Risk of cutaneous melanoma in relation to the numbers, types and sites of naevi: a case-control study.

The atypical mole syndrome (AMS) phenotype, characterised by a large number of common naevi as well as atypical naevi, has been described in families with a genetic susceptibility to melanoma. However, the importance of this phenotype for melanoma in the general population has not been conclusively determined. This study was designed to examine the types and distribution of naevi as well as the prevalence of the AMS phenotype in melanoma patients in England compared with controls. A total of 426 cutaneous melanoma cases (61% of all incident cases) aged 16-75 years were recruited between 1989 and 1993 from the north-east Thames region of the UK and 416 controls from the same age group were recruited over the same period and from the same region. Each subject answered a questionnaire covering demographic details, sun exposure history and other risk factors and underwent a skin examination with total body naevus count performed by a dermatologist. The AMS phenotype was defined using a scoring system. Atypical naevi gave the highest relative risk for cutaneous melanoma, with an odds ratio (OR) of 28.7 (P < 0.0001) for four or more atypical naevi compared with none. Many common naevi were also an important risk factor: the OR for 100 or more naevi 2 mm or above in diameter compared with 0-4 naevi was 7.7 (P < 0.0001). Melanoma was also associated with naevi on sun-exposed sites but also with naevi on non-sun-exposed sites such as the dorsum of the feet, buttocks and anterior scalp. Sixteen per cent of the cases had the AMS phenotype compared with 2% of the controls (OR 10.4, P < 0.0001). The AMS phenotype was more common in males than females (P = 0.008). The odds ratio for the presence of the AMS phenotype was dependent on age, with an odds ratio of 16.1 (95% CI 4.6-57.5) for the presence of the AMS phenotype if aged less than 40 compared with an odds ratio of 6.9 (95% CI 2.9-16.6) if aged 40 or more. The AMS phenotype was strongly predictive of an increased risk of melanoma outside the familial context.