1678PDImmune classification of soft tissue sarcoma predicts clinical outcome

Abstract Background Soft tissue sarcomas (STS) form a group of rare cancers which accounts for around 1% of tumours. Although up to 15% of patients respond in immunotherapy trials, there are no biomarkers predicting response of STS to checkpoint blockade therapies yet. Methods We analysed transcriptomic data of 4 publicly available cohorts, accounting for more than 600 STS. We used MCP-counter, a deconvolution method to estimate the tumour microenvironment (TME) composition Based on MCP-counter estimates, we established a robust immune classificationof STS tumors into 5 Sarcoma Immune Classes, labelled A, B, C, D and E. These classes exhibited different type and extents of TME. We validated the profiles of these 5 groups on a 72-patients cohort using immunohistochemichal (IHC) stainings for CD3, CD8, CD20 and CD34. Results One group (A: 23.3% of tumours) exhibits a very low to low immune infiltrate for all TME cell types. Another class (C: 14.5% of all tumours) displays moderate immune infiltrate and a strong presence of endothelial cells. Finally, another group (E: 15.6%) is highly infiltrated by all immune cell types. The two remaining groups (B: 27.4% and D: 19.4%) are heterogeneous, respectively rather highly and lowly infiltrated. The immune high group E is associated with an overexpression of several immune checkpoint genes: PDCD1 (PD-1), CD274 (PD-L1), PDCD1LG2 (PD-L2), LAG3, HAVCR2 (TIM-3), CTLA4. On 72 patients, we showed that the immune-high group could be identified by the IHC-visible presence of tertiary lymphoid structures (TLS), defined as T cell aggregates juxtaposing B cell aggregates. The immune-high group also exhibited prolonged overall survival as compared with other groups. Using data from a phase II clinical trial with pembrolizumab, we show that responders can be identified as class E tumours, therefore allowing patient selection. Conclusions We have defined a novel immune-based classification of STS into 5 classes, among which an immune-high group characterized by a strong infiltration by all immune cell and expression of immune checkpoints, presence of TLS and longer overall survival. This class groups responders to PD-1 blockade in a phase II clinical trial. Legal entity responsible for the study INSERM. Funding Institut National de la Sante et de la Recherche Medicale, the University of Paris, Sorbonne University, the Programme Cartes d’Identite des Tumeurs (CIT) from the Ligue Nationale Contre le Cancer, Institut National du Cancer (HTE-INSERM plan cancer, C16082DS), Association pour la Recherche sur le Cancer (ARC), Cancer Research for Personalized Medecine programme, “FONCER contre le cancer” programme, Labex Immuno-Oncology, the National Institute of Health, Moon Shot program at MD Anderson Cancer Center, Ministry of Education and Ministry of Science and Technology of Taiwan, National Taiwan University, Merck, Inc, SARC, Sarcoma Foundation of America, and the QuadW Foundation. Disclosure T.W. Chen: Advisory / Consultancy, Research grant / Funding (self): Eisai; Advisory / Consultancy: Lilly. M.A. Burgess: Advisory / Consultancy: EMD Serono; Advisory / Consultancy: Immune Design; Advisory / Consultancy: Eisai. J. Wargo: Advisory / Consultancy: Merck; Advisory / Consultancy: BMS; Advisory / Consultancy: Novartis; Advisory / Consultancy: AstraZeneca; Advisory / Consultancy: Roche; Advisory / Consultancy: Genentech; Advisory / Consultancy: Illumina. H.A. Tawbi: Advisory / Consultancy, Research grant / Funding (self): BMS; Advisory / Consultancy, Research grant / Funding (self): Merck; Advisory / Consultancy, Research grant / Funding (self): Genentech; Research grant / Funding (self): Celgene; Research grant / Funding (self): GSK. W.H. Fridman: Advisory / Consultancy: MedImmune; Advisory / Consultancy: Novartis; Advisory / Consultancy: Servier; Advisory / Consultancy: Pierre Fabre. All other authors have declared no conflicts of interest.