Development and validation of apoptosis‐related signature and molecular subtype to improve prognosis prediction in osteosarcoma patients

BACKGROUND: Previous evidence has shown that apoptosis performs integral functions in the tumorigenesis and development of various tumors. Therefore, this study aimed to establish a molecular subtype and prognostic signature based on apoptosis‐related genes (ARGs) to understand the molecular mechanisms and predict prognosis in patients with osteosarcoma. METHODS: The GEO and TARGET databases were utilized to obtain the expression levels of ARGs and clinical information of osteosarcoma patients. Consensus clustering analysis was used to explore the different molecular subtypes based on ARGs. GO, KEGG, GSEA, ESTIMATE, and ssGSEA analyses were performed to examine the differences in biological functions and immune characteristics between the distinct molecular subtypes. Then, we constructed an ARG signature by LASSO analysis. The prognostic significance of the ARG signature in osteosarcoma was determined by Kaplan–Meier plotter, Cox regression, and nomogram analyses. RESULTS: Two apoptosis‐related subtypes were identified. Cluster 1 had a better prognosis, higher immunogenicity, and immune cell infiltration, as well as a better response to immunotherapy than Cluster 2. We discovered that patients in the high‐risk cohort had a lower survival rate than those in the low‐risk cohort according to the ARG signature. Furthermore, Cox regression analysis confirmed that a high risk score independently acted as an unfavorable prognostic marker. Additionally, the nomogram combining risk scores with clinical characteristics can improve prediction efficiency. CONCLUSION: We demonstrated that patients suffering from osteosarcoma may be classified into two apoptosis‐related subtypes. Moreover, we developed an ARG prognostic signature to predict the prognosis status of osteosarcoma patients.