In contrast to the more frequent cancers of the pancreatic head prognostic factors after resection of non-pancreatic periampullary cancers are less well established. In pancreatic cancer lymphatic involvement (evaluation of nodal disease per se, lymph node-ratio, number of examined nodes or microinvolvement) is a strong prognostic factor. We, therefore, evaluated the outcome after resection of more than 100 periampullary cancers with analyses of different parameters defining lymphatic spread.. Methods: We could analyze the outcome of 106 patients who underwent pancreaticoduodenectomy for ampullary (51%), distal bile duct (37%) or duodenal cancer (12%). Median age at time of surgery was 66 years (63% male). Evaluation was performed by retrospective analysis of our prospective pancreatic database. Survival analyses were done by the Kaplan-Meier- and by the Cox-regressions-methods. Results: In the 106 patients free resection margins were achieved in 81%. Tumor grading was G1/2 in 62% and G3/4 in 38%. Sixty-one percent of the tumors were node-positive. The median number of examined lymph nodes was 17 (range 4-44), the median number of involved nodes one (0 -14). The median lymph node-ratio was 0.07 (0-0.83). Actuarial five-year survival (5-ySurv) of the entire group was 31% and almost identical in the three tumor entities (p=0.89). By univariate analysis positive nodes (5-ySurv 19% vs. 45% in node negative patients; p<0.03), a positive margin (5-ySurv 13% vs. 37% with negative margins; p<0,01) and poor grading (5-ySurv in G3/4 10% vs. 41% in G1/2.; p<0,001) were identified as negative prognostic factors. The lymph node-ratio (after classification into groups with cut-off of 0.1 or 0.2) showed the same univariate influence as nodal disease alone. The number of examined nodes, gender and age did not influence survival. In multivariate survival analysis, however, only poor tumor grading was an independent prognostic factor (p<0.001; relative risk 2.5). Compared to patients with G1- or G2-grading patients with poor tumor grading had a significantly higher rate of nodal disease, a higher lymph node-ratio and a higher frequency of positive margins (31% vs. 10% in G1/2 tumors; p<0.01) although median tumor diameters were not different (22 vs. 20 mm). Conclusions: Prognosis after resection of non-pancreatic periampullary cancers is better than in pancreatic cancer. Prognosis is strongly influenced by tumor grading and therefore tumor biology which also correlates with further parameters like nodal disease and resection margins. Tumor grading, therefore, should be considered in further therapy studies in patients with periampullary cancers.