Please note: In 2003, the CTF updated its Grades of Recommendations to include an "I Recommendation" for situations where insufficient evidence exists to allow a recommendation to be made.  (Formerly, these situations were captured under a "C Recommendation".)  This change is not retroactive, and all "C Recommendations" made prior to 2003 have not been reevaluated in light of the new "I" recommendation grade.  For a discussion of these recommendation grades, please link to the 2003 article in the Canadian Medical Association Journal here.

Prevention of Skin Cancer

Overview
Routine screening for skin cancer by primary care providers is not recommended for the general population (C Recommendation). For individuals with significantly increased risk (family melanoma syndrome, first degree relative with malignant melanoma) it would seem prudent to monitor regularly by physical examination and dermatologists may be the most appropriate assessors. Currently there is insufficient evidence to recommend either for or against counselling patients to perform periodic skin self-examinations. Clinicians should advise patients with increased sun exposure or at increased risk of skin cancer to protect their skin from solar rays (B Recommendation). Persons with a prior history of solar keratosis who cannot avoid sun exposure should use sunscreens that block both UV-A and UV-B radiation (B Recommendation), although there is insufficient evidence to support its use for the prevention of squamous cell carcinoma, basal cell carcinoma or malignant melanoma.

Burden of Suffering
In Canada in 1992, there were approximately 51,000 new cases of skin cancer. Of these, 47,000 were either basal cell carcinoma (BCC) and squamous cell carcinomas (SCC), commonly referred to as nonmelanomatous skin cancer (NMSC).<1> These skin cancers rarely metastasize and are generally easily treated. However, extensive lesions can cause marked tissue destruction and disfigurement as well as functional impairment if the tumours are not detected early. An increased risk for NMSC is associated with: a personal history of NMSC, older age, light eyes, skin or hair, poor ability to tan, a high density of freckles, and a marked cumulative lifetime exposure to sun.

Malignant melanoma (MM) differs considerably from these other two skin cancers. While less common, it is far more lethal. In 1993, there were an estimated 2,950 new cases of MM in Canada and in that same year, there were 560 deaths from MM. Using 1989 data, it is estimated that MM ranks 14th amongst all cancers in potential years of life lost. Over the last two decades there has been an increase in age-standardized incidence (6% in men, 4% in women) and an increase in mortality rates (3.4% in men, 1.6% in women). A number of attempts have been undertaken to identify important risk factors associated with MM. These have included melanocytic, precursor or marker lesions (e.g. atypical moles, certain congenital moles), increased numbers of common moles, immunosuppression, and a family or personal history of skin cancer, particularly MM. A recent case-control study in Ontario identified four host risk factors associated with MM: hair colour (red haired individuals are at greater risk than those with black hair), freckled and nevus density (greater risk for those with large numbers), and a propensity to sunburn after repeated sun exposure compared to those who tan.

Some individuals are at markedly increased risk for MM. Individuals with the rare condition of "familial atypical mole and melanoma" (FAM-M) syndrome, show an increased risk of 100-fold or more. Concerns in the past have been raised around the identification of risk in individuals solely with personal or family history of extensive dysplastic nevi. However, both the clinical and pathological diagnoses of dysplastic nevi show considerable inter-observer variation, making accurate diagnosis difficult. Hence, one must be cautious to guard against inappropriate labelling.

The etiology of both NMSC and MM have been linked to ultraviolet light, particularly UV-B light. The population studies evaluating the association between NMSC and sun exposure are consistent in supporting the hypothesis that cumulative sun exposure over a lifetime is a key etiologic component. The role of ultraviolet light in the etiology of MM has received considerably more attention and debate.<2-6> Depending on the choice of confounding variables, case-control studies have shown association in conflicting directions between episodes of sunburn (particularly in early life), and the subsequent development of MM in later life. While not yet conclusive, these data plus data from migrant studies lend support to the concern that exposure to extensive sunlight in earlier years may play an etiologic role. The results, however, are not yet conclusive. Unlike NMSC, it appears as though the association with UV-B sunlight is one of "intermittent exposure" rather than cumulative exposure.

Maneuver
There are two basic approaches to skin cancer prevention, primary prevention including avoidance of sun during peak hours, use of protective clothing, and use of sunscreens; and secondary prevention, which relies on the early detection of MM or NMSC either by the patient themselves or by a physician. The primary prevention maneuvers will be discussed in the section on effectiveness.

Physical Exam by Physicians
The detection of a suspicious lesion constitutes a positive screening test which then needs to be confirmed by skin biopsy. Studies have not been able to identify the true sensitivity and specificity of skin examination because for the most part only individuals testing "positive" have gone on to biopsy of the lesions. One study has attempted to estimate sensitivity using population incidence rates as an estimate for false negative rates and calculated sensitivities of 97% for MM, 94% for BCC and 89% for SCC.<7> The use of population incidence rates to calculate false negatives, however, is suspect in this study because over 78% of those screened had two or more risk factors for skin cancer and, hence, are not representative of the general population. In individuals with positive examinations at screening clinics, studies have indicated histologic confirmation at the rate of 40% for MM, 43% and 57% for BCC, and 14% and 75% for SCC.<7,8> The likelihood of histologic confirmation for MM in examinations by dermatologists in skin clinics ranges from 38-64% and from 72-84% for examinations by skin cancer specialists. In a randomized community study evaluating screening by expert dermatologists, histologic examination confirmed the clinical diagnosis of SCC in 38% of the cases and of BCC in 59%.<9>

Primary care physicians and other generalists who have not been trained in dermatology tend to make fewer correct diagnoses in studies using colour photographs.<10-12>

The yield in screening for skin cancer is influenced by the proportion of body surface examined. For MM, only 20% of the lesions normally occur on exposed body surfaces in contrast to 85 to 90% of NMSC. It has been estimated that the detection of MM would increase two to six times with total body skin examination (TSE). There are no data regarding the appropriate frequency of maneuvers such as TSE and as a result, those who advocate this procedure recommend annual or biennial intervals based solely on clinical judgement. This maneuver can be markedly influenced by poor patient compliance, a concern raised in one study where only 4.2% of the patients returned for a one year follow-up for TSE.<13>

There are no identified serious adverse effects associated with TSE and while theoretically complications can arise with biopsy, the actual rate of complications is small. There are, however, no controlled studies to evaluate either the detection rate or any adverse effects associated with TSE.

Patient Self-Examination
This is an option which some have advocated particularly using a seven point checklist to evaluate skin lesions. One which evaluated a seven point checklist found it had a sensitivity of 71% and a specificity of 99% with a predictive value of 7% for MM diagnosis, using a dermatologist’s clinical diagnosis as the gold standard.<14> There are no data which appropriately evaluate the ability of patients to accurately detect suspicious lesions in general nor are there data regarding the accuracy of periodic skin self-examination or the efficacy of instructions in reducing self-examination errors.

Effectiveness of Prevention and Treatment

Early Detection of Lesions
Basal cell and squamous cell skin carcinomas are very common but are slow growing and rarely metastasize. While early identification and treatment of early lesions might reduce morbidity and disfigurement, no studies exist which have evaluated such early detection efforts.

The principle potential benefit of routine skin examination lies in discovering early MM. However, the sensitivity and specificity of skin examination by primary physicians and the optimal frequency of such examinations is unknown. Moreover, there are no randomized controlled trials evaluating early detection and subsequent treatment. Hence, one cannot rule out the impact of lead time and length biases on studies which employ a before/after design or use inappropriate control groups. The second challenge arises because of the relatively low lifetime risk (1%) of MM in the general population. Hence, 99% of patients from the general population would be examined with routine screening and would never be found to have MM. This raises questions, as yet unevaluated, regarding the impact of potential adverse effects as well as the cost-benefit ratio of early detection efforts for MM in the general population.

At the population level, education campaigns have been mounted to encourage MM screening by primary care providers, particularly in Scotland. One study found a trend towards a reduction in the thickness of tumours (p<0.05) and a trend (which was not statistically evaluated) showing a decrease in mortality in women.<15> However, there was no control group and the before/after design does not allow important confounding factors to be ruled out.

One of the difficulties in before/after studies arises when the thickness of MM is used an outcome. Studies have indicated that survival and the likelihood of recurrence after resection is correlated to the thickness of lesions. A MM <1 mm thick is associated with an eight year disease-free survival for 90% of patients compared to a rate of 74% for lesions of 1-2 mm thick. While this association exists as a prognostic variable in studies without control groups, one cannot determine whether the identification of thinner lesions is a result of early detection efforts or simply represents a natural trend in the population. Indeed, some data indicate that in certain geographic regions, there is a population trend towards thinner lesions, a phenomenon which would appear to exist quite apart from early detection efforts.

Likewise data for individuals who may be at high risk are limited. Two large case series of individuals with atypical moles, screened regularly by dermatologists indicated a tendency towards the detection of thinner lesions. Likewise before/after studies for individuals with the FAM-M syndrome and in individuals with prior MM, report that regular screening by dermatologists results in the detection of significantly thinner tumours compared to historical and other non-concurrent controls.

Primary Prevention
The approaches to the primary prevention of skin cancer include the following: limiting exposure to sunlight (by sun avoidance or by wearing protective clothing) or by applying sunscreen preparations. Such an approach has been promoted through public education campaigns including the "slip, slap, slop" program popularized in Australia (slip on protective clothing, slap on a hat and slop on sunscreen).

There are no randomized controlled trials or other appropriate comparative studies to evaluate the avoidance of sun exposure or the use of protective clothing to prevent MM and NMSC. The strongest evidence available to support what may be a prudent approach, comes from the literature which identifies an association between ultraviolet light and MM and NMSC.<2-6> Such studies, however, do not include prospective cohort studies or randomized controlled trials and do not directly address the specific prescription of avoiding sun or using protective clothing. The etiologic evidence, however, can be used to argue for the prudence of such an approach.

The evidence for the use of sunscreens is less clear cut with the exception of their ability to reduce the rate of recurrence or development of new solar keratoses. A recent randomized controlled trial evaluated the regular use of sunscreens which block UV-A and UV-B in a population of individuals over age 40 who had previous solar keratoses.<16> Over a six month period, the mean rate of solar keratoses in the control group increased by 1 per subject and decreased by 0.6 per subject in the sunscreen group. While solar keratoses are considered a precursor to squamous cell carcinoma, the risk of progression in any one year is less than 1 per 1,000 and of all squamous cell carcinomas, only 60% arise from previous solar keratoses and 40% of SCC arise from normal skin. Hence, the evidence from this study suggests sunscreen may have some impact on subsequent SCC, but the actual clinical impact is unclear. Hence, there is evidence for the prevention of solar keratoses but there is debate about the appropriateness of this as an intermediate outcome measure for squamous cell carcinoma in terms of its clinical significance. There are no randomized controlled trials of the benefits of the sunscreen in BCC or MM. However, concern has been raised by the findings from case-control and cohort studies which have indicated either no effect or a significant increased risk of BCC and MM in sunscreen users.<5,17,18> While debate continues around the potential effect of residual confounders on these results, the concerns remain unanswered. Concerns have also been raised regarding the importance of blocking UV-A as well as UV-B.

Any of the side effects of sunscreen are generally mild to moderate and occur in only 1 to 2% of users of sunscreen but these include contact and photocontact dermatitis, contact urticaria, and comedogenicity, although these are readily reversible with discontinuing the use of sunscreens.

Counselling
There are few data on the effectiveness of counselling patients to protect themselves from sunlight. One before/after study evaluating counselling at the time of skin cancer removal and on a yearly basis thereafter, reported increased use of protective clothing and sunscreen and reduced deliberate tanning in a 2-6 year follow-up.<19> The follow-up, however, involved only the two-thirds of patients who complied with follow-up visits and it was impossible to determine how much of the effect of the study was the result of surgery alone. Evidence from before/after studies indicates that public education can increase knowledge and beliefs about the health risk of sunlight, but it is not clear that individuals act on their knowledge and beliefs in this specific situation. Community and work site education interventions to reduce the risk of skin cancer, including one with a concurrent control group, have demonstrated significantly increased use of sun protection measures such as hats, shirts, and staying in the shade after the intervention. Whether the results of such education interventions can be generalized to physician counselling is not known.

Recommendations of Others
In 1989, the U.S. Preventive Services Task Force recommended routine screening for skin cancer for persons at high risk and that clinicians advise all patients with increased outdoor exposure to use sunscreen preparations and other measures to protect their skin from ultraviolet rays. There was no evidence for or against counselling patients to perform skin self-examination.<20> These recommendations are currently under review.

The (U.S.) National Institutes of Health (NIH) Consensus Panel recommended regular screening visits for skin cancer and patient education concerning periodic skin self-examinations. The NIH Consensus Panel also recommended that some family members of patients with MM be enrolled in surveillance programs. The American Cancer Society recommends monthly skin self-examination for all adults, and physician skin examination every three years in persons 20 to 39 years old and annually in persons over 40 years old. The Academy of Dermatology and the National Cancer Institute (U.S.) recommend regular screening visits for skin cancer and patient education for skin self-examination. These same agencies plus the American Cancer Society and the American Medical Association recommend patient education concerning sun avoidance and sunscreen protection.

The Canadian Dermatologists Association advocates the use of sunscreens of SPF 15 or higher for routine use by individuals who spend time outdoors exposed to the sun.

Conclusions and Recommendations
Routine screening for skin cancer by primary care providers using total-body skin examination is not recommended for the general population. Clinicians should remain alert for skin lesions with malignant features (i.e. asymmetry, border irregularity, colour variability, diameter greater than 6 mm, or rapidly changing lesions) when examining patients for other reasons, particularly in those with established risk factors. Such risk factors include clinical evidence of melanocytic precursor or marker lesions (i.e. atypical moles, certain congenital moles), large numbers of common moles, immunosuppression, a family or personal history of skin cancer, substantial cumulative lifetime sun exposure, intermittent intense sun exposure or severe sunburns in childhood, or light skin, hair, and eye colour, freckles, or poor tanning ability. Appropriate biopsy specimens should be taken of suspicious lesions (C Recommendation).

Currently, there is insufficient evidence to recommend for or against counselling patients to perform periodic self-examination of the skin. Clinicians may wish to educate patients with established risk factors for skin cancer (see above) concerning signs and symptoms suggesting cutaneous malignancy and the possible benefits of periodic self-examination (C Recommendation).

Persons with Family Melanoma Syndrome are at substantially increased risk for malignant melanoma. Clinicians examining these patients should be particularly alert to skin lesions with malignant features and should consider referral to skin cancer specialists for evaluation. For this very select subgroup there is fair evidence to offer total body skin examination (B Recommendation).

Clinicians may find it prudent to counsel persons and parents of children with established risk factors for skin cancer (including those with light skin, eyes and hair, or poor ability to tan), to avoid excessive sun exposure, especially between the hours of 10:00 a.m. and 3:00 p.m., and to use protective clothing such as shirts and hats when they are out in the sun. This recommendation is based on the etiologic evidence for UV exposure, the potential for large health benefits, low cost, and low risk of adverse effects from such counselling, even though the effectiveness of counselling is not well established (B Recommendation).

The routine use of sunscreens that block both UV-A and -B radiation is recommended for persons with prior evidence of solar keratosis who cannot avoid sun exposure (B Recommendation). There is insufficient evidence to recommend for or against counselling patients to use sunscreens to prevent malignant melanoma or basal cell carcinoma (C Recommendation).

Unanswered Questions (Research Agenda)
There is a need for continued research into the etiology of malignant melanoma. There is as well a need for further identification of individual risk factors to allow targeting of early detection and primary prevention efforts.

There is also a need for the proper evaluation of sun avoidance and the use of protective clothing as well as the effectiveness of counselling individuals to comply with these behaviours.

Finally, there is a need for further evaluation of the effectiveness of sunscreens, particularly in the prevention of malignant melanoma; and there is a need to evaluate the effect of counselling patients on the use of such agents should they be demonstrated to be effective.

Evidence
The literature was identified through a MEDLINE search for 1988 to March 1993 using the search words cancer, skin neoplasm, melanoma, dysplastic nevus, sunscreening agents, isotretinoin, and sunlight. Additional literature was identified by reviewing the citations used in articles identified in the MEDLINE search. This review was initiated in 1993 and the recommendations finalized by the Task Force in March 1994.

Full Citation
Feightner J.W. Prevention of skin cancer. In: Canadian Task Force on the Periodic Health Examination. Canadian Guide to Clinical Preventive Health Care. Ottawa: Health Canada, 1994; 850-59.