Mo1547 Usefulness of Narrow-Band Imaging in Delineating Early Gastric Cancer

Usefulness of Narrow-Band Imaging in Delineating Early Gastric Cancer Sayoko Nakayama, Hirohisa Machida, Kazunari Tominaga, Masatsugu Shiba, Yasuaki Nagami, Satoshi Sugimori, Masami Nakatani, Natsuhiko Kameda, Hirotoshi Okazaki, Hirokazu Yamagami, Tetsuya Tanigawa, Kenji Watanabe, Toshio Watanabe, Yasuhiro Fujiwara, Tetsuo Arakawa Osaka City University Graduate School of Medicine, Osaka, Japan; Department of Gastroenterology, Osaka City General Hospital, Osaka, Japan [Background & Aims]: Endoscopic submucosal dissection (ESD) has been widely accepted as an endoscopic therapy for early gastric cancer (EGC), to obtain a precise pathological diagnosis and promising curability for EGC. Appropriately margining resection at will is one of distinctive features of ESD, therefore, the precise estimation for the lateral extent of the lesion has an important role in successful ESD. The aim of this study is to investigate the usefulness in delineating the lateral extent of EGC using narrow-band imaging (NBI) endoscopy with magnification. [Methods]: February 2004 to October 2010, a total of 503 patients who underwent ESD for EGC achieving en bloc resection were enrolled in this study from. Patients were assigned to two groups, according to whether or not they underwent NBI endoscopy with magnification before ESD; Group A: 37 lesions evaluated with chromoendoscopy, as a historical control and group B: 465 lesions evaluated with NBI endoscopy. As outcome measures, complete resection rate (defined as en bloc resection and cancer free margin), location, size, macroscopic type, histology and presence of ulcer scar (UL) were compared between groups. In estimating the lateral extent, the margin was determined according to surface structure of mucosa in chromoendoscopy with indigo-carmine dye spraying and the presence of irregular microvascular pattern and demarcation line which were specific for EGC in NBI endoscopy. [Results]: Complete resection rate in group B was superior to that in group A (98.3% (8/ 465) vs. 91.9% (34/37), p 0.01). The incidence of incomplete resection was as followed; 1) Location; upper/ middle/ low: 1.5% (2/132)/ 3.5% (5/144)/ 0.5% (1/ 189), 2) Size; less than 20mm/ over 20mm: 1.3% (4/305)/ 2.5% (4/160), 3) Macroscopic type; elevated/ flat/ depressed: 0% (0/244)/ 6.7% (1/15)/ 3.4% (7/ 206), 4) Histology; differentiated/ undifferentiated: 1.3% (6/446)/ 5.3% (2/38), 5) Presence of an ulcer scar; 6.2%(5/81)/ 0.8% (3/384). As a result of univariate analysis in group B, complete resection rate was significantly inferior in the point of macroscopic type (p 0.01), presence of UL (p 0.01) and histology (p 0.027), in comparison with cancer positive and negative in margin. And it was not significant in location and size. In multivariate analysis, it was significant in the point of histology (p 0.025) and presence of UL (p 0.025). [Conclusions]: NBI endoscopy with magnification might be usefulness in an appropriate delineating the lateral extent of EGC. Features of histology or ulcer scar may affect the accurate endoscopic marginal estimation for EGC even though using NBI endoscopy.