Current statement and safe implementation of minimally invasive surgery in the pancreas

Abstract Minimally invasive pancreatic resection has become very popular in modern pancreatic surgery. Evidence of the benefits of a minimally invasive approach is accumulating thanks to prospective and randomized controlled studies. Minimally invasive surgery provides advantages to the surgeon due to the high definition of the surgical field and the freedom of fine movement of the robot, but should be considered only in selected patients and in high‐volume centers. Minimally invasive distal pancreatectomy for benign and low‐grade malignant tumors has established a secure position over open distal pancreatectomy, since it is associated with a shorter hospital stay, reduced blood loss, and equivalent complication rates. Minimally invasive distal pancreatectomy for pancreatic ductal adenocarcinoma appears to be a feasible, safe, and oncologically equivalent technique in experienced hands. On the other hand, the feasibility and safety of minimally invasive pancreaticoduodenectomy are still controversial compared with open pancreaticoduodenectomy. The choice of either technique among open, laparoscopic, and robotic approaches depends on surgeons' experience and hospital resources with a focus on patient safety. Further studies are needed to prove the perioperative and oncological advantages of minimally invasive surgery compared to open surgery in the pancreas. Here, we review the current status of minimally invasive pancreatic surgery and its safe implementation.

hospital stay, reduced blood loss, and equivalent complication rates.
Both laparoscopic and robotic DP can be safe and feasible options.

| MINIMALLY INVA S IVE D IS TAL PAN CRE ATEC TOMY (MIDP)
Distal pancreatectomy has traditionally been performed using an open approach. In the past decade, the minimally invasive approach using laparoscopic surgery or robot-assisted surgery has become increasingly popular. 5,6 Pooled data of observational studies, generally from single, high-volume expert centers, have suggested that MIDP is associated with a shorter length of hospital stay compared with ODP. 7,8 Despite this potential benefit, MIDP is only used in about one-third of patients, according to a recent analysis of the National Surgical Quality Improvement Program (NSQIP) database. 9 In Japan, a multicenter prospective registration study including 1197 LDPs revealed that postoperative morbidity and 90-day mortality rates after LDP were 17% and 0.3%, respectively. 10 Thus, LDPs are performed safely in Japan, especially in experienced institutions.
The first multicenter, patient-blinded, randomized controlled trial (LEOPARD study) demonstrated enhanced functional recovery after MIDP compared with ODP. 11 MIDP also reduced operative blood loss, delayed gastric emptying, hospital stay, and adverse effect on postoperative quality of life. 11 Additionally, MIDP (excluding the robotic approach) has been reported to be at least as cost-effective as ODP. 12 Cosmesis and quality of life were similar in MIDP and OPD 1 year after surgery. 12 On the other hand, the clinical application of MIDP for PDAC is still controversial. Radical (R0) resection and enough lymph node Clavien-Dindo grade ≥3 complications and 90-day mortality were comparable for MIDP and ODP. Although the R0 resection rate was significantly higher in MIDP, Gerota's fascia resection was less frequent and lymph node retrieval was lower after MIDP compared to ODP. Median survival time in a propensity score-matched cohort was comparable for MIDP and ODP (29 and 31 months, respectively). 13 To compare oncologic outcomes between MIDP (laparoscopic or robot-assisted) and ODP in patients with PDAC, another systematic review screened 1760 studies and then included 21 studies with 11 246 patients. 14 In this review, although overall survival, R0 resection rate, and use of adjuvant chemotherapy were comparable for MIDP and ODP, the retrieved lymph node was significantly lower in MIDP. Additionally, patients undergoing MIDP were more likely to have smaller tumors, less perineural invasion, and less lymphovascular invasion, reflecting earlier stage disease as a result of treatment allocation bias. The systematic review 14 concluded that MIDP for PDAC was associated with comparable survival, R0 resection, and use of adjuvant chemotherapy, but a lower lymph node retrieval as compared to ODP. To overcome the issue of Gerota's fascia resection and lymph node retrieval during MIDP, the laparoscopic radical antegrade modular pancreatosplenectomy procedure for left side PDAC may be applicable. 15 MIDP has spread internationally, and the oncological outcomes are mainly elucidated by prospective, observational series. 13,[16][17][18][19][20] Due to the absence of RCTs, the oncologic efficacy of MIDP for PDAC remains unclear. In the current status, MIDP for PDAC provides equivalent results in terms of lymph node retrieval and positive margin status and comparable survival compared to ODP (Table 1). A further RCT to confirm the oncological safety of MIDP for PDAC is needed.
The recently developed robotic surgical system has overcome the limitations of laparoscopic technology by providing an isometric 3D view and a high level of flexibility for manipulation. Robotic approaches for MIDP have been increasingly applied throughout the world. In Japan, robotic application for MIDP is covered by health insurance from April 2020. Although studies addressing the robotic benefits of MIDP are still few, it has been reported that robotic DP is as feasible and as safe as the laparoscopic and the conventional open approaches. 21 Lai and Tang 23 reported that RDP required a longer operative time than LDP, but there were no marked differences in blood loss, spleen-preservation rate, postoperative hospital stay, or overall morbidity rate between the two groups. In a comparison of surgical outcomes among three types of DPs (RDP in 21, LDP in 25, and ODP in 43 cases) for benign and malignant diseases, operative time was longest in RDP (ODP < LDP < RDP) and blood loss was lowest in RDP (RDP < LDP < ODP). 24 The rate of patients with Clavien-Dindo ≥ grade III was lowest in RDP (RDP < LDP < ODP) and length of hospital stay was also shortest in RDP (RDP < LDP < ODP). Thus, RDP can provide less invasiveness in certain aspects.

| MINIMALLY INVA S IVE PAN CRE ATODUODENEC TOMY (MIPD)
Minimally invasive pancreatoduodenectomy (MIPD) is a challenging surgery because of its technical difficulty. The greatest difference between MIDP and MIPD is the presence in MIPD of anastomotic reconstruction such as cholangiojejunostomy, pancreatojejunostomy, and gastrojejunostomy. The laparoscopic procedure is unsuitable for the reconstruction process of LPD because of the difficulty in adjusting the axis of the forceps to adequate suturing lines of pancreatic or biliary reconstruction. 10 Immature reconstruction substantially leads to postoperative complication such as clinically relevant pancreatic fistula/anastomotic leakage, resulting in prolonged hospital stay and even surgery-related mortality. These situations also apply to OPD.
According to the Miami international evidence-based guidelines, insufficient data exist to recommend MIPD over OPD. 4 LPD has been associated with less delayed gastric emptying, decreased blood loss and shorter hospital stay compared to OPD, without increasing overall costs. 7,25, 26 Lai and Tang 23 comprehensively reviewed the LPDs and reported equivalent outcomes with respect to perioperative morbidity and mortality rates compared to OPDs, although the laparoscopic approach tended to have a longer operation time and less blood loss. These results may be difficult to generalize to other centers, because all these procedures were performed by experienced laparoscopic surgeons. Lai and Tang 23 concluded that LPD was feasible and safe in well-selected patients in experienced hands.
Three RCTs comparing LPD and OPD have been published. [27][28][29] Two single-center RCTs reported a shorter hospital stay in LPD. 28,29 On the other hand, a recent report from the Dutch Pancreatic   including various pancreatic pathologies. [44][45][46] The oncologic benefit on PDAC derived from RPD remains uncertain, because of insufficient pooled data. [40][41][42][43] According to limited cohort studies, 40-43 the survival outcomes in PDAC are comparable between RPD and OPD (Table 1). In addition, there is no evidence of superiority between RPD and LPD. Although RPD may overcome some of the technical difficulties in LDP, its feasibility and safety need to be verified by the accumulation of clinical data. Depending on the surgical outcomes (mainly in the operative time and blood loss) that were used to assess the learning curve, 10-20 cases have been proposed to be required to reach proficiency in LDP [53][54][55][56][57][58][59][60] (Table 2). For LPD, the surgeons in the three RCTs were required to have had experience of LPD, such as 25 LPD ≥ in the PLOT trial, 20 LPD ≥ in the PADULAP trial, and 20 LPD ≥ in the LEOPARD trial, prior to participation in the trials. [27][28][29] In the LPD learning curve, related improvement in surgical outcomes of the operative time and blood loss was seen after 30-50 cases 53,61-65 (Table 2) 29 In the PADULAP trial, the single expert surgeon, who had done 20 LPDs before the start of the trial, performed LPDs, but also had extensive experience in laparoscopic gastric bypass surgery with hand-sewn anastomosis (more than 250 procedures). 28 The necessity for comprehensive and continuous surgical training to achieve patient safety is clear. In Japan, the endoscopic surgical skill qualification system (ESSQS) was es- and III) was reported to be significantly lower than that in the first 100 cases (phase I) (15.1% and 30.0%, respectively). 43  Hospital procedural volume has been reported to be important to decrease postoperative complications. 70

| CON CLUS IONS
Minimally invasive pancreatic resection has spread internationally and has yielded less invasiveness in certain aspects. MIDP for benign and low-grade malignant tumors has established a secure position over ODP in its reduced invasiveness, although the long-

ACK N OWLED G EM ENT
None.