Multivariable analysis comparing outcome after sentinel node biopsy or therapeutic lymph node dissection in patients with melanoma.

BACKGROUND: Sentinel node (SN) status is the most important prognostic factor for overall survival in stage I or II melanoma. Yet SN-positive tumours with submicroscopic involvement of the SN (clusters of cells smaller than 0.1 mm) have shown a distant recurrence rate of only 9 per cent at 5 years, as good as that in SN-negative patients. This study compared the outcome after completion lymph node dissection (CLND) in SN-positive tumours with elective total lymph node dissection (TLND) in patients with palpable nodes. METHODS: A total of 188 patients were identified; 124 had TLND and 64 had CLND. Median follow-up was 56 and 37 months respectively. There were no significant differences between the groups regarding tumour Breslow thickness, ulceration and site of the primary tumour. Survival rates were calculated from date of primary excision. All patients with primary melanomas on extremities or trunk were included. RESULTS: On univariable analysis, the site of the primary tumour (extremity versus trunk) (P < 0.001), Breslow thickness (P = 0.005) and ulceration (P < 0.001) were prognostic for overall survival. There was a non-significant 13 per cent difference in overall survival at 5 years between CLND and TLND (P = 0.115). Excluding 15 patients who had SN disease with submicrometastases reduced the difference to 6 per cent (P = 0.415). CONCLUSION: This study showed no significant survival benefit for SN-positive CLND compared with TLND, especially when patients with nodes containing submicrometastases were excluded.