Background: The 1-year disease-related mortality after resection for pancreatic cancer is approximately 30%. This study examined potential preoperative parameters that would help avoid unnecessary surgery.
Methods: Among the patients resected at our institution from 1997 to 2006, a total of 228 underwent pancreatic resection for ductal adenocarcinoma. By means of a survival cutoff of 12 months, two groups were created: early death (ED) and long survivors. A logistic regression analysis was performed to identify perioperative predictors of ED.
Results: Among 228 resected patients, postoperative mortality occurred in four cases (1.8%) that were excluded from the study. In the remaining 224 patients, 43 (19.2%) died of disease within 12 months from surgery (ED), and the remaining 181 (80.8%) had a longer survival. Multivariate analysis selected duration of preoperative symptoms > 40 days, CA 19-9 > 200 U/mL, pathological grading G3-G4, and R2 resection as independent predictors of ED.
Conclusions: Duration of symptoms, CA 19-9 serum level, and pathological grading possibly retrieved by endoscopic ultrasound-guided biopsy can be preoperatively used to identify patients with disease that is not suitable for up-front surgery, even if deemed resectable by high-quality imaging.