Integration of Programmed Cell Death Ligand 1 and Neutrophil to Lymphocyte Ratio to Predict Response to Nivolumab in Non-Small Cell Lung Cancer

We will retrospectively analyze patients with advanced NSCLC who have received at least one cycle of nivolumab (3mg/kg intravenously every 2 weeks) within the early access program (EAP) or after the drug approval. All patients have been treated with immunotherapy, in second or further line, in two Italian Institution (the S. Maria delle Croci Hospital in Ravenna and the S. Maria della Misericordia Hospital in Perugia) between February 2015 and May 2017.

Patient data and laboratory values will be recorded in an electronic anonymized database and personally collected by one of the investigators (C.B.). PD-L1 expression will be assessed by immunoistochemistry (IHC) on available archival tissue samples with clone E1L3N (monoclonal rabbit; Cell Signaling technology, Danverd, MA) or SP263 (monoclonal rabbit; Ventana Medical System, Tucson, AZ), in a Ventana automated stainer according to the manufacturer's protocol and using proprietary reagents. Since a specific PD-L1 level is not mandatory for prescribing immunotherapy in second or further lines, except for pembrolizumab, we established the PD-L1 positivity as expression on ≥ 1% of tumor cells. To calculate NLR, the absolute neutrophil count will be divided by the lymphocytes value measured in peripheral blood within 4 weeks prior to the first infusion of nivolumab. Patients will be dichotomized according to a pre-specified cutoff value as high (NLR ≥ 3) or low (NLR < 3), since a ratio greater than 3 has been associated with poor outcome in many types of cancer. Patients will be divided in two cohorts according to combined PD-L1/NLR value: cohort1 with PD-L1 positive and low NLR and cohort 2 with PD-L1 negative and high NLR. Primary end point will be ORR, calculated as the percentage of responses among all treated patients. Response to treatment will be assessed by computed tomography and classified according to RECIST 1.1 criteria [15 Eisenauer Cancer 2009]. The influence of the combined PD-L1/NLR on overall response rates (ORR), will be analyzed with univariable logistic regression. Secondary outcome will be OS and PFS, calculated from the start of nivolumab treatment to death and radiographic or clinical progression, respectively, with deterioration of performance status classified as disease progression. We will also examine whether PD-L1, NRL, or combined PD-L1/NRL contribute to the prediction of OS or PFS, through multi-variable analyses. Since data collected included many variables, we will perform multivariable models to examine the dependence of OS and PFS on known prognostic factors: Eastern Cooperative Oncology Group Performance Status (ECOG PS) (0-1 vs. ≥2) and brain involvement at time of initiating nivolumab, smoking history [never (<100 cigarettes per lifetime) vs. former/current smokers], histology (squamous vs. non-squamous), molecular profiling for epidermal growth factor receptor (EGFR), Kirsten rat sarcoma viral oncogene homolog (K-RAS) and anaplastic lymphoma kinase (ALK) when available. Patients with squamous-cell cancers will be counted as wild type for targetable mutation, which are considered to be infrequent in this population. The study was approved by the local Research Ethics Committee.