Abstract:BACKGROUND:Homeobox A (HOXA) family is involved in the development of malignancies as either tumor suppressors or oncogenes. However, their roles in glioblastoma (GBM) and clinical significance have not been fully elucidated.METHODS:HOXA mutation and expressions in pan-cancers were investigated using GSCA and Oncomine, which in GBM were validated by cBioPortal, Chinese Glioma Genome Atlas (CGGA), and The Cancer Genome Atlas (TCGA) datasets. Kaplan-Meier analyses were conducted to determine prognostic values of HOXAs at genetic and mRNA levels. Diagnostic roles of HOXAs in tumor classification were explored by GlioVis and R software. Independent prognostic HOXAs were identified using Cox survival analyses, the least absolute shrinkage and selection operator (LASSO) regression, quantitative real-time PCR, and immunohistochemical staining. A HOXAs-based nomogram survival prediction model was developed and evaluated using Kaplan-Meier analysis, time-dependent Area Under Curve, calibration plots, and Decision Curve Analysis in training and validation cohorts.RESULTS:HOXAs were highly mutated and overexpressed in pan-cancers, especially in CGGA and TCGA GBM datasets. Genetic alteration and mRNA expression of HOXAs were both found to be prognostic. Specific HOXAs could distinguish IDH mutation (HOXA1-7, HOXA9, HOXA13) and molecular GBM subtypes (HOXA1-2, HOXA9-11, HOXA13). HOXA1/2/3/10 were confirmed to be independent prognostic members, with high expressions validated in clinical GBM tissues. The HOXAs-based nomogram model exhibited good prediction performance and net benefits for patients in training and validation cohorts.CONCLUSION:HOXA family has diagnostic values, and the HOXAs-based nomogram model is effective in survival prediction, providing a novel approach to support the treatment of GBM patients.

Development and Validation of a Predictive Model in Diagnosis and Prognosis of Primary Glioblastoma Patients Based on Homeobox A Family.