Abstract 459P
Background
Gastric adenocarcinoma (GA) is a global health concern with poor prognosis despite advancements in treatment. Although molecular classifications, mainly The Cancer Genome Atlas (TCGA), provide valuable insights, their clinical utility has been limited. We conducted a multi-layered functional analysis using TCGA RNA sequencing data to delineate molecular subtypes and explore therapeutic implications.
Methods
We reanalyzed the TCGA RNA-seq data of a cohort of 142 GA patients with localized disease who received adjuvant chemotherapy, using a functional approach including probabilistic graphical models and a recurrent sparse k-means / consensus cluster algorithms-based layer analysis.
Results
Our analysis revealed significant differences in survival outcomes among TCGA groups, with the GS subtype exhibiting the poorest prognosis. We defined twelve functional nodes and seven biological layers, each with specific biological functions. The Combined Molecular Layer (CML) classification identified three prognostic groups that align with TCGA subtypes. Notably, CML2 (GS-like) showed differential gene expression related to lipid metabolism, correlating with poorer survival outcomes. Transcriptomic heterogeneity within the CIN subtype was evident, along with distinct clusters related to proteolysis and lipid metabolism. Within this subtype, we identified a subset of tumors with transcriptome profiles and prognoses resembling MSI tumors, which we termed CIN-MSI-like. Importantly, claudin-18, a critical gene in the proteolysis classification, was found to be overexpressed across various TCGA subtypes, suggesting its potential as a target for therapy.
Conclusions
Our study provides insights into GA biology, facilitating refined patient stratification and personalized treatment strategies. Integration of molecular subtyping into clinical decision-making can improve patient outcomes in GA management. Further validation and prospective studies are warranted to translate these findings into clinical practice.
Legal entity responsible for the study
Instituto de Investigación Sanitaria del Hospital Universitario La Paz (IdiPAZ), 28046 Madrid, Spain.
Funding
Has not received any funding.
Disclosure
All authors have declared no conflicts of interest.