関口 裕美

OBJECTIVES: Early gastric cancers (EGCs) of the elevated type or with submucosal invasion are easily found by routine endoscopy. However, most early cancers are challenging to detect because of subtle morphological or color differences from surrounding atrophic mucosa and intestinal metaplasia. Linked color imaging (LCI) enhances mucosal color difference, making it easier to detect EGCs. The aim of this study is to clarify the advantages and possible disadvantages of LCI for screening for obscure EGC. METHODS: A total of 665 malignant gastric lesions resected using endoscopic submucosal dissection between January 2015 and April 2018 were retrospectively reviewed. Obviously detectable lesions were not included in the main analysis when determining the target lesion. White light imaging (WLI)/LCI images of 508 endoscopically obscure malignant lesions were included in the final analysis and evaluated by three non-expert and three expert endoscopists using visibility scores for detection and extent. RESULTS: The detection visibility scores using LCI were significantly higher than those using WLI regardless of lesion characteristics including location, size, histological type, depth of invasion, and Helicobacter pylori status. The detection score improved in 46.4% cases and deteriorated in 4.9% when the modality changed from WLI to LCI. A mixed-effects multivariate logistic regression analysis showed that use of LCI (odds ratio [OR] 2.57), elevated type (OR 1.92), invasion to submucosa (OR 2.18) were significantly associated with improved visibility of EGC. CONCLUSIONS: Linked color imaging significantly improves visibility of EGC regardless of differences in lesion morphology, histology, location, depth of invasion, and H. pylori status compared to conventional WLI.

BACKGROUND: Duodenal ulcers are classified into bulbar and post-bulbar ulcers. The aim of this study is to compare the long-term outcomes of patients with post-bulbar ulcer bleeding and those with bulbar ulcer bleeding. METHODS: A total of 272 patients with hemorrhagic duodenal ulcers requiring hospitalization were included. Their medical records were retrospectively reviewed. RESULTS: All patients were categorized as bulbar or post-bulbar bleeding ulcer groups. The post-bulbar ulcer group had more patients of advanced age, concurrent malignancy, diabetes mellitus, hypertension, cirrhosis, and chronic kidney disease undergoing hemodialysis. We performed long-term follow-up for an average of 2.6 years. The mortality rate during the follow-up period in the post-bulbar ulcer group was significantly higher than that in the bulbar ulcer group (p < 0.001). The PNED score was a better predictor of 30-day mortality compared to the complete Rockall score and the Glasgow-Blatchford Score. Predictors of mortality were evaluated using a Cox proportional hazards regression model. In multivariate analysis, post-bulbar ulcer, concurrent malignancy, cirrhosis, antiplatelet/anticoagulant use, and transfusion were significant predictors of mortality. CONCLUSIONS: Patients with post-bulbar ulcers have a poorer prognosis than those with bulbar ulcers. After the diagnosis of hemorrhagic post-bulbar duodenal ulcer, close follow-up is necessary.