Background: Evidence is growing that carcinomas of the pancreatic head infiltrating major blood vessels (portal vein, superior mesenteric vein, hepatic or mesenteric artery) can be safely treated by pancreatoduodenectomy with en-bloc resection of the infiltrated vessels. Prognosis of these patients is a matter of debate.
Methods: Morbidity, mortality, and long-term survival in patients undergoing pancreatoduodenectomy with or without vascular resections (VR) were investigated. Between 1994 and 2005, 524 pancreato-duodenectomies were performed for malignancies. One-hundred patients (19.1%) had additional VR. 58 patients underwent venous resection with tangential excision with direct suture, 8 patients tangential excision with patch, 14 segmental resection with end-to-end anastomosis, 14 segmental resection with graft interposition and 6 patients different arterial resections. Additionally, 20 patients had vascular resections for benign and borderline dignities.
Results: Long-term survival was analyzed in 77 patients with pancreatic, in 12 patients with ampullary and in 11 patients with distal bile duct cancers using univariate (log-rank) and multivariate (Cox regression) methods. In 68 among 100 patients with VR, histopathology showed true tumor involvement. In the complete study group, negative resection margins (R0) were obtained in 85 of patients with VR and in 86% of patients without VR. No statistical difference was detected between both groups when comparing median operative time (350 VR versus 360 minutes w/o VR), and median volume of intraoperatively transfused blood units (3 units in both groups). Postoperative surgical complications/mortality occurred in 26%/7% (VR) versus 31%/5% (no VR, n.s.). Analysis of long-term survival in all 424 patients without VR revealed five-year survival probability of 16% compared to 4% in patients with VR (median overall survival 14 vs. 13 months; p=0.27). In multivariate modelling, resection margin and histological grading significantly influenced survival.