Search

Recent studies have emphasized the close relationship between macrophages and tumor immunity, and the prognosis of lung adenocarcinoma (LUAD) patients is intimately linked to this. Nonetheless, the prognostic signature and classification of different immune patterns in LUAD patients based on the macrophages is largely unexplored. Two sc-RNAseq datasets of LUAD patients were collected and reprocessed. The differentially expressed genes (DEGs) related to macrophages between LUAD tissues and normal lung tissues were then identified. Based upon the above genes, three distinct immune patterns in the TCGA-LUAD cohort were identified. The ssGSEA and CIBERSORT were applied for immune profiling and characterization of different subtypes. A four-gene prognostic signature for LUAD patients was established based on the DEGs between the subtypes using stepwise multi-Cox regression. TCGA-LUAD cohort was used as training set. Five GEO-LUAD datasets and an independent cohort containing 112 LUAD samples were used for validation. TIDE (tumor immune dysfunction and exclusion) and drug sensitivity analyses were also performed. Macrophage-related differentially expressed genes were found out using the publicly available scRNA-seq data of LUAD. Three different immune patterns which were proved to have distinct immune infiltration characteristics in the TCGA-LUAD cohort were recognized based on the above macrophage-related genes. Thereafter, 174 DEGs among the above three different immune patterns were figured out; on the basis of this, a four-gene prognostic signature was constructed. This signature distinguished the prognosis of LUAD patients well in various GSE datasets as well as our independent cohort. Further analyses revealed that patients which had a higher risk score also accompanied with a lower immune infiltration level and a worse response to several immunotherapy biomarkers. This study highlighted that macrophage were significantly associated with TME diversity and complexity. The four-gene prognostic signature could be used for predicting outcomes and immune landscapes for patients with LUAD. [Display omitted] • Depicting the LUAD landscape and acquiring macrophage-related differentially expressed genes from scRNAseq data of LUAD. • Identifying and depicting 3 distinct LUAD subtypes with distinct immune landscapes which were termed as "immune-infiltrated phenotype", "immune-excluded phenotype" and "immune-desert phenotype". • Establishment of a four-gene prognostic model for LUAD patients and validation of the signature using publicly attainable LUAD cohorts and an independent cohort containing 112 LUAD samples. [ABSTRACT FROM AUTHOR]