A Cancer-Associated Fibroblast Prognostic Signature in Hepatocellular Carcinoma

Abstract

Background: Although the number of therapies for hepatocellular carcinoma has increased, the survival rate is still unsatisfactory. Cancer-associated fibroblasts have been reported to regulate hepatocellular carcinoma progression via various mechanisms. We aimed to investigate an effective prognostic tool related to cancer-associated fibroblasts for decision-making in patients with hepatocellular carcinoma. Methods: Bioinformatics analyses were used to identify cancer- and fibroblast-associated genes based on data from the Cancer Genome Atlas and Gene Expression Omnibus datasets. Following Cox and least absolute shrinkage and selection operator analyses, the optimal prognostic genes were identified, and a prognostic cancer-associated fibroblast signature was established based on these genes. Receiver operator characteristic analysis was used to validate the performance of the signature. In addition, the correlation between the cancer-associated fibroblast signature and clinical, immune, and mutational features was analyzed. Finally, a prognostic nomogram was developed and evaluated. Results: Six cancer- and fibroblast-associated prognostic genes (PZP (pregnancy zone protein), TSPYL5 (testis-specific protein Y-encoded-like 5), ADAMTSL2 (a disintegrin and metalloproteinase with thrombospondin motifs like 2), SAMD12 (sterile alpha motif domain containing 12), PNMA2 (paraneoplastic Ma antigens family member 2), and N4BP3 (NEDD4 binding protein 3)) in hepatocellular carcinoma were identified to construct the cancer-associated fibroblast risk score. The receiver operating characteristic curve showed that the signature exhibited good performance in predicting the survival of patients with hepatocellular carcinoma (with an area under the curve >0.75). Furthermore, the low-risk group showed better stromal and immune scores and a lower tumor mutation burden. Finally, a nomogram model was constructed to predict the survival of hepatocellular carcinoma patients. Conclusions: This study shows a promising cancer-associated fibroblast signature that might be useful in predicting the survival and personalized management of patients with hepatocellular carcinoma in the clinic.