Construction of a novel platelet‑related gene risk model to predict the prognosis and drug response in virus‑related hepatocellular carcinoma.

Abstract

Platelet activity in the tumor microenvironment (TME) is crucial for the development of tumors. However, the roles and clinical potential of platelet activity in the TME for virus-related hepatocellular carcinoma (HCC) remain unclear. The present study aimed to identify a novel signature based on platelet activity for prognostic prediction and treatment decisions in virus-related HCC. First, a novel platelet signature score (PSS) for each patient with virus-related HCC from The Cancer Genome Atlas was calculated using gene set variation analysis, and the patients were divided into two subgroups (high and low PSS). It was demonstrated that the patients with a high PSS had a worse prognosis, higher platelet activity, stronger inflammation and immunosuppression in TME than patients with a low PSS. Furthermore, 137 differentially expressed genes (DEGs; fold change >2; P<0.05) were identified using 'DESeq2' and 'edgeR' software. Subsequently, 3 genes (cyclin-J-Like protein, nuclear receptor subfamily 0 group B member 1 and tripartite motif containing 54) were identified from DEGs using univariate Cox and least absolute shrinkage and selection operator (LASSO) analyses. Risk score (RS) was calculated based on gene expression and coefficients from LASSO. Patients were divided into high and low RS groups according to the median value, and the 3-gene model was used to predict prognoses and drug responses. Notably, it was demonstrated that patients with a low RS may be better candidates for immune therapy due to lower levels of tumor immune dysfunction and exclusion scores. Moreover, patients with a high RS may be better candidates for nonimmune therapy due to lower half-maximal inhibitory concentration values of drugs (such as AKT inhibitors and gemcitabine). Finally, it was demonstrated that patients with a high PSS and RS had a higher platelet activity, inflammation status, tumor hallmarks and the worst prognosis than patients with a low PSS and RS. This helped to better find patients with these characteristics and suitable treatments using this method. Collectively, the findings of the present study indicate that PSS combined with RS has great potential to evaluate the prognosis of patients with virus -related HCC and assist in deciding treatment strategies.