Laparoscopic versus open pancreatic resection for ductal adenocarcinoma: separate propensity score matching analyses of distal pancreatectomy and pancreaticoduodenectomy

Pancreatic duct adenocarcinoma (PDAC) is currently the fourth leading cause of cancer-related deaths in developed countries and may rank second by the year 2030 [1, 2]. Surgical resection is considered the only method to radically cure this type of cancer [3]. The surgical extent depends on the tumour location: left-sided PDAC should be treated by distal or subtotal pancreatectomy (DP), and PDAC on the pancreatic head should be seeking pancreaticoduodenectomy (PD). Minimally invasive surgery (MIS), which is characterised by laparoscopic interventions, has become the standard of care for many surgical procedures across different specialities. The selection of MIS is the professional objective of surgeons and the most acceptable treatment for patients [4]. Regarding pancreatectomy, while the safety and effectiveness of laparoscopic DP (LDP) has been gradually evaluated [5], laparoscopic PD (LPD) for PDAC is still in its infancy due to the complexity of the operation and the steep learning curve required for its introduction [6, 7]. The Miami International Evidence-based Guidelines suggested that LPD should be exclusive to experienced surgeons in high-volume centres [8]. Additionally, to date, data on laparoscopic pancreatic resection (LPR) for PDAC and oncological outcomes are limited. The varying preferences between surgeons, difficulties in the intracorporeal hand-sewn technique, and expected cumulative and standardised outcomes are challenges in conducting LPR for PDAC [9]. In other words, the therapeutic role of LPR for PDAC has not yet been established [10, 11]. Considering the different natures of LDP and LPD, it is reasonable to separately analyse LDP and LPD for the treatment of PDAC. We first proposed LDP as early as 2003 [12] and conducted LPD for the treatment of PDAC in 2012 after extensive laparoscopic experience [13, 14]. In this study, we evaluated the safety and effectiveness of LDP and LPD by separately comparing their short- and long-term clinical outcomes with those of open DP (ODP) and PD (OPD).

The use of LDP for PDAC is still debatable, although LDP provides a significant alternative for benign or low-grade tumour on pancreatic bodies and tails [5, 8]. Two surveys suggested that 19–31% of surgeons expected LDP to be inferior to ODP in PDAC treatment [26, 27]. On the other hand, approaching PDs laparoscopically for diseases on pancreatic heads was less frequent owing to the intricacy of the dissection and the complexity of the pancreatoenteric and biliodigestive anastomoses [6, 10]. Moreover, LPD for PDAC is still in its infancy because of concerns about the safety and oncological efficacy [10]. Therefore, it is necessary to separately evaluate the efficacy of LDP and LPD for PDAC treatment. This study suggests that both the procedures are technically feasible and safe, and both consistently exhibit clear benefits (less blood loss and short hospital stay). LDP and LPD have no obvious advantages in decreasing postoperative morbidity. LDP appears advantageous for the retrieval of lymph nodes. Additionally, long-term survival outcomes were closely matched between laparoscopic and open surgery.

The feasibility of LPD was the major concern in the adoption of this surgical technique, as a high conversion rate was reported, especially in the initial period. In ITT analysis, the conversion rates were only 3.5% (3/86) for LDP cases but 10 (9.9%) conversions for LPD cases. Two recently published randomised clinical trials (RCTs) reported that the conversion rates for LPD were more than 20% [28, 29]. A high conversion rate indicates that LPD for PDAC remains a challenging procedure. The most common reasons for conversions were uncontrollable haemorrhages or suspected vessel involvement, which was similar to other publications on LPD for PDAC treatment [30, 31]. Nickel et al. reviewed six studies focusing on the learning curve of LPD and revealed that the volume to reach a technical competency ranged from 10 to 60 cases depending on the surgeon’s expertise [32]. Our conversion cases were also mainly in the initial period, which may partially explain the high conversion rate of the RCTs, although surgeons had an experience of handling at least 10 cases of LPD before anticipating the RCTs. In our centre, the surgical team was competent in performing advanced procedures, including reconstruction of the gastrointestinal tract, intracorporeal suture, and emergency haemostasis. We believe that training together in a relatively constant surgical team contributes to better surgical outcomes, especially in emergencies during LPD, which may turn into conversion. Interestingly, conversion was found to be associated with higher patient BMI. The advantage of less blood loss in LPD was more obvious when data on conversion were omitted, and the difference in overall morbidity between LPD and OPD was statistically significant (17/91 [18.7%] vs. 32/101 [31.7%], p = 0.04). We believe that part of the reason for high conversion is that PDAC frequently induces substantial pancreatic inflammation in the pancreatic remnant, which is difficult to resect due to pronounced adhesions or infiltrations to the surrounding tissues or vessels. Portomesenteric vein (PV) involvement is a common clinical finding in PDACs, but it is difficult to diagnose prior to surgery [33]. However, researchers strongly recommend that the PV should be resected, once detected as contributing to a tumour by surgeons with considerable proficiency in vascular resection and reconstruction [34, 35]. As for LPD, approaching appropriate cases without vessel involvement or severe adhesions laparoscopically, avoiding obese or overweight patients in the learning curve would be helpful in reducing conversion [28].

According to our data, LDP had a reduced time of operation, but an extended operative time in LPD in comparison with their open counterparts. The former is largely attributed to the simplicity of LDP, and fast management of the trocar incision could reduce the time required for laparoscopic resection. Comparatively, LPD poses a tougher challenge for surgeons because it involves not only complex dissection but also complex gastrointestinal, pancreatic, and biliary anastomoses, all of which present a technical challenge. Our initial LPD for PDAC lasted for nearly 600 min [16]. Currently, this can be completed in approximately 300–350 min [15]. Kendrick et al., in one of the largest single series studies, described initial LPD duration to be 460 min, which improved to 320 min after about 50 cases [36]. Stauffer et al. suggested an average operation time of 518 min, which was clearly more than that required for an open surgery (140 min) [31]. The learning curve can be overcome in high-volume centres, with average LPD operative times reduced to lower than 400 min [37]. Nevertheless because of tumour biology and the progression of the disease, LPD for PDAC treatments is not regarded as a common option; as a result, it is challenging to address the related learning curve [6]. A longer operation time was related to the worse perioperative outcomes following pancreatic resections, according to a research conducted under the American College of Surgeons National Surgical Quality Improvement Program [38]. Therefore, we believe that a long duration is a definite disadvantage of LPD for PDAC treatment.

We found that both LDP and LPD showed a trend of less overall morbidity without statistical significance in contrast to ODP and OPD (DP: 9.3% vs. 16.3%, p = 0.17; PD: 21.8% vs. 31.7%, p = 0.11). POPFs are commonly viewed as the most ominous of complications after pancreatic resection, and the most effective management for the pancreatic stump is still under debate, although various surgical procedures of pancreatic stump management after DP and anastomosis after PD have been devised to prevent POPFs in conventional open surgery [11, 39]. As laparoscopic surgical instruments are developed and the operative experience is accumulated, it is possible to do open surgery performed laparoscopically using the same reported methods [6, 11]. DGE is not life threatening, but can have significant consequences such as patient discomfort, prolonged hospital stays, diminished nutritional status, and delays in the initiation of adjuvant therapy [40, 41]. The pathogenesis of DGE is multifactorial. Given improved access and visualisation, as well as the meticulous attention, a laparoscopic approach could theoretically reduce DGE because of the following reasons [42, 43]: 1) it can mitigate the impact on the organs and peritoneum, leading to less seroperitoneum which helps alleviate gastric dysrhythmias, 2) it can ameliorate pyloric or antral ischaemia as a result of reserved small vessels, and 3) it can mitigate pylorospasms secondary to denervation of the stomach and duodenum or jejunum. Additionally, open procedures are reported to portend a higher risk of pleural effusions, pulmonary infections, and atelectasis than minimally invasive ones [44, 45].

Before the widespread application of a new surgical approach, oncologic safety and effectiveness should be verified. The long-term survival outcomes of MISs for common malignancies have conflicting results [46‐48], leading to a constant controversy over MIS for cancer treatments. Nassour et al. used the NCDB database to compare the long-term oncologic outcomes of LPD and LDP to open surgery in patients with PDAC, and found that MIS was associated with similar long-term survival for PD, and improved survival for DP [49]. Our study revealed that R0 resection of LDP and LPD are similar to open surgery. It is worth noting that preoperative serum levels of CA 19–9 predict resectability and survival [50]. Patients with CA19–9 levels > 4000 U/mL had a resection rate of 38% [50]. In the present study, a majority of patients had mildly elevated CA199 levels, except those combined with diseases of the biliary tract, while preoperative examination showed no signs of metastasis. Additionally, our data showed that LNs retrieved in LPD patients were not inferior to those of OPD (22.6 ± 6.5 vs. 21.0 ± 6.2, p = 0.07), and the LNs harvest of LDP was superior to ODP (14.4 ± 5.2 vs. 12.7 ± 5.0, p = 0.03). Noticeably, the DIPLOMA research revealed that LDP was linked to a higher R0 resection rate (67% vs. 58%) and a smaller number of LNs (14 vs.22) [51]. However, this pan-European PSM study discovered that lower LN retrieval with LDP does not make a noticeable difference to the average OS.(28 vs. 31 months) [51]. In general, it was revealed by studies, including meta-analyses, that the long-term outcomes of LDP for PDAC are promising [52‐55]. As for LPD, a single-centre study conducted by Asbun and Stauffer reported similar long-term survival rates at 1, 2, 3, 4, and 5 years for OPD (68, 40, 24, 17, and 15%) and LPD (67, 43, 43, 38, and 32%), respectively [31]. Kuesters et al. conducted a series of LPD procedures and reported a similar 5-year survival rate between LPD (20%) and OPD (14%) for PDACs [56]. Some publications have demonstrated that LPD has a positive effect on long-term oncologic outcomes in patients with PDAC [30, 57]. It was hypothesised that the enhanced recovery following laparoscopic surgery was conducive to activating multimodality therapies in advance, thus improving survival [30]. Nevertheless, according to a retrospective analysis of the NCDB, MIS failed to improve the use or initiation of adjuvant chemotherapy for patients with PDAC [58]. Moreover, the survival effect of the activation time of adjuvant chemotherapy in patients with resected PDAC remains unknown, as studies have indicated contradictory outcomes [59, 60]. A technically similar oncologic resection is worth performing, irrespective of the open or laparoscopic approach if the principles of radical resection are complied with. Accordingly, Lee et al. compared resected PDAC in a study, which included both laparoscopic and open cases, and found that according to the Yonsei criteria (a preoperative CT-based determined method) can predict excellent short-term and long-term oncologic outcomes [61]. In other words, if surgery follows the oncologic principle, the oncologic impact is not influenced by differences in the surgical approach [61].

The limitations of this study include its retrospective design, small sample size, the absence of randomisation, and a short follow-up period. Although PSM was performed to balance the covariates, thus reducing selection bias, other factors cannot be ignored. In the PD arm, the follow-up period of LPD was shorter than that of OPD since LPD was initially conducted in late 2012. Additionally, ODP has been largely performed in early years, in contrast to LDP. This can give rise to bias, taking postoperative management as an example. Recently, surgeons are unavoidably influenced by the concept of enhanced recovery after surgery, which includes early mobilisation, oral feeding, midthoracic epidural analgesia, and premature removal of abdominal drain; it can reduce the length of hospitalisation and recovery in postoperative management [62], causing bias in favour of the LDP and LPD group. The limited follow-up period of LDP and LPD is insufficient to provide enough information on long-term outcomes. Additionally, the sample size hinders the effort to arrive at reliable conclusions, especially regarding several variables distinct between the groups, but with no significance revealed. Considering a standard approach for borderline resectable or locally advanced PDAC [63], neoadjuvant chemotherapy for resectable PDAC was applied in some high-volume centres but was not applied in our centre before 2017, so oncological outcomes of pancreatectomy for PDAC patients after neoadjuvant chemotherapy still need further analysis.

Laparoscopic pancreatectomy was safe and effective for the treatment of PDAC; LDP and LPD were associated with less blood loss and shorter hospital stay. However, the short-term surgical advantage of LPD is not as obvious as those of LDP, mainly due to the surgical conversions. The oncological outcomes of LDP and LPD were not inferior to those of traditional open procedures for the treatment of PDAC. Future studies may consider longer follow-up periods and larger patient samples to validate our findings.