Despite a proportion of renal cancer patients can experiment marked and durable responses to immunecheckpoint inhibitors, the treatment efficacy is widely variable and identifying the patient who will benefit from immunotherapy remains an issue. We performed a prospective study to investigate if soluble forms of the immune-checkpoints PD-1 (sPD-1), PD-L1 (sPD-L1), pan-BTN3As, BTN3A1, and BTN2A1, could be candidate to predict the response to immune-checkpoint blockade therapy. We evaluated the plasma levels in a learning cohort of metastatic clear cell renal carcinoma (mccRCC) patients treated with the anti-PD-1 agent nivolumab by ad hoc developed ELISA's. Using specific cutoffs determined through ROC curves, we showed that high baseline levels of sPD-1 (>2.11 ng/ml), sPD-L1 (>0.66 ng/ml), and sBTN3A1 (>6.84 ng/ml) were associated with a longer progression-free survival (PFS) to nivolumab treatment [median PFS, levels above thresholds: sPD-1, 20.7 months (p < .0001); sPD-L1, 19 months (p < .0001); sBTN3A1, 17.5 months (p = .002)]. High sPD-1 and sBTN3A1 levels were also associated with best overall response by RECIST and objective response of >20%. The results were confirmed in a validation cohort of 20 mccRCC patients. The analysis of plasma dynamic changes after nivolumab showed a statistically significant decrease of sPD-1 after 2 cycles (Day 28) in the long-responder patients. Our study revealed that the plasma levels of sPD-1, sPD-L1, and sBTN3A1 can predict response to nivolumab, discriminating responders from non-responders already at therapy baseline, with the advantages of non-invasive sample collection and real-time monitoring that allow to evaluate the dynamic changes during cancer evolution and treatment.