Background Second-line immunotherapy (IO) has shown an overall survival benefit. However, only 18% to 20% of patients with advanced non–small-cell lung cancer (aNSCLC) will respond, with a median progression-free survival (PFS) of 2 to 4 months. Thus, biomarkers to select those patients most likely to benefit from IO are greatly needed. Patients and Methods We conducted a retrospective analysis of 154 patients with aNSCLC who had received anti–programmed cell death 1 therapy as second line or further treatment. We assessed the absolute neutrophil, lymphocyte, monocyte, and eosinophil counts at baseline (T0) and the second (T1) and third (T2) cycles. The neutrophil/lymphocyte ratio (NLR), derived-NLR (dNLR), lymphocyte/monocyte ratio (LMR), and their percentage of change at T1 and T2 compared with T0 were evaluated. The clinical characteristics and lactate dehydrogenase (LDH) level were also considered. Univariate and multivariate analyses were performed. Significant biomarkers for PFS on multivariate analysis were combined in a prognostic score. Results For overall survival, the negative prognostic biomarkers were Eastern Cooperative Oncology Group (ECOG) performance status (PS) 2, NLR at T0, and dNLR at T1; the LMR at T0, T1, and T2 was identified as a positive prognostic biomarker. For PFS, the negative prognostic biomarkers were ECOG PS 2, liver metastases, NLR at T0, dNLR at T1 and T2, and ≥ 30% increase of NLR from T0 to T1; the positive prognostic biomarkers were heavy smoking, LDH, and LMR at T2. The ≥ 30% increase of LMR from T0 to T1 and T0 to T2 correlated with the overall response rate. A prognostic score (EPSILoN score; smoking, ECOG PS, liver metastases, LDH, NLR) identified 3 prognostic groups (median PFS, 10.2, 4.9, and 1.7 months, respectively; P Conclusions The EPSILoN score combines 5 baseline clinical and blood biomarkers and can help to identify patients with aNSCLC who will most likely benefit from second-line IO. Further studies are warranted.
