Single-centre comparison of robotic and open pancreatoduodenectomy: a propensity score-matched study.

Abstract

Pancreatoduodenectomy for pancreatic head and periampullary cancers is still associated with high perioperative morbidity and mortality. The aim of this study was to compare the short-term outcomes of robot-assisted pancreatoduodenectomy (RAPD) and open pancreatoduodenectomy (OPD) performed in a high-volume centre. A single-centre, prospective database was used to retrospectively compare the early outcomes of RAPD procedures to standard OPD procedures completed between January 2014 and December 2018. Of the 121 included patients, 78 underwent RAPD and 43 underwent OPD. After propensity score matching (PSM), 35 RAPD patients were matched with 35 OPD patients with similar preoperative characteristics. There were no statistically significant differences in most of the baseline demographics and perioperative outcomes in the two groups after PSM optimization with the exception of the operative time (530min (RAPD) versus 335min (OPD) post-match, p < 0.000). No differences were found between the two groups in terms of complications (including pancreatic leaks, 11.4% in both OPD and RAPD), perioperative mortality, reoperations or readmissions. Earlier refeeding was obtained in the RAPD group vs. the OPD group (3 vs. 4days, p = 0.002). Although the differences in the length of the hospital stay and blood transfusions were not statistically significant, both parameters showed a positive trend in favour of RAPD. The number of harvested lymph nodes was similar and oncologically adequate. RAPD is a safe and oncologically adequate technique to treat malignancies arising from the pancreatic head and periampullary region. Several perioperative parameters resulted in trends favouring RAPD over OPD, at the price of longer operating time. Data should be reinforced with a larger sample to guarantee statistical significance.