Risk stratification for gastric cancer after Helicobacter pylori eradication: A population-based study on Matsu Islands.

Abstract

The reliable method to stratify the gastric cancer risk after Helicobacter pylori eradication remains an elusive goal. Mass eradication of H.pylori began in 2004 in a high-risk population. After eradication, a screening program involving first-stage serological tests (pepsinogen-I, pepsinogen-II, H.pylori immunoglobin G, and gastrin-17) and second-stage endoscopic examination was launched in 2015-2018. Index lesions included gastric cancer or extensive premalignant lesions. We evaluated the performance of the serological tests to "rule in" and "rule out" the risk based on positive and negative likelihood ratios, respectively. The methylation levels of microRNA-124a-3 in the stomach were measured to indicate genetic damage. Among 6512 invited subjects, 3895 (59.6%) participated. Both gastrin-17 and pepsinogen tests were normal in 3560 (91.4%) subjects; 206 (5.3%) gastrin-17 and 129 (3.3%) pepsinogen tests were abnormal. Years after eradication, the severity of gastritis had fallen greatly, and extensive premalignant lesions or gastric cancer frequently occurred in newly non-atrophic-appearing mucosa. Pepsinogen testing could moderately predict atrophic gastritis (positive likelihood ratio: 4.11 [95% confidence interval: 2.92-5.77]; negative likelihood ratio: 0.14 [0.10-0.19]). Gastrin-17 was not useful (0.66 and 1.20, respectively). However, pepsinogen testing poorly predicted the index lesions (2.04 [1.21-3.42] and 0.57 [0.34-0.95]). DNA methylation levels in the post-eradication mucosa were more discriminative for predicting index lesions (3.89 [2.32-6.54] and 0.25 [0.15-0.42]). After eradication, pepsinogen false-negative results become more frequent because histology is improved but genetic damage may persist. Direct testing for genetic damage offers better discrimination.