Abstract
Background Ovarian cancer (OC) is the top of the aggressive malignancies in females with a poor survival rate. However, the roles of immune-related pseudogenes (irPseus) in the immune infiltration of OC and the impact on overall survival (OS) have not been adequately studied. Therefore, this study aims to identify a novel model constructed by irPseus to predict OS in OC and to determine its significance in immunotherapy and chemotherapy. Methods In this study, with the use of The Cancer Genome Atlas (TCGA) combined with Genotype-Tissue Expression (GTEx), 55 differentially expressed irPseus (DEirPseus) were identified. Then, we constructed 10 irPseus pairs with the help of univariate, Lasso, and multivariate Cox regression analysis. The prognostic performance of the model was determined and measured by the Kaplan–Meier curve, a time-dependent receiver operating characteristic (ROC) curve. Results After dividing OC subjects into high- and low-risk subgroups via the cut-off point, it was revealed that subjects in the high-risk group had a shorter OS. The multivariate Cox regression performed between the model and multiple clinicopathological variables revealed that the model could effectively and independently predict the prognosis of OC. The prognostic model characterized infiltration by various kinds of immune cells and demonstrated the immunotherapy response of subjects with cytotoxic lymphocyte antigen 4 (CTLA4), anti-programmed death-1 (PD-1), and anti-PD-ligand 1 (PD-L1) therapy. A high risk score was related to a higher inhibitory concentration (IC50) for etoposide (P=0.0099) and mitomycin C (P=0.0013). Conclusion It was the first study to identify a novel signature developed by DEirPseus pairs and verify the role in predicting OS, immune infiltrates, immunotherapy, and chemosensitivity. The irPseus are vital factors predicting the prognosis of OC and could act as a novel potential treatment target.
Highlights
Ovarian cancer (OC) is a common highly aggressive malignancy of the female reproductive system, with the 4th highest morbidity and the 3rd highest mortality around the world [1]. e report of Global Cancer Statistics 2020 estimated that the incidence and mortality of OC accounts for 1.6% and 2.1%, respectively [2]. e current regular treatment strategy for OC is to quickly resect the primary lesion or followed by adjuvant chemotherapy
We carried out the LASSO algorithm after 1000 iterations to shrink the count of variables in the risk signature (Figure 2(a) and 2(b)), and totally 10 immune-related pseudogenes (irPseus) pairs were identified (Figure 2(c)) using multivariate Cox regression analysis, which included 18 pseudogenes
For the first time, we used the strategy of irPseus pairs pairing and aimed to develop a reliable signature with 10pseudogene combinations and not requiring specific expression levels of the pseudogene
Summary
Ovarian cancer (OC) is a common highly aggressive malignancy of the female reproductive system, with the 4th highest morbidity and the 3rd highest mortality around the world [1]. e report of Global Cancer Statistics 2020 estimated that the incidence and mortality of OC accounts for 1.6% and 2.1%, respectively [2]. e current regular treatment strategy for OC is to quickly resect the primary lesion or followed by adjuvant chemotherapy. An open-label, randomized, phase 3 trial investigated progression-free survival and incidence of adverse events after chemotherapy with or without avelumab followed by avelumab maintenance versus chemotherapy alone in 998 patients with previously untreated epithelial OC, while the results failed to support the use of avelumab in the frontline treatment setting [3]. Multiple studies demonstrated the performance of immunotherapy, such as spontaneous cancer regressions in OC and revealed the possible procedures of immune evasion and responses to immune checkpoint inhibitors in patients with OC [5,6,7]. The roles of immune-related pseudogenes (irPseus) in the immune infiltration of OC and the impact on overall survival (OS) have not been adequately studied. After dividing OC subjects into high- and low-risk subgroups via the cut-off point, it was revealed that subjects in the high-risk group had a shorter OS. e multivariate Cox regression performed between the model and multiple clinicopathological variables revealed that the model could effectively and independently predict the prognosis of OC
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
