Independence of dysplastic nevi from total nevi in determining risk for nonfamilial melanoma

In the determination of risk for melanoma, relatively little is known about the possible independence of two important predictors, total nevi and clinically dysplastic nevi. From a study conducted in Sydney, Australia [see J. J. Nordlund et al., Cancer Res. 45, 1855-1861 (1985)], 246 cases of melanoma (excluding 7% of targeted patients with a history of melanoma in a first-degree relative) were compared with 134 nonmelanoma controls. Participants had been examined by a dermatologist and an oncologist. Logistic regression analysis was used and included an age-sex interaction term in computing all estimates of relative risk in this report. Relative risk for melanoma in those with 16+ total nevi was significantly elevated at 1.8 but declined to a statistically nonsignificant level of 1.2 (95% confidence limit (CL): 0.7, 2.0) after adjustment for dysplastic nevi. In contrast, relative risk for melanoma in those with any dysplastic nevi was 7.6 (95% CL: 3.6, 16.0) but was maintained at a similarly elevated and statistically significant level of 7.7 (95% CL: 3.5, 17.1) after adjustment for total nevi. These patterns were even more evident in the younger half of the study population. The analyses suggest that much of the association between TN and nonfamilial melanoma is explained by the presence of dysplastic nevi and, conversely, they imply that dysplastic nevi represent a clinically distinct, qualitative disorder rather than simply a quantitative disorder wherein dysplastic nevi stem merely from an increase in total nevi. The dysplastic nevus syndrome accounts for 32% of all nonfamilial melanomas.