[Expression of Glutathione Peroxidases and Its Effect on Clinical Prognosis in Glioma Patients].

Objective To investigate the relationship between the expression of glutathione peroxidase(GPX)genes and the clinical prognosis in glioma patients,and to construct and evaluate the model for predicting the prognosis of glioma. Methods The clinical information and GPX expression of 663 patients,including 153 patients of glioblastoma(GBM)and 510 patients of low-grade glioma(LGG),were obtained from The Cancer Genome Atlas(TCGA)database.The relationship between GPX expression and patient survival was analyzed.The key GPX affecting the prognosis of glioma was screened out by single- and multi-factor Cox's proportional-hazards regression models and validated by least absolute shrinkage and selection operator(Lasso)regression.Finally,we constructed the model for predicting the prognosis of glioma with the screening results and then used concordance index and calibration curve respectively to evaluate the discrimination and calibration of model. Results Compared with those in the control group,the expression levels of GPX1,GPX3,GPX4,GPX7,and GPX8 were up-regulated in glioma patients(all <0.001).Moreover,the expression levels of other GPX except GPX3 were higher in GBM patients than in LGG patients(all <0.001).The Kaplan-Meier curves showed that the progression-free survival of GBM with high expression of GPX1(=0.013)and GPX4(=0.040),as well as the overall survival,disease-specific survival,and progression-free survival of LGG with high expression of GPX1,GPX7,and GPX8,was shortened(all <0.001).GPX7 and GPX8 were screened out as the key factors affecting the prognosis of LGG.The results were further used to construct a nomogram model,which suggested GPX7 was the most important variable.The concordance index of the model was 0.843(95%=0.809-0.853),and the calibration curve showed that the predicted and actual results had good consistency. Conclusion GPX7 is an independent risk factor affecting the prognosis of LGG,and the nomogram model constructed with it can be used to predict the survival rate of LGG.