Prognostic value of anoikis-related genes revealed using multi-omics analysis and machine learning based on lower-grade glioma features and tumour immune microenvironment

Abstract

Abstract Background: Lower-grade glioma (LGG) is a prevalent glial cell-derived brain tumor that is aggressive and infiltrative. Anoikis, a new and distinct form of cell death, is a catch-all phrase describing cells losing their ability to adhere to the extracellular matrix (ECM) and nearby cells, followed by the inducing of apoptosis. However, what role the mechanisms associated with anoikis play in LGG have not been thoroughly discovered. Methods: The Cancer Genome Atlas (TCGA), Gene Expression Omnibus (GEO), and Chinese Glioma Genome Atlas (CGGA) are three large databases that provide sequencing information for LGG patients, as well as the corresponding clinical data, were included in this study as the training set and multi-group validation set for the data. Application of ConsensusClusterPlus Consensus Clustering for molecular subtype classification of LGG patients based on anoikis-related genes (ARGs)with prognostic value. Subsequently, we screened genes significantly associated with patient prognosis using different machine learning algorithms. Risk profiles are constructed and assessed based on these screened genes. Results: Patients with LGG were classified into two distinct molecular subtypes based on a clustering approach, each characterized by their prognosis, clinical features, and tumor microenvironment. A 6-ARG prognostic signal (EGFR, SIX1, SP1, ANGPTL2, PDCD4, and BMP2) was subsequently constructed, and the signature genes showed good predictive performance not only in the training set but also in multiple validation sets. Additionally, we go into great depth about how high-risk and low-risk groups differ from one another in terms of attributes, including immune characteristics, tumor mutation characteristics, and drug sensitivity showing significant differences in the risk subgroups. Finally, this risk score is combined with multiple LGG clinicopathological features to create an at-a-glance nomogram for quantitatively predicting the probability of clinical survival in individuals with LGG, and the AUC values and decision curve analysis (DCA) of this nomogram suggest that the model can benefit patients from clinical treatment strategies. Conclusion: Overall, ARG signs can be used as a valid indicator of prognosis prediction and response to immunotherapy in patients with LGG.