Results of pancreaticoduodenectomy for pancreatic cancer: extended versus standard procedure.

In Western experience, the long-term survival benefit after extended pancreaticoduodenectomy (EPD) in patients with pancreatic ductal adenocarcinoma is still controversial. The aim of this work was to evaluate weather EPD for pancreatic ductal adenocarcinoma prolongs long-term survival compared to standard pancreaticoduodenectomy (SPD). From November 1992 to September 1996, we performed pancreatic resections in 30 patients affected by stage I-III pancreatic ductal adenocarcinoma: 13 patients underwent SPD and 17 patients underwent EPD, consecutively. The two groups of patients were similar for all the demographic, clinical, and pathological characteristics, and all the intraoperative factors considered except the number of resected lymph nodes (mean number per case = 34.2 +/- 15.5 in the EPD group versus 12.8 +/- 3.6 in the SPD group, p <0.001) and the operative time (median time per case = 375 minutes in the EPD group versus 270 minutes in the SPD group, p = 0.009). Patients in the two groups experienced a similar postoperative course. The estimated survival probability at 1 and 3 years after operation was 0.76 (95% confidence interval [CI]: 0.49 to 0.90) and 0.24 (95% CI: 0.07 to 0.45) in the EPD group; 0.31 (95% CI: 0.09 to 0.55) and 0.08 (95% CI: 0.00 to 0.29) in the SPD group (p = 0.014). According to a Cox model, the treatment was associated with R0 patients' long-term survival (SPD versus EPD: hazard ratio (HR) = 4.82, 95% CI: 1.66 to 14.00, p = 0.004). Grading of tumor differentiation was confirmed to be a relevant prognostic factor (poor versus moderate: HR = 4.33, 95% CI: 1.49 to 12.61, p = 0.007), whereas type of resection had no significant effect (pylorus-preserving versus hemigastrectomy: HR = 1.49, 95% CI: 0.56 to 3.95, p = 0.42). The proportion of R0 patients with local recurrence was lower in the EPD group (20.0% versus 70.0%, p = 0.034).