Establishment of an Autophagy-Related Clinical Prognosis Model for Predicting the Overall Survival of Osteosarcoma

PURPOSE: Osteosarcoma is the most common primary and highly invasive bone tumor in children and adolescents. The purpose of this study is to construct a multi-gene expression feature related to autophagy, which can be used to predict the prognosis of patients with osteosarcoma. Materials and methods. The clinical and gene expression data of patients with osteosarcoma were obtained from the target database. Enrichment analysis of autophagy-related genes related to overall survival (OS-related ARGs) screened by univariate Cox regression was used to determine OS-related ARGs function and signal pathway. In addition, the selected OS-related ARGs were incorporated into multivariate Cox regression to construct prognostic signature for the overall survival (OS) of osteosarcoma. Use the dataset obtained from the GEO database to verify the signature. Besides, gene set enrichment analysis (GSEA) were applied to further elucidate the molecular mechanisms. Finally, the nomogram is established by combining the risk signature with the clinical characteristics. RESULTS: Our study eventually included 85 patients. Survival analysis showed that patients with low riskScore had better OS. In addition, 16 genes were included in OS-related ARGs. We also generate a prognosis signature based on two OS-related ARGs. The signature can significantly divide patients into low-risk groups and high-risk groups, and has been verified in the data set of GEO. Subsequently, the riskScore, primary tumor site and metastasis status were identified as independent prognostic factors for OS and a nomogram were generated. The C-index of nomogram is 0.789 (95% CI: 0.703~0.875), ROC curve and calibration chart shows that nomogram has a good consistency between prediction and observation of patients. CONCLUSIONS: ARGs was related to the prognosis of osteosarcoma and can be used as a biomarker of prognosis in patients with osteosarcoma. Nomogram can be used to predict OS of patients and improve treatment strategies.