Baseline neutrophils and derived neutrophil-to-lymphocyte ratio: prognostic relevance in metastatic melanoma patients receiving ipilimumab

Abstract

Clinical responses to ipilimumab are variable in terms of onset, magnitude and duration. Upfront identification of patients who are more likely or unlikely to benefit from treatment is a major need. Prospectively collected data from 720 advanced melanoma patients treated with ipilimumab 3 mg/kg within the Italian expanded access program were analyzed. The derived neutrophil-to-lymphocyte ratio (dNLR) was calculated from baseline peripheral blood cell counts, and receiver operating characteristic curve was used to evaluate the best cutoff for this marker. Patients were stratified according to dichotomized baseline absolute neutrophil counts (ANC), dNLR and their combination. The prognostic values of ANC and dNLR for survival were assessed using multivariate Cox proportional hazard models. A subgroup analysis including LDH in the models was also carried out. The median follow-up was 16.5 months. The optimal cutoff for dNLR was 3. Baseline ANC and dNLR were significantly associated with the outcome of ipilimumab-treated melanoma patients, in terms of disease progression and death (P < 0.0001 for all). Furthermore, for each elevated variable, prognosis worsened. Patients with both ANC ≥ 7500 and dNLR ≥ 3 had a significantly and independently increased risk of death [hazard ratio(HR) = 5.76; 95% confidence interval (CI) 4.29-7.75] and of progression (HR = 4.10; 95% CI 3.08-5.46) compared with patients with both lower ANC and dNLR. Patients with one of the two factors elevated displayed an intermediate risk of progression and death. The 1- and 2-year survival rates were 2% and 0%, respectively, for patients with ANC ≥ 7500 and dNLR ≥ 3, and 43% and 24%, respectively, for patients with both lower ANC and dNLR. Although these findings need to be confirmed and validated, we suggest that a neutrophil-based index may help risk-group stratification and assist disease-management strategies. Furthermore, the potential predictive value of this index for response to ipilimumab should be investigated in randomized clinical trials.