Abstract 5703: The immune cell state atlas analysis predicts therapeutic benefits with immune checkpoint inhibitors

Abstract Introduction: In this study, we investigated the prognostic role of the immune cell state atlas in predicting therapeutic benefits of patients treated with immune checkpoint inhibitors (ICI) within the ORIEN network of 18 collaborating cancer centers under the Total Cancer Care protocol. Methods: We utilized RNA-seq data of 926 samples generated from 875 individuals. Gene expression data were deconvoluted for immune cell states using the Carcinoma EcoTyper software. We then conducted a series of survival analyses to test the association between survival outcomes and predicted cell types and states in five malignant tumors: Genitourinary (GU), Gastrointestinal (GI), Thoracic (THO), Cutaneous (CUT), Head & Neck (H&N). The regularized Cox regression model in R package ‘glmnet’ was then applied to select the complementary pathway signatures (including gene ontology and KEGG pathways) to the immune cell states in predicting survival outcomes. We also explored the immune-related long non-coding RNAs (lncRNA) as potential biomarkers for cell states and patient outcomes. Results: EcoTyper analysis revealed that 692 (~80%) of patients were assigned to the 10 pre-identified Carcinoma Ecotypes (CE1 to CE10) or cell state atlas group. Overall, two immune deficiency ecotype patient groups (CE1 and CE2) pre-identified based on the independent training data were linked to worse survival, while two proinflammatory ecotype groups (CE9 and CE10) were associated with favorable surxvival. Those ecotype groups showed strong prognostic significance in predicting OS in melanoma and H&N. Meanwhile, CE6, a non-neoplastic tissue enriched cell subtype, was also found to be highly associated with longer OS in H&N and GU. CE7, an age-related mutation patient subgroup, contributed to shorter survival in both melanoma and GI. We also found that a subset of activated B cell state and the exhausted/effector CD4 T cell state were significantly associated with patient survival in melanoma and GU, respectively. The penalized Cox regression model revealed that β-catenin signaling pathway, P53 pathway and heme metabolism in the MSigDB Hallmark gene sets are the most complementary pathways to the ecotype scores in multiple cancer types. In additional, multiple pathways in KEGG such as endocytosis were found to jointly contribute to the ecotype-pathway composite prognostic model. In anazlying immune-related lncRNA biomarkers, we highlighted the prognostic role of NKILA in our dataset, which has been studied to promote tumor immune evasion. Conclusion: Our analysis has successfully established the utility of immune cell state atlas in predicting therapeutic benefits with ICIs. We expect that the discovered complementary signatures in the cancer-cell compartment will also lead to a novel spectrum of tumor-based biomarkers to ICI. Citation Format: Tingyi Li, Vineeth Sukrithan, Aakrosh Ratan, Martin McCarter, John Carpten, Howard Colman, Alexandra P. Ikeguchi, Igor Puzanov, Susanne Arnold, Michelle Churchman, Patrick Hwu, Paulo C. Rodriguez, William S. Dalton, George J. Weiner, Ahmad Tarhini, Xuefeng Wang. The immune cell state atlas analysis predicts therapeutic benefits with immune checkpoint inhibitors. [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 5703.