Our institutional experience with MIPD began nearly 2 decades following the first reported LPD in 1994 (6). Unlike other forms of minimally invasive surgery (MIS), the trajectory of MIPD took on a slow inception due its inherent technical complexity, particularly in the reconstructive phase. Intuitively, the natural evolution of MIPD over time would suggest a tendancy towards improvement in certain post-operative outcomes. While our study identified a reduction in grade-C POPF over time, there was an increase in the number of clinically relevant grade-B POPF. This may likely be attributed to the increasing frequency of patients over time undergoing MIPD for non-PDAC pathology or with intermediate/high original fistula risk scores. Furthermore, the vast majority of grade-B POPF in this series were a result of prolonged drainage as opposed to a consequence of intervention. A study by Maggino et al proposed subclassifying grade-B POPF into 3 distinct groups based on the possible number of therapeutic combinations, with patients requiring interventional procedures having poorer outcomes compared to those receiving isolated prolonged drainage (19). This level of granularity permits for more precise reporting given the influence each sub-classification has on the post-operative course. While exact data regarding each subclassification was not available at the time of this study, a review of all major post-operative complications revealed that only 25% (n = 6) of all grade B-POPF in groups 2 and 3 were a result of interventional-radiology guided drainage, with the remaining a consequence of prolonged drainage.
Performing a secure pancreato-enteric anastomosis is perhaps the Achilles heel of most surgeons endeavouring in MIPD. Numerous techniques and approaches for reconstruction have been described (15, 20, 21) but none have proven to be superior. The use of hybrid assistance with open reconstruction in the early period represents an important transition in MIPD, allowing surgeons to acquaint themselves with laparoscopic dissection while maintaining the use of an open anastomotic technique which they have been familiar with (22). Subsequently, the use of laparoscopic or robotic approaches to complete the anastomosis can be attained once the initial hurdles of MIS dissection have been overcome. Presently, there is scarcity of data comparing the operative approaches to the pancreato-enteric anastomosis and their eventual impact on post-operative outcomes. Several key findings may be gleaned from our current study; firstly, while there were no differences in mortality, major morbidity and POPF rates when comparing open vs MIS pancreatic anastomoses, we identified an increase in blood loss and intraoperative blood transfusions in the open pancreatojejunostomy (PJ) group. Given that the dissection phase for all approaches were performed via an MIS approach, it is peculiar that open reconstruction alone would result in significant blood loss. One possibility lies in the underlying pathology for which the PD was performed for: the higher number of PDAC operated on in the open reconstruction arm likely contributed to difficult dissection given the prevailing desmoplasia. Consequently, this increases the bleeding and transfusion risk. Secondly, when comparing laparoscopic and robotic approaches to the PJ, we identified an increase in blood loss and pseudoaneurysms in the laparoscopic arm without any differences in clinically significant POPF or mortality. Some of the reasons for this observation relate to the advantages of the robotic platform: greater stability, precision in instrument handling, comfortable ergonomics and elimination of operator tremors allow for better tissue manipulation, dissection, and intra-corporeal suturing (23, 24). In fact, while it is presumed that pseudoaneurysms occur in response to a POPF, iatrogenic injury sustained to vessels during improper tissue manipulation or vascular skeletonization can occur in the absence of overt POPF and may contribute to subsequent pseudoaneurysm formation (25). To date, there have been several studies comparing RPD and LPD. Liu et al found that when compared to the laparoscopic group, the robotic group exhibited reduced blood loss, operative time, and shorter hospital stays (26). More recently, Khachfe et al demonstrated reduced complications and improved optimal outcomes with robotic over laparoscopic adoption for MIPD (27). Similarly, Ricci et al. also concluded in their systematic review and network meta-analysis that the safest MIS PD techniques to improve outcomes involve the usage of a robotic assistance (28). In fact, after the LEOPARD-2 trial (29), MIPD embarked on a shift from LPD to RPD in the Netherlands where robotic training in the form of a robust curriculum involving centres with sufficient PD volume demonstrated outcomes similar to our present study (30). Additionally, the learning curves for RPD were considerably shorter in these training units than those initially from “pioneering” centres, highlighting the natural evolution towards better outcomes and shorter training periods as the collective experience in RPD accumulates (31). However, despite its advantages, widespread adoption of the robot remains hampered by its cost effectiveness and lack of strong evidence proving its superiority over conventional laparoscopy or even open surgery (7). The need for high quality randomized trials, coupled with the entry of cheaper and improved robotic systems may potentially improve the presence of robotic surgery globally.
There are several limitations in this study. Firstly, the retrospective nature of this study renders it susceptible to issues of bias and confounding. In lieu of randomized controlled trials which remain as gold standard, we attempted the use of propensity score matching to create a methodologically rigorous alternative in order to minimize the effect of known confounders. Secondly, the effect of multiple surgeons, each with different levels of MIS experience, could potentially implicate the degree of post-operative outcomes such as POPF. This is particularly so given that several surgeons only began MIPD during periods 2 or 3. The resultant heterogeneity contributed by varying individuals with different MIS experience may therefore impact the external validity of our findings, especially so given the influence of surgeon experience on outcomes following MIPD (7, 32).
In conclusion, our initial experience demonstrates that MIPD may be embarked on safely for the management of peri-ampullary tumours. Comparison with international benchmark outcomes for open PD demonstrated that the although MIPD during the initial experience was associated with a higher-Grade B POPF rate; the postoperative morbidity, major morbidity, Grade C POPF and in-hospital mortality were within international benchmark cutoffs. Robotic reconstruction of the pancreatoenteric anastomoses may reduce the rate of arterial pseudoaneurysm compared to laparoscopy.