What role does hand-assistance have in minimally invasive pancreatic surgery?

© The Author(s) 2021. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, sharing, adaptation, distribution and reproduction in any medium or format, for any purpose, even commercially, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.

operative field visualization with 3D imaging. Nonetheless these advantages have to be balanced with drawbacks such as lack of haptics and high cost in terms of both initial investment in purchasing the robotic platform and single operation cost.

MIS IN THE PANCREATIC SURGERY FIELD: A STEEP PATH
Despite all the hype around these new technologies, the implementation and diffusion of MIS have been hampered by the large amount of time and dedication necessary to master the techniques to have results comparable to the open approach. The concept of learning curve became particularly popular with the advent of minimally invasive surgery, when surgeons needed to completely rethink their abilities and adapt them to new techniques and technologies. It has also been postulated that the learning curve appears to be longer in MIS relative to open surgery, and that the curve becomes steeper and steeper with the increasing complexity of surgical procedures [5] . For complex major abdominal surgeries, a great number of procedures is required to master the technique, and there may be dangerously high morbidity and mortality rates at the beginning of the learning curve. This has been particularly the case of pancreatic surgery.
Despite the appeal of MIS and its widespread adoption in several fields of surgery, the attitude of pancreatic surgeons has been initially tepid. On the one hand, there was the conceptual problem of whether in complex and demanding surgical operations such as pancreatic resections the size of the incision can truly be considered the main contributor to surgical trauma. On the other hand, some peculiar aspects of pancreatic surgery have initially hampered the widespread diffusion of the minimally invasive approach in this field: the peculiar retroperitoneal location of the pancreas, its delicate texture and proximity to major vessels, the complexity of the dissection, the concerns regarding oncological safety in the case of malignancy, the difficulty of the anastomotic components, and the still relatively high morbidity and mortality that characterize pancreatic resections [6][7][8][9][10] . Another more practical matter is the relative rarity of pancreatic diseases and the complexity of most cases, which make them not suitable to be approached minimally invasively by surgeons at the beginning of their learning curve; the result is an even longer time to reach proficiency and an acceptable morbidity and mortality rate [11] .
Reports of the initial experience with totally laparoscopic pancreatic surgery showed no apparent advantage for pancreaticoduodenectomy, with no improvement in postoperative outcomes and increased morbidity. Conversely, the results are promising for distal pancreatectomy, since it was associated with acceptable operative time and reduced morbidity and length of stay (LOS) [12] .

Rationale and limitations
To overcome the difficulties in adaptation of complex procedures from an open approach, some hybrid techniques have been developed for laparoscopy.
One of the proposed approaches is hand-assisted laparoscopic surgery (HALS): a mini-laparotomy is planned through which the surgeon can insert his or her hand covered by a glove or a hand port that prevents the loss of the pneumoperitoneum. This allows for the surgical operation to be performed via laparoscopy but with the help of an intra-abdominal hand. At the beginning, this technique was greeted with skepticism because of the need to perform a laparotomic incision, which is in direct contrast with the principle of minimal invasivity and because of the lack of adequate instruments able to maintain the pneumoperitoneum with an intra-abdominally inserted hand [13] . However, with the development of appropriate instruments, HALS found its niche in enabling the surgeon to start approaching major abdominal operations in laparoscopy, with as safety net the familiarity and the expertise of having a hand directly in contact with the structures. Clear pros of this approach are restoration of the tactile feedback and better manipulation of tissues, such as better organ retraction, finger blunt dissection, exposure and control of possible unexpected intraoperative bleeding and complications [14,15] , and a shorter operative time than laparoscopy [16] , while maintaining some of the advantages of MIS over the open approach, notably a lower estimated blood loss and a shorter LOS. Among the cons, there is clearly the additional surgical trauma posed by the mini-laparotomy, although this problem may be partially mitigated by using this technique in operations which would already require an incision to retrieve the resected specimen. Moreover, despite the handiness of having a direct access to the abdominal cavity, the presence of the hand may reduce the space and range of movements of laparoscopic instrumentation and impair vision [13] .

Fields of use
After its introduction, this technique was initially adopted in several different fields of surgery, in which a pure laparoscopic approach was still striving to be undertaken. In esophagogastric surgery, HALS was applied to both trans-hiatal esophagectomy and total and partial gastrectomy with good results in terms of postoperative and oncological outcomes [17][18][19] . A trial was also made in bariatric surgery, but no advantages were found over the open approach for gastric bypass in terms of incidence of incisional hernia and reduction of LOS despite an increased cost [20] . One of the areas in which HALS has had greater success is colorectal, surgery in which an incision is needed anyway, no matter the approach, to extract the specimen and possibly perform the anastomosis. HALS has been used for partial or total colectomy, anterior rectum resection, and abdominoperineal resection, and it maintains the advantages of laparoscopy in terms of bowel movements, refeeding, and hospital stay [21][22][23][24] . Another application of HALS was in the living-donor nephrectomy, where it showed a shorter warm ischemic time than pure laparoscopy, while offering a smaller incision and faster recovery than the open approach [25][26][27][28] . From initial reports, HALS appeared to facilitate the laparoscopic approach, increasing the level of subjective safety and thus shortening the learning curve.

HALS in the pancreatic surgery field
In pancreatic surgery, preliminary data were presented by Cuschieri [29] and Gagner and Gentileschi [30] , in the early era of pancreatic laparoscopy, presenting the advantages of the hand-assisted technique over the totally laparoscopic approach for such major procedures in terms of safety, exposure, and oncological appropriateness. Furthermore, HALS can provide particular advantages in the case of malignancy, allowing for palpation of the tumor and manual staging, and in the case of voluminous cystic lesions, which can be more effectively removed en-bloc [31][32][33][34][35] .
The hand-assisted pancreatic resections were performed with the insertion of trocars along with a subcostal mini-laparotomy, through which the non-dominant hand was inserted to provide traction and direct palpation, while the demolition and reconstruction phase were both accomplished via laparoscopic instrumentation by the dominant hand. In the case of Pancreaticoduodenectomy (PD), all three anastomosis were performed intracorporeally [30] , which is also because mini-laparotomy is usually located in a position not favorable to be exploited for an open pancreatic anastomosis [36] .
The HALS approach was mostly used to perform Distal Pancreatectomy (DP) because it is a relatively easier procedure without need for complex anastomosis and therefore a greater effort has been put in trying to make this procedure as less invasive as possible. Initial experience with totally laparoscopic DP has been encouraging, stating a marked reduction of LOS, but, at the same time, relevant limitations were identified, such as a long operative time and a high conversion rate [37,38] . At the beginning, trying to transition from a purely open approach to a totally minimally-invasive procedure, HALS appeared to be a good compromise, and several reports have been published stating its advantages [39,40] . Postlewait et al. [41] reported a lower intraoperative blood loss and shorter hospital stay than open surgery and comparable perioperative and oncological outcomes. Gamboa et al. [16] showed similar results and additionally reported a shorter operative time than totally minimally invasive approach, a similar LOS, and a lower conversion rate, even though patients undergoing hand-assisted distal pancreatectomy (HADP) had more comorbidities and a higher number of previous abdominal operations. Kneuertz et al. [42] reported the outcomes of laparoscopic DP (LDP) at their institution over an 11-year period; a reduced use of hand-assistance was observed with growing experience and a reduced LOS in TLS relative to HALS. A similar trend in reduction of HALS use over time was reported by Jayaraman et al. [43] and Nakamura et al. [44] . A relevant piece of literature includes HADP in the laparoscopic cases, and it is therefore difficult to extrapolate data on specific HADP outcomes [45][46][47][48][49][50][51][52] . The current available literature on the topic is summarized in Table 1; articles where the surgical technique is not specified were excluded. Placement of trocars and hand-port is shown in Figure 1.
Some reports have postulated a non-inferiority of the hand-assisted approach for PD relative to open, but its usefulness has been questioned [29,30,[62][63][64] . In PD, the advantage of hand assistance does not appear to be striking. This is probably ascribable to the fact that the complex reconstruction phase, in HAPD, is performed intracorporeally, and, if a surgeon has enough laparoscopic skills to perform the reconstructive part, he conceptually should not need the help of the hand in the demolition phase [36] . Accordingly, recent literature reports a very limited adoption (0.6%) of the hand-assisted approach for PD [36] . Some hybrid approaches have been proposed, with the demolition phase performed with a hand-assisted approach and the reconstruction phase with an open approach via a mini-laparotomy [65] .

LAPAROSCOPIC-ASSISTED SURGERY
A similar but somewhat different hybrid approach that appeared to be more suitable for PD is laparoscopicassisted surgery (LAS). In LAS, the preparation and part of the demolition phase of the surgical operation is managed via laparoscopy, while the reconstruction part is performed out of the body via a small laparotomic incision [66] . With this approach, we are able to take advantage of the improved vision of secluded spaces given by the laparoscopy, sparing a large incision to the patient and granting a faster postoperative recovery, while assuring an adequate anastomosis technique and hemostasis through a small incision that can also be used for the retrieval of the resected specimen [67,68] . Several authors, in the initial phase of approaching minimally invasive PD, used a laparoscopic-assisted PD (LAPD) approach and reported their case series, proposing the feasibility of LAPD [69][70][71][72][73][74] . LAPD showed non-inferior results to open surgery in terms of perioperative and oncological outcomes (comparable number of harvested lymph nodes and higher R0 rates) [75] . Similar results were also reported by Tan et al. [76] and Mendoza et al. [77] , who showed no differences in oncological and perioperative outcomes between open PD and LAPD. Tian et al. [68] reported a lower estimated blood loss and shorter time to first flatus and Wang et al. [67] described again a lower intraoperative blood loss and a shorter LOS. Additionally, a lower rate of anastomosis related complications has been reported compared to totally laparoscopic PD performed by experienced pancreatic surgeons at the beginning of their learning curve [78] . Similarly promising results were reported by Deichmann et al. [79] . No differences in intraoperative characteristics and postoperative outcomes were found between LAPD and robotic-assisted PD by Piedimonte et al. [80] . Patel et al. [81] reported a shorter LOS and lower severe morbidity rate and reoperation rate in LAPD compared to TLS, although a progressive shift from LAPD to TLS was observed over time. Somewhat similar results were published by Wang et al. [82] , reporting an increased operative time and blood loss in LAPD relative to TLS but similar LOS, morbidity rate, and postoperative pancreatic fistula (POPF) rate, with LAPD adopted by more inexperienced surgeons. In addition, Goh et al. [83] reported a more frequent adoption of the hybrid technique during their early experience to allow for a safer transition to totally MIS. van Hilst et al. [84] compared postoperative outcomes in LAPD and TLS without finding any significant difference; similar results were reported by The literature search was conducted on the PubMed database. The search terms used were "laparoscopy" OR "hand-assisted" AND "pancreatic resection" OR "distal pancreatectomy" OR "pancreatectomy" individually or in combination. A manual search of reference lists of included articles was conducted. Case reports were excluded from the  Dulucq et al. [85] . Speicher et al. [86] tracked the evolution of PD procedure over time at their institution, observing a progressive increase in the use of TLS over LAPD with growing experience and a parallel decrease of OT and complication rate; analogous findings were reported by Kim et al. [87] and Lu et al. [88] . The literature appears rather inhomogeneous, and it is difficult to draw definitive conclusions; however, in light of the reported data, the hybrid method appears to be safe and not inferior to the open approach [69,70,89,90] . It also seems to provide some advantages over TLS in the early phase of the learning curve, but this may lose relevance in the case of surgeons with extensive experience in laparoscopy. A relevant piece of literature includes LAPD in the laparoscopic cases, and it is therefore difficult to extrapolate data on specific LAPD outcomes. The current available literature on the topic is summarized in Table 2; articles where the surgical technique is not specified were excluded. Placement of trocars and mini-laparotomy is shown in Figure 2.

2019
PD LAS 21 -- The literature search was conducted on the PubMed database; the search terms used were "laparoscopy" OR "laparoscopic-assisted" OR "hybrid" OR "hand-assisted" AND "pancreatic resection" OR "pancreaticoduodenectomy" OR "Whipple" or "pancreatectomy" individually or in combination. A manual search of reference lists of included articles was conducted. Case reports were excluded from the table.
Articles with no full text available were excluded [97] .  MIDP for the treatment of pancreatic ductal adenocarcinoma, available data suggest the oncological appropriateness of the procedure, but high-level evidence is still lacking. Oncological outcomes were comparable in terms of resection margins, disease free survival, and overall survival, while the number of harvested lymph nodes was found to be lower in one metanalysis and comparable in a second one [101,109,110] . The DIPLOMA trial [111] showed a higher R0 resection rate for MIDP, a less frequent Gerota's fascia resection, a lower number of harvested lymph nodes, and a comparable median survival. Randomized clinical trials are ongoing, trying to give a definitive answer. Regarding the choice of the type of MIS technique, several observational studies have been published comparing the robotic versus laparoscopic approach. Theoretically, the robotic platform should provide advantages in terms of improved dexterity and vision, allowing for completion of more complex procedures, but whether this translates into better outcomes and cost-effectiveness in clinical practice is still controversial [112] . Reported outcomes in the literature are heterogenous: recent metanalyses showed a higher rate of splenic vessel preservation and a lower conversion rate, but higher cost in Robotic DP compared to LDP [113,114] . Another metanalysis reported a shorter LOS and an increase of spleen preservation rate at the expense of increased cost [115] . Oncological and postoperative outcomes, such as POPF rate and overall morbidity, were comparable. Other studies showed no major differences in perioperative outcomes [116][117][118] . Therefore, the Miami Guidelines conclude that both laparoscopic and robotic DP are considered valuable and equivalent options, and the choice between the two depends on the preference of the surgeon and his familiarity with the technique [108] .

Pancreaticoduodenectomy
Pancreaticoduodenectomy is still performed in the majority of centers with an open approach due to its technical difficulty and the complex reconstructive phase. Available data on safety and feasibility of MIPD are conflicting. Reports from low-volume centers showed an increased morbidity and mortality after MIPD [119,120] , while experience in high-volume centers demonstrated a similar rate of mortality and morbidity compared to OPD. Moreover, in high-volume centers, LPD showed a lower rate of DGE, decreased blood loss, and a shorter hospital stay but a longer operative time [36,121,122] . Three randomized clinical trials have been published with mixed results. Palanivelu et al. [123] showed similar oncological and perioperative outcomes in OPD and LPD. Conversely, the LEOPARD-2 trial was interrupted early because of safety concerns due to a disproportionally high number of deaths in the LPD arm [124] , while the PADULAP trial reported a lower major complication rate and a shorter LOS and similar oncological outcomes [125] . No major differences in outcomes have been reported between LPD and RPD [126,127] . In view of existing evidence, the Miami Guidelines concluded that insufficient data exist to recommend MIPD over OPD. MIPD appears to be safe and feasible but only if performed by surgeons who have completed the learning curve and if set in high-volume centers experienced in both pancreatic surgery and MIS.

HALS: DOES IT STILL HAVE A ROLE IN PANCREATIC SURGERY PRACTICE TODAY?
Analysis of trends in the use of MIS in pancreatic surgery showed how, with time, we had a steep increase of MIDP, and the increase in number was paralleled by increasing complexity of procedures and a decrease in conversion rate and operative time [42] . Moreover, the proportion of procedures performed with hand assistance decreased with time as surgeons became more skilled in MIS. It is worth noting that a recent analysis showed that MIDP is only used in one third of eligible patients [128] . Therefore, on the one hand, HADP plays a very marginal role in high-volume centers, where surgeons have finished their learning curve, while, on the other hand, there are still centers in the process of implementation of MIS where HADP may play a fundamental role as a bridge to totally MIDP, easing the transition and shortening the learning curve. Moreover, HADP, with its shorter operative time, may be preferred in patients with multiple cardiological, pulmonary, and renal comorbidities who would not tolerate well the effects of prolonged anesthesia and pneumoperitoneum [16] . Furthermore, HADP may be used as an intermediary step in conversion from MIDP to open in complex cases where manual assistance or tactile feedback is required or in the case of intraoperative complications because it appears that converted hand-assisted cases have a lower estimated blood loss and a shorter LOS than open [16,30] .
The role of MIS in PD is still not defined; MIPD can be performed in high-volume centers by experienced surgeons with acceptable outcomes, but the results are difficult to be generalized. In the process of the implementation of MIPD, LAPD may play a role as a bridge to totally laparoscopic PD allowing for a safer transition [129,130] .
In conclusion, the choice of the right approach needs to be tailored to the patient with a focus on his or her safety and to the surgeon keeping in mind his or her limits and expertise.

Authors' contribution
Conception, design, drafting and revision of the manuscript: Donisi G, Zerbi A