2596 Background: Given that most of cancer patients treated with anti-PD1/PD-L1 immune checkpoint blockers (ICB) do not derive benefit, there is a crucial need to identify reliable predictive biomarker of response. Besides PD-1, several key immune checkpoints, such as CTLA4, LAG3, TIM3 and TIGIT, are associated with a T cell exhausted phenotype and play a crucial role in leading to cancer immune evasion. The impact of simultaneous expression by T cells of distinct inhibitory receptors on outcome of patients treated with ICB is still unknown. Methods: We analyzed the tissue samples, collected before ICB initiation, from patients with solid tumors and included in an institutional molecular profiling program (NCT02534649). We used multiplexed-immunohistofluorescence with the following panel CD3/PD1/TIM3/LAG3/TIGIT/CTLA4, and performed immune cell characterization using multispectral images analysis. We then investigated the correlation between coexpression of T cell-associated exhaustion markers, clinical response rate, progression-free survival (PFS) and overall survival (OS) by Cox proportional hazards models. Results: Four hundred thirty five patients were included in the analysis (NSCLC: n=207, 47.6%; sarcoma: n=42, 9.7%; urothelial: n=30, 6.9%; others: n=156, 35.9%). Digital pathology analysis allowed us to identify a population of “super-exhausted” T cells characterized by the co-expression of PD1, LAG3, TIGIT and TIM3 which was enriched in 125 cases (28.7%), and was significantly associated with better PFS (HR 1.60, CI95 1.26-2.04, p<0.001) and OS (HR 1.42, CI95 1.07-1.89, p=0.016) in the whole cohort. Patients with super-exhausted high tumors had higher objective response rate (38.4%) compared to super-exhausted low tumors (19.7%, p<0.001). The presence of super-exhausted T cells was significantly higher in responders (10%) versus non responders (4%, p<0.001). Correlation with better outcome was observed whatever the subgroup considered (NSCLC vs other tumors, CD8 T cells density and presence of tertiary lymphoid structure [TLS]). In multivariate analysis (n=372, 85.5%), increased tumor infiltration by super-exhausted T cells (>1 %) was significantly associated with better PFS (HR 0.61, CI95 0.46-0.81, p<0.001, Table) and OS (HR 0.68, CI95 0.48-0.97, p=0.033, Table). Conclusions: The presence of super-exhausted T cells may represent a new predictive biomarker of response to ICB and pave the way for the development of effective ICB combinations. Data from an independent validation cohort will be presented at the meeting. [Table: see text]