Prognostic and Therapeutic Significance of Cancer‐Associated Fibroblasts Genes in Osteosarcoma Based on Bulk and Single‐Cell <scp>RNA</scp> Sequencing Data

Osteosarcoma (OS) is the most frequent primary solid malignancy of bone, whose course is usually dismal without efficient treatments. The aim of the study was to discover novel risk models to more accurately predict and improve the prognosis of patients with osteosarcoma. The single-cell RNA sequencing (scRNA-seq) data was obtained from the GEO database. Bulk RNA-seq data and microarray data of OS were obtained from the TARGET and GEO databases respectively. A clustering tree was plotted to classify all cells into different clusters. The "cellchat" R package was used to establish and visualise cell-cell interaction networks. Then Univariate COX regression analysis was used to determine the prognostic CAF-related genes, followed by the Lasso-Cox regression analysis to build a risk on the prognostic CAF-related genes. Finally, from multiple perspectives, the signature was validated as an accurate and dependable tool in predicting the prognosis and guiding treatment therapies in OS patients. From the single-cell dataset, six OS patients and 46,544 cells were enrolled. All cells were classified into 22 clusters, and the clusters were annotated to 14 types of cells. Subsequently, CAFs were observed as a vital TME components. In cell-cell interaction networks in OS cells, CAFs had a profound impact as four roles. Via the Univariate COX regression analysis, 14 CAF-related genes were screened out. By the Lasso-Cox regression analyses, 11 key CAF-related genes were obtained, based on which an 11-gene signature that could predict the prognosis of osteosarcoma patients was constructed. According to the median of risk scores, all patients were grouped in to the high- and low-risk group, and their overall survival, activated pathways, immune cell infiltrations, and drug sensitivity were significantly differential, which may have important implications for the clinical treatment of patients with osteosarcoma. Our study, a systematic analysis of gene and regulatory genes, has proven that CAF-related genes had excellent diagnostic and prognostic capabilities in OS, and it may reshape the TME in OS. The novel CAF-related risk signature can effectively predict the prognosis of OS and provide new strategies for cancer treatment.