A prognostic model based on immune cells in tumor microenvironment to predict prognosis in endometrial cancer.

OBJECTIVE: The interaction between the tumor immune microenvironment (TIME) and malignant tumor cells plays a crucial role in tumor initiation and progression. This study aimed to establish and validate a prognostic model based on TIME characteristics to predict prognosis and guide personalized treatment in patients with endometrial cancer (EC). METHODS: A total of 67 EC patients who underwent surgery and TIME profiling at Peking University People's Hospital between January 2018 and December 2022 were included in this study. A prognostic model was developed based on the densities of stroma CD3(+)cell and stroma CD8(+)cells. To validate the model, an independent cohort of 200 EC patients was used, in which immunohistochemical (IHC) staining for CD3(+) and CD8(+) cells was performed to assess the model's predictive accuracy. RESULTS: (1) Multiplex immunofluorescence (mIF) analysis of the 67 EC patients revealed significant differences between the Recurrence and Non-Recurrence groups in the densities of stroma PD-L1(+) cell, CD8(+) cell, CD68(+)CD163(-) cell, CD3(+) cell and CD56(+) cell, with stroma CD3(+) cell showing the most significant difference (P = 0.004); (2) In 514 EC patients from The Cancer Genome Atlas (TCGA) database, significant differences were observed between the Recurrence and Non-Recurrence groups in the abundance of CD8(+) cell, regulatory T cells (Tregs), and activated dendritic cells (DCs), with CD8(+) cell showing the strongest association (P < 0.001); (3) Stroma CD3(+) and CD8(+)cells were selected as modeling variables to construct the prognostic model, which stratified patients into three clusters: Cluster 1 (n = 17), Cluster 2 (n = 39), and Cluster 3 (n = 11). Survival analysis demonstrated significant differences among the three clusters (P = 0.006); (4) The three clusters also exhibited distinct immune cell compositions, molecular subtypes, and clinicopathological characteristics; (5) In the validation cohort of 200 EC patients, clustering based on IHC-measured CD3(+) and CD8(+) cells densities produced three clusters (Cluster 1, Cluster 2, and Cluster 3) with significantly different survival outcomes (P < 0.001), confirming the predictive accuracy of the proposed model. CONCLUSION: This study identified two immune cell types-Stroma CD3(+) and CD8(+)cells-significantly associated with the prognosis of EC and established a TIME-based prognostic model with robust predictive performance and accuracy.