B Cell-Related Prognostic Model in Gastric Cancer via Integration of Single-Cell Sequencing and Transcriptome Sequencing Data

Background & Objectives Tumor-infiltrating B cells (TIBs) are a critical component of the tumor microenvironment (TME) in gastric cancer (GC). Their infiltration levels and functional phenotypes are closely correlated with the antitumor immune response and clinical prognosis of patients with GC. This study aimed to screen hub genes associated with TIBs in GC and establish a prognostic prediction model with robust performance. Methods Single-cell RNA sequencing (scRNA-seq) data, bulk RNA sequencing (bulk RNA-seq) data, and complete clinical follow-up data of patients with GC were retrieved from the Gene Expression Omnibus (GEO) and The Cancer Genome Atlas (TCGA) databases. Quality control (QC), principal component analysis (PCA)-based dimensionality reduction, and cell clustering were performed on the scRNA-seq data using the Seurat package, followed by the identification of differentially expressed genes (DEGs) among B cell subtypes. The limma algorithm was applied to screen DEGs from the bulk RNA-seq data. Key genes were selected from the intersection of the two DEG sets via multivariate Cox regression analysis, and a risk scoring model was constructed by integrating the corresponding regression risk coefficients. The independent prognostic value of the risk score was validated using univariate Cox regression in combination with clinical data. Patients were stratified into high- and low-risk groups according to the median risk score. The model was comprehensively evaluated in both internal and external datasets using Kaplan-Meier (K-M) survival curves, receiver operating characteristic (ROC) curves, and nine machine learning algorithms. Results Two hub genes, CTHRC1 and CD44, were identified in this study, based on which a prognostic risk scoring model for GC was developed. Validation with K-M survival curves, ROC curves, and nine machine learning algorithms consistently showed that the model could effectively discriminate between high- and low-risk groups in both internal and external datasets, with stable predictive performance. Conclusions By integrating multi-omics data from scRNA-seq and bulk RNA-seq, this study identified hub genes related to B cell infiltration. The constructed B cell immunity-associated prognostic risk model for GC can provide a reliable research reference for the prognostic evaluation of GC.