Combination of radiosensitivity index and PD-L1 status predicts overall survival of endometrial cancer patients treated with radiotherapy: A study based on TCGA data

Identifying biomarker for the clinical outcome of patients with endometrial cancer (EC) after radiotherapy is helpful to optimize treatment. The predictive efficacy of the combination of radiosensitivity index (RSI) and programmed cell death ligand 1 (PD-L1) should be investigated. The endometrial cancer cohort contained 225 patients received radiotherapy, and 285 patients who did not receive radiotherapy was downloaded from The Caner Genome Atlas (TCGA) and divided into radiosensitive (RS) and radioresistant (RR) groups according RSI. The EC patients were further classified into PD-L1-high and PD-L1-low groups according to the median value of CD274 mRNA expression. Cox proportional hazards regression and Kaplan-Meier analysis for overall survival were performed. The differently expressed analysis, gene set enrichment analysis and immune cell infiltration were performed. Among EC patients in the TCGA dataset, a group characterized by radioresistance and low PD-L1 status (PD-L1-high-RR group) was identified. Kaplan-Meier analysis showed EC patients in the RR-PD-L1-low group had significantly worse overall survival than those in the other groups when received radiotherapy (p=0.0006). Cox proportional hazards regression showed patients in the RR-PD-L1-low group were significantly associated with a poor of overall survival (HR=3.79; 95% CI=1.43-6.23; p=0.007) when received radiotherapy. MAPK signaling pathway was differently enriched in RR-PD-L1-high group. CD8+ T cell and M1 macrophage were lower infiltration in RR-PD-L1-high group, whereas M2 macrophage was more infiltration in RR-PD-L1-high group. The combination of RSI and PD-L1 may serve a predictive tool to identify EC patients who benefit from radiotherapy, especially combined immunotherapy.