Survival after surgery of pancreatic cancer is still poor even after curative resection. Some prognostic factors like the status of the resection margin, lymph node (LN) status or tumor grading were identified. However, only few data have been published regarding the prognostic influence of the LN-ratio (number of LN involved to number of examined LN), with sometimes conflicting results. We, therefore, evaluated potential prognostic factors in 182 patients after resection of pancreatic cancer.
Methods: Since 1994 204 patients underwent pancreatic resection for ductal pancreatic adenocarcinoma. Survival was evaluated in 182 patients with complete follow-up evaluations (54% female). Of those 182 patients 88% had cancer of the pancreatic head, 5% of the body and 7% of the pancreatic tail. Patients underwent pancreatoduodenectomy (86%), distal resection (11%) or pancreatectomy (3%). Survival was analyzed by the Kaplan-Meier- and Cox-methods.
Results: In all 204 resected patients mortality was 3.9% (n=8). In the 182 patients with follow-up 70% had free resection margins, 62% had G1 or G2-classified tumors and 70% positive LN. Median tumor size was 30 (7-80) mm. The median number of examined LN was 16, median number of involved LN one (range 0-22). Median LN-ratio was 0.1 (0-0.79). Cumulative five-year survival (5-y SV) in all patients was 16%. In univariate analysis a LN-ratio ≥ 0.2 (5-y SV 6% vs. 19% with LN-ratio < 0.2; p=0.003), LN-ratio ≥ 0.3 (5-y SV zero% vs. 18% with LN-ratio < 0.3; p<0.001), a positive resection margin (p<0.01) and poor differentiation (G3/G4; p<0.03) were associated with poorer survival. In multivariate analyses a LN-ratio ≥ 0.2 (p<0.01; relative risk RR 1.7), LN-ratio ≥ 0.3 (p<0.001; relative risk RR 2.4), positive margins (p<0.03; RR 1.6) and poor differentiation (p<0.05; RR 1.4) were independent factors predicting poorer outcome. The LN-ratio as a continuous variable also significantly correlated with survival whereas the conventional nodal status or the number of involved LN per se had no significant influence on survival. Patients with one LN involved had the same outcome than patients with negative nodes but prognosis worsened significantly in patients with two or more LN involved.
Conclusions: Not the lymph node involvement per se but especially the LN-ratio is an independent prognostic factor after resection of pancreatic head cancers. In our series this LN-ratio was even the strongest predictor of survival (cutoff 0.2). The routine estimation of this LN-ratio may be helpful not only for the individual prediction of prognosis but also in the planning of adjuvant therapy and further outcome and therapy studies.