147 Background: Inhibitors of the immune checkpoint PD-1/PD-L1 have become a standard of care in NSCLC. Patient selection, currently based on PD-L1 expression in tumor tissue, is limited by its temporal and spatial heterogeneity. We hypothesized that monitoring PD-L1 staining of circulating tumor cells (CTCs) could represent a valuable non-invasive biomarker. Methods: Up to 3 blood samples were prospectively collected from patients with advanced NSCLC: i) pretreatment (nivolumab), ii) first follow-up, iii) progression. CTCs were isolated from 10 mL of blood using cell size-based technology (ISET, Rarecells). PD-L1 expression was assessed by immunofluorescence on CTCs and immunohistochemistry on tissue. Results: 162 samples from 96 patients were collected. PD-L1 expression could be assessed pretreatment on tissue and CTCs in 72% and 93%, respectively; and was ≥1% in 37% and 83%; ≥5% in 35% and 79%, respectively. No correlation between tissue and CTCs PD-L1 expressions was observed (Spearman coefficient correlation = 0.04, p = 0.77). At baseline, each 10/7.5 ml increase in CTCs count was associated with increased risk of death and progression (HR[95%CI]: 1.06 [1.005;1.117] p = 0.03 for OS and HR[95%CI]: 1.05 [1.01;1.10] p = 0.02 for PFS). The presence of PD-L1(+)CTC (≥1%) had no prognostic impact (OS: p = 0.89 and PFS: p = 0.55), but pretreatment PD-L1(+)CTCs were more frequent in the “non-responders” group (PFS < 6 months) (p = 0.04). Median CTC count was 30 (n = 96), 68.3 (n = 44) and 50.3 (n = 22) pretreatment, at the first follow-up and at progression, respectively. The changes in CTC count at the first follow-up had no impact on PFS (p = 0.45) or OS (p = 0.68). 96% of patients had PD-L1(+)CTCs at progression. Further analyses are ongoing to assess the prognostic value of the persistence of CTCs and PD-L1(+)CTCs at the first follow up and at progression. Conclusions: Analysis of PD-L1 expression on CTCs is highly feasible. PD-L1 expressions on tissue and CTCs are discordant, CTCs being more likely positives, suggesting false negatives occurring in small biopsies. Further analyses of the kinetics of this biomarker throughout ICI treatment, along with other circulating biomarkers, are ongoing. Clinical trial information: NCT02827344.