Re: Determinants of BRAF mutations in primary melanomas.

I read with interest the paper by Maldonado et al. ( 1 ) investigating the distribution of BRAF mutations across different melanoma types. One of the questions addressed by the authors was whether cutaneous melanomas with a BRAF mutation arose from preexisting melanocytic nevi. In an analysis of 46 lesions, they found melanomas with an associated nevus to have only a slightly higher prevalence of BRAF mutation (55%; n = 11) than did melanomas exhibiting no evidence of nevus (43%; n = 35); the difference was not statistically signifi cant. On the basis of these fi ndings, the authors concluded that not all melanomas with an associated nevus may arise from melanocytes with BRAF mutations. When one is assessing melanomas for histologic evidence of a preexisting nevus, thick tumors are not necessarily informative, given the possibility that evidence of a preexisting nevus may have been obliterated by the growing tumor ( 2 ) . The misclassifi cation of tumors with respect to evidence of a preexisting nevus would be expected to attenuate or obscure actual differences in tumor characteristics between melanomas that arise from a precursor nevus and melanomas that arise de novo. The level of misclassifi cation in the study of Maldonado et al. may be fairly high, given that the cutaneous melanomas included in their analysis were relatively thick (median thickness = 3.6 mm; range = 1 – 15 mm). It would be informative if the authors addressed this issue by restricting their analytic sample to thinner melanomas to minimize the effects of such misclassifi cation. In addition, it may be advisable to exclude acral lentiginous melanomas from the analysis, because the etiology of this tumor subtype appears to be quite different from that of other cutaneous melanomas ( 3 ) .