared with only 12% of patients with a single primary melanoma.27

Age and gender – men aged above 50 years

Melanoma tends to occur at a younger age as compared with other major cancers, with a median age of 62 years in men and 54 years in women.49 The risk increases with advancing age, especially in men (Fig. 2).50 Compared to age-matched women, melanomas in older men are more likely to be rapidly growing, thick (≥ 3 mm), and predominantly of the nodular type.11,51 Because nodular melanomas tend to be uniform in color, symmetric in shape, and amelanotic, they may evade detection because of lack of ABCD criteria.51 Men over 50 years should be targeted and screened by a dermatologist and receive education on the importance of self-examination.52–54 An analysis of the database of the skin cancer screenings conducted by the AAD during 1992–1994 (n = 242 374 examinations) revealed that men who were over 50 years comprised 44% of those with a confirmed diagnosis of melanoma but only 25% of the total screenees. Similar discrepancies have been reported from Australia, where one large screening (n = 16 383 whole-body skin examinations by primary physicians) resulted in 2302 referrals to dermatologists for suspect lesions, of which only 15.5% were men and 18.2% were 50 years or older. Histopathology confirmed 33 melanomas, of which 23 (70%) were in men and 24 (73%) in participants aged 50 years and above.54

Figure 2.

 Lifetime risk of melanoma for men and women in the United States. Table shows a general increase in lifetime risk with advancing age in men and women. For women, risk increases at the age of 40–59 years but declines at 60–69 years and increases again. Elderly men have the highest risk of melanoma at 1/64. Males are more likely than females to develop melanoma (Adapted from: Jemal A, Siegel R, Ward E, et al. Cancer Statistics, 2008. CA Cancer J Clin. Feb 20 2008)

New, changing, or symptomatic moles

Public education efforts should stir patients with rapidly growing moles to seek medical advice in a timely manner. The current shortage of dermatologists in parts of the U.S., combined with other factors, has contributed to an access differential based on reason for visit, such that patients seeking appointments for a changing mole face increasingly lengthy wait times (median = 26 d) compared to those seeking cosmetic procedures.17,55 With an average growth rate of 0.5 mm per month for nodular melanoma, such long wait times are not ideal. A patient who expresses concern over a new skin lesion or a change in the appearance of an existing mole should be offered an expedited appointment. A study of the signs and symptoms reported by melanoma patients at the time of their initial visit found that an increase in size (50%) and change in color (40%) were the most common complaints/concerns, followed by bleeding (26%), lumps (23%), itching (22%), skin breakdown (18%), and pain (7%).56 In multivariate analysis, factors associated with increased Breslow depth were bleeding (OR = 7.5), pain (3.3), lump (2.2), itching (1.9), and change in size (1.7).56 Appropriate telephone triage has been used to identify patients with truly concerning lesions.57

Prior removal of suspicious moles

Nevi, particularly DN, are known risk markers for melanoma. Patients may need to be specifically asked about prior removal of suspicious “moles.” Some patients are not sure about their history of melanoma; interrogation of prior mole removal and obtaining the pathology report may reveal a history of prior melanoma. We have encountered patients with few DN on skin examination, but with multiple scars that subsequently revealed removal of multiple DN (Fig. 3). For this reason, presence of a scar has been shown to be a risk factor.58 These patients are still at risk for melanoma, as most melanomas in patients with DN arise de novo and the DN only serve as markers of heightened risk.

Figure 3.

 Patient with multiple dysplastic nevi and scars. Scars from prior removals of moles should raise suspicion and be considered a risk factor for melanoma. The scars are likely due to removal of dysplastic nevi

Sun exposure

Solar UV exposure is the most important environmental risk factor for melanoma, and the only one that is potentially modifiable. Those residing at lower latitudes and in sunny areas are generally the most exposed, while those with fair skin are the most susceptible. Questions on skin phototype and sun exposure history are presented in Table 1. Interestingly, the role of sunlight in melanoma development does not follow a linear dose–response relationship. It is specifically a pattern of intense intermittent exposure (vs. a chronic continuous exposure) that confers the greatest risk (RR = 1.46, 1.19–1.79), especially for superficial spreading melanoma.59,60 The intense intermittent pattern of exposure occurs during recreational activities and on vacation, in skin that is not accustomed to the sun. This pattern has also been shown to raise one’s risk for MPMs.61 By contrast, the chronic continuous pattern of exposure, seen in occupational settings, is associated more with actinic keratosis, SCC, and lentigo maligna melanoma.62,63

Table 1.   Skin cancer screening questionnaire
Past history of cancer
   Have you ever been diagnosed with melanoma or any other skin cancer? If so, when?
   Have you ever been diagnosed with breast, pancreatic, brain, or any other cancer?
Family history of cancer
   Has anyone in your family been diagnosed with melanoma or any other skin cancer?
   Name and age at diagnosis?
   Has anyone in your family been diagnosed with breast, pancreatic, brain, or any other cancer?
   Are you a “moley” person?
   Did you have many moles or “beauty marks” as a child?
   Do your moles change over time? Have you noticed any recent changes?
   Do you worry about your moles?
Phenotypic traits
   What is your natural hair color?
   Eye color?
   Have you ever had freckles?
   What is the color of your untanned skin?
   Do you tan easily? Do you burn easily?
   Approximately how many painful sunburns have you experienced in your lifetime?
Sun exposure
   How would you describe your current sun exposure: mild/moderate/intense?
   Does your job require you to be outdoors for most of the day?
   When outdoors on a sunny day, do you:
    wear sunscreen?
    wear clothing/hat to protect yourself?
    seek shade?
    stay indoors during noon?
   How would you describe you childhood sun exposure: mild, moderate, or intense?
   Have you ever been to a tanning salon? If so, how many times and when?
   Do you intentionally seek ultraviolet exposure (sun or tanning bed) so as to get a tan?
Skin health history
   Have you ever had any problems with your skin?
   Were you born with any large “birthmarks”?
   Have you had any moles removed?
   Have you noticed any changes in your skin recently?
Attitude towards prevention
   Do you know how to perform a proper skin self-examination? How often?
   Do you always wear sunscreen when outdoors?
   When was the last time you consulted a dermatologist? What was the reason for the visit?
Are you worried about getting skin cancer, and why?

Sunburn history may be a marker of intense intermittent exposure. While retrospective data on childhood sun exposure and history of burns in childhood are difficult to ascertain, a history of sunburns in childhood may double one’s risk (RR = 1.99, 1.45–2.74) and in adulthood increases it, albeit not dramatically (RR = 1.53, 1.26–1.86).60

Artificial UV exposure

The prevalence of artificial tanning use has been increasing steadily since the 1970s, and each day over 1 million Americans visit a tanning salon.64 The UVA exposure in doses per unit of time for a typical tanning session is significantly higher than the exposure during ordinary outdoor activities or even during sunbathing. A meta-analysis of 19 studies on indoor tanning and melanoma found a relative risk of 1.15 (1.00–1.31) for “ever use” of sunbeds, although a linear dose–response was not seen.65 Exposure to sunbeds before 35 years of age significantly increased the risk (RR = 1.75; 1.35–2.26).

Patients with longstanding severe psoriasis who have received over 250 treatments with psoralen–UVA radiation have a 5-fold increase in the incidence rate of melanoma beginning 15 years after treatment.66 Melanoma can have a long latency period in these patients, and, at 25 years of follow-up, the incidence rate of melanoma is approximately 10-fold higher.67

Chronic immunosuppression

Immunosuppression due to cancer and AIDS has been shown to be associated with melanoma. Melanoma risk is slightly increased in patients with a history of chronic lymphocytic leukemia or non-Hodgkin’s lymphoma, independent of treatment.68–72 Melanoma has also been shown to develop after organ transplants, and accounts for 6% of post-transplantation skin cancers in adults and 15% in children.73,74

Genetic susceptibility

In addition to familial melanoma, as noted above, rare genetic syndromes can predispose to melanoma. Xeroderma pigmentosum (XP) carries a 1000-fold increase in risk for melanoma and NMSC above the general population.75 In addition to regular skin self-examination and strict sun avoidance and protection, XP patients should be followed up by a dermatologist, preferably one with expertise in the disease, on a regular and frequent basis. Interestingly, melanoma has also been found in breast–ovarian cancer families, one study reporting a 2.6-fold risk in carriers of BRCA2 mutations (RR = 2.58, 1.28–5.17).76 Furthermore, an increased incidence of melanoma among female breast cancer survivors and breast cancer among female melanoma patients have been observed.77 Finally, the association between melanoma and pancreatic cancer, mentioned previously, is associated with mutations in CDKN2A.

Risk factors identified during skin examination

Melanocytic nevi are both precursors to and markers indicating an increased risk for melanoma and may be used to identify high risk patients. Although approximately 20–40% of melanomas occur in association with a pre-existing nevus,78 the risk of a particular nevus becoming melanoma is very low. The lifetime risk of any single nevus in a 20-year-old person transforming into melanoma by the age of 80 years is 1/3000 for men and 1/11 000 for women.79 Although prophylactic removal of atypical moles has been advocated by some,80 most believe this to be an impractical and irrational approach because over 65% of melanomas in DN patients arise de novo.81 Observational management for the detection followed by selective removal of lesions suspicious for melanoma with assiduous effort not to lose patients to follow-up is an alternative and more practical approach. A summary of melanoma risk gathered by physical examination follows.

Common nevi

Total nevus count is positively correlated with melanoma risk (Fig. 4).57,82–90 In a meta-analysis, Gandini et al. found that patients with more than 100 nevi have a 7-fold increase in risk for melanoma (Table 2).91 It is important to survey the entire skin surface, as patients tend to give inaccurate mole counts; one study showed poor agreement between self-assessment and physician-assessment of nevus counts (kappa = 0.32).22“Moley” patients should be followed by a dermatologist for complete skin examinations on a regular basis. Photographically assisted follow-up may facilitate detection of new and changing nevi in patients with many nevi.

Figure 4.

 Melanoma risk according to number of common nevi. Studies have validated that total common nevus count is a strong positive predictor of melanoma risk. Earlier studies have generally shown a higher risk as compared with later studies

Table 2.   Risk factors for melanoma
CategoryFeatureRisk (given as relative risk [RR] with 95% CI unless indicated otherwise)
  1. *Estimates are for cases of familial melanoma.

GenotypeXeroderma pigmentosumApproximately one in every five patients will develop melanoma122
CDKN2A35–70 fold37
Suspected in:67% lifetime risk
 – families with multiple melanomas or pancreatic cancer
CDK4 (very rare)Significantly elevated123
BRCA2 – Familial Breast and Ovarian Cancer2.58 (1.28–5.17)76
Suspected in:
 – Caucasians with red hair, freckling phenotype in family
OCA2Slightly increased risk, independent of phenotype
Suspected in:
– patients with a family history of albinism
Personal history of skin cancerMelanoma1–8% of melanoma patients will be diagnosed with another melanoma26
Nonmelanoma skin cancer (including actinic damage)4.28 (2.80–6.55)34
Family historyAny1.74 (1.41–2.14)34
Parent2.40 (2.10–2.72)47
3.2 (2.6–3.9) for North American males48*
4.4 (3.5–5.2) for North American females
Sibling2.98 (2.54–3.47)47
3.2 (2.6–3.9) for North American males48
4.4 (3.5–5.2) for North American females48
Two first-degree relatives8.92 (4.25–15.31)47
Parent with multiple primary melanomas61.78 (5.82–227.19)47
Atypical nevi9001.00
11.45 (1.31–1.60)
22.10 (1.71–2.54)
33.03 (2.23–4.06)
44.39 (2.91–6.47)
56.36 (3.80–10.33)
Common nevi900–151.00
16–401.47 (1.36–1.59)
41–602.24 (1.90–2.64)
61–803.26 (2.55–4.15)
81–1004.74 (3.44–6.53)
101–1206.89 (4.63–10.25)
Phenotypic traits34
Hair colorRed vs. Dark3.64 (2.56–5.37)
Blond vs. dark1.96 (1.41–2.74)
Light brown vs. dark1.62 (1.11–2.34)
FrecklesHigh vs. low density2.10 (1.80–2.45)
Eye colorLight (green, hazel, blue) vs. dark∼1.5
Sun exposureIntermittent2.35 (1.78–3.09)34
Chronic0.98 (0.85–1.12)
Sunburn history2.02 (1.73–2.34)
Tanning SalonEver vs. never, 1.15 (1.00–1.31)65

Atypical nevi

An atypical or dysplastic nevus (DN) is the one that is at least 5 mm in size with a flat component and has at least two of the following: variable pigmentation, irregular asymmetric outline, and indistinct border. DN count is strongly correlated with melanoma risk (Fig. 5).57,82–88,90 Gandini et al. showed that the presence of even a single atypical nevus conferred a high risk of sporadic melanoma, increasing six-fold with presence of five atypical nevi (Table 2).91 Risk of melanoma is high for patients whose nevi display at least moderate histological dysplasia.92,93 Atypical moles are also associated with MPM.30,94,95 One study found that 38% of MPM patients had DN compared with 18% of those with single primary melanoma (P < 0.001).27

Figure 5.

 Melanoma risk according to number of atypical nevi. There is a large body of evidence for clinical dysplastic nevi as melanoma risk markers independent of the total nevus count. Some studies have correlated melanoma risk to the degree of histolgical dysplasia

Congenital melanocytic nevi

Melanoma can develop within any congential melanocytic nevus (CMN) but the risk is the largest for large lesions (> 20 cm). Although the absolute lifetime risk for melanoma developing in patients with large CMN is less than 10%, the relative risk is high and is reported to be between 52 and 465.96,97 Management options include close follow-up, monthly self-examination, and prophylactic removal. If desired by the patient, prophylactic removal should be carried out early in life because nearly 70% of the melanomas developing in patients with large CMN occur before the child’s tenth birthday.98

Small (< 1.5 cm) and medium (1.6–19.9 cm) CMNs have a reported absolute lifetime risk of developing melanoma that ranges 0–5%. The relative risk, however, is estimated to be approximately 10.96 The melanomas that develop in association with smaller CMN tend to originate at the dermo-epidermal junction and present in later life (past puberty).98 Thus, management should include clinical follow-up, monthly self-examination, and prophylactic removal, if desired. As melanomas developing in small CMN occur later in life, the timing of prophylactic excision is not as important as it is for large CMN. Photography and dermoscopy can prove very useful in following these patients.

Changes found on examination

Skin examinations performed by the patient and by the physician are complementary. After evaluating the lesions that the patient may bring to his attention, the physician should thoroughly examine the rest of the skin for new or subtly changing lesions. Total body photography (TBP) assists in detecting new or grossly changed moles by providing overview images of the skin’s surface. It helps to single out those lesions that may warrant further inspection. Although change is a sensitive sign for melanoma, it is not very specific, because benign nevi also change, especially during peak-adolescence and, in women, during pregnancy. In fact, fewer than 3% of changing nevi will prove to be melanoma upon histopathologic analysis.99–101 Thus, it should come as no surprise that the odds of a changing lesion turning out to be a melanoma is far less for younger people as compared with adults (i.e. > 50 years).99 Finally, nevi may also change following intense UV exposure;102–106 however, these changes have been reported to revert to normal after 1 month following cessation of exposure.104

One approach to identify melanoma, while at the same time limiting the removal of changing but benign nevi, is to first use TBP to detect new and changing lesions, and then include dermoscopy and/or short-term mole monitoring to help determine whether a biopsy is warranted. Short-term (i.e. at 3 months) mole monitoring (STMM) utilizes baseline dermoscopic images to detect subtle changes that may be suggestive of malignancy.107 Kittler et al. have shown that most featureless melanomas eventually display melanoma-specific features if followed over time.108 Using dermoscopy, they followed melanocytic lesions lacking any melanoma-specific features and found that, before 4.5 months, only 38% of melanomas manifested melanoma-specific features. This value increased to 55% at 4.5–8 months and to 65% after 8 months. Whereas melanomas showed asymmetrical enlargement, changes in color and architecture, and regression features, nevi grew symmetrically and did not change otherwise. STMM enables the timely identification of these initially featureless melanomas (i.e. lacking melanoma-specific features). The premise behind STMM is that biologically relevant lesions such as melanoma will manifest dermoscopic change within 3 months.106 Research has verified these observations showing that 94% of melanomas manifested dermoscopic change, albeit most of the melanomas still did not reveal any melanoma-specific features, over a 3-month follow-up period.109 A majority of the remaining 6% of melanomas were detected at longer follow-up intervals, but these melanomas proved to be biologically indolent, slow-growing lesions such as lentigo maligna melanoma.110 Altamura et al. also showed that 17% of the dermoscopically changed lesions found with STMM proved to be melanoma, which is significantly better than the rate of 3% based on routine follow-up examination.

Pigment phenotype

Individuals with red hair, fair skin, numerous freckles, and light eyes are considered at moderate risk for melanoma. Red hair, compared with dark hair, carries the highest relative risk (RR = 3.64, 2.56–5.37), while blond hair confers a moderately increased risk (RR = 1.96, 1.41–2.74).34 Patients with high-density freckling have double the risk compared with patients with little or no freckling34 especially when the freckles are found on the upper back. The risk is similarly doubled for Fitzpatrick phototype I (i.e. burns easily, cannot tan) compared with phototype IV (i.e. tans easily, rarely burns).34 Light eye color (i.e. blue, hazel, green) raises one’s risk by approximately 50%.34 Of note, patients with “fair” phenotypes often develop featureless or amelanotic melanomas that may quite easily escape detection. It is for this reason that they should be followed by a dermatologist independent of the presence of other risk factors, and the threshold for biopsy of lesions in these patients should be lower.

Evidence of sun damage

A patient with extensive sun-damaged skin should be assumed to be at a greater risk for melanoma, especially lentigo maligna melanoma.62 Clinical signs of sun damage include uneven pigmentation, solar and labial lentigines, solar elastosis, poikiloderma, cutis rhomboidalis on the back of the neck, and actinic keratoses.

Grading the risk

The strongest risk factors for melanoma are personal and family history of melanoma, presence of dysplastic or atypical nevi, intense intermittent sun exposure, and the skin’s response to UV radiation as manifested at the phenotypic and molecular levels. A summary of the aforementioned risk factors is shown in Table 2. Using advanced statistical methods, mathematical risk models to predict the likelihood of melanoma developing in the setting of various risk factors have been attempted.111–114 Using data from three large prospective studies of 175 000 patients (535 melanoma cases), Cho et al. created one such model, which yielded relative risk estimates similar to those cited by widely accepted studies and a reasonable overall concordance statistic of 60% (Table 3).111

Table 3.   Example of a mathematical risk model for melanoma for white patients
ParameterNo. casesRegression coefficientSEP-valueRelative risk 95% CI
  1. Melanoma incidence model fitted to the combined dataset of the Nurses’ Health Study II and Health Professionals Follow-Up Study (adapted from Cho E, Rosner BA, Feskanich D, Colditz GA. Risk factors and individual probabilities of melanoma for whites. J Clin Oncol. Apr 20 2005; 23(12): 2669–2675).

  2. Risk score is calculated by adding the regression coefficient for the intercept (−9.2523) and for each risk factor. For example, a 60-year-old woman with no family history of melanoma, 3–5 episodes of severe and painful sunburn, 3–5 moles, and red hair will get a score of: −9.2523 + 0.0165(60) + 0.4833 + 0.9882 + 0.7200 = −6.0708. Her estimated relative risk of melanoma is 8.9 (= exp[−6.0708 + 8.2623]) compared with a 60-year-old woman with no family history, no severe or painful sunburns, no moles, and light brown hair (risk = −9.2523 + 0.0165(60) = −8.2623).

Intercept −9.25230.2456< 0.0001  
 Women444   1.00 
 Men910.55270.1273< 0.00011.741.35–2.23
 Age (for 1 year increase) 0.01650.00413< 0.00011.021.01–1.03
Family history of melanoma
 No492   1.00 
 Yes430.78330.1597< 0.00012.191.60–2.99
History of severe and painful sunburn
 No64   1.00 
 1–2 burns1090.33580.15790.03341.41.03–1.91
 3–5 burns1070.48330.15990.00251.621.19–2.22
 6–9 burns890.66830.167< 0.00011.951.41–2.71
 10+1660.85740.1549< 0.00012.361.74–3.19
No. moles > 3 mm in diameter on arms or lower legs
 No227   1.00 
 3–5910.98820.1247< 0.00012.692.10–3.43
 6–9441.04260.1665< 0.00012.842.05–3.93
 10+741.12380.14< 0.00013.082.34–4.05
Hair color
 Dark brown170–0.24380.10310.0180.780.64–0.96
 Light brown213   1.00 
 Red460.720.1641< 0.00012.051.49–2.83

Epidemiological results from the AAD Skin Cancer Screening Program indicate that five factors, remembered by the acronym HARMM, independently increase the likelihood of suspected melanoma in screenees: history of previous melanoma (OR = 3.3, 2.9–3.8), age > 50 years (1.2, 1.1–1.3), regular dermatologist absent (1.4, 1.3–1.5), mole changing (2.0, 1.9–2.2), and male gender (1.4, 1.3–1.5). Five or more risk factors mean “high risk.” The total number of screenees considered at high risk was approximately 6% of the total population, yet they accounted for 14% of presumptive melanoma cases and were 4.4 times more likely to be diagnosed with suspected melanoma than those at the lowest risk (i.e. 0–1 factors).115

Risk models for melanoma generally do not account for phenotypic variation because of geographic location and ethnicity. A case–control study from Vienna assessing conventional melanoma risk factors (family history, phenotype, previous sunburns etc.) found that the “high risk” group accounted for only 40% of the melanoma patients.23 Sixty percent of the melanoma patients in this study were not distinct from the controls (i.e. patients without melanoma, based on these risk factors). This may indicate that conventional risk factors, including skin phototype, hair and eye color, and history of sunburns, to name a few, do not adequately distinguish cases from controls in Vienna, because the general population is, on average, fair skinned. By contrast, the heterogeneous nature of the U.S. population enables conventional phenotypic risk factors to be adequate in melanoma risk stratification. For example, a 55-year-old man with red hair and fair skin whose father had melanoma is probably at higher risk than a 40-year-old Hispanic woman with no family history of the disease. The model, however, is crude, and until molecular profiling and genotyping are possible, we rely on risk factors to weed out the patients who are at higher risk relative to the population at large.

Future risk stratification

Many exciting developments in the dermoscopic characterization of nevi are underway. Scope et al. have shown that roughly 80% of individuals have one of four predominant global dermoscopic patterns that describes most of their nevi, with the reticular pattern seen half of the time.116 Thus, physicians may compare a suspicious mole with the patient’s predominant mole patterns. Furthermore, Zalaudek et al. have shown that the global dermoscopic pattern varies with skin type and age in whites, such that a lesion that defies the pattern expected according to the patient’s skin type and age may warrant investigation.117 Lipoff et al. observed a complex dermoscopic pattern (i.e. globular and reticular patterns in the same lesion) three times more frequently in melanoma patients compared with age- and sex-matched controls (Fig. 6).118 Taken together, these studies represent advances in risk assessment on the basis of dermoscopic patterning with two major points. First, a complex dermoscopic pattern within individual nevi may be a robust marker of melanoma risk. Secondly, greater variability in mole pattern within an individual (e.g. a patient who has dark moles with a globular pattern, other pink moles with homogenous pattern and some brown moles with reticular-globular pattern) may also imply greater risk. Examination of nevi in their context, both on the dermoscopic and clinical levels, can help weed out lesions that are truly “atypical,” not just histologically but atypical by virtue of appearing different from the surrounding nevi. A small trial recently demonstrated that the “ugly duckling sign” had 90% sensitivity for detecting melanoma, further validating this in-context approach.119

Figure 6.

 Dermoscopic patterns in nevi. Four dermoscopic patterns have been recognized: (a) reticular or network, (b) globular, (c) homogeneous (neither globular nor reticular), and (d) complex (globular and reticular). A recent case–control study found that a complex pattern was found three times more frequently in melanomas than age- and sex-matched controls

Recommendations for evaluation and follow-up of the high-risk patient

High-risk members of familial melanoma families should undergo total skin examinations every 3–12 months starting during adolescence until nevi become stable, and then annually thereafter.120 TBP and dermoscopy have been shown to aid in the management of high-risk patients by obviating the need for many biopsies and by enabling early detection of melanoma.99 These patients should also be taught how to perform monthly skin self-examinations and be encouraged to involve a partner to screen areas such as the scalp and back that are difficult to visualize. The importance of self-examination must be emphasized. TBP and dermoscopy should also be recommended for the following groups: patients with complex skin examinations, i.e. patients with numerous moles, particularly multiple DN, patients who underwent numerous nevus excisions, and, finally, patients with high anxiety regarding developing melanoma.

A patient presenting with many atypical moles (> 5) is considered to be at moderate risk and should be followed every 6–12 months depending on degree of clinical heterogeneity in the nevi. If all atypical nevi look similar to one another, then annual visits should suffice; however, if the nevi look different from each other and/or have a complex dermoscopic pattern (Fig. 6), then follow-up intervals of 6 months or shorter should be considered. At each visit, changes in lesions can be identified by comparing the examination with baseline photography and these lesions can then be further evaluated via dermoscopy to determine if a biopsy is necessary or not. Lesions that have changed but lack any specific dermoscopic features of melanoma can either be excised or subjected to STMM. If subjected to STMM, any dermoscopic change observed at the 3 months follow-up visit, even if no melanoma-specific features are seen, warrants a biopsy.

Patients with a previous melanoma should be followed more frequently during the first 5 years after diagnosis, and then annually. To promote prevention and early detection, all first-degree relatives of the melanoma patient should be screened starting in their teens and taught sun protective behavior and how to perform a skin examination. A recent randomized controlled study found that siblings of melanoma patients who participated in an educational skin cancer program were more likely than controls to perform skin self-examination (OR = 1.76, 1.06–2.91).121 Counseling patients on sun safety practices should recommend judicious sun exposure, while enjoying an active lifestyle with sun protective measures.


Singling out patients at high risk for melanoma is a challenge that may save many lives. It is imperative that dermatologists we continue to take a proactive approach to identify and target patients who are at high risk for melanoma for screening. Questionnaires, history, and the physical examination can and should be used to identify high-risk patients. A multi-modal approach to intensive surveillance utilizing TBP and dermoscopy can help patients and detect subtle changes that may be indicative of melanoma. High-risk patients should be educated on sun safety practices and the importance of regular skin self-examination. Family members should similarly be targeted for melanoma screening. Additional studies are needed to assess the impact of skin cancer screenings in curbing the overall mortality from this disease.