At Hanover Medical School 466 patients were operated for ductal pancreatic cancer from 1971 to 1993. In 192 cases the pancreatic tumor could be resected, which amounts to a resection rate of 41%. These patients were subjected to uni- und multivariate factorial analysis in order to evaluate factors of significant prognostic value. Extended pancreatic resection because of vascular involvement or invasion of adjacent organs was performed in 40% (n = 77) of the cases. A curative R0 resection was possible in 85% of patients. The operative lethality dropped with time and from 1985 to 1993 was 3.5%. Patient survival at 5 years was 13.4% with a median survival time of 10.9 months. According to the univariate analysis the prognosis deteriorated with increasing tumor size and lymphatic and hematogenic metastases. There was a significant correlation between tumor stage according to the UICC and prognosis. The prognosis after resection was also significantly influenced by the residual tumor state and tumor grading. Extended pancreatic resections, however, had no significant prognostic impact. The multivariate analysis showed that only three factors were of independent significant relevance: residual tumor state, tumor size and tumor grade. The presence or absence of lymphatic and hematogenic metastases had no independent prognostic significance after resection of ductal pancreatic carcinomas.