The incidence of cutaneous melanoma in Queensland is the highest reported in the world and rates have increased steadily during the past few decades, accompanied by changes in anatomical distribution. In order to address the enigma of the site distribution of melanoma, a population-based case-control study was conducted which combined classical epidemiological methods with laboratory based techniques to identify phenotypic, cellular and molecular determinants of site-specific melanoma.
150 male cases of melanoma aged over 50 and diagnosed between 1/7/1993 and 30/6/1994 were randomly ascertained from the Queensland Cancer Registry, and age-matched to 150 controls randomly selected from the Queensland Electoral Roll. Data were collected through structured face-to-face interviews with all participants, asking about demographic information, medical history, phenotypic characteristics and occupational and environmental exposures. Observational data were also obtained, including counts of naevi and solar keratoses on the face, left upper limb, shoulders and back. Skin impressions using silicon moulds were taken as a record of solar damage to the dorsum of the hand. From consenting participants, 2mm skin biopsies were taken from the dorsum of the left hand and from a site adjacent to the melanoma (matched in controls). Transverse epidermal sections of the biopsied tissues were stained with a melanocyte-specific monoclonal antibody, and the density of melanocytes was determined by a validated method. Finally, paraffin sections of melanoma tissue were obtained from the cases and assessed for expression of p53 tumour suppressor gene product using immunohistochemical techniques. Factors associated with abnormal expression of p53 were identified.
Determinants of melanoma were observed to vary according to the site of the lesion. With regard to phenotypic factors, melanomas of the head/neck were more strongly associated with measures of facultative pigmentation, whereas melanomas of the back/shoulders were associated with innate pigmentation. The number of benign melanocytic naevi was strongly associated with melanoma at all sites, except the upper limb.
Measures of ultraviolet radiation exposure were inconsistently associated with risk of sitespecific melanoma, probably partly due to the very high levels of sun exposure among controls as well as cases. Overall, melanomas of the head/neck and lower limbs were more strongly associated with sun exposure than melanomas of the back/shoulders, although the association was not significant. Similarly, a past history of non-melanoma skin cancer was most strongly associated with melanoma of the head/neck, less strongly at other sites, and not at all for the upper limb.
The density of the target cell was observed to vary widely across individuals, as well as across anatomical sites. Systematic differences were observed in the mean density of melanocytes at specific sites, with highest densities observed on the back/shoulders, followed by the upper limbs and lower limbs. High melanocyte densities on the upper limbs and anterior trunk were associated with non-significantly increased risk of site-specific melanoma. These findings suggested that site-specific melanocyte densities may partially explain the observed distribution of melanoma, and that risk of melanoma is highest in those with high melanocyte densities. Abnormal expression of the p53 tumour suppressor gene was observed in 23 (18.3%) melanomas. Strongest predictors of p53-immunopositive melanoma were inability to tan, history of non-melanoma skin cancer and occurrence on the head/neck or lower/limbs. By contrast, p53-immunonegative melanomas were strongly associated with naevus density and facial freckling.
Taken together, these findings indicate that the component causes of melanoma vary according to the site of the lesion. This study also demonstrated that the distribution of the target cell is not uniform across body sites, but rather mirrors the site-specific incidence of melanoma. It seems likely that at least two independent pathways may lead to melanoma, characterised by environmental induction on the one hand, and pigment cell instability on the other.