Abstract Soft tissue sarcomas (STS) form a group of rare cancers which accounts for around 1% of non-pediatric solid tumors. Recent clinical trials with check-point blockade immunotherapies reported a response in up to 15% of patients, but there are no biomarkers predicting response of STS to checkpoint blockade therapies yet. We analyzed transcriptomic data of publicly available cohorts, accounting for more than 600 complex genomic STS, including leiomyosarcoma (LMS, 35.4%), dedifferentiated liposarcoma (DDLPS, 33.9%) and undifferentiated pleomorphic sarcoma (UPS, 30.8%). Using a deconvolution method to estimate the tumor microenvironment composition with accurate scores depicting the immune profile of tumors, we established a robust immune classification that is consistent through different datasets. We divided the patients into 5 Sarcoma Immune Classes, labelled A1, A2, B, C1 and C2, which exhibit different extents and composition of immune infiltrate. Two groups (A1: 23.3% and A2: 27.1% of all tumors) exhibit a very low to low immune infiltrate for all cell types. Another class (B: 14.5% of all tumors) displays moderate immune infiltrate and a strong presence of endothelial cells. Finally, the remaining two groups (C1: 19.4% and C2: 15.6%) are highly infiltrated by all immune cell types. All histologies were found in all five immune classes, although LMS were enriched in classes A1 and A2. The immune-high groups exhibit a significantly prolonged overall survival compared to the other groups. Notably, the immune high group C2 is characterized by an extremely rich immune infiltrate containing both lymphocytes and myeloid cells, and is associated with a striking overexpression of several immune checkpoints genes: PDCD1 (PD-1), CD274 (PD-L1), PDCD1LG2 (PD-L2), LAG3, HAVCR2 (TIM-3), CTLA4. This group also contains tumors with a strong expression of the B cell attractant chemokine CXCL13, often associated with the presence of tertiary lymphoid structures (TLS). We validated the relevance of TLS as a biomarker of this group by quantitative immunohistochemistry (IHC) using stainings for CD3, DC-Lamp, CD8, and CD20 on a 95-patient STS cohort (LMS, 34%; DDLPS, 33%; UPS, 34%). Tumors with visible TLS had a higher density of CD3+, CD8+ and CD20+ cells in the tumor core and the invasive margin, hence showing the possibility to identify the immune-high group by using IHC. The expression of immune checkpoint molecules was also detected by multiplex immunofluorescence for CD3, CD20, PD-1 and Lag3 in these tumors, confirming the presence of checkpoint molecules in the immune-high group. To check whether the immune-high group C2 was associated with enhanced response to checkpoint blockade therapy, we intend to analyze tumors from a recently published phase 2 clinical trial in which patients underwent treatment with anti-PD-1 pembrolizumab. Citation Format: Florent Petitprez, Tom Wei-Wu Chen, Cheng-Ming Sun, Julien Calderaro, Li-Ping Hsiao, Laetitia Lacroix, Ivo Natario, Maud Toulmonde, Carlo Lucchesi, Yec'han Laizet, Antoine Italiano, Aurélien de Reyniès, Catherine Sautès-Fridman, Wolf H. Fridman. A novel transcriptomic-based immune classification of soft tissue sarcoma (STS) and its association with molecular characteristics, clinical outcome and response to therapy [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 4045.