  |
|
|
||||||||
|
||||||||||
| Journal of the New Zealand Medical Association,
22-September-2006, Vol 119 No 1242
|
||||||||||
|
Mary Jane Sneyd, Brian Cox
The incidence and mortality rates of malignant melanoma have
shown large increases in New Zealand over the past 30 years.1,2 Similar trends
have also been observed in many other developed countries,3,4 but New Zealand
and Australia still have the highest incidence rates in the world. While smaller
increases in both incidence and mortality have been observed in people born
after about 1950 in New Zealand, they have continued to increase for older
people and in particular those aged over 60 years.5
Until the Cancer Registry Act 1993 came into force in New
Zealand in July 1994, the Cancer Registry had been based primarily on public
hospital records and so had missed many tumours excised outside hospital and
many patients treated privately. Several studies6–8 have shown that
incidence rates of melanoma estimated directly from pathology reports were
considerably higher than was apparent from registered cases provided by the New
Zealand Cancer Registry. Since the introduction of the Cancer Registry Act 1993,
statutory notification has greatly increased the numbers of melanomas
registered.
Multiple strategies will be required to combat the
increasing burden of melanoma. A cancer control strategy encompasses all aspects
of cancer: prevention, screening, early detection, diagnosis, treatment,
rehabilitation, and palliative care. It also includes cancer surveillance and
research.9 The results in this report focus on the likely effects of four
different interventions (prevention, screening, early diagnosis and treatment)
in reducing the burden of melanoma in New Zealand.
MethodsCancer control plans and management guidelines, both
from New Zealand and overseas, were reviewed to identify potential activities
that could reduce the burden of melanoma in New Zealand. Their effects on the
incidence of and mortality from melanoma in New Zealand were estimated.
Service provision depends on actual numbers of cases or
deaths, not rates of disease or death, whereas comparisons over time or place
require standardised rates of disease. Absolute numbers and age-standardised
rates,10,11 standardised to Segi’s world population, have been presented.
Registration rates are used as the closest approximation to the national
incidence rate available in New Zealand.
Estimates of projections in incidence and mortality
have been used to estimate the future burden of disease. The projection models
separately relate cancer incidence and mortality data to three time dimensions:
age, period, and cohort.12 Estimates of the future burden used the average
projection from a set of models rather than relying on any individual model
alone.12
The base year for the estimations of changes in
incidence and mortality resulting from cancer control activities was the data
published for 2002.10,11 The number of deaths prevented by cancer control
activities has been calculated for 2002 as if the nominated interventions were
already in existence. This, in effect, standardises their impact to the 2002
calendar year.
Life expectancy tables were used to calculate
person-years of life lost from death due to melanoma.
Population attributable risk percent (PAR%) was
calculated in the usual way:
...where Pe = proportion exposed in the population and
RR = relative risk for the exposure of interest.
Further details of the methods are available in another
report of this work.13
ResultsIncidence and
mortality—In 2002, melanoma accounted for 1842 new cancer
registrations, of which 933 were in men and 909 in women (Figure 1). The
registration of melanoma increased from 1037 new registrations in 1993 to 1487
in 1994 and showed a peak at 1759 in 1995 due to the effects of compulsory
cancer registration. Registrations of melanoma then decreased in 1996 and 1997,
but in 2001 and 2002 the numbers of registrations increased, to approximately
the same level as in 1995 for women and slightly higher than in 1995 for
men.
In 2002 there were 235 deaths from melanoma (Figure 1); 149
in men and 86 in women. In this year, death from melanoma accounted for 3.6% of
all male cancer deaths and 2.3% of all female cancer deaths. Numbers of deaths
from melanoma have increased since 2002, to 174 in men and 111 in women in
2003.
Figure 1. Melanoma registrations and deaths by
year
 Age-standardised registration rates (ASR) for melanoma show
a very similar pattern (Figure 2). ASRs peaked in 1995 (40.2 per 100,000
population for men and 37.5 per 100,000 population for women) after the
introduction of the Cancer Registry Act, declined in the following few years and
then recently have shown a slight increase. Prior to 1992, men had a lower
registration rate than women, but since 1993, men have had higher
age-standardised registration rates than women (Figure 2).
Melanoma is reasonably common in younger age groups with
significant numbers of melanomas diagnosed between 25 and 39 years of age in
both men and women (Figure 3). Melanoma is the commonest cancer in adolescence.
In 2002, the greatest number of registrations in women occurred in those aged
45–49 years and in men occurred at 70–74 years of age. However, the
highest age-specific rate for men occurred in those aged 85 years or more, and
in women occurred in those aged 75–79 years. The average age at diagnosis
in 2002 for men was 61.2 years and for women was 57.1 years.
Many deaths from melanoma occur at a younger age than for
most other solid tumours, with the average age at death of 65.5 years for men
and 66.7 years for women in 2002. In the same year, 2,354 person-years of life
were lost for men and 1,573 person-years of life were lost for women due to
melanoma.
Figure 2. Age-standardised registration (ASR) rates for
melanoma by year
 Melanomas were considerably under-registered before 1 July
1994 when the Cancer Registry Act 1993 came into force, so estimations of
projections have used adjusted rates, where the adjustor was the average
incidence to mortality ratio pre- and post-1994.12 Between 2001 and 2011, the
absolute number of registrations for men over the age of 15 was expected to
increase by 32%, to 1,148 per year and the number of deaths was expected to
increase by 17%, to 183 per year.12
When these calculations were made it was estimated that, for
women over the age of 15 years, the absolute number of registrations would
increase to 799 per year and the number of deaths was estimated to increase to
113 per year by 2011.12 However, in 2002, melanoma registrations for women were
already higher than the projection, at 909 new cases, and in 2003, deaths in
women had almost reached the projection for 2011.
Cancer control
activities—The estimated impact of cancer control activities for
melanoma are summarised in Table 1.
Primary
prevention—The main causal factor for the development of melanoma
is exposure of the skin to ultra-violet (UV) light. Using the relative risk for
ever versus never being sunburnt (with blisters) as 1.4, and the prevalence of
ever being sunburnt (with blisters) as 54%,14 the population attributable risk
per cent (PAR%) due to being sunburnt with blisters was 17.8%. So it is possible
that 328 of the 1842 new cases of melanoma in 2002 were directly attributable to
severe sunburn.\
Figure 3. Melanoma registrations and deaths by age,
2001
 If severe sunburn (with blisters) in the population is
decreased by 10%, to a prevalence of 48.6%, then the PAR% decreases to 16.3%. If
we apply this PAR% to the 2002 incidence rate, 300 cases of melanoma could be
directly attributable to severe sunburn. Thus a reduction of 10% in the
prevalence of severe sunburn could result in approximately 28 fewer cases of
melanoma per year. With a mortality to incidence ratio of 12.8% in 2002, this
could result in a reduction of about 4 deaths from melanoma each year.
Screening—Population
screening by skin examination has the potential to reduce mortality but there is
currently no data to assess the potential benefits of population screening in
New Zealand.
Surveillance (including screening by skin examination) of
high-risk people is possible. If 80% of the melanomas that occur in high-risk
individuals were found early enough to prevent death and we assume that 2% of
melanoma cases and 2% of deaths occur in these individuals, approximately 4
deaths per year would be prevented by screening of high-risk groups. That is,
about 1.6% of all deaths each year from melanoma.
Early
diagnosis—Survival decreases with increasing melanoma depth, but
melanoma has a very good prognosis (about 95% 10-year survival) for tumours less
than 1 mm thick.15 Prognosis is poor for tumours thicker than 3.5 mm: the 5-year
disease-free survival is less than 50%.16
In 1998 and 1999, approximately 50% of invasive melanomas in
New Zealand were diagnosed at ≤0.75 mm. Using survival data from the USA17
and Australia15 (comparable figures for survival by depth in New Zealand are not
available), if patients with melanomas diagnosed at ≤0.75 mm depth (the
previous cut-off point for ‘thin’ melanomas) have a 10-year
melanoma-specific survival rate of about 96% and an average survival rate of 64%
for depths greater than 0.75mm, then a 10% shift in the depth distribution from
>0.75 mm into the ≤0.75 mm depth category would result in about 29
deaths prevented per year, based on the melanoma registration figures for 2002.
Treatment—A
reduction in death rate from improvements in treatment is possible but likely to
be small in the near future. Surgical excision of early lesions is currently the
main curative treatment for melanoma.18 Elective lymph node dissection is no
longer recommended, and the value of sentinel node biopsy is currently still
under investigation.19–21 There are as yet no adjuvant therapies of proven
benefit for melanoma.19,21 Interferon-α treatment has been shown to
increase disease-free survival but is also associated with severe side effects.
Melanoma is usually responsive to radiation but only in certain circumstances is
radiation the treatment of choice.
DiscussionReducing the impact of melanoma in New Zealand requires a
planned, systematic, and coordinated approach to multiple activities.
Underpinning this approach is the requirement to collect information on
incidence, prevalence, mortality, diagnosis, stage, and survival of melanoma
patients.9 Research seeks to identify and evaluate means of reducing melanoma
morbidity and mortality, and thus research is a fundamental element for the
production of evidence needed for effective prevention and control of
melanoma.
Prior to the Cancer Registry Act which made notification of
cancer compulsory, many melanomas had not been notified to the Cancer Registry
and this appears to have been greater for men than women. The system for
registration of cancer in New Zealand prior to mid-1994 was a voluntary system,
and notifications came predominantly from public hospitals. Thus, melanoma was
considerably under-reported in New Zealand until 1995–1996 and the
interpretation of trends over time is thereby restricted.
As expected, both the numbers of registrations of melanoma
and the ASRs increased dramatically after the introduction of the Cancer
Registry Act. However, in 1996, the rates declined, but not to pre-Cancer
Registry Act levels. The reason for this decrease, which was greater in women,
is not clear, although it appears that prevalent rather than incident cases were
initially being registered in 1995. It is possible that before 1994, recurrences
were registered as new cancers because of a failure to register the original
diagnosis in earlier years.
After the Cancer Registry Act, recurrences were easier to
identify because of more complete registration and thus less likely to be
registered as incident cases. Since 1997, the registration rates in men have
remained reasonably stable albeit with a suggestion of an increase in recent
years, whereas the registration rates in women increased slightly in 2000, 2001
and 2002. Incidence rates of melanoma in both men and women have also increased
in Queensland22 and South Australia23 from 1998 to 2002.
Cancer prevention should be a key element in all cancer
control programmes; it is often the most cost-effective form of cancer control.9
The main aetiological factor for melanoma is exposure of Caucasian skin to
ultra-violet (UV) light, particularly intermittent exposure and particularly
during childhood. The best avenue currently for melanoma prevention is believed
to be by encouraging protection against sunburn, particularly in children, and
in fair-haired and fair-skinned people.18
The efficacy of sunscreens in reducing exposure to sunlight
has not been proven. A randomised controlled trial (RCT) of sunscreen use and a
review by the International Agency for Research on Cancer have shown that
sunbathers often use sunscreen to extend their time in the sun, thus increasing
their exposure to UV light and their risk of melanoma.24,25 Frequent sunbed use
is also suspected of increasing the risk of melanoma.
Coordinated public health policies and comprehensive
interventions are needed to encourage and support healthy environments.9 There
is reasonable evidence that knowledge about skin cancer can be increased by
health education and health promotion, but there is no evidence that sun
exposure behaviour can be easily altered.18, 26 Neither is there data to support
the suggestion that health promotion of sun avoidance has substantially altered
the incidence of melanoma.18
Even Australia, which has had comprehensive health promotion
messages about skin cancer prevention for more than 30 years, shows little
decrease in the incidence of melanoma, except possibly in younger women in New
South Wales and Queensland.27 Moreover, a recent Australian study28 of
adolescent sun exposure and sun protection behaviours showed a significant
increase in sun exposure and sunburns from 1993 to 1999.
Melanoma meets many of the criteria of a cancer whose
outcome could be improved by screening. Population screening by skin examination
has the potential to identify a high proportion of people with early melanomas
and reduce mortality by early treatment, but there is as yet no conclusive
evidence of improvement in survival.18,29
The efficacy of early detection or screening programmes for
melanoma has not been tested by randomised trials anywhere in the world.30 A
systematic review of papers published between 1994 and 1999 on screening for
skin cancer31 found no direct evidence that screening by physicians reduced
morbidity or mortality from melanoma. Nevertheless, melanoma prevention and
control programmes, including education campaigns and screening, have started in
many other countries over the past decade although data on their effects are
only beginning to be collected.32
Organised population screening has not been introduced into
Australia as there is little evidence for its survival benefit, and its
cost-effectiveness is poor. A cluster RCT underway in Queensland, Australia, is
investigating a community-based screening programme versus normal practice, but
any effects on mortality may not be evident for 10 years.33
Approximately 10% of melanoma patients in Australia had a
first-degree relative who had had a confirmed melanoma.18 Many of these familial
clusters will be due to chance, but about 2% of all melanoma cases occur in
high-risk kindreds. Assuming the same proportion in New Zealand, high-risk
kindreds are thus relatively rare, so surveillance, including screening, of
these high-risk individuals will (while of potential benefit to these
individuals) make little impact on the overall burden of melanoma.
It is also important to educate health professionals and the
public about early signs of melanoma and to encourage early presentation.
Development of techniques of skin surface microscopy may help health
professionals diagnose pigmented skin lesions early.34 The recent decrease in
mortality in younger cohorts in Australia35 may be due to earlier presentation
and improvements in early diagnosis.
Breslow thickness at diagnosis has decreased from the
1980–1986 to 1994–2000 time periods.36,37 However, about half the
improvement in survival from melanoma was unexplained by the change in depth
distribution.36 There is no comparable data series yet available for New
Zealand. However, in 1998 and 1999, only 50% of invasive melanomas in New
Zealand were diagnosed at ≤0.75 mm depth (thin melanoma), whereas in South
Australia between 1994 and 2000, 57.8% were thin at diagnosis.
Optimising survival and quality of life for patients with
melanoma requires having access to treatments that (on the basis of current
evidence) are known to provide the best outcomes.9 The use of guidelines,
protocols, and interdisciplinary management of melanoma patients may achieve
consistent treatment standards. Surgical excision of early lesions is currently
the main curative treatment for melanoma.18 Melanomas <1 mm deep are
treated definitively by excision with a 1 cm margin.21
In situ tumours and those <0.76 mm
deep, and with no vertical growth phase, are commonly excised with a 0.5 mm
margin or less.21 More recent guidelines19,20 have recommended that lesions up
to 2 mm deep be excised with a 1 cm margin.
Elective lymph node dissection is no longer
recommended.19–21 Although recommended by some groups19,20 to be carried
out in specialist centres, the value of sentinel node biopsy is currently still
under investigation. There are as yet no adjuvant therapies of proven benefit
for melanoma.19,21 Interferon-α therapy has been shown to increase
disease-free survival but is associated with severe side-effects.
The UK guidelines21 state there is no place for isolated
limb perfusion whereas the Swiss guidelines19 and the European Society for
Medical Oncology20 recommend its use in specialist centres. Melanoma is usually
responsive to radiation. It may be used for large unresectable lesions or for
lentigo maligna in elderly frail patients, but it is not usually the first
treatment of choice.
The number of melanoma cases and deaths from melanoma have
been projected to significantly increase in the following years and there is
evidence that the number of deaths in women have already exceeded the
projections. If future mortality from melanoma in New Zealand is to be
controlled, then it is important that a greater proportion of new cases are
diagnosed when they are thin and when the chances of a complete cure are
high.
Author information:
Mary J Sneyd, Research Fellow; Brian Cox, Research Associate Professor; Hugh
Adam Cancer Epidemiology Unit, Department of Preventive and Social Medicine,
University of Otago, Dunedin
Acknowledgement:
Mary J Sneyd and Brian Cox are supported by the Directors’ Cancer Research
Trust.
Correspondence: Dr
Mary Jane Sneyd, Hugh Adam Cancer Epidemiology Unit, Department of Preventive
and Social Medicine, University of Otago Medical School, PO Box 913, Dunedin.
Fax (03) 479 7164; email: mary-jane.sneyd@stonebow.otago.ac.nz
References:
|
|
||||||||
| | Current
issue | Search journal |
Archived issues | Classifieds
| Hotline (free ads) Subscribe | Contribute | Advertise | Contact Us | Copyright | Other Journals |
|