Cutaneous malignant melanoma : body site, sun exposure, genetic factors and prognosis

Abstract

Background: The interplay between tumor site, ultraviolet radiation (UVR) exposure, genetic factors and various epidemiological parameters of cutaneous melanoma is complex. In this thesis, we aimed to investigate the impact of detailed body site of the primary tumor beyond the conventional and strictly anatomic division of the head-neck, trunk, upper- and lower extremities, focusing on body site division according to UVR exposure patterns in relation to risk, outcome and other factors.

Methods: We retrospectively reviewed medical records of patients with cutaneous melanoma to obtain detailed site information of the primary tumor (studies I–IV). We then entered detailed site on a three-dimensional anatomic model in a computer software, and classified body areas according to UVR exposure patterns, visibility upon skin self-examination, and anatomic classifications more detailed than the ICD (International Classification of Diseases) codes. Through linkage with regional and/or national registers, we obtained data on incidence and survival, as well as established confounders and (when applicable) losses to follow-up. We also reviewed medical records for sentinel node location and -status (study I), and performed PCR and pyrosequencing for mutation analysis of proto-oncogenes BRAF and NRAS (study IV).

Results: Trunk melanoma was associated with multiple but not uncommon sentinel node locations, compared to extremity melanoma. Multiple or uncommon sentinel node locations were not found to explain the association between trunk melanoma and reduced patient survival (study I). Melanomas assigned to intermittent UVR exposure patterns according to detailed anatomic site of the primary tumor displayed a higher incidence increase in the Stockholm-Gotland region during the past decades compared to melanomas on sites with assigned chronic UVR exposure patterns (study II). Site-assigned highly intermittent UVR patterns and poorly visible sites upon skin self-examination were associated with reduced patient survival compared to chronically UVR exposed and easily visible sites, respectively (study III). Intermittently UVR exposed sites were associated with BRAF mutations, and chronically UVR exposed sites with NRAS mutations (study IV).

Conclusions: Primary tumor sites with assigned intermittent UVR exposure patterns are presumably related to the incidence increase as well as reduced patient survival and genetic BRAF mutations of cutaneous melanoma, whereas sites assigned to chronic UVR exposure patterns contribute less to the incidence increase, are prognostically more favorable, and predominantly display NRAS mutations. Multiple or uncommon sentinel node locations do not explain the adverse prognosis of trunk melanoma.