656 Background: Anti-PD-1 inhibitor nivolumab and combination checkpoint inhibition (nivolumab and ipilimumab) are approved for advanced clear cell renal cell cancer (ccRCC), with objective response rates between 25% to 40%. PD-L1 expression is a poor predictor of anti-PD-1/PD-L1 therapy response in ccRCC. Better predictive biomarkers are required. The CLEAR score (CS) is a multigene prognostic expression score for ccRCC (Eur Urol 2015) previously associated with TKI outcomes. Given association of prognosis and clinical benefit, CS was explored as a biomarker for response to anti-PD-1/PD-L1 therapy in this study. Methods: CS is derived from the expression of 8 genes in ccRCC tumor tissue and intervening stroma and includes components of chemokine signaling. Based on this, immunohistochemistry of ccRCC tumor tissue (n = 24; matched CS also derived from same tissue), was done for tumor-infiltrating lymphocytes (CD8+) and PD-L1. Staining intensity was also correlated to CS. In another set of RCC patients who received nivolumab or pembrolizumab after progression on anti-angiogenic therapy (n = 12), correlation of their tumor CS and response outcome to anti-PD-1 therapy was done. Results: CS was significantly higher for ccRCC tumors with higher intratumor and/or stromal CD8+ cell infiltration (p < 0.001). A similar correlation with PD-L1 staining was not observed. Among RCC patients receiving anti-PD-1 therapy, clinical benefit (stable disease or partial response) was characterized by higher CS, regardless of treatment line. When analysis was limited to ccRCC histology, CS were significantly higher (p < 0.05) for responders (n = 2) compared to non-responders (n = 4) to nivolumab treatment. Conclusions: This is the first study reporting the correlation of a multi-gene score (CS) with immune phenotypes in ccRCC. CS correlates well with intratumor/stromal CD8+ cell infiltration, and is a potential biomarker for response to anti-PD-1 therapy. Consistent with clinical trial data, high CS (associated with poor prognostic tumors) correlated with better responses. Studies with a larger cohort of patients are underway to establish the utility of CS in predicting response to anti-PD-1/PD-L1 therapy.