Attachment to the peritoneum of cancer cells appears to be the most characteristic step in the peritoneal implantation. It is likely that the interaction between gastric cancer cells and the peritoneum is mediated by specific adhesion molecules. CD44 has been shown to recognize both the major cell-surface receptor for hyaluronic acid and the extra cellular matrix. The purpose of the present study is to elucidate the binding capacity of highly disseminated cells of gastric carcinoma and to investigate the ability to mediate the binding of gastric scirrhous carcinoma ceils to peritoneum. MATERIALS and METHODS: OCUM2M cell line (poorly developed peritoneal dissemination by peritoneal inoculation) was established from human gastric scirrhous carcinoma. OCUM-2MD3 cell line (highly developed peritoneal dissemination by peritoneal inoculation) was established from parental cell line OCUM2M. The adherent capacity of two gastric cancer cell lines to mesothelial cells, hyaluronate, matfigel ,fibronectin, laminin and type I collagen w as investigated by adhesion assay. Then, the effect ofanti-CD44 antibody on the adhesion of OCUM-2MD3 to mesothelial cells, hyaluronate, fibronectin and type I collagen was examined. RESULT: OCUM-2MD3 had higher adhesiveness to monolayer mes'othelium, hyaluronate, matrigel, fibronectin, laminin and type I collagen than OCUM-2M. Adhesion activity of OCUM-2MD3 to mesothelium, hyaluronate and fibronectin was inhibited partly by anti-CD44 antibody. CONCLUSION: Highly disseminated cell line OCUM-2MD3 demonstrated significant binding to peritoneal components and it was likely that the interaction between a gastric cancer cell line OCUM-2MD3 and the peritoneum is partly mediated by CD44. ENDOSCOPIC ASPIRATION MUCOSAL RESECTION FOR GASTRIC LESIONS. H. Nishisaki , K. Yasu take , T. Nakashima, H. Hasegawa, M. Oya. Depar tment of Gastroenterology, Hyogo Medical Center for Adults , A k a s h i , Japan ((Introduction)) Endoscopic ~.gastric mucosal resect ion (EMR) is used widely in Japan for d iagnos is and t r ea tmen t of gas t r i c lesions. But lesions; o n the' lesser cu rva tu re of upper Corpus are diff icul t to remove wi th EMR. We eva lua ted an endoscopic a sp i r a t i on mucosa l resect ion technique tha t ut i l ized a p las t ic clear hood (EAMR). {Subjects and Method} Gastr ic mucosal lesions Of less t h a n 10mm i n d iameter 'were obta ined by biopsy from 21 p a t i e n t s ; 17 pa t i en t s had borderl ine lesions (adenoma ; e levated l e s ion / dysp las ia ; depressed lesion), 3 had gasl;ric cancers and 1 had carcinoid. The locat ion of the gas t r ic lesions were as fo l lows ; 2 fornix lesions, 10 corpus lesions, 2 angle lesions and 9 a n t r u m lesions. EAMR was done W i t h an endoscopy (Q-20, O lympus Optical Co. or EG400HR, FUJINON Co.) covered w i t h a p las t i c clear h o o d (Olympus Optical Co., i 3 m m outer, 1 ) m m inner diameter, 10ram depth). Our procedure was as fo l lows: the endoscope t ip was covered wi th the plas t ic h o o d and an anglo ca the ter (Hanaco Co.) was placed on the side o f the endoscope as snare guide tube. A s m a l l s i z e snare was inser ted t h rough the c a t h e t e r and ol~ened outs ide the hood f i t ted a t the t ip of the scope. F i f ty percent glucose was : injected into the submucosa l layer to lift an. observed t a rge t mucosa l lesion. The lesion Was then eas i ly asp i ra ted inside the hood, snared and cut. (Results} In 20 pa t i en t s lesions were, complete ly resected. But in one pa t ien t resect ion was incomplete as the lesion was in fornix direct ly under EC-junct ion. Diameter of resected areas were 25I0 x l l 7 m m ~ (12.0 -+ 3.6 x 9.5 +2.1mm~). Complicat ions included one bleedirlg and one perforat ion. Pa thological f ind ings were as follows ; 2 erosions, 8 adenomas, 7 d y s plasias, 3 g a s t r i c cancers and one carcinoid. <(Conclusion} EAMR is useful for r emoving gas t r i c neoplast ic lesions qu ick ly and eas i ly and can contribiate to decreasing the morb id i ty of ad vanced gas t r i c cancer.