Development and validation of a novel prognostic signature for osteosarcoma.

Background: Osteosarcoma (OS) is a primary malignant bone tumor known for its aggressive nature and high potential for metastasis, primarily impacting adolescents and young adults. Few studies have focused on discovering gene signatures that effectively guide treatment strategies and predict outcomes in OS. The study aimed to create a prognostic model to predict OS patient survival. Methods: Candidate genes were identified by intersecting genes from survival analyses of two patient datasets. Unsupervised consensus clustering was employed with these candidate genes to investigate potential tumor types. Following this, a gene signature was developed and validated within one of the datasets. Finally, the immune microenvironment was assessed using the CIBERSORT and ESTIMATE algorithms. Results: A total of 123 candidate genes were identified, and samples from the Gene Expression Omnibus (GEO) dataset were subsequently partitioned into two clusters through unsupervised cluster analysis utilizing these candidate genes. Univariate Cox regression was employed to select 46 genes, from which a ten-gene signature was developed using the least absolute shrinkage and selection operator (LASSO) regression to predict the prognosis of OS. The resulting novel gene signature demonstrated significant predictive accuracy for the overall survival of patients with OS. Conclusions: In conclusion, we have discovered a ten-gene signature that serves as a new prognostic predictor for OS patients.