Portal vein resection during pancreaticoduodenectomy for pancreatic neuroendocrine tumors. An international multicenter comparative study.

BACKGROUND: The role of portal vein resection for pancreatic cancer is well established but not for pancreatic neuroendocrine neoplasms. Evidence from studies providing information on long-term outcome after venous resection in pancreatic neuroendocrine neoplasms patients is lacking. METHODS: This is a multicenter retrospective cohort study comparing pancreaticoduodenectomy with vein resection with standard pancreaticoduodenectomy in patients with pancreatic neuroendocrine neoplasms. The primary endpoint was to evaluate the long-term survival in both groups. Progression-free survival and overall survival were calculated using the method of Kaplan and Meier, but a propensity score-matched cohort analysis was subsequently performed to remove selection bias and improve homogeneity. The secondary outcome was Clavien-Dindo ≥3. RESULTS: Sixty-one (11%) patients underwent pancreaticoduodenectomy with vein resection and 480 patients pancreaticoduodenectomy. Five (1%) perioperative deaths were recorded in the pancreaticoduodenectomy group, and postoperative clinically relevant morbidity rates were similar in the 2 groups (pancreaticoduodenectomy with vein resection 48% vs pancreaticoduodenectomy 33%). In the initial survival analysis, pancreaticoduodenectomy with vein resection was associated with worse 3-year progression-free survival (48% pancreaticoduodenectomy with vein resection vs 83% pancreaticoduodenectomy; P < .01) and 5-year overall survival (67% pancreaticoduodenectomy with vein resection vs 91% pancreaticoduodenectomy). After propensity score matching, no significant difference was found in both 3-year progression-free survival (49% pancreaticoduodenectomy with vein resection vs 59% pancreaticoduodenectomy; P = .14) and 5-year overall survival (71% pancreaticoduodenectomy with vein resection vs 69% pancreaticoduodenectomy; P = .98). CONCLUSION: This study demonstrates no significant difference in perioperative risk with a similar overall survival between pancreaticoduodenectomy and pancreaticoduodenectomy with vein resection. Tumor involvement of the superior mesenteric/portal vein axis should not preclude surgical resection in patients with locally advanced pancreatic neuroendocrine neoplasms.

The initiation, standardization and proficiency (ISP) phases of the learning curve for minimally invasive liver resection: comparison of a fellowship-trained surgeon with the pioneers and early adopters.

BACKGROUND: Using the Ideal Development Exploration Assessment and Long-term study (IDEAL) paradigm, Halls et al. created risk-adjusted cumulative sum (RA-CUSUM) curves concluding that Pioneers (P) and Early Adopters (EA) of minimally invasive (MI) liver resection obtained similar results after fewer cases. In this study, we applied this framework to a MI Hepatic-Pancreatic and Biliary fellowship-trained surgeon (FT) in order to assess where along the curves this generation fell. METHODS: The term FT was used to designate surgeons without previous independent operative experience who went from surgical residency directly into fellowship. Three phases of the learning curve were defined using published data on EAs and Ps of MI Hepatectomy, including phase 1 (initiation) (i.e., the first 17 or 50), phase 2 (standardization) (i.e., cases 18-46 or 1-50) and phase 3 (proficiency) (i.e., cases after 46, 50 or 135). Data analysis was performed using the Social Science Statistics software ( www.socscistatistics.com ). Statistical significance was defined as p < .05. RESULTS: From November 2007 until April 2018, 95 MI hepatectomies were performed by a FT. During phase 1, the FT approached larger tumors than the EA group (p = 0.002), that were more often malignant (94.1%) when compared to the P group (52.5%) (p < 0.001). During phase 2, the FT operated on larger tumors and more malignancies (93.1%) when compared to the Ps (p = 0.004 and p = 0.017, respectively). However, there was no difference when compared to the EA. In the phase 3, the EAs tended to perform more major hepatectomies (58.7) when compared to either the FT (30.6%) (p = 0.002) or the P's cases 51-135 and after 135 (35.3% and 44.3%, respectively) (both p values < 0.001). When compared to the Ps cases from 51-135, the FT operated on more malignancies (p = 0.012), but this was no longer the case after 135 cases by the Ps (p = 0.164). There were no statistically significant differences when conversions; major complications or 30- and 90-day mortality were compared among these 3 groups. DISCUSSION: Using the IDEAL framework and RA-CUSUM curves, a FT surgeon was found to have curves similar to EAs despite having no previous independent experience operating on the liver. As in our study, FTs may tend to approach larger and more malignant tumors and do more concomitant procedures in patients with higher ASA classifications than either of their predecessors, without statistically significant increases in major morbidity or mortality. CONCLUSION: It is possible that the ISP (i.e., initiation, standardization, proficiency) model could apply to other innovative surgical procedures, creating different learning curves depending on where along the IDEAL paradigm surgeons fall.