Identification of an Immune Cell Infiltration-Related Prognostic Model to Guide Immunotherapy in Endometrial Carcinoma

Background: Uterine corpus endometrial carcinoma (UCEC) is a prevalent gynaecologic malignancy. It has been demonstrated that the immune cell infiltration (ICI) pattern plays a critical role in the tumour progression of UCEC. Methods: To further investigate the immune microenvironment landscape of UCEC, we analysed the gene expression data of 539 UCEC patients from The Cancer Genome Atlas (TCGA) database using CIBERSORT and ESTIMATE for consensus clustering of immune cells. We used the limma package to compare differentially expressed genes (DEGs) among ICI patterns and constructed a prognostic model using Cox regression to calculate the risk score of UCEC patients. The immunophenoscore was downloaded to explore the immunotherapeutic effect between low- and high-risk score patients. Finally, the tumour mutation burden (TMB) was calculated using the somatic mutation data. Results: We identified two different immune infiltration patterns in 539 UCEC samples, the immune-desert and immune-inflamed phenotypes, which had distinct prognostic and biological features. We obtained 29 DEGs to construct the ICI-related prognostic model and established a four ICI-related gene-based prognostic model comprising LINC01871, CXCL13, IGKJ5, and LINC01281. The risk score was associated with distinct clinical outcomes, ICI, and immunotherapeutic effects. Patients with a low risk score had higher effective immune cells, which could be classified into the immune-inflamed phenotype. Additionally, patients with a low risk score had a significantly higher immunophenoscore, suggesting a better immunotherapeutic outcome. Finally, TMB was confirmed to be associated with prognosis, which was synergistic with the risk score. Conclusions: This study comprehensively analysed the ICI pattern in UCEC patients and established a four ICI-related gene-based prognostic model to predict prognosis and guide precise immunotherapy strategies.