Background: Despite the remarkable success of immunotherapy, only ~20-40% of cancer patients display an extended durable response. Pre-existing biomarkers for immunotherapy outcome are significantly limited in their predictive power, e.g., PDL-1 expression by tumor cells (AUC ~ 0.6-0.75). This requires additional efforts to search for new approaches to discover new biomarkers in immune-oncology. While mice are the most widely used and cost-effective model to study human disease, translating preclinical biomarkers into clinical practice faces significant obstacles – in part due to the lack of diverse or appropriate models. In this study we have taken a holistic approach to search for new biomarkers for immunotherapy outcomes, with the aim to improve translatability of the findings and their validity in cancer patients. Methods: Several pre-clinical models were used, each capturing one possible mechanistic aspect of immunotherapy response, such as tumor- and host-dependency, in order to reflect the variability seen in human cancers. These models have been analyzed for myeloid cell composition within tumors and peripheral blood using high throughput techniques such as single-cell RNA sequencing and mass cytometry, followed by rigorous bioinformatic tools. Peripheral blood mononuclear cells (PBMCs) were obtained from non-small cell lung cancer (NSCLC, n = 34) and melanoma (n = 38) patients prior to immune-checkpoint inhibitor backbone therapy as monotherapy or in combination with chemotherapy. These results were validated on additional human bulk mRNA datasets (n = 417). Results: Using the preclinical models, we identified interferon-stimulated, Ly6E^hi neutrophils as a pre-treatment, blood-borne biomarker for anti-PD1 response. These results were validated in PBMCs of advanced metastatic NSCLC and malignant melanoma patients predominately treated with immune checkpoint inhibitor-based therapy, with a high degree of accuracy (AUC ~ 0.9). These results were also validated in independent cohorts of immunotherapy-treated patients with other cancer types obtained from bulk mRNA datasets including renal cell carcinoma (n = 109), urothelial carcinoma (n = 228) and additional melanoma patients (n = 80). Functional studies revealed that Ly6E^hi neutrophils sensitize otherwise resistant tumors to anti-PD1 therapy, in part by directly activating cytotoxic T cells and contributing to tumor cell killing, while operate upstream of T cells in the immunotherapy response. Conclusions: Our study demonstrates a pragmatic approach to search for new, clinically relevant and functionally active cellular biomarkers for immunotherapy outcomes in humans. This study paves the way for a conceptual framework to advance the field of immuno-oncology by utilizing this approach for successful clinical trials.