[Prognostic advantage for defined risk groups by lymphocyte dissection. Long-term study of 3,616 melanoma patients]

Nine medical centres with different practices in elective lymph node dissection (ELND) but comparable standards regarding diagnosis, excision of the primary tumour, classification, and follow-up, have collected their data on 3616 patients with primary melanoma of the skin (tumour category pT 2 to pT 4a, N 0, M 0 [UICC 1987] with the aim of producing an unbiased analysis of the prognostic benefit of ELND. The multivariate risk analysis (Cox's proportional hazard model) revealed tumour thickness (Breslow or alternative pT categories), sex, anatomic site of the primary tumour, and ELND therapy ("yes" or "no") as independent prognostic factors. Observed survival curves (Kaplan-Meier) show a significant difference of prognosis with regard to ELND therapy in the following risk groups: women with melanomas over 2.5 to 4 mm thick on head, neck, thorax, and in acral locations; men with melanomas over 1.5 to 4 mm thick on head, neck, thorax, and in acral locations; and finally men with melanomas over 2.5 to 4 mm thick on abdomen and extremities. Further investigations and the discovery of additional prognostic factors would help in more precisely formulation of guidelines for ELND.