Abstract
BACKGROUND/AIMS: Effective implementation of population-based esophagogastroduodenoscopy (EGD) screening requires careful evaluation of its diagnostic performance and potential harms. This study aimed to assess the diagnostic performance and potential harms of EGD for gastric cancer screening under the National Cancer Screening Program. METHODS: In this retrospective study, asymptomatic individuals aged ≥40 years who underwent screening EGD between 2017 and 2023 were included. Diagnostic yield and potential harms were compared between middle-aged and older adults. Univariable and multivariable Poisson regression analyses were used to identify factors associated with detection of high-risk gastric lesions. RESULTS: Among 12,413 participants, a mean of 35.4 endoscopic images per examination was obtained, with a mean procedure time of 4.2 minutes. Precancerous conditions and high-risk gastric lesions were identified in 64.3% and 0.7% of participants, respectively, both of which were more prevalent in older adults (p<0.001 and p=0.043, respectively). Procedure-related adverse events (1.1%) and false-positive findings (2.8%) were comparable between age groups. However, false-negative results were significantly higher among older adults than middle-aged individuals (3.4% vs 1.1%, p<0.001). In multivariable analysis, older age (≥65 years), male sex, more EGD images, and longer procedure time were independently associated with the detection of high-risk lesions. CONCLUSIONS: In population-based EGD screening, older age, male sex, and high-quality procedural metrics (e.g., more images and longer examination time) were independently associated with the detection of high-risk gastric lesions. These findings may inform risk-stratified screening strategies and quality benchmarks for EGD screening programs.