Abstract 1432P
Background
Advanced Gastric Cancer (GC) is characterized by poor prognosis. Beyond intrinsic molecular features, tumor microenvironment (TM) has emerged as an important factor for tumor development and progression. This study aimed to define the TM characteristics of advanced GC through transcriptomic profiling to understand their influence on treatment response and to better predict an immunotherapy effect.
Methods
An in-silico analysis of the transcriptomic data of advanced GC patients from TCGA was first performed. A Principal Component Analysis (PCA) was carried out to generate a gene signature (GS) to define GC microenvironment. Progression-free survival (PFS) of patients grouped into high and low (first vs. forth quartile) immune profile was described using proportional hazard Cox model and Kaplan-Meier. To validate the results, fresh GC biopsies were prospectively collected and characterized with an immunohistochemistry specific customized immune-panel and an RNA-seq analysis.
Results
Transcriptomic data from 209 advanced GC patients from TCGA were studied. A customized 32-GS was identified and its optimal threshold led to characterize the TM into two different profiles: pro and anti-inflammatory. In the in-silico cohort, patients with a pro-inflammatory profile presented a significantly prolonged PFS (HR: 0.48; 95%CI: 0.24-0.96) p = 0.039) over those with an anti-inflammatory signature. A consecutive series of 26 fresh biopsies from stage IV primary gastric tumors was collected (12 diffuse; 14 intestinal, of them 3 MSI-H and 2 HER2+). The RNA-seq analysis of these samples identified a similar classification in two different groups. Patients defined by our transcriptome analysis as anti-inflammatory mostly presented higher PDL-1 expression, while those defined as proinflammatory were enriched in CD8 and macrophages (90% accuracy between the two evaluations).
Conclusions
Our model was able to identify two different groups on the immune profile across advanced GC patients. Transcriptomics could help in better defining those patients who could benefit from checkpoint inhibitors beyond PDL1 expression. Further validation in a wider prospective cohort is ongoing.
Clinical trial identification
Editorial acknowledgement
Legal entity responsible for the study
The authors.
Funding
Instituto de Salud Carlos III.
Disclosure
A. Cervantes: Other, Personal and Institutional, Advisory Role: Merck Serono; Financial Interests, Institutional, Funding: Merck Serono; Other, Personal and Institutional, Invited Speaker: Merck Serono; Other, Personal and Institutional, Advisory Role: Roche; Financial Interests, Personal and Institutional, Funding: Roche; Other, Institutional, Invited Speaker: Roche; Other, Personal and Institutional, Advisory Role: Beigene; Financial Interests, Institutional, Funding: Beigene; Other, Personal and Institutional, Advisory Role: Bayer; Financial Interests, Institutional, Funding: Bayer; Other, Personal and Institutional, Invited Speaker: Bayer; Other, Personal and Institutional, Advisory Role: Servier; Financial Interests, Institutional, Funding: Servier; Other, Personal and Institutional, Invited Speaker: Servier; Other, Personal and Institutional, Advisory Role: Lilly; Financial Interests, Institutional, Funding: Lilly; Other, Personal and Institutional, Advisory Role: Novartis; Financial Interests, Institutional, Funding: Novartis; Other, Personal and Institutional, Advisory Role: Takeda; Financial Interests, Institutional, Funding: Takeda; Other, Personal and Institutional, Advisory Role: Astelas; Financial Interests, Institutional, Funding: Astelas; Other, Personal and Institutional, Advisory Role: Pierre Fabre; Financial Interests, Institutional, Funding: Genentech; Financial Interests, Institutional, Funding: Fibrogen; Financial Interests, Institutional, Funding: Amcure; Financial Interests, Institutional, Funding: Sierra Oncology; Financial Interests, Institutional, Funding: AstraZeneca; Financial Interests, Institutional, Funding: Medimmune; Financial Interests, Institutional, Funding: BMS; Financial Interests, Institutional, Funding: MSD; Other, Personal and Institutional, Invited Speaker: Foundation Medicine. All other authors have declared no conflicts of interest.