In a clinical evaluation, thirty-one patients with thoracic esophageal cancer who were treated from 1997 to 2000 were selected as subjects including those who underwent hand sewn anastomosis method, circular stapler method and triangular anastomosis method. Incidence of anastomosis failure was 27.3%, 25.0% and 8.3% for the hand sewn anastomosis method, circular stapler method and triangular anastomosis method, respectively. Anastomosis stenosis was found in 32.4%, 45.6% and 8.3% for the hand sewn anastomosis method, circular stapler method and triangular anastomosis method, respectively. In the basic examination, reduction rate of anastomosis caliber was 22.1+/-4.8%, 14.9+/-1.4% and 7.37+/-0.9%, for the hand sewn anastomosis method, circular stapler method and triangular anastomosis method. Microscopic evaluation revealed serious problems with the circular stapler method. The cause of anastomotic stenosis may include the fact that if anastomosis is performed by a circular stapler method, all the layers of gastrointestinal tract are punched out at the anastomosis portion, and mucosal conjugation is not observed and the muscular layer is exposed in the inner lumen of the gastrointestinal tract. Taking that the ulcer is formed circularly at the anastomotic portion into account, it is easily understood that this circular ulcer develops into stenosis in the healing process. The advantage of triangular anastomosis for esophago-gastric anastomosis is less suture failure, and is extremely advantageous for prevention of stenosis at the anastomotic portion when compared with other anastomosis methods. However, with regard to the healing process of eversion anastomosis in gastrointestinal tract instrumental anastomosis, detailed examination is expected hereafter.