Robot-assisted surgery offers clear advantages for both the patient and the surgeon, including better instrument handling and angulation, excellent visualization (3D, magnification), and reduced physical strain [4, 5].

These benefits have, among other things, led to broader acceptance of robotic techniques in minimally invasive pancreatic surgery for complex procedures, especially in comparison to laparoscopic techniques [11].

Propensity score–matched and single-center studies consistently demonstrate that robot-assisted distal pancreatectomy (RDP) offers several perioperative advantages over the open approach (ODP). These include a significantly shorter operative time (120 vs. 175 min; p < 0.001), reduced estimated blood loss (50 vs. 200 mL; p < 0.001), higher spleen preservation rates (approximately 63% vs. 26–33%; p < 0.001), lower infection rates (4.6% vs. 12.3%; p = 0.006), and faster recovery of gastrointestinal function. In addition, the length of hospital stay is significantly shorter following RDP (8 vs. 12 days; p < 0.01) [12]. Importantly, no significant differences have been observed regarding postoperative pancreatic fistula (POPF) or overall complication rates [13]. These findings are supported by systematic reviews and network meta-analyses, which report significantly lower blood loss (mean difference −426 mL; p < 0.001) and a shorter hospital stay (−4.06 days; p < 0.001) for RDP, while demonstrating no increase in major morbidity or clinically relevant POPF [14].

When compared to laparoscopic distal pancreatectomy (LDP), RDP appears to provide specific technical advantages. Meta-analyses indicate that RDP is associated with higher spleen preservation rates (OR 3.52; p < 0.0001), lower conversion rates to laparotomy (OR 0.41; p < 0.00001), and a slightly shorter postoperative hospital stay (−0.57 days; p = 0.002), as well as with lower 30-day mortality (0.1% vs. 1.0%; p = 0.03). However, these benefits come at the cost of significantly longer operative times (approximately 283 vs. 231 min; p = 0.01) and increased overall costs, while postoperative complication rates remain comparable. From an oncological perspective, the DIPLOMA trial demonstrated equivalent radical resection rates as well as comparable 1‑ and 2‑year survival and recurrence rates between minimally invasive and open distal pancreatectomy [15, 16]. The results of two ongoing RCTs are awaited to complement the already existing evidence: the randomized controlled trial to compare minimally invasive distal pancreatectomy to open resection (DISPACT‑2 trial) [17] and the international, multicenter, patient-blinded DIPLOMA III trial examining outcomes after laparoscopic versus robot-assisted left-sided pancreatectomy for benign and premalignant lesions [18].

Taken together, current evidence supports RDP as a safe and advantageous alternative to ODP, with comparable outcomes to LDP, which is reflected in current international guideline recommendations favoring minimally invasive approaches in appropriate patients [8, 19, 20].

For central pancreatectomy, a randomized controlled trial comparing robotic (RCP) and open procedures (OCP) demonstrated significant perioperative benefits associated with the robotic approach [21]. These included a shorter hospital stay (15.6 vs. 21.7 days; p = 0.002), reduced operative time (160 vs. 193 min; p = 0.002), lower intraoperative blood loss (50 vs. 200 mL; p < 0.001), lower rates of clinically relevant POPF (18% vs. 36%; p = 0.043), and faster postoperative recovery (3.1 vs. 4.6 days; p < 0.001). Retrospective analyses confirm reductions in blood loss and operative time with RCP, although trends toward higher overall morbidity (51.8% vs. 40%; p = 0.140) and POPF rates (34.5% vs. 22%; p = 0.105) did not reach statistical significance [22]. Furthermore, meta-analyses of minimally invasive central pancreatectomy, including both robotic and laparoscopic approaches, demonstrate reduced transfusion rates (0% vs. 8%; p < 0.01), lower rates of exocrine pancreatic insufficiency (1% vs. 5%; p = 0.032), and fewer readmissions (2% vs. 11%; p = 0.010), while showing no significant differences in POPF, major morbidity, or new-onset diabetes [23]. Despite these promising findings, the overall evidence base remains limited, and guideline recommendations are heterogeneous, with several calling for further high-quality studies before widespread implementation [8, 19, 20].

A direct comparison of the outcomes of robot-assisted pancreaticoduodenectomy (RPD) and open pancreaticoduodenectomy (OPD) was conducted in two randomized controlled studies [23, 24]. Liu et al. found a significantly shorter postoperative hospital stay after robot-assisted pancreatoduodenectomy (RPD median 11.0 days vs. OPD median 13.5 days; p = 0.029) and showed no differences in postoperative 90-day mortality (RPD 1% vs. OPD 1%; p = 0.999) and the incidence of severe complications (Clavien–Dindo ≥ 3; RPD 18 [22%] vs. OPD 19 [24%]; p = 0.82) between the two procedures. The time to first adjuvant therapy was significantly shorter for RPD (RPD 42 days vs. 48.5 OPD; p = 0.0025) [23]. Equally, in the EUROPA trial [24], no significant difference in 90-day postoperative morbidity after robotic and open pancreatoduodenectomy was found. However, significantly more patients experienced delayed gastric emptying after RPD than after OPD (RDP 10 [34.4%] vs. OPD 2 [6.0%]; p = 0.05) [24]. In the recently published DIPLOMA-2 trial, RPD was noninferior to OPD in terms of all items tested. However, this trial was only performed by surgeons who had long surpassed the learning curve. As already shown for RDP, robot-assisted pancreatoduodenectomy also caused significantly higher procedure-related costs (RDP 4744 € ± 1254 vs. OPD 866 € ± 459; p < 0.001) and longer operation time (RPD 431 min ± 103 vs. OPD 367 min ± 106; p = 0.021) than open procedures [24].

No RTC aiming for a direct comparison between laparoscopic and robotic pancreatoduodenectomy has been published so far [25]. In very experienced hands, RPD seams feasible, without a higher risk of adverse outcomes for the treated patients.

In summary, robot-assisted distal pancreatectomy offers clear perioperative advantages over the open approach and demonstrates comparable outcomes to laparoscopic surgery. Robotic central pancreatectomy appears promising but requires further validation due to limited evidence [8, 19, 20]. Robot-assisted pancreatoduodenectomy is feasible and oncologically safe in experienced centers, although it is associated with longer operative times and higher costs. Across all procedures, surgeon experience and treatment in high-volume centers are key determinants of patient safety and outcomes, highlighting the importance of structured training programs and careful consideration of learning curves.

For the clinical phase of training in minimally invasive pancreatic surgery, the three key steps competency/feasibility, proficiency, and mastery were identified [26, 27]. After each phase, patient selection concerning risk factors and technical difficulty changes, leading to increasingly complex cases [28].

Accordingly, the first phase, competency/feasibility, is reached when a procedure with average technical difficulty can be performed in patients without risk factors without supervision. The associated learning curve for feasibility is often based on operative time [26, 27].

Results for reaching the so-measured competency/feasibility learning curves show a mean cut-off of 33 cases for robot-assisted distal pancreatectomy and of 23 cases for robot-assisted pancreatoduodenectomy [28].

For proficiency, textbook outcomes need to be reproduced for procedures with increased technical difficulty or patients with higher risk profiles [28]. Often the rates of major complications (Clavien–Dindo grade ≥ III) are used to define cut-offs on the learning curve [26, 27]. In their review, Preuschka et al. reported a mean number of 80 cases to reach proficiency in robot-assisted distal pancreatectomy and 64 cases for robot-assisted pancreatoduodenectomy [28].

Mastery means reaching textbook outcomes in highly complex cases such as vessel and multi-visceral resection and operation on patients with high risk profiles [28]. Müller et al. defined textbook outcome as hospital stay shorter inside the 75th percentile (i.e., < 20 days), no mortality, no complication requiring (medium) intensive care unit admission, and no reoperation [26]. Recently, Preuschka et al. published mean case numbers of 122 robot-assisted distal pancreatectomies and 143 robot-assisted pancreatoduodenectomies as cut-offs for reaching mastery [28].

In the latest international consensus guidelines [20] surgeons are expected to reach the safety proficiency learning curve for robot-assisted distal and central pancreatectomy after 20 cases each and for robotic-assisted pancreatoduodenectomy after 50 cases. According to the European guidelines on minimally invasive pancreatic surgery, competency is attained after 15–21 procedures for robot-assisted distal pancreatectomy and after 25–40 cases for robot-assisted pancreatoduodenectomy [8], and a yearly center volume of 20 is recommended to ensure safe implementation [8].

The results of the Dutch multicenter training program in robotic pancreatoduodenectomy (LAELAPS-3) and the Pittsburgh Medical Center (UPMC) program underscore the importance and success of structured training programs for pancreatic surgery [27, 29].