Prognostic model for predicting recurrence-free survival in hepatocellular carcinoma using integrated analysis of...

Globally, primary liver cancer ranks as the sixth most prevalent cancer and was the third leading cause of cancer-related deaths in 2022. Hepatocellular carcinoma (HCC) accounts for 75-85% of all cases. In total, ~70% of patients with HCC experience recurrence within 5 years, which impacts their long-term survival outcomes. Therefore, the development of a reliable predictive model for the probability of HCC recurrence represents a crucial clinical need. However, studies that integrate bulk RNA sequencing (RNA-seq) and single-cell (sc)RNA-seq to construct prognostic models are lacking. The present study analyzed bulk RNA-seq and scRNA-seq datasets of patients with HCC to identify differentially expressed genes (DEGs) that affect HCC recurrence-free survival (RFS). Subsequently, least absolute shrinkage and selection operator Cox penalized regression analysis was performed to construct a prognostic model. Enrichment analysis and immune infiltration analysis were applied to identify the underlying mechanisms involved. Univariate and multivariate Cox regression analyses were subsequently performed. Finally, independent dataset and reverse transcription-quantitative PCR (RT-qPCR) experiments were used to evaluate the prognostic model. A total of 5,586 DEGs were obtained from the bulk RNA-seq dataset and 2,320 DEGs from the scRNA-seq dataset. Moreover, 53 DEGs associated with the RFS of patients with HCC were identified. A total of 6 of these genes (cyclin-dependent kinase inhibitor 2A, complement factor H-related 3, cytochrome P450 family 2 subfamily C member 9, high mobility group box 2, immunoglobulin λ constant 2 and Jupiter microtubule-associated homolog 1) were incorporated into the prognostic model. Patients in the high-risk group had significantly worse RFS time than those in the low-risk group. Furthermore, the cell cycle and immunosuppression were identified as possible factors affecting RFS. In addition, the prognostic signature retained independent predictive value for HCC RFS, and it was validated successfully by another publicly available dataset and RT-qPCR experiments using patient tissues. In conclusion, the present study constructed a prognostic model for predicting RFS in patients with HCC via integrated analysis of scRNA-seq and bulk RNA-seq data that could serve as a valuable reference tool for clinicians.