A prognostic pyroptosis-related LncRNA classifier associated with the immune landscape and therapy efficacy in glioma

Background: Glioma has the highest fatality rate among intracranial tumours. Besides, the heterogeneity of gliomas leads to different therapeutic effects even with the same treatment. Developing a new signature for glioma to achieve the concept of “personalised medicine” remains a significant challenge. Method: The Cancer Genome Atlas (TCGA) and the Chinese Glioma Genome Atlas (CGGA) were searched to acquire information on glioma patients. Initially, correlation and univariate Cox regression analyses were performed to screen for prognostic pyroptosis-related long noncoding RNAs (PRLs). Secondly, 11 PRLs were selected to construct the classifier using certain algorithms. The efficacy of the classifier was then detected by the “timeROC” package for both the training and validation datasets. CIBERSORT and ESTIMATE packages were applied for comparing the differences (variations) in the immune landscape between the high- and low-risk groups. Finally, the therapeutic efficacy of the chemotherapy, radiotherapy, and immunotherapy were assessed using the “oncoPredict” package, survival analysis, and the tumour immune dysfunction and exclusion (TIDE) score, respectively. Results: A classifier comprising 11 PRLs was constructed. The PRL classifier exhibits a more robust prediction capacity for the survival outcomes in patients with gliomas than the clinical characteristics irrespective of the dataset (training or validation dataset). Moreover, it was found that the tumour landscape between the low- and high-risk groups was significantly different. A high-risk score was linked to a more immunosuppressive tumour microenvironment. According to the outcome prediction and analysis of the chemotherapy, patients with different scores showed different responses to various chemotherapeutic drugs and immunotherapy. Meanwhile, the patient with glioma of WHO grade Ⅳ or aged >50 years in the high risk group had better survival following radiotherapy. Conclusion: We constructed a PRL classifier to roughly predict the outcome of patients with gliomas. Furthermore, the PRL classifier was linked to the immune landscape of glioma and may guide clinical treatments.