Robot-assisted pancreaticoduodenectomy with vascular resection: technical details and results from a high-volume center

Abstract

Background: Pancreaticoduodenectomy with vein resection (PD-VR) is widely accepted as a standard procedure to achieve a higher rate of R0 resections in borderline resectable pancreatic tumors. Thanks to the availability of newer technologies, such as the da Vinci Surgical System, several high-volume centers are reporting small series of minimally invasive PD-VR. Methods: A retrospective review of a prospectively maintained database was performed to identify patients who underwent robot-assisted PD-VR (RAPD-VR) between May 2011 and December 2019. The following factors were specifically analyzed: intraoperative results, post-operative complications, mortality at 90 days, patency of vascular reconstructions, overall survival (OS) and disease-free survival (DFS). Results: During the study period 184 patients underwent RAPD, including 22 who received a RAPD-VR (12.0%). The superior mesenteric vein was resected in 9 patients (40.9%), the portal vein in 3 patients (13.6%) and the spleno-mesenteric junction in 10 patients (45.5%). Based on the classification provided by the International Study Group on Pancreatic Surgery these procedures were classified as follows: 1 type I (4.5%), 3 type II (13.6%), 10 type III (45.5%) and 8 type IV (36.4%). In no patient the splenic vein was ligated and left behind. The splenic vein was always reimplanted either on the porto-mesenteric axis or in the inferior vena cava. All but one procedure, were completed under robotic assistance (conversion rate 1/22; 4.5%) after a mean operative time of 610.0±83.5 minutes. Median estimated blood loss was 899.7 mL (719.4–1,430.2 mL), with 2 patients (9.1%) receiving intraoperative blood transfusions. Sixteen patients developed post-operative complications (72.7%), graded ≥III (according to Clavien-Dindo) in 5 patients (22.7%). Two patients died within 90 days, accounting for a postoperative mortality of 9.1%. Interestingly, post-operative pancreatic fistula (grade B) occurred in only 1 patient (4.5%). Repeat surgery was required in 4 patients (18.2%) and hospital readmission in 1 patient (4.5%). At the longest available follow-up, vein reconstruction was patent in 19 patients (86.4%). Eighteen patients had a final diagnosis of pancreatic ductal adenocarcinoma (81.8%). After circumferential study of resection margins, microscopic tumor residual ≤1 mm was found in 11 patients (50.0%). The mean number of examined lymph nodes was 42.2 (±16.3), and vascular infiltration was confirmed in 13 patients (59.1%). Median OS was 39.7 (27.5–not available) and DFS 32.9 (11.5–45.8). Tumor recurrence was identified in 6 patients (27.3%). One patient (4.5%) developed isolated local recurrence. Conclusions: We have shown the feasibility of RAPD-VR. The results reported herein need to be confirmed in larger series and their generalizability remains to be established.