Project Summary/Abstract The proposed research will characterize the prevalence and management strategies of gastric intestinal metaplasia (GIM), a lesion of the digestive tract that affects about 12.1 million adults in the United States (US) and is a precursor to gastric adenocarcinoma, or gastric cancer (GC). While mass surveillance of GIM is unlikely to be effective in regions of low-prevalence for GC such as the US, we hypothesize that risk- stratification for targeted screening and surveillance will be cost-effective and can improve outcomes. Extensive GIM has been identified by gastrointestinal society guidelines as a major risk factor for progression to GC, but there is currently a dearth of evidence in understanding the prevalence of GIM overall and by subtype (limited and extensive) in the US population, or how it should be managed. Furthermore, the recently proposed guidelines from the American Gastroenterological Association suggest surveillance of GIM should be considered in racial/ethnic minorities, foreign-born individuals, or those with a history of Helicobacter pylori (H. pylori) infection; however, evidence on the threshold to initiate such interventions and the intervals at which they should be continued is lacking. The research proposed in this K08 application will accomplish three interrelated Specific Aims. In Aim 1, we will utilize data from across the Veterans Affairs health care system to characterize the prevalence of subtypes of GIM and risk factors for progression of GIM to dysplasia or malignancy in the US context. These data will provide a platform for Aim 2, in which we will build a simulation model of the natural history of progression from precancerous gastric lesions to GC to assess which individual- level risk factor profiles could benefit most from screening and surveillance of GIM. The model will utilize 64 different phenotypic profiles, stratified by: gender (male/female), race/ethnicity (non-Hispanic white, Hispanic, black, and Asian), extent of GIM (limited vs. extensive), foreign-born status (immigrant vs. US-born), and H. pylori infection status (previous/current infection vs. no infection). Outcomes will be reported based on number of cancers prevented, survival, and incremental cost-effectiveness ratios (ICERs). Once an optimal profile and screening regimen is identified on the patient level, in Aim 3 we will assess the population-level impact of implementing such a strategy on health care outcomes and costs in the US. The long window of progression from GIM to malignancy and the low incidence of GC make sufficiently powered randomized controlled trials difficult. Simulation modeling allows for the integration of available knowledge to test multiple surveillance regimens of GIM across a broad range of risk factor combinations, which would not be feasible in clinical studies. By award period end, the proposed research will utilize national pathology data coupled with simulation modeling to identify a specific subgroup of high-risk patients that may benefit most from surveillance in the US. The proposal, mentorship and advisory committee, career development plans and institutional support will lay the basis for an independent NIH-funded career.

Project Narrative Data are limited regarding risk factors for progression of gastric intestinal metaplasia to gastric cancer in the Unites States (US) population, and regarding which individuals may benefit from preventative interventions such as endoscopic screening/surveillance. The proposed research will utilize pathology data from across the US to characterize the prevalence of extensive and limited gastric intestinal metaplasia, and the respective risks of progression of these lesions to gastric cancer. Subsequently, we will utilize simulation modeling to assess how targeted screening and surveillance can impact health outcomes in patients across 64 specific, diverse risk-factor profiles, both on a patient level and a population level.