The impact of surgery on the course of melanoma.

Skin melanoma, unlike other cancers, occurs on the body surface: it can be detected and treated before it reaches the stage where it can metastasise; the impact of surgery is unrivalled, but only while it is at this early stage. In melanomas more than 0.75-1.00 mm thick, an increasing proportion acquire metastatic properties. There is today evidence showing that wide excision does not help, and that the effect of surgery is limited to local control of the disease. According to randomised trials, the territory of early spread--without concomitant distant micrometastases--that can be eradicated by surgery is shrinking. It has been demonstrated that resection margins of 3-4 cm are no better in terms of recurrence and survival than margins of 1 or 2 cm. Most melanomas can now be adequately resected without skin grafting. Regional elective lymph node dissection for high-risk melanoma (1.5 mm thick or more) does not improve survival over that obtained with delayed lymph node dissection performed when clinical metastases appear. By analogy, prophylactic isolated limb perfusion with melphalan reduces the rate of in-transit metastases but does not improve survival. Sentinel node biopsy allows early detection of regional lymph node metastases with minimally invasive surgery. On-going randomised studies will show whether it can have any impact on survival. Considering the experience with elective lymph node dissection, it seems unlikely that selective--as opposed to elective--lymph node dissection, of positive sentinel nodes, will influence survival. The already extensive experience with sentinel node biopsy provides a death risk hierarchy: one N2 node (with clinical metastasis), one N1 node--or sentinel node--with micrometastasis and one N0 node with no histologically detectable micrometastasis but PCR positive give, respectively, 50%, 60% and 70% 5-year survival rates. In other words, the earlier the detection of metastasis, the longer the survival. In terms of growth kinetics, the earlier the detection of metastasis, the longer the time to death, with no evidence that surgery would have an impact. There is just one, as yet unpredictable, subset of pa- tients with lymph node-confined disease in whom surgery might have an impact. It is hoped that, in the future, gene expression profiles of primary melanoma will help to pick out these patients. Multivariate analysis has shown that the sentinel node status is the most powerful prognostic factor in primary melanoma. Sentinel node biopsy is a valuable tool for selecting patients for adjuvant treatments within the frame of clinical trials, in which micrometastatic and clinically involved lymph nodes are entered separately. In-transit metastases can be eradicated in 50% of cases by isolated limb perfusion with melphalan under mild hyperthermia. When in-transit metastases are recurrent, deep seated, or bulky, the combination of tumour necrosis factor (TNF) with melphalan and interferon gamma yields a complete response rate of around 80%. This is the first antiangiogenic treatment of cancer that is effective in clinical practice, but it has no effect on survival. Current better knowledge of melanoma biology indicates that local, limited surgery has an impact on local or regional spread only.