A proportion of cutaneous melanomas display neval remnants on histologic examination. Converging lines of epidemiologic and molecular evidence suggest that melanomas arising from nevus precursors differ from melanomas arising de novo. In a large, population-based study comprising 636 cutaneous melanomas subjected to dermatopathology review, we explored the molecular, host, and environmental factors associated with the presence of neval remnants. We found that nevus-associated melanomas were significantly associated with younger age at presentation, non-brown eye color, trunk site, thickness of less than 0.5 mm, and BRAF mutation. Compared with patients with de novo melanomas, those with nevus-associated tumors were more likely to self-report many moles on their skin as a teenager (odds ratio = 1.94, 95% confidence interval = 1.01-3.72) but less likely to report many facial freckles (odds ratio = 0.49, 95% confidence interval = 0.25-0.96). They also had high total nevus counts (odds ratio = 2.18, 95% confidence interval = 1.26-3.78). On histologic examination, nevus-associated melanomas exhibited less dermal elastosis in adjacent skin compared with de novo melanomas (odds ratio = 0.55, 95% confidence interval = 0.30-1.01). These epidemiologic data accord with the emerging molecular paradigm that nevus-associated melanomas arise through a distinct sequence of causal events that differ from those leading to other cutaneous melanomas.

Cutaneous melanomas are postulated to arise through at least two causal pathways, namely the "chronic sun exposure" and "nevus" pathways. While chronic sun exposure probably causes many head/neck melanomas, its role at other sites is unclear. In a population-based, case-case comparison study conducted in Brisbane, Australia, we determined the prevalence and epidemiologic correlates of chronic solar damage in skin adjacent to invasive, incident melanomas on the trunk (n 5 418) or head/ neck (n 5 92) among patients aged 18-79 in 2007-2010. Participants self-reported information about environmental and phenotypic factors, and a dermatologist counted nevi and actinic keratoses. Dermatopathologists assessed solar elastosis adjacent to each melanoma using a four-point scale (nil, mild, moderate, marked), and noted the presence or absence of adjacent neval remnants. We measured associations between various factors and solar elastosis using polytomous logistic regression. Marked or moderate solar elastosis was observed in 10% and 27%, respectively, of trunk melanomas, and 60% and 17%, respectively, of head/neck melanomas. At both sites, marked elastosis was positively associated with age (p trend < 0.0001) and inversely associated with neval remnants (p trend < 0.001). For trunk melanomas, marked elastosis was associated with highest quartiles of total sun exposure [odds-ratio (OR) 5 5.47, 95% confidence interval (CI) 5 1.08-27.60] and facial freckling (OR 5 2.98, 95% CI 5 1.17-7.56), and inversely associated with deeply tanning skin (OR 5 0.29, 95% CI 5 0.08-1.11) and high nevus counts (OR 5 0.08, 95% CI 5 0.01-0.66). Mostly similar associations were observed with moderate solar elastosis. About one in three trunk melanomas in Queensland have evidence of moderate-to-marked sun damage, and they differ in risk associations from those without.Cutaneous melanoma is a potentially lethal cancer, the incidence of which has risen considerably in Caucasian populations worldwide during recent decades. 1 While sun exposure was recognized early as the likely environmental cause of melanoma, 2 ecological and analytical epidemiological studies were somewhat inconsistent in their findings, leading to

Background: Cutaneous melanomas have been hypothesized to arise through different pathways according to phenotype, body site, and sun exposure. To further test this hypothesis, we explored associations between phenotype and melanoma at different sites using a case-case comparative approach.Methods: Melanoma patients (n ¼ 762) aged 18 to 79 years and diagnosed from 2007 to 2010 were ascertained from pathology laboratories in Brisbane, Australia. Patients reported phenotypic information and a dermatologist counted melanocytic nevi and solar keratoses. We compared data for patients with trunk melanoma (n ¼ 541, the reference group), head/neck melanoma (n ¼ 122), or lentigo maligna melanoma (LMM) of the head/neck (n ¼ 69). ORs and 95% confidence intervals were calculated using classical or polytomous logistic regression models.Results: Compared with trunk melanoma patients, those with head/neck melanoma were significantly less likely to have high nevus counts (135: OR ¼ 0.27; P trend ¼ 0.0004). Associations between category of nevus count and LMM head/neck were weaker and significantly different (135: OR ¼ 1.09; P trend ¼ 0.69; P homogeneity ¼ 0.02). Patients with head/neck melanoma were more likely than those with truncal melanoma to have high solar keratosis counts (7: OR ¼ 1.78, P trend ¼ 0.04). Again, associations with LMM head/neck were weaker, albeit not significantly different (7: OR ¼ 1.61; P trend ¼ 0.42; P homogeneity ¼ 0.86).Conclusion: Trunk melanomas are more strongly associated with nevus counts than head/neck melanomas, but are less strongly associated with number of solar keratoses, a marker of chronic sun exposure.Impact: These findings underscore the notion that melanomas on the trunk typically arise through a causal pathway associated with nevus propensity, whereas melanomas on the head/neck arise through a pathway associated with cumulative sun exposure. Cancer Epidemiol Biomarkers Prev; 22(12); 2222-31. Ó2013 AACR.