ORIGINAL RESEARCH article
Front. Immunol.
Sec. Cancer Immunity and Immunotherapy
Volume 16 - 2025 | doi: 10.3389/fimmu.2025.1680517
Integrative Multi-Omics Analysis of Gastric Cancer Evolution from Precancerous Lesions to Metastasis Identifies a Deep Learning-Based Prognostic Model
Provisionally accepted
- Department of Medical Oncology, Affiliated Cancer Hospital of Zhengzhou University, Zhengzhou, China
Select one of your emails
You have multiple emails registered with Frontiers:
Notify me on publication
Please enter your email address:
If you already have an account, please login
You don't have a Frontiers account ? You can register here
Background: Gastric cancer progression involves complex interactions among tumor cells, immune components, and stromal elements within the tumor microenvironment. However, a comprehensive understanding of cellular heterogeneity, spatial organization, and cell-cell communication in gastric cancer remains incomplete. Methods: Single-cell RNA sequencing was performed on 252,399 cells from six tissue types, spanning gastritis, intestinal metaplasia, primary tumors, adjacent normal tissue, and metastatic lesions. Integration with spatial transcriptomics enabled spatial mapping of cellular interactions. Pseudotime, cell-cell communication, and transcriptional heterogeneity analyses were conducted. Tumor stage-associated gene modules were identified using Weighted Gene Co-expression Network Analysis (WGCNA) of The Cancer Genome Atlas (TCGA) data. Finally, a deep learning-based prognostic model was developed and externally validated. Results: Our analysis revealed dynamic remodeling of the tumor microenvironment during gastric cancer progression, characterized by the expansion of dysfunctional CD8+ T cells, pro-tumorigenic fibroblasts (e.g., ITGBL1+, PI16+, and ITLN1+), and altered myeloid populations. Stromal-immune crosstalk, particularly fibroblast-driven immunosuppressive signaling, was prominent. Spatial transcriptomics revealed the colocalization of immune and stromal cells, supporting spatially organized cellular interactions. WGCNA identified a gene module (657 genes) associated with T cell, myeloid, and stromal alterations, as well as tumor stage. A deep learning model based on this gene set accurately stratified patients according to survival in both TCGA and independent validation cohorts. Risk scores were correlated with clinical features, including tumor stage and therapeutic response. Conclusions: Our integrative single-cell, spatial, and computational analysis provides a high-resolution map of gastric cancer microenvironment remodeling. We identified key stromal and immune subpopulations, extensive cellular communication networks, and spatial structures that collectively drive tumor progression and metastasis. The derived gene signature and prognostic model have the potential for clinical risk stratification and therapeutic targeting in gastric cancer.
Keywords: gastric cancer, single-cell RNA sequencing, Tumor Microenvironment, WGCNA, Deep learning prognostic model
Received: 06 Aug 2025; Accepted: 14 Oct 2025.
Copyright: © 2025 Ren, Li, Xu, Qiao, Li, Zhang and Liu. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
* Correspondence: Yulin Ren, renyulin1124@163.com
Disclaimer: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.