A Macrophage/Monocyte‐Related Four‐Gene Signature for Prognostic Assessment of Uveal Melanoma: BTBD6 , C2CD4B , CCL24 , and S100A4

Background Uveal melanoma (UVM) is a highly malignant ocular tumor with a poor prognosis. Macrophages and monocytes in the tumor microenvironment promote immune escape, angiogenesis, and metastasis. Thus, exploring their roles may provide insights into UVM progression. Methods The Cancer Genome Atlas (TCGA) was accessed to obtain the data of mRNA expression and follow‐up data of UVM, and UVM single‐cell profiles were downloaded to cluster cells by annotation of single‐cell marker genes. The differentially expressed genes (DEGs) in macrophage/monocyte cells compared to other cell types were revealed. ssGSEA was applied to compute the score of DEGs and to reveal the genes for WGCNA in UVM. A prognostic risk model for UVM was constructed by uni/multivariate Cox and LASSO regression analyses to reveal the differential overall survival status. Further cellular validations were conducted to examine the effects of core genes in UVM. TIMER tool was applied for the analysis of immune cell infiltration levels in UVM. Chemotherapeutic drug sensitivity in UVM was assessed with the pRRophetic package. Results Six cell subpopulations were identified in the UVM samples, among which macrophage/monocyte cells were more predominant. Kaplan–Meier curves showed that UVM patients in the group of high RiskScore (consisting of the genes BTBD6 , C2CD4B , CCL24 , and S100A4 ) presented a poorer prognosis, higher infiltration of monocytic lineage, T cells, CD8 T cells, cytotoxic lymphocytes, and higher expression of immune checkpoint–related genes. A significant negative correlation between RiskScore and the IC 50 of XMD8‐85, lapatinib, roscovitine, salubrinal, bexarotene, LFM‐A13, FTI‐277, and TGX221 chemotherapeutic agents was further noticed. Conclusion In this study, we computationally identified genes associated with both disease progression and macrophage/monocyte‐related characteristics in UVM and constructed a prognostic risk model with predicted immune infiltration patterns. These findings generate testable hypotheses that may inform future experimental studies on the immune mechanisms underlying UVM.