Background
The function of liquid-liquid phase separation (LLPS) in the progression of hepatocellular carcinoma (HCC) is still elusive. We aimed to explore the predictive value and immuno-therapeutic responses of LLPS-related signature (LLPSRS) in HCC.
Methods
In our study, we delved into the characteristics of LLPS at multiple omics levels. By utilizing single-cell and transcriptome analysis, we applied the lasso-cox to identify LLPSRS. In order to enhance the practicality of LLPSRS, we established and externally validated a LLPSRS nomogram, providing a quantitative prognostic tool for HCC patients. Furthermore, we investigated the mechanism of LLPSRS according to transcriptome, genomic, and single-cell levels, revealing important connections between LLPSRS, HCC prognosis, and immune landscape. Finally, we examined the different responses of the risk subgroups to immune checkpoint inhibitors and their sensitivity to major LLPSRS targeted drugs.
Results
Using TCGA data and LASSO-COX regression analysis, A risk-predictive scoring model for 9 LLPSRS was developed. The high-risk group exhibited notably lower overall survival compared to the low-risk group. High AUC values from time-dependent ROC curves demonstrated the model's robust performance. A nomogram that integrated the risk score and clinical features showed excellent prognostic ability. GO and KEGG analyses identified the potential pathways of the gene signature. The LLPSRS signature's associations with clinicopathological characteristics, tumor microenvironment, immunotherapy response, and chemotherapy sensitivity underscored its significant clinical relevance.
Conclusion
This model accurately predicts the outcomes of HCC patients and uncovers the potential mechanisms of LLPSRS in HCC, and paves the new avenues for personalized treatment and immuno-therapy development.