Xuefeng Lv ^1,2 Xiaohan Ma ^1,2* Shu Deng ^1,2* Jinming Xie ^3* Enwu Yuan ^1,2*

• ¹ Department of Laboratory Medicine, Third Affiliated Hospital of Zhengzhou University, Zhengzhou, Henan Province, China
• ² Independent researcher, Zhengzhou, China
• ³ Independent researcher, Deyang, China

Numerous studies have demonstrated acute myeloid leukemia (AML) is one of the malignancies with high mortality worldwide. Immunogenic cell death (ICD) is a form of cell death that is specialised in that it triggers the body's immune response, particularly the adaptive immune response. Recent evidence has confirmed that pseudogenes are implicated in multiple human tumorigenesis and progression although lacking the function of coding protein. However, the roles of ICD-associated genes in AML remain largely unascertained. TCGA-AML and GSE71014 cohorts were picked out and we combined them into a merged dataset by removing the batch effect using the sva package in the R project. A consensus clustering analysis of the ICD genes in AML was performed to define two subgroups (C1 and C2 clusters) with distinct ICD patterns in AML. Most ICD-related genes were significantly up-regulated in the C2 cluster. The single sample gene set enrichment analysis (ssGSEA) revealed that the immune cell infiltration and immune checkpoint gene expression of the C2 cluster was strongly high, suggesting that the C2 population responded well to immune checkpoint blockade (ICB) therapy and had better survival. The C1 group was sensitive to chemotherapy, including Cytarabine, Midostaurin, and Doxorubicin. On the other hand, 15 ICD-related pseudogenes were identified to be associated with AML prognosis. Based on the expression of 15 prognostic-related pseudogenes, we developed a prognostic model and categorized AML samples into low and high-risk groups. The receiver operator curve (ROC) analysis and nomogram manifested that our prognostic model had high accuracy in predicting survival. However, the high-risk group was sensitive to ICB therapy and chemotherapy such as Methotrexate, Cytarabine, and Axitinib while the low-risk group benefited from 5-Fluorouracil, Talazoparib, and Navitoclax therapy. In summary, we defined two subgroups relying on 33 ICD-related genes and this classification exerted a decisive role in assessing immunotherapy and chemotherapy benefit. Significantly, a prognostic signature identified by critical ICDrelated pseudogene was created. The pseudogene prognostic signature had a powerful performance in predicting prognosis and therapeutic efficacy, including immunotherapy and chemotherapy to AML. Our research points out novel implications of ICD in cancer prognosis and treatment approach choice.