Initially, we conducted a consensus meeting to learn ME-NBI features of EGC after H. pylori eradication associated with diagnostic difficulty. For this consensus meeting, we consulted an EGC cohort in our retrospective study . This cohort consisted of 71 EGCs from 61 patients after H. pylori eradication (eradication group) and 115 EGCs with current H. pylori infection (control group). All patients attended the endoscopy center of Fujita Health University for the ESD between April 2011 and December 2016. All endoscopic photographs of EGC were reviewed by two expert endoscopists (TT and NH) and the ME-NBI features of EGCs were evaluated using the vessel plus surface (VS) classification system . In this system, endoscopic diagnosis of EGC is performed in terms of microvascular (MV), microsurface (MS) patterns and presence of demarcation line (DL). Diagnosis of EGC using the ME-NBI was based on the presence of irregular MV and/or MS with clear DL, while no or poor evidence of such findings was defined as the lesion with difficult diagnosis. Factors associated with diagnostic difficult was further explored using the ME-NBI photographs. Fujita Health University School of Medicine approved the protocol of the retrospective study, and written informed consent was obtained from all participating subjects.
Based on the consensus meeting, we prospectively evaluated the real-time diagnostic yield of combining WLE, chromoendoscopy (CE) using indigo carmine, and ME-NBI in undiagnosed gastric lesions in patients after successful H. pylori eradication. For this prospective study, 219 patients who had at least 6 months of post-eradication were enrolled.
Between December 2016 and September 2018, 298 upper endoscopic examinations were performed for these patients by two expert endoscopists (TT and NH). For all examinations, endoscopy was performed with the GIF-H260Z or the GIF-H290Z (Olympus Corporation, Tokyo, Japan). After inserting the endoscopy into the stomach, screening was performed using WLE. If the newly detected and undiagnosed gastric lesions were identified in the screening endoscopy, lesion was diagnosed by 1) WLE, followed by 2) CE using indigo carmine dye (0.2%) and 3) ME-NBI. If an endoscopist could diagnosed the lesion as either neoplastic or non-neoplastic, we considered it as a diagnosed case and diagnostic results (neoplastic or non-neoplastic) were recorded based on the real-time observation. As a rule, these results were not changed if a different diagnosis was made by other modalities. Diagnostic criteria of EGC using the conventional WLE and the CE was based on the presence of well–demarcated, depressed, or elevated lesions with an irregular margin and an irregular mucosal area with a color change (reddish or whitish), while no evidence of such findings was considered to be a noncancerous lesion by the conventional WLE and the CE. Although, the diagnostic criteria of EGC using ME-NBI was made by the VS classification system , based on the consensus meeting, we especially tried to evaluate irregular MV using high power magnification. After endoscopic observation, at least biopsy was taken from the lesion. The biopsy specimens were evaluated using H&E staining. The diagnostic pathological criteria were based on the Japanese Classification of Gastric Carcinoma, 14th edition . Groups 3, 4 and 5 were diagnosed as neoplastic lesions and group 1 was diagnosed as non-neoplastic lesions. If the biopsy result was group 2 (indefinite for neoplasia), additional examination was scheduled for the conclusive pathological diagnosis. Fujita Health University School of Medicine approved the protocol of the prospective study, and written informed consent was obtained from all participating subjects. This study was registered with the University Hospital Medical Information Network (UMIN000033100).
Diagnostic yields of WL, CE and M-NBI were evaluated in terms of sensitivities, specificities, and diagnostic accuracies were calculated by reference to the pathological diagnosis of endoscopic biopsy. Diagnostic accuracy was defined as (true positive+ true negative)/total number of cases). Statistical differences of continuous and categorical variables between the two groups were determined using the Student‘s t-test and the Chi-square Test, respectively. P<0.05 was considered significant.