Background: Surgery is the only therapy with potentially curative intention in pancreatic cancer. This analysis aimed to determine prognostic parameters in a patient cohort with resected pancreatic adenocarcinoma with a special focus on the revised R1-definition.
Methods: Between October 2001 and August 2009, data from 1071 consecutively resected patients with pancreatic adenocarcinoma were prospectively collected in an electronical database. Parameters tested for survival prediction in univariate analysis included patient, tumor, and resection characteristics as well as adjuvant therapy. The parameters with significant results were used for multivariate survival analysis. Identified parameters with positive or negative prognostic effect were used to define risk groups and to assess the effects on patient survival.
Results: Age, ASA-score, CEA and CA19-9 levels, preoperative insulin-dependent diabetes mellitus, T-, N-, M-, R-, G-tumor classification, advanced disease, and LNR were all significant in univariate analysis, whereas gender, NYHA score, BMI, insurance status, type of surgical procedure, and adjuvant therapy were not. In multivariate analysis, age ≥70 years, preoperative insulin-dependent diabetes, CA19-9 ≥400 U/mL, T4-, M1- or G3-status, and LNR > 0.2 were independent negative predictors, whereas Tis/T1/T2-status, G1-differentiation, and R0-status (revised definition) were independently associated with good prognosis. Using these risk factors, patients were stratified into 4 risk-groups with significantly different prognosis; 5-year survival varied between 0% and 54.5%. Risk stratification resulted in improved survival prognostication within the predominant AJCC IIA and AJCC IIB stages.
Conclusions: A newly defined prognostic profiling including the revised R1-definition discriminates survival of patients with resectable pancreatic adenocarcinoma better than the AJCC staging system, and may be of particular relevance for patient-adjusted therapy in the heterogeneous group of AJCC stage II tumors.