Abstract
Exosomes are progressively being detected as an indicator for the diagnosis and prognosis of cancer in clinical settings. Many clinical trials have confirmed the impact of exosomes on tumor growth, particularly in anti-tumor immunity and immunosuppression of exosomes. Therefore, we developed a risk score based on genes found in glioblastoma-derived exosomes. In this study, we used the TCGA dataset as the training queue and GSE13041, GSE43378, GSE4412, and CGGA datasets as the external validation queue. Based on machine algorithms and bioinformatics methods, an exosome-generalized risk score was established. We found that the risk score could independently predict the prognosis of patients with glioma, and there were significant differences in the outcomes of patients in the high- and low-risk groups. Univariate and multivariate analyses showed that risk score is a valid predictive biomarker for gliomas. Two immunotherapy datasets, IMvigor210 and GSE78220, were obtained from previous studies. A high-risk score showed a significant association with multiple immunomodulators that could act on cancer immune evasion. The exosome-related risk score could predict the effectiveness of anti-PD-1 immunotherapy. Moreover, we compared the sensitivity of patients with high- and low-risk scores to various anti-cancer drugs and found that patients with high-risk scores had better responses to a variety of anti-cancer drugs. The risk-scoring model established in this study provides a useful tool to predict the total survival time of patients with glioma and guide immunotherapy.