Comparison of surgical outcomes among open, laparoscopic, and robotic pancreatoduodenectomy: a single-center retrospective study

Pancreatoduodenectomy (PD) was first officially reported by Whipple in 1935 [1]. Since then, PD has been progressively acknowledged as a standard procedure for removing benign and malignant tumors from the head of pancreas and the periampullary region. The convoluted procedure involves broad resection and reconstruction of the digestive tract; therefore, it has invariably been identified as one of the most complicated and risky operations.

Laparoscopic pancreatoduodenectomy (LPD)has been developed for approximately 27 years since this technology was reported by Gagner and Pomp for the first time in 1994 [2]. It is still controversial in the academic community because of its initially unacceptable feasibility and security. In recent years, LPD has been shown to provide obvious advantages such as less trauma, less delayed gastric emptying, less transfusion, faster postoperative recovery, and shorter hospital stay compared with conventional open pancreatoduodenectomy (OPD), benefiting from the breakthroughs of laparoscopic instruments and technologies [3‐7]. Nevertheless, because it is complex and time-consuming, LPD is likely to tire surgeons, making it an unappealing treatment option.

Giulianotti et al. reportedrobotic pancreatoduodenectomy (RPD) for the first time in 2003, and it has since played a revolutionary part in the advancement of minimally invasive PD [8]. RPD can dramatically overcome some limitations of LPD by providing excellent ergonomics, instrument joints with 540° of movement, automatic tremor removal, and generation of high-definition of 3D images at 10–15 × magnifications, all of which aid in the delicate and complex procedure and reduce the surgeon’s fatigue [9, 10]. However, RPD has not been adopted universally as a result of the high expense of the instrument, technical challenges, risks of surgery, and oncological outcomes [9‐12].

This study focuses on exploring the safety, feasibility, and justification of RPD by comparing the perioperative outcomes among OPD, LPD, and RPD groups based on PD data from the same pancreatic treatment team at the First Affiliated Hospital of Guangxi Medical University over the same time period.

RPD has been used in removing various benign and malignant tumors from the head of the pancreas and the periampullary region in recent years, but its adoption remains limited to a few large academic centers, so the availability of associated data is extremely poor. There is no general consensus on the feasibility and safety of RPD and whether it increases surgical risks.

In this study, the RPD, LPD, and OPD groups were fairly similar with regard to most baseline information, such as age, sex, BMI, and tumor size. Accordingly, we believe that the comparison of outcomes among the groups is reliable. Of note, no statistically significant difference in operative time was observed between RPD and LPD, which agrees with the result reported by Zimmerman et al. [22]. Moreover, we noticed that OPD required the least operative time and RPD required the most operative time. The high operative time of RPD was caused by technical difficulties such as the installation and adjustment of robotic arms, especially when the assistants and/or scrub nurses were not well trained or unacquainted with the setup. As robotic technology is developed, the operating time of RPD should be remarkably reduced in the future. We found that blood loss was lower in the RPD group than the OPD and LPD groups, but the difference was not statistically significant. This was similar to the results of Shi et al., who demonstrated that RPD was superior to OPD in estimated blood loss (mean: 297.3 ml vs. 415.2 ml), with a statistically significant difference (p = 0.002) [23]. Furthermore, in an American multi-center study, scientists discovered that RPD offered a mean reduction in blood loss of 181 ml compared to OPD [24]. Interestingly, when Orti-Rodriguez et al. compared an LPD group of 147 patients and an RPD group of 284 patients, the estimated blood loss was significantly higher in the RPD group than in the LPD group (median: 346.44 ml vs. 172.93 ml, p < 0.05) [25]. This discrepancy from our results might stem from patient selection bias and a learning curve.

Additionally, a lower incidence of conversion to OPD was found in the RPD group compared with the LPD group (6.5% vs. 38.1%). This has been corroborated by other published series comparing robotic and laparoscopic approaches [26‐28]. More specifically, advantages offered by the robotic approach, including 3D visualization and increased flexibility of the instrument, facilitate the completion of complex resection and digestive tract reconstruction when compared with the laparoscopic approach.

A comparative study between an RPD group with 304 cases and an OPD group with 172 cases by Shyr et al. illustrated that 4.8% of all RPD cases were converted OPD. Remarkably, in the first 100 RPD cases, the conversion rate was roughly 10%, but this figure declined to 7% after 200 cases, and 2.9% after 300 cases, which indicated that the accumulation of experience in surgery could be a dependable predictor of conversion rate [29, 30]. It is widely believed that one of the advantages of robotic and laparoscopic approaches is the reduction of inpatient time [31]. In fact, in a trial by Adam et al., no difference was discovered in hospital stay among 7061 RPD, LPD and OPD patients, including 6078 OPD patients, from the National Cancer Database, which is consistent with our results [32]. The probable reason for this finding in our study is that only a small proportion of the hospital stay is paid for by patients, as most of the stay is covered by medical insurance; hence, patients prefer to remain in the hospital until there is no need to worry about their health issues. We think this could be one of the reasons why the length of hospital stay was higher in our center as compared with most centers in other countries. Additionally, in a recent meta-analysis, Doula et al. [33] reported their findings to be inconclusive as to whether LPD or RPD decreased hospitalization stay. Furthermore, in our study, RPD required less time to remove the abdominal drainage tube than the OPD group (12.5 d vs. 17.3 d, p = 0.001), which may be due to the fact that Da Vinci robotic surgery could decrease inflammatory stress response and peritoneal exudate caused by surgical trauma, as reported by Shibata and his co-workers in 2015 [34]. We also found that there was higher ADL on postoperative day 3 in the RPD group than the LPD group (35.8 vs. 25.7, p = 0.0017). The low conversion rate from RPD to OPD may lead to quicker recovery [35], which could be one of the reasons for this result. In our center, the indications for abdominal drain removal are related to the characteristics and volume of drainage fluid, biochemical examination results (amylase, bilirubin), and bacterial culture results, but we need to be cautious about the removal of the abdominal drainage to prevent the catastrophic consequences of delayed biliary or pancreatic leakage, so the time of the abdominal drainage removal should be decided according to the actual situation.

No significant differences were found in the postoperative complications, such as POPF, PPH, bile leakage, reoperation, and perioperative mortality, among OPD, LPD and RPD. Therefore, RPD appeared not to compromise surgical outcomes, although this retrospective study has limitations due to unavoidable selection bias.

Of note, there are some undeniable limitations in this study. This was a single-center retrospective study with a small sample size, and it lacked randomized controlled trials (RCTs). Additionally, the surgeons in this study were likely still in the early phase of the learning curve of RPD. Further high-quality validation studies are urgently needed.

This study demonstrates that not only is RPD feasible and reliable for the removal of benign and malignant tumors, but it also offers additional benefits in short-term operative outcomes without increasing the surgical risks. In the future, studies involving numerous subjects and large-scale RCTs are required to establish solid conclusions.