Endoplasmic Reticulum Stress-related Classification for Prognosis Prediction in Hepatocellular Carcinoma

Abstract Background: Cancer cells under ER stress are common in hepatocellular carcinoma (HCC) and ER stress is strongly associated with poor prognosis. The aim of this study was to discover credible biomarkers for predicting prognosis of HCC based on ER stress-related genes (ERSRGs). Methods: Univariate Cox regression was performed to calculate the association between ERSRGs and survival outcomes of HCC patients in TCGA. Then LASSO-Cox regression strategy and stepwise Cox regression examination were performed to investigate the quality and establish the prognostic characteristics associated with prognosis. Finally, the model was subsequently validated in two additional independent HCC cohorts.Results: A novel seven-gene prognostic risk model based on ERSRGs was constructed and exhibited superior accuracy in forecasting the survival outcomes and 1-, 2-, 3- year survival rate of HCC patients. qRT-PCR was performed to validate the prognostic risk model in an independent clinical cohort containing 59 HCC patients and the results revealed that this signature had a good prognostic performance. Moreover, we found ER stress could affect the immune microenvironment in HCC and immune checkpoint inhibitors (ICIs) treatment was more effective for patients in high-risk subgroup. In addition, we identified 103 tumor-sensitive drugs in the CellMiner database that may be available for the treatment of HCC patients targeting ER stress and constructed a nomogram combining ER stress-related feature, TNM stage, age and gender. Conclusions: Our seven genetic risk model associated with ER stress can accurately predict survival outcome in HCC patients and facilitate the selection of the best individualized treatment targeting ER stress.