Find in Library

Growing evidence indicates a link between retinoic acid (RA) metabolism and sarcoma progression or immunity in laboratory studies. However, a comprehensive analysis of RA abnormality in the sarcoma population is still lacking. Herein, we systematically analyzed the molecular features of 19 retinoic acid metabolism-related enzymes and sarcoma patients’ clinical information based on TCGA/TARGET/GSE datasets. We identified two RA expression subtypes, which were related to distinct clinical survival outcomes and exhibited different biological features. Gene set enrichment analysis indicated a set of immune pathways were enriched in G1 while oncogenic pathways were enriched in G2. Immune cell infiltration analysis using the TIMER algorithm revealed more CD4+ and CD8+ T cell infiltration in G1 subgroups than in G2. Moreover, we generated a seven genes signature to predict the RA metabolism index based on the LASSO-penalized Cox regression model. Survival analysis demonstrated the significant prognostic differences between high- and low-risk groups among different bone and soft tissue datasets. A higher risk index was associated with less T cell CD8+ infiltration. The predictive ability of the RA risk score was validated in 71 bone or soft tissue sarcoma clinical samples. These results indicated that RA-based classification could distinguish sarcoma patients with different clinical outcomes and immune statuses, which may help to explore better treatment decision-making for sarcoma patients.