| Objective: Osteosarcoma is the most common primary malignant bone tumour and affects mainly young people.The prognosis of patients with unrespectable or recurrent osteosarcoma is poor.The aim of this study was to find the prognostic molecular features of osteosarcoma and to construct risk models to predict prognosis.Methods: Transcriptomic and clinical data of osteosarcoma patients were downloaded from the TARGET database.Genes differentially expressed in the metastatic and non-metastatic groups were analyzed and enrichment analysis was performed.Univariate and multivariate Cox regression analysis were performed to screen for prognostic genes and to develop risk prediction models.A nomogram incorporating clinical information was then developed and validated to predict overall patient survival at 2,3 and 5 years.Result: A total of 51 differential genes were identified by transcriptome analysis.GO enrichment analysis revealed that differential genes were mainly involved in monocyte proliferation,cytokine regulation,IL-10 regulation and lymphocyte proliferation pathways.Univariate Cox analysis yielded 10 prognosis-related differential genes.And a multivariate Cox regression model based on five prognosis-related differential genes(AOC3,GPI,MYC,SLC27A5,RPL22L1)was constructed and successfully classified patients into high and low risk groups,which showed significantly different prognosis.In addition,clinical information such as gender,age,risk score,metastasis and tumour primary site were incorporated into the Nomogram based on a risk prediction model to predict overall survival of patients at 2,3 and 5 years with a model C index of 0.77.Conclusion: Our study provided 10 genes as molecular markers of osteosarcoma prognosis,and a risk model based on prognostic genes was effective in predicting survival outcomes of osteosarcoma patients. |
