Potentially curative radical pancreaticoduodenectomy for ampullary adenocarcinoma was performed in 24 patients over a 35-year period. The overall operative mortality was 12.5%. Actuarial survival rate at 5 years was 61% f 13.4 standard error of the mean (SEM) and subsequently remained unchanged. In the same time period, 21 patients underwent potentially curative radical pancreaticoduodenectomy for periampullary tumors of pancreatic origin. Similar analysis showed an overall operative mortality of 23.8% and a survival rate at 5 years of 27% k 12.5 SEM. The results of radical pancreaticoduodenectomy for ampullary carcinoma in the most recent years (1976 to 1988) were compared with those of former years (1953 to 1975). There were no statistically significant differences in the 5-year survival rate; however, the operative mortality decreased from 25% in the former period to 6.3% in the recent period. Survival was dependent on nodal status. The 5-year survival rate was 78% k 11.5 SEM in the absence of nodal metastasis versus 50% k 25 SEM in the presence of regional nodal metastasis. These findings support the concept that radical pancreaticoduodenectomy offers a realistic probability for cure in a selected group of patients with carcinomas of the ampulla of Vater. Cancer 66:1717-1720,1990.
ARCINOMAS of the ampulla of Vater have definitive C distinguishing characteristics. Ampullary carcinomas have a significantly higher resectability rate and better long-term patient survival rates than pancreatic ductal adenocarcinoma originating in the periampullary area. I The close proximity of three separate visceral mucosae (duodenum, common bile duct, pancreas, and pancreatic duct) makes it difficult to clearly identify the origin of periampullary neoplasms. Likewise, it has been exceedingly difficult to characterize the natural history of these lesions and to establish the effectiveness of the various therapeutic modalities used in these tumors. This dilemma is amplified by the paucity of information on which therapeutic decisions must be made, including uncertainty of the histologic origin and extent of peri-