Association of baseline neutrophil-to-lymphocyte ratio (NLR) with response and survival in advanced melanoma (MEL) receiving PD-1 inhibitors.

Abstract

9571 Background: Inflammation is an adverse prognostic factor in cancer. Neutrophil-to-lymphocyte ratio (NLR) is an easily derived biomarker of systemic inflammation. Several studies have demonstrated that elevated NLR is linked with adverse prognosis in patients (pts) receiving immunotherapy including PD-1 inhibitors. To evaluate the prognostic utility of NLR, we performed a retrospective evaluation of NLR and other covariates in stage IV cutaneous MEL. Methods: Stage IV cutaneous MEL pts who received anti PD-1 therapy at the University of Pittsburgh between 2014-2018 were included in this analysis. PD-1 blockade was continued until progression or intolerable toxicity. Tumor assessment was performed at baseline and every 12 weeks and response classified per RECIST v1.1. Clinical and demographic data were obtained. Baseline NLR was defined based on values at the first treatment date. Descriptive statistics were created for all covariates. Kaplan Meier and Cox proportional hazard regression were performed to assess how variables related to response (ORR), overall survival (OS) and progression free survival (PFS) measured in months (mos). Results: 172 pts with advanced MEL were evaluated. Elevated NLR was associated with poorer PFS and OS and ORR at all cutoffs (NLR≥2 to NLR≥5) with NLR≥5 having the greatest discriminative value. ORR steadily declined with increasing NLR: NLR≥1 (ORR 64%), NLR≥2 (ORR 61%), NLR≥3 (ORR 52%), NLR≥4 (ORR 43%), NLR≥5 (ORR 43%). Elevated NLR ( < 5 vs. ≥5) was associated with poorer PFS (median 21.5 mths vs. 5.2 mos; p = 0.00041) and OS (median 35.4 os vs. 10.6 mos; p < 0.0001). In a multivariate model, elevated NLR ( < 5 vs. ≥5) was independently associated with poorer OS/PFS separate from ulceration, performance status and elevated LDH. There was no evidence of an age-related increase or decrease in NLR. Conclusions: Baseline NLR was independently associated with response, PFS and OS in the largest retrospective series of advanced MEL pts treated with PD-1 blockade. NLR independent of other factors predicted poorer PFS and OS at NLR cutoffs (NLR≥3 to NLR≥5), although NLR≥5 segregated pts best. NLR is an inexpensive and easily obtained real-world biomarker that has a high value in predicting outcomes to PD-1 blockade.