A development and validation of predictive model based on novel immune-related gene-based subtypes for the risk...

Cutaneous melanoma (CM) exhibits considerable heterogeneity, and the immune status of patients can serve as a prognostic indicator. The increasing significance of immune-related markers in cancer prognosis provides clinicians with valuable tools for risk stratification and management decisions. The objective of this study was to develop a predictive model for assessing the risk of CM based on novel subtypes delineated according to immune-related genes.

This study included a cohort from The Cancer Genome Atlas (TCGA). Immune-related genes were carefully selected, and a comprehensive analysis was performed to characterize the molecular alterations and clinical implications linked to these genes. From this, an immune-related risk scoring system aimed at predicting the survival outcomes of patients diagnosed with CM was developed.

In this study, using an unsupervised consensus clustering algorithm, the study identified two subtypes-Cluster 1 (C1) and Cluster 2 (C2)-within the TCGA melanoma (MEL) cohort based on 1,959 immune-related genes. Survival analysis indicated that C1 was linked to poorer overall survival (OS) as compared to C2. We found significant correlations between these subtypes and clinical variables including tumor-node-metastasis (TNM) classification, new tumor events, and radiation therapy. Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis revealed that 161 genes upregulated in C1 were associated with tyrosine metabolism, melanogenesis, and the p53 signaling pathway, while downregulated genes in C1 were linked to hematopoietic cell lineage, cytokine-cytokine receptor interactions, and cell adhesion molecules. Immune-related genes in CM were optimized and assessed using univariate Cox regression and a protein-protein interaction (PPI) network, with 20 genes being identified, including CXCL10, CCL5, CXCR4, CXCR3, IL10, CCR5, CCR7, STAT1, TNF, CD4, CD8A, ITGB2, FCGR3A, ITGAM, PTPRC, CD19, LCK, B2M, TYROBP, and IFNG. From these, four key prognostic markers (CXCL10, IL10, B2M, and IFNG) were selected via a least absolute shrinkage and selection operator (LASSO) regression penalty approach and multivariate Cox analyses. For the prediction of the 1-, 3-, and 5-year survival rates, the immune-related risk score yielded area under the curve (AUC) values of 0.671, 0.667, and 0.676, respectively.

CM was divided into two subtypes based on immune gene expression, with the C1 subtype associated with poor prognosis. A prognostic risk model was developed using these classifications to predict patient outcomes.