Background: A sizable proportion of PD-L1≥50% NSCLC patients (pts) do not respond to single agent immune checkpoint inhibitors (ICI) and no biomarker is able to identify non responders. A three levels plasma microRNA signature classifier (MSC), reflecting an immunosuppressive profile of immune cell subsets, has already shown its ability to identify pts treated with ICI with a worse prognosis independently from PD-L1 expression. Aim of this trial is to prospectively test the ability of MSC to identify at diagnosis PD-L1≥50% non responders to ICI. Methods: we prospectively collected baseline plasma samples from 41 consecutive advanced EGFR/ALK/ROS1 wild-type NSCLC pts with PD-L1 TPS ≥50% treated with ICI as first (n = 28) or further line treatment to run the MSC molecular test. Pts were stratified in high (H) vs intermediate/low (I/L) risk levels. Overall response rate (ORR) and the relative risk of response (RR) were evaluated by 2x2 contingency table using Fisher exact test. Cox proportional hazard models were used to define crude and adjusted hazard ratio (HR). Results: According to RECIST 1.1 criteria, 14 (34%) pts respond to ICIs. With a global median follow-up of 9.8 months, median progression free survival (PFS) and overall survival (OS) were 7.9 months and not reached, respectively. Ten (24%) NSCLC pts were MSC H risk level. ORR was 0% in MSC H vs 45% in MSC I/L risk pts (RR = 0.10; 95%CI = 0.00-0.90; p = 0.0080). Median PFS was 11.4 months for MSC I/L pts vs 2.3 months for MSC H risk (HR = 0.26 95%CI = 0.11-0.62; p = 0.0021). Median OS was not reached for MSC I/L vs 2.7 months for MSC H risk pts (HR = 0.17 95%CI = 0.06-0.48; p = 0.0008). Data remained significant adjusting for age, sex, pack-years and ECOG performance status at the baseline: PFS HR = 0.24 (95%CI = 0.09-0.66; p = 0.0054) and OS HR = 0.15 (95%CI = 0.05-0.51; p = 0.0023). Conclusions: plasma MSC shows promising results for treatment selection with ICI. So far, MSC is the only molecular test able to identify a group of NSCLC pts with PD-L1≥50% who do not respond to single agent immunotherapy. Ongoing trials are validating these results and testing the possible predictive effect of MSC in chemotherapy plus immunotherapy combinations.