The study was conducted at Fukuoka University from January 2009 to January 2018. The Medical Ethics Committee of Fukuoka University approved this study (IRB number: U20-01-011).
1. Primary Endpoint
Primary endpoint of this study to investigate the association between the clinicopathological characteristics and endoscopic features of EGCs diagnosed after H. pylori eradication. We enrolled 397 consecutive patients who underwent endoscopic submucosal dissection (ESD) treatment of their EGC. Of these, 225 patients were excluded as follows: 89 patients were not examined for H. pylori status; in 124 patients, H. pylori was not detected, and these patients did not have a past history of receiving H. pylori eradication therapy; and in another 12 patients, EGCs were resected using ESD within 1 year after H. pylori eradication. The remaining 172 patients were divided into two groups: An H. pylori-eradication group (H. pylori-EG) that comprised 36 EGCs in 32 patients who had undergone H. pylori eradication therapy more than 1 year before and on whom ESD was performed to treat their EGC, and an H. pylori-positive group (H. pylori-PG; the control group) that comprised 156 EGCs in 140 patients with an active H. pylori infection and on whom ESD was performed to treat their EGC (Fig. 1).
2. Secondary Endpoint
It remains unclear whether the mucosal atrophy affect the characteristics of EGC after eradication of H. pylori. Therefore, we compared the clinicopathological characteristics and endoscopic features between the two groups classified by the degree of gastric mucosal atrophy in the H. pylori-EG group. To investigate the association between EGC characteristics and the degree of mucosal atrophy after H. pylori eradication, 29 lesions of EGC in the H. pylori-EG group, in which the degree of mucosal atrophy had been confirmed in ESD specimens, were divided into two subgroups according to endoscopic and histological examinations. The mild atrophic mucosa subgroup comprised 6 lesions that were none to mild atrophic mucosa around the EGC, and the moderate to severe atrophic mucosa subgroup comprised 23 lesions that were moderate to severe atrophic mucosa around the EGC (Fig. 1).
Evaluation of H. pylori status
Evaluation of H. pylori eradication treatment was based on the 13C-urea breath test (UBT) or serum immunoglobulin (Ig) G antibody test (E-plate, Eiken, Tokyo, Japan) and on histological assessment using endoscopic biopsy specimens. When both examinations were negative, we determined that H. pylori had been eradicated. One hundred and forty patients in the control group were H. pylori-positive based on UBT or serum IgG antibody and histological assessment, and had no history of receiving H. pylori eradication therapy.
Clinicopathological findings such as size, location, macroscopic type, histological type, and depth of tumor invasion were reviewed for gastric carcinomas according to both the Japanese  and World Health Organization classification . The extent and degree of atrophic gastritis was evaluated endoscopically and histologically and classified into six categories according to the Kimura and Takemoto classification system (C-1 to O-3) .
Endoscopic examinations were performed by three experienced endoscopists (H.I, T.T, N.K) using a magnifying endoscope (GIF-H260Z, H290Z, Olympus Medical System, Tokyo, Japan). We used structural enhancement levels of A-8 for conventional endoscopy and B-8 for narrow-band imaging with magnifying endoscopy (NBI-ME). NBI-ME for diagnosing EGC was performed using a systematic classification system based on microvascular patterns and microsurface patterns (the “VS classification”). An irregular microvascular pattern (IMVP) and/or an irregular microsurface pattern (IMSP) with clear demarcation lines are the hallmarks of EGC . EGCs were confirmed using histopathological findings of biopsies and ESD samples.
Data were analyzed using JMP® 15 statistical software (SAS Institute Inc., Cary, NC, USA). Continuous variables between two groups were evaluated using the Mann–Whitney 𝑈 test. Categorical variables were evaluated using the chi-squared test. P values of <0.05 were considered statistically significant.