Bum-Sup Jang1, In Ah Kim2. 1. Department of Radiation Oncology, Seoul National University Hospital, Republic of Korea. 2. Department of Radiation Oncology and Cancer Research Institute, Seoul National University, College of Medicine, Republic of Korea; Department of Radiation Oncology, Seoul National University Bundang Hospital, Seongnamsi, Republic of Korea. Electronic address: inah228@snu.ac.kr.
Abstract
PURPOSE: Identifying predictive factors for the clinical outcome of patients with lower grade gliomas following radiotherapy could help optimize patient treatments. Here, we investigate the predictive efficacy of both a previously identified "31-gene signature" and programmed death ligand-1 (PD-L1) expression. MATERIAL AND METHODS: We identified 511 patients with lower grade glioma (Grade 2 and 3) in The Cancer Genome Atlas dataset and divided them into two clusters: radiosensitive (RS) and radioresistant (RR). Patients were also classified as PD-L1-high or PD-L1-low based on CD274 mRNA expression. Five-year survival rates were compared across patient groups, and differentially expressed genes were identified via a gene enrichment analysis. RESULTS: Among 511 patients with lower grade glioma in The Cancer Genome Atlas dataset, we identified a group that was characterized by radioresistant and high PD-L1 (the PD-L1-high-RR group). Multivariate Cox models demonstrated that the membership in the PD-L1-high-RR can predict overall survival regarding to RT. Differentially expressed genes associated with the PD-L1-high-RR group were found to play a role in the immune response, including the T-cell receptor signaling pathway. CONCLUSION: We tested the predictive value of a "31-gene signature" and PD-L1 expression status in a dataset of patients with lower grade glioma. Our results suggest that the patient population classified as the PD-L1-high-RR may benefit most from radiotherapy combined with anti-PD-1/PD-L1 treatment. Prospective clinical trial is necessary to validate the findings in a homogenous treated patient cohort.
PURPOSE: Identifying predictive factors for the clinical outcome of patients with lower grade gliomas following radiotherapy could help optimize patient treatments. Here, we investigate the predictive efficacy of both a previously identified "31-gene signature" and programmed death ligand-1 (PD-L1) expression. MATERIAL AND METHODS: We identified 511 patients with lower grade glioma (Grade 2 and 3) in The Cancer Genome Atlas dataset and divided them into two clusters: radiosensitive (RS) and radioresistant (RR). Patients were also classified as PD-L1-high or PD-L1-low based on CD274 mRNA expression. Five-year survival rates were compared across patient groups, and differentially expressed genes were identified via a gene enrichment analysis. RESULTS: Among 511 patients with lower grade glioma in The Cancer Genome Atlas dataset, we identified a group that was characterized by radioresistant and high PD-L1 (the PD-L1-high-RR group). Multivariate Cox models demonstrated that the membership in the PD-L1-high-RR can predict overall survival regarding to RT. Differentially expressed genes associated with the PD-L1-high-RR group were found to play a role in the immune response, including the T-cell receptor signaling pathway. CONCLUSION: We tested the predictive value of a "31-gene signature" and PD-L1 expression status in a dataset of patients with lower grade glioma. Our results suggest that the patient population classified as the PD-L1-high-RR may benefit most from radiotherapy combined with anti-PD-1/PD-L1 treatment. Prospective clinical trial is necessary to validate the findings in a homogenous treated patient cohort.