Natural history and biomarker discovery in gastric preneoplasia: Insights from the TUMMIE longitudinal study

Background: Gastric cancer remains a leading cause of cancer mortality worldwide, with marked racial and ethnic disparities in incidence and outcomes. Gastric intestinal metaplasia (GIM) is a key preneoplastic lesion, yet its natural history and progression biomarkers are poorly understood, particularly in diverse U.S. populations. Methods: The TUMMIE study is a longitudinal cohort of patients with gastric preneoplasia at UC Davis Health. In a pilot phase, we enrolled 69 individuals (mean age 66 years) with histologically confirmed GIM. The cohort demographic composition was 67% female and 33% male, with race/ethnicity distributed as 67% non-Hispanic White, 15% Asian, 7% Black, 7% Hispanic White, and 4% multiracial, Native Hawaiian/Other Pacific Islander, or missing. Twenty-two patients (32%) reported a personal history of cancer and nearly half (48%) reported a family history of gastric conditions.. Among these, 39 participants provided stool and saliva samples, which were analyzed for Helicobacter pylori (H. pylori) strain detection using digital droplet PCR (ddPCR). We assessed concordance between clinical testing and ddPCR-based detection in both sample types. Additionally, we conducted a scoping review of published biomarkers associated with gastric preneoplasia progression. Results: Of the 39 biospecimen donors, 8 tested high positive for H. pylori by ddPCR. Strain typing revealed clear variation in H. pylori virulence profiles, with ddPCR detecting both CagA-positive and CagA-negative strains across participants, including several samples with strong high-droplet signals (>100 to >1000) thereby indicating strain diversity in the cohort. Clarithromycin resistance mutations were detected by ddPCR, this indicates the presence of both antibiotic-resistant and antibiotic-sensitive H. pylori strains. Several samples carried VacA toxin-associated alleles and EPIYA motifs. Preliminary analyses show correlations between ddPCR detection and clinical test results. Demographic characteristics and infection rates varied across racial/ethnic groups, highlighting potential disparities in risk profiles. The scoping review identified candidate biomarkers in pathways related to inflammation, immune response, and epithelial transformation. Conclusions: This pilot phase demonstrates feasibility of biospecimen collection and molecular characterization in a multiethnic cohort with GIM. Integration of ddPCR-based H. pylori detection and biomarker discovery may inform risk stratification and surveillance strategies for gastric cancer prevention. Expanded longitudinal follow-up will enable validation of progression biomarkers and clarify racial/ethnic differences in disease trajectory.