Background: The purpose of this study was to determine predictors of survival after resection for periampullary neoplasms.
Methods: Over a 15-year period, 208 patients underwent laparotomy for periampullary neoplasms. Data were analyzed to assess predictors of survival.
Results: Pathologic examination showed pancreatic cancer (n = 136; 65%), ampullary cancer (n = 28; 13%), distal common bile duct cancer (n = 10; 5%), duodenal cancer (n = 4; 2%), neuroendocrine tumor (n = 11; 5%), cystadenocarcinoma (n = 4; 2%), cystadenoma (n = 5; 2%), and other (n = 10; 5%). A total of 129 patients underwent pancreatic resection (71 Whipples, 35 total pancreatectomies, 21 distal pancreatectomies, and 2 partial pancreatectomies) whereas 79 patients were found to be unresectable and underwent palliative bypass and/or biopsy. Median survival was 20.4 months for resectable patients versus 4.5 months for unresectable patients (P<0.001). Of the 129 resected patients, factors significantly (P<0.05) favoring long-term survival on univariate analysis included well-differentiated histology, common bile duct or ampullary adenocarcinoma, early stage, tumor diameter <2 cm, negative margins, and absence of lymph node metastases, perineural, or vascular invasion. Age, sex, race, and type of procedure had no influence on survival. On multivariate analysis, only tumor differentiation appeared independently related to survival. Using Kendall's tau analysis, tumor type and grade correlated significantly with all other predictors.
Conclusions: Of all variables studied, tumor type and poor tumor differentiation in periampullary neoplasms appear to be markers that predict a constellation of other adverse findings.