BackgroundChronic inflammation is considered the most critical predisposing factor of hepatocellular carcinoma (HCC), with inflammatory cell heterogeneity, hepatic fibrosis accumulation, and abnormal vascular proliferation as prominent features of the HCC tumor microenvironment (TME). Cancer‐associated fibroblasts (CAFs) play a key role in HCC TME remodeling. Therefore, the level of abundance of CAFs may significantly affect the prognosis and outcome in HCC patients.MethodsUnsupervised clustering was performed based on 39 genes related to CAFs in HCC identified by single‐cell RNA sequencing data. Patients of bulk RNA were grouped into CAF low abundance cluster and high abundance clusters. Subsequently, prognosis, immune infiltration landscape, metabolism, and treatment response between the two clusters were investigated and validated by immunohistochemistry.ResultsPatients in the CAF high cluster had a higher level of inflammatory cell infiltration, a more significant immunosuppressive microenvironment, and a significantly worse prognosis than those in the low cluster. At the metabolic level, the CAF high cluster had lower levels of aerobic oxidation and higher angiogenic scores. Drug treatment response prediction indicated that the CAF high cluster could have a better response to PD‐1 inhibitors and conventional chemotherapeutic agents for HCC such as anti‐angiogenic drugs, whereas CAF low cluster may be more sensitive to transarterial chemoembolization treatment.ConclusionsThis study not only revealed the TME characteristics of HCC with the difference in CAF abundance but also further confirmed that the combination therapy of PD‐1 inhibitors and anti‐angiogenic drugs may be more valuable for patients with high CAF abundance.

Tumor-associated fibroblasts (CAFs) are heterogeneous constituents of the tumor microenvironment that play a crucial role in tumorigenesis, progression, and therapeutic responses. Through single-cell RNA-seq analysis of breast cancer, we identified four distinct CAF subtypes, among which matrix CAFs (mCAFs) demonstrated significant involvement in tumor matrix remodeling and a strong correlation with the TGF-beta signaling pathway. Based on consensus clustering of TCGA breast cancer data utilizing mCAFs single-cell characteristic gene signatures, we segregated samples into high-Fibrotic and low-Fibrotic groups, where patients in the high-Fibrotic group exhibited a significantly worse prognosis. Subsequent analysis of bulk RNA-seq data using WGCNA and univariate Cox analysis revealed 17 differential genes of significant prognostic value. The mCAFs risk feature prognosis model (mRPS) was developed using ten different machine learning algorithms. When compared to the conventional TNM staging system, mRPS demonstrated improved accuracy for predicting clinical outcomes. The mRPS was found to be closely associated with the infiltration level of antitumor effector immune cells. Using six machine learning algorithms based on consensus prognostic genes, BRCA samples were classified into two subtypes IFN-gamma dominant (Immune C2) and TGF-beta (Immune C6) with an accuracy of more than 90%. The mRPS has significant clinical value, as Low mRPS were found to be associated with better patient prognosis, suggesting a higher likelihood of benefiting from immunotherapy. Taken together, the mRPS model offers a stable prediction of BRCA prognosis, reflects the local immune status of the tumor, and can help inform clinical decision-making regarding tumor immunotherapy.

No abstract