Cutaneous melanoma histologically associated with a nevus and melanoma de novo have a different profile of risk: Results from a case-control study

Abstract

Background: Histopathologic association between melanocytic nevus and melanoma has been reported in approximately 10% to more than 50% of melanoma cases. Whether melanomas in contiguity with a nevus have a different natural history and pathogenesis from melanomas without a nevus is still to be determined. Objective: The present study was undertaken to clarify whether melanocytic nevus–associated melanomas (MN[+]) have a different risk factor profile from cases without histopathologic evidence of melanocytic nevus association (MN[–]). Methods: The study population consisted of 131 invasive melanoma cases with a thickness of 4.00 mm or less and 174 control cases without melanomas. The whole series was evaluated for the following risk factors: phenotypic traits; the number of common, atypical, and congenital nevus–like nevi; and freckling and history of sunburns. Melanoma cases were revised for the presence of associated melanocytic nevi. The analysis of risk factors was performed by a case-control approach comparing cases, classified by histologic association with nevus, to the group of controls. Possible differences in risk factor distribution between MN(+) cases and MN(–) cases were evaluated with a polychotomous logistic regression model and a likelihood ratio test for heterogeneity. Results: Histopathologic association between melanocytic nevus and melanoma was found in 27 cases (20.6%). Phenotypic traits were shown to be more powerful predictors of risk for MN(–) than for MN(+) cases (blond/red hair; odds ratio, 7.4 and 1.2, respectively; likelihood ratio test for heterogeneity, 4.13; P < .05). Conversely, history of frequent sunburn was a risk factor only in MN(+) cases (more than 5 sunburns; odds ratio, 6.7; 95% confidence interval, 1.3-33.7), but not in MN(–) cases (odds ratio, 1.2; 95% confidence interval, 0.3-4.0; likelihood ratio test for heterogeneity, 4.2; P < .05). Where melanocytic nevi are concerned, an increased number of common nevi was a predictor of melanoma risk in both MN(+) and MN(–) cases, but with a different magnitude of risk, higher for MN(+) cases (number of common nevi, 10-30; odds ratio, 14.4 and 4.7, respectively; likelihood ratio test for heterogeneity, 3.7; P = .055). Conclusion: This study showed that, although MN(+) and MN(–) melanomas share many risk factors, there is a different strength of association between the 2 groups. The effect of a history of sunburn as a predictor of risk was found only for nevus-associated melanomas, suggesting a possible role of sunburns in the neoplastic transformation of nevi. (J Am Acad Dermatol 1999;40:549-57.)