In order to examine the feasibility of endoscopic microwave coagulation therapy (MCT) for the treatment of common bile duct cancer in humans, we investigated the safety and effectiveness of endoscopic MCT with the assistance of intraductal ultrasonography (IDUS). We performed MCT in the canine small intestine as a model because of its similarity in size to the obstructed human common bile duct. Multiple coagulation lesions were made in anesthetized dogs, with the endoscope inserted through a jejunostomy. Under condition "A", the tip of the MCT probe was kept in a fixed position in contact with the mucosa of the intestine, and coagulation was performed with an output of 30 or 50 W for 30 or 10 s, respectively. Under condition "B", the tip of the MCT probe was moved along the intestinal mucosa, and coagulation was performed with an output of 50, 40, or 30 W, all for 10 s. After coagulation, a thin-caliber ultrasonic probe was inserted endoscopically for observation of the effects of coagulation. Each coagulation under condition "A" resulted in degeneration of all layers of the intestine. Coagulation under condition "B" resulted in more localized degeneration, extending from the mucosal layer to the proper muscle layer. By comparison with histologic observations, the accuracy of IDUS in assessing depth of degeneration was 87.5%. Endoscopic microwave coagulation was considered safe and effective, and IDUS was able to accurately assess the effects of microwave coagulation.