Minimally invasive versus open radical antegrade modular pancreatosplenectomy for pancreatic ductal adenocarcinoma: an entropy balancing analysis

Background: The safety and efficacy of minimally invasive radical antegrade modular pancreatosplenectomy (MI-RAMPS) remain to be established in pancreatic cancer (PDAC) METHODS: Eighty-five open (O)-RAMPS were compared to 93 MI-RAMPS. The entropy balance matching approach was used to compare the two cohorts, eliminating the selection bias. Three models were created. Model 1 made O-RAMPS equal to the MI-RAMPS cohort (i.e., compared the two procedures for resectable PDAC); model 2 made MI-RAMPS equal to O-RAMPS (i.e., compared the two procedures for borderline-resectable PDAC); model 3, compared robotic and laparoscopic RAMPS. Results: O-RAMPS and MI-RAMPS showed "non-small" differences for BMI, comorbidity, back pain, tumor size, vascular resection, anterior or posterior RAMPS, multi-visceral resection, stump management, grading, and neoadjuvant therapy. Before reweighting, O-RAMPS had fewer clinically relevant postoperative pancreatic fistulae (CR-POPF) (20.0% vs. 40.9%; p = 0.003), while MI-RAMPS had a higher mean of lymph nodes (25.7 vs. 31.7; p = 0.011). In model 1, MI-RAMPS and O-RAMPS achieved similar results. In model 2, O-RAMPS was associated with lower comprehensive complication index scores (MD = 11.2; p = 0.038), and CR-POPF rates (OR = 0.2; p = 0.001). In model 3, robotic-RAMPS had a higher probability of negative resection margins. Conclusion: In patients with anatomically resectable PDAC, MI-RAMPS is feasible and as safe as O-RAMPS.