Abstract 690: Rethinking molecular analysis of gastric cancer: Insight into morphologic phenotype

Abstract Introduction: Advanced gastric cancer (GC) is one of the most challenging malignancies, with high mortality rates and a lack of effective treatment strategies. Histomorphological features, termed ‘morphologic phenotypes’, predict therapeutic response, with lymphoepithelial carcinoma showing immunotherapy responsiveness while diffuse-type GC exhibits resistance to immunotherapy and targeted therapies such as HER2 and CLDN18 inhibitors. Despite these associations, comprehensive molecular understanding based on meticulous morphological assessments remains elusive. Herein, we aim to identify robust biomarkers and elucidate the molecular mechanisms underlying GC phenotypes, providing a more precise reflection of the spatial phenotype of GC. Methods: TCGA stomach adenocarcinoma RNA-sequencing (n=412), WSIs (n=398), and clinical data were retrieved from the GDC Data Portal. Expert gastrointestinal pathologists reviewed H&E-stained WSIs for consensus morphological classifications. RNA-seq data were analyzed using DESeq2 for normalization and differential expression testing. Overall single-cell RNA sequencing (scRNA-seq) analysis was conducted using Seurat to validate findings at the cellular level. Functional analyses, including Gene Set Enrichment Analysis (GSEA) and Over-Representation Analysis (ORA), utilized the MSigDB Hallmark and Gene Ontology Biological Process databases to identify key pathways and biological processes. Results: A total of 4,342 differentially expressed genes (DEGs) were identified, revealing novel distinctions between diffuse and intestinal GC subtypes. A calcium-dependent adhesion signature showed coordinated upregulation of CDH2 and ACTA2, with CDH1 downregulation in diffuse-type tumors. Notably, elevated ACTA2 expression, a marker of CAF, suggested enhanced tumor-stromal interaction and potential therapy resistance mechanisms. Further analysis indicated that CDH1 suppression could be attributed to somatic mutations or epigenetic regulation. Single-cell level transcriptome profiling revealed tumor heterogeneity, with distinct cell populations exhibiting unique molecular signatures in diffuse-type tumors. These findings highlight the CDH2/CDH1 axis as a key determinant of aggressive tumor phenotypes in diffuse-type gastric cancer, offering potential therapeutic insights. Conclusion: This study underscores the importance of standardized pathological annotations for uncovering phenotype-specific molecular biomarkers in GC. By integrating RNA-seq and scRNA-seq data, we highlighted the CDH1/CDH2 adhesion axis as a critical determinant of diffuse-type tumor aggressiveness. These molecular insights emphasize the significance of phenotype-specific biomarker discovery and reveal potential mechanisms of therapy resistance, providing a foundation for novel therapeutic strategies and improved patient stratification. Citation Format: Harim Chun, Kwangil Yim, Sangjeong Ahn, Sung Hak Lee. Rethinking molecular analysis of gastric cancer: Insight into morphologic phenotype. [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 690.