Lymph node ratio as a prognostic factor in melanoma: results from European Organization for Research and Treatment of Cancer 18871, 18952, and 18991 studies.

Abstract

The aim of this study was to assess the prognostic importance of lymph node ratio (LNR) in stage III melanoma after complete lymph nodal dissections. From European Organization for Research and Treatment of Cancer randomized trials 18871, 18952, and 18991, 2358 patients had full information on positive and examined lymph nodes (LNs) and were included. Cox proportional hazards models stratified by trial were used to assess the prognostic impact of LNR adjusted for confounders on melanoma-specific survival. Optimal cutoff values for LNR were calculated for each LN dissection site (axillary, inguinal, and neck). LNR (≥ vs. <35%: hazard ratio=1.44, 95% confidence interval: 1.23-1.69) and number of positive LNs appeared to be of independent strong prognostic importance. Dissection sites impacted the optimal LNR cutoff: 35% for axillary, 40% for inguinal, and 50% for neck dissections. Combining these into one 'high versus low LNR' resulted in a highly significant multivariately adjusted hazard ratio of 1.48 (95% confidence interval: 1.26-1.74). In subgroup analyses, LNR was only significant in advanced disease (American Joint Committee on Cancer stage N2b, N3; IIIC). LNR was most significant for inguinal dissections, followed by axillary dissections, but seemed less useful in neck dissections. LNR is an independent significant prognostic factor in stage III melanoma patients. Our study showed higher than previously reported cutoffs that differed per dissection site. However, because of conflicting results compared with other studies and apparent limited prognostic impact confined to subgroups, the practical use of LNR seems limited.