ISSN: 1837-9664Journal of Cancer
Global reach, higher impact
J Cancer 2020; 11(15):4297-4307. doi:10.7150/jca.43805 This issue Cite
Research Paper
Survival Risk Prediction Models of Gliomas Based on IDH and 1p/19q
1. Xiangya School of Medicine, Central South University, 172 Tongzipo Road, Changsha, Hunan 410013, China
2. Department of Neurosurgery, Xiangya Hospital, Central South University, 87 Xiangya Road, Changsha, Hunan 410008, China
3. Hunan International Scientific and Technological Cooperation Base of Brain Tumor Research, Xiangya Hospital, Central South University, No. 87, Xiangya Road, Changsha, Hunan 410008, China
*These authors contributed equally to this work.
Abstract
Gliomas have been classified into different molecular subtypes based on their molecular features. To explore the prognostic factors of different subtypes of gliomas, we performed a univariate survival analysis based on the RNA-seq data of 653 patients obtained from The Cancer Genome Atlas. We identified 12205 (20.18%), 6125 (10.13%) and 5206 (8.61%) genes associated with the overall survival (OS) of the IDH-wildtype, IDH-mutation 1p/19q intact and IDH-mutation 1p/19q codeletion gliomas, respectively. Pathway enrichment analysis revealed that OS related genes were mainly involved in alcoholism, systemic lupus erythematosus, hematopoietic cell lineage and diabetes. The OS related genes were further selected using Lasso regression, and three prognostic risk score models were constructed to effectively predict the OS of the patients with different subtypes of gliomas. In total, 76 signature genes were identified and were selected to construct the three models. Moreover, neither of the 76 genes overlapped between different models, which suggested the enormous difference among the three subtypes, although some signature genes (SERPINA5, RP11.229A12.2 and RP11.62F24.2) were also identified as the OS related genes in different glioma subtypes. Interestingly, five genes (RP11.229A12.2, RP11.62F24.2, C3orf67, RP11.275H4.1 and TBX3) played opposing roles (protective or risk factor) in different subtypes. Additionally, the prognosis models consisted of a substantial proportion of non-coding RNA (58.74%, 70.13% and 58.11% in the IDH-wildtype, IDH-mutation 1p/19q intact and IDH-mutation 1p/19q codeletion). Furthermore, multivariate analysis integrating clinical variables demonstrated that risk group predicted by the prognostic models was an independent prognostic factor for gliomas. In conclusion, we have constructed and validated three models that have the potential to predict the prognosis of glioma patients. The genes and pathways identified in this study require further investigation for their underlying mechanisms and potential clinical significance in improving the OS of the glioma patients.
Keywords: survival analysis, prediction model, ncRNA, molecular features, glioma.
Citation styles
