IDDF2024-ABS-0168 Integrating genomics and metabolomics to inform helicobacter pylori eradication for targeted gastric cancer prevention

<h3>Background</h3> <i>Helicobacter pylori (H.</i><i>pylori)</i> contributes to the pathogenesis of gastric cancer (GC). Although eradication therapy is fundamental for GC prevention, its varying efficacy among individuals suggests the need for targeted approaches. Herein, We propose a population-based strategy integrating genomics and metabolomics information to decode heterogeneous <i>H.pylori</i> treatment efficacy for targeted GC prevention. <h3>Methods</h3> We employed four independent cohorts: the UK Biobank (UKB, n=145,938), the Upper Gastrointestinal Cancer Early Detection Program (UGCED, n=370), the Shandong Intervention Trial (SIT, n=2,755), and a nested-case control cohort from the Mass Intervention Trial in Shandong (MITS, n=2,804). We developed a causal machine learning framework to derive genetically influenced metabotypes (GIMs) from blood metabolome across multi-ethnicities, leveraging the genetic basis of the metabolome associated with GC from UKB. We evaluated the predictive capacity of GIMs for GC risk and gastric lesion progression within the UGCED, SIT, and MITS cohorts. Based on predictive GIMs, we identified genetic regulators that may modify GC risk and <i>H.pylori</i> treatment efficacy. <h3>Results</h3> GIMs are consistently associated with blood metabolome across multiple ethnicities and centers, aligning with external validation sources (IDDF2024-ABS-0168-Figure 1. Stable causal learning framework robustly captured genetically influenced metabotypes across multiple ethnicities and health assessment centers). In the UGCED cohort, GIMs distinguished GC risk, with 28 biomarkers associated with gastric lesion progression over multiple endoscopic follow-ups. GIMs enabled GC risk stratification in the MITS and SIT cohorts over 11.8 (hazard ratio [HR]:2.95, 95%CI:2.50-3.48) and 22.3 years (HR:3.58, 95%CI:2.43-5.27) respectively, and pinpointed population subgroups with optimal <i>H.pylori</i> treatment efficacy in GC prevention (IDDF2024-ABS-0168 Figure 2. Genetically influenced metabotypes pinpointed population subgroups with optimal <i>H pylori</i> treatment efficacy in GC prevention). Further analyses pinpointed pleiotropic genes (<i>KPNB1, NPEPPS, APOB</i>) linking GC risk to blood metabolome and identified 128 tissue-specific effector genes colocalized with blood metabolome. We mapped these regulation patterns to metabolic networks, highlighting glutathione metabolism and glycosphingolipid biosynthesis pathways potentially affecting <i>H.pylori</i> treatment efficacy. Additionally, 40 genetic variants identified via GIMs were integrated into a user-friendly risk assessment tool, enabling risk-and-benefit stratification for GC prevention (IDDF2024-ABS-0168 Figure 3. Integrative analyses revealed key genetic variants and metabolic pathways that modify <i>H pylori</i> treatment efficacy in gastric cancer prevention). <h3>Conclusions</h3> Through geno-metabolomics integration, we identified genetic variants that may modify <i>H.pylori</i> treatment efficacy in GC prevention, offering novel and translational insights for targeted primary prevention.