Multi-omics sequencing of gastroesophageal junction adenocarcinoma reveals prognosis-relevant key factors and a novel immunogenomic classification.

Background: Gastroesophageal junction adenocarcinoma (GEJAC) exhibits distinct molecular characteristics due to its unique anatomical location. We sought to investigate effective and reliable molecular classification of GEJAC to guide personalized treatment.

Methods: We analyzed the whole genomic, transcriptomic, T-cell receptor repertoires, and immunohistochemical data in 92 GEJAC patients and delineated the landscape of genetic and immune alterations. In addition to COSMIC nomenclature, the de novo nomenclature was also utilized to define signatures and investigate their correlation with survival. A novel molecular subtype was developed and validated in other cohorts.

Results: We found 30 mutated driver genes, 7 novel genomic signatures, 3 copy-number variations, and 2 V-J gene usages related to prognosis that were not identified in previous study. A high frequency of COSMIC-SBS-384-1 and De novo-SV-32-A was associated with more neoantigen generation and a better survival. Using 19 molecular features, we identified three immune-related subtypes (immune inflamed, intermediate, and deserted) with discrete profiles of genomic signatures, immune status, and clinical outcome. The immune deserted subtype (27.2%) was characterized by an earlier KRAS mutation, worse immune reaction, and prognosis than the other two subtypes. The immune inflamed subtypes exhibited the highest levels of neoantigens, TCR/pMHC-binding strength, CD8 + T-cell infiltration, IFN-α/γ response pathways, and survival rate.

Conclusions: These results emphasize the immune reaction and prognostic value of novel molecular classifications based on multi-omics data and provide a solid basis for better management of GEJAC.