Epidemiology of Skin Cancer

Melanoma incidence

Melanoma incidence has increased rapidly over the past 65 years. Between 1935 and 1939, the incidence of melanoma in Connecticut was 1.0 per 100,000 (age-standardized, 1970). In 2006, melanoma incidence had increased to 20.9 per 100,000 (age-standardized, 1970) ( Fig. 5.1 ). Similar trends have been noted in the SEER registry. In 1973, the melanoma incidence rate was 6.8 per 100,000, but by 2006, this rate had climbed to 21.1 per 100,000. It is estimated that in the US in 2010, there will be 68,130 new cases of melanoma diagnosed and an additional 46,770 cases of melanoma in situ. The increase in invasive melanoma incidence was 6.1% per year between 1973 and 1981, and 1.6% per year between 1997 and 2006 (based on rates age-adjusted to 2000 US standard population). It is currently estimated that melanoma will be the fifth most common cancer (other than keratinocyte carcinoma) diagnosed in men, with 38,870 cases (5%), behind prostate, lung, colon, and urinary bladder. In women it is estimated to be the seventh most common cancer, with 29,260 cases (4%), behind breast, lung, colon, uterine, thyroid, and non-Hodgkin lymphoma. Melanoma is the fourth most common cancer in Australia, New Zealand and Sweden, the tenth most common in Scandinavia, and the eighteenth most common cancer in most of the United Kingdom (England, Scotland and Wales) (other than keratocyte carcinoma).

Studies continue to examine incidence and mortality trends to further determine the contribution of increasing age or birth cohort (all individuals born within a specific time period, and who subsequently may share similar early exposures and experiences). There have been attempts to distinguish a ‘cohort’ from a ‘period’ effect ( time trends are influenced by more recent exposures or events, such as changes in diagnostic criteria or cancer screening). Melanoma incidence, however, has generally been found to follow cohort patterns of changes. For example, assume incidence started to increase. With a cohort effect, this change would be noted first among more recent birth cohorts (younger people) and would not be observed in prior birth cohorts (older people). A change in middle-age incidence first would be noted when these more recent birth cohort populations reached middle-age.

It has been suggested that a ‘leveling off’ of the melanoma incidence is occurring in many countries. In the US, more recent generations of men have similar incidence rates compared to prior generations, even though these rates are still increasing in these older generations of men. However, incidence appears to be increasing in more recent generations of women. This is consistent with a cohort effect ( Figs 5.2 and 5.3 ). Incidence also appears to have leveled off in Australia, especially in younger cohorts born after 1960, supportive of a birth cohort effect. Incidence appears to even have fallen significantly in the last 20 years in young women (age groups 15–34 and 35–49) in New South Wales (annual percent change −3% and −0.9% respectively). Trends in melanoma incidence in Europe (1953–1997) show a plateauing of incidence rates in younger age groups in Scandinavian countries, whose rates are expected to remain stable or decrease further in the future, though these trends in incidence were less distinct in young people in other areas of northern and western Europe.

The stabilization of melanoma incidence in recent cohorts may be related to the educational programs implemented in the last 35 years. Increased quality and potency of sunscreens in common use, albeit at times ineffective use, might suggest reduced exposures. However, the prevalence of sunburns, at least in the US, appears to be increasing and there is continued migration to sunnier climates and exposures to indoor tanning as well. Regardless of the apparent recent trend toward stabilization in the US and several other countries, the incidence rate of melanoma has increased faster than the mortality rate. Improved surveillance may have increased detection and resulted in the potential surgical removal of earlier ‘cancers’ that may never have progressed to become lesions of clinical significance, or may have increased removal of potentially fatal lesions at a curable point in their evolution, or a combination of both.

Because the frequency of skin cancer is significantly lower in non-white populations, epidemiological information is more limited. SEER provides detailed estimates of incidence by race only for blacks. In 2006, SEER data showed an incidence rate in whites of 25.6 and in blacks of 0.9 per 100,000 (age-standardized, 2000). Melanoma incidence from 2002 to 2006 was: white Hispanic, 4.6 cases per 100,000; Asian/Pacific Islander, 1.4 cases per 100,000; American Indian/Alaskan native, 3.3 cases per 100,000; and black, 1.0 cases per 100,000 (age-standardized, 2000).

SEER incidence data indicate that melanoma incidence in the US trends upward with age ( Fig. 5.4 ). Melanoma incidence peaks at 80–84 years of age at a rate of 82.2 cases per 100,000. Similar age-trends have been seen in other countries, including Australia, Italy and the Netherlands.

In the US, melanoma is more common in men than in women. In 1973, the incidence rates were 7.3 per 100,000 in males and 6.4 per 100,000 in females. In 2006, the incidence rate had risen in men and women, to 26.1 and 17.6 cases per 100,000, respectively ( Fig. 5.5 ). Gender differences have been noted in other countries, including Australia, where in 2005, the age-adjusted incidence in males was 60.9 cases per 100,000 population while the female incidence was 42.5 cases per 100,000 (age-standardized, Australia, 2001) (see Fig. 5.6 ).

Gender differences appear to vary with age as well. In 2002 to 2006, in SEER registrants under age 45, melanoma incidence was higher in women, whereas after age 45, men had higher incidence rates. Between 2002 and 2006 in the US, the age-adjusted incidence in males over age 65 was 105.5 per 100,000 while the incidence in elderly females was 41.0 per 100,000 (age-standardized, 2000).

Incidence rates vary substantially worldwide ( Table 5.1 ). New Zealand and Australia have the highest incidence. In Queensland, Australia, the incidence rate is 41.1–55.8 cases per 100,000, more than four times that in the US SEER-14 program (10.5–14.8 cases per 100,000) (age-standardized, world population). These high rates have led to a large public health and economic burden which has motivated extensive and successful public health campaigns. Incidence in European nations ranges from 2.4–2.6 cases per 100,000 in Bulgaria to 11.5-19.6 per 100,000 in Switzerland. In Africaand Central and South America, incidence is low. It is also low in Asia, with India and China reporting less than one case per 100,000. In Israel, incidence varies with ancestry and place of birth. It is higher in Israeli Jews (10.5–12.2 cases per 100,000) than non-Jews (0.8–1.4 cases per 100,000).

Trends in international incidence suggest that melanoma is continuing to increase, Between the mid 1960s and the mid 1980s, the average annual percent increase in melanoma incidence generally ranged from 3% to 6%, with the highest rate of increase noted in white residents of Hawaii, who had over a 9% increase. Kricker and Armstrong examined international data for trends in age-specific rates, and found that incidence rates have stabilized or begun to fall in young people (less than age 55) in some populations, including Denmark, Canada, the US, Australia, New Zealand, Norway and the UK, but are continuing to rise in other countries such as Poland, Spain and Yugoslavia. There is evidence suggesting a latitudinal effect world wide in melanoma incidence, with generally higher incidence reported nearer the equator. In New Zealand data from 1968 to 1989, a latitudinal trend from north to south existed for each gender and across time. New Zealanders living in the northern region of the country may have at least a 37% higher incidence than those living in the south. There is a higher incidence in Scandinavian compared to Mediterranean countries, which is attributable to gradients in sun sensitivity in these populations.

Melanoma mortality

Nearly 75% of skin cancer deaths in the US are attributable to melanoma. Melanoma mortality has increased substantially in the US over the last 30 years, although it is now stabilizing. Mortality increased 4.3% annually in the white population between 1973 and 1977, 1.5% annually between 1977 and 1990, and 0.2% annually between 1990 and 1999 and from 1997 to 2006. It is estimated that in 2010, 8700 Americans will die of the disease.

An analysis of World Health Organization (WHO) Cancer Mortality Data Bank data (examining Australia, Canada, Czechoslovakia, France, Italy, Japan, UK, US, and a combined Denmark, Finland, Sweden and Norway) examined mortality rates and recent trends from 1960 to 1994. In 1960, some of the lowest mortality rates for the 30–59 age group were seen in France, Italy and Czechoslovakia (fewer than 0.5 deaths per 100,000 (world standard population)). However, over the last 30 years the highest rates of increase in mortality were found in these same three countries, with death rates increasing annually by 9–16%. Age-adjusted mortality in Japan remained low (less than one death per 100,000 in all age groups) over the entire time period.

There are marked differences in mortality with increasing age in the US. The mortality rate in 2002–2006 for men and women younger than age 65 was 1.7 and 0.9 deaths per 100,000 respectively. However, for those aged 65 and older, the mortality rate for men was 19.5 per 100,000 and for women was 7.6 per 100,000. The highest mortality is seen in men over age 80: 128.9–134.9 cases per 100,000.

There is a higher mortality rate in men compared with women of the same age in the US. In men, there was a significant increase in annual percent change in melanoma mortality of 2.4% between 1975 and 1987, 0.7% from 1987 to 1998, and 2.0% from 2002 to 2006, with no significant change during 1998–2002. In women, the annual percent change in melanoma mortality was 0.8% between 1975 and 1988; however, in recent years (1988–2006) it appears to be decreasing (−0.6%) significantly ( Fig. 5.7 ). It is estimated that there will be almost twice as many deaths in men (5670) as in women (3030) in the US in 2010. The mortality rate in Australia is also higher for men than for women ( Fig. 5.8 ).

The age-adjusted death rate from melanoma in 2006 was 3.1 per 100,000 for the white population and 0.4 per 100,000 in the black population. The 5-year relative survival rate of melanoma is subsequently lower in blacks compared to whites (all stages: 77% vs. 91%). Examination of stage of diagnosis reveals that a higher percentage of blacks diagnosed with melanoma between 1999 and 2005 had regional or distant disease-stage than did the white population (34% vs. 12%).

The mortality rate appears to be stabilizing in portions of the world, including the US, Australia and parts of Europe. Unlike incidence, US mortality has declined in more recent birth cohorts. Death certificate or histopathology criteria changes are not felt to have a significant impact on this trend. The stabilization of melanoma mortality in more recent cohorts may be the result of a combination of a slower increase in melanoma incidence and a lower case-fatality, presumably due at least in part to earlier detection ( Figs 5.9 and 5.10 ).

Cohort analysis of WHO mortality data have demonstrated several different patterns. In Australia, the US, and the Scandinavian countries, there appears to be an increasing mortality rate for the generation born prior to 1940, followed by a decrease in mortality in younger cohorts. In the UK and Canada, the rates increase in generations born between 1920 and 1950, with a stabilization in more recent cohorts. In France, Czechoslovakia and Italy, there has been a step increase in melanoma mortality that appears linear with little change in trend. In Sweden, mortality has plateaued in men in the last 10–15 years and slightly decreased in women (−2.3%). Mortality has decreased in women among all age groups but in men only for those younger than 60, and analysis of trends has been suggestive of a period effect.