Extra survival tree model based on genes associated with genomic instability for improving clinical outcomes and making treatment recommendations for gliomas

Genomic instability is a hallmark of cancer and may encourage the formation, heterogeneity, and spread of tumors. Several recent studies have shown that genes associated with genomic instability can aid in predicting the prognoses of patients with tumors. However, the gene screening methodology and the process of prognostic modeling are incomplete. We applied a quantitative integrated genomic instability framework to identify genes associated with genomic instability in The Cancer Genome Atlas glioma cohort. A comprehensive machine learning-based technique was used to create a genomic instability-related signature (GIRS). GIRS aided in predicting patient prognosis with greater accuracy and clinical benefits than classical clinical and molecular pathological features of gliomas. Compared with the 126 published models, GIRS had better predictive and generalization properties. In addition, GIRS reflects the degree of genomic instability and can be used as a new indicator of genomic instability. In addition, the GIRS can be used to recommend treatments for patients undergoing radiotherapy and chemotherapy to obtain the best therapeutic effects. We successfully identified genomic instability-related genes in gliomas and developed a GIRS risk model, a novel marker of genomic instability, to effectively predict the prognosis of patients with glioma.