Analysis of predictive factors for the outcome of complete lymph node dissection in melanoma patients with metastatic sentinel lymph nodes

Abstract

Sentinel lymph node biopsy (SLNB) is a widely accepted procedure to accurately stage patients with melanoma. However, there is no consensus concerning the practical consequences of a positive SLN, since a survival benefit of a complete lymph node dissection (CLND) has not yet been demonstrated. We wondered whether we could identify a subgroup of patients with metastatic involvement of the SLN who could be excluded from the recommendation to undergo CLND. At the Department of Dermatology at the University of Munich, a total of 213 patients with metastatic SLNs (24.9%) were identified among 854 patients who had undergone SLNB between 1996 and 2007. All SLN-positive patients had been advised to have CLND. Survival analyses were performed by using the Kaplan-Meier approach. A total of 176 (82.6%) of 213 SLN-positive patients underwent CLND. In this group, 26 patients (14.8%) showed metastatic disease in non-sentinel lymph nodes (NSLN). The 5-year overall survival (OS) was 26.1% in NSLN-positive patients and 74% in NSLN-negative patients. SLN-positive patients who refused CLND had a better prognosis than patients with CLND. Breslow tumor thickness was significantly associated with positive CLND status with higher median values in CLND-positive than CLND-negative patients (3.03 vs 2.22 mm). The subgroup of patients with metastatic disease in CLND may have been too small to reach statistical significance for other tumor- or patient-related parameters. Mitotic indices of the primary melanomas had not been determined in this retrospective study; thus a possible correlation with lymph node status could not be tested. Among SLN-positive patients, the presence of metastatic NSLN is a highly significant poor prognostic factor. Tumor thickness is a significant prognostic parameter for positive CLND status and might be considered in the decision to perform CLND in case of metastatic SLN.