Definitions

LPD: LPD involved the entire procedure, resection, and reconstruction performed laparoscopically except for feeding jejunostomy, which was conducted following the extraction of the specimen through 5 cm mini-laparotomy through the umbilical port.

Conversion: Conversion from laparoscopic to open surgery was defined as any unplanned switch from laparoscopy to laparotomy for reasons other than specimen extraction.

Hybrid PD: A planned conversion where the entire resection was performed laparoscopically and the reconstruction was conducted through an approximately 8 cm mini-laparotomy

Morbidity: Procedure or non-procedure related complications requiring medical or surgical intervention.

Perioperative mortality: All deaths within 30 days of surgery or during the same admission, irrespective of the cause.

Postoperative pancreatic fistula: Postoperative pancreatic fistula (POPF) was defined in accordance with the definition of the International Study Group of Pancreatic Fistula [15].

Delayed gastric emptying: Delayed gastric emptying was defined based on the guidelines of the International Study Group for Pancreatic Surgery (ISGPS) [16].

Postpancreatectomy hemorrhage: This involved grade B or C post pancreatectomy hemorrhage as proposed by the ISGPS [17].

Results

During the study period, LPD was performed in 13 patients. Three cases had to be converted to open surgery: one due to bleeding during pancreatic stump mobilization for PG, one due to bleeding during uncinate dissection in the second case, and one due to unsatisfactory PG, which was converted and a redo PG was performed. All conversions were completed through minilaparotomy. Finally, only 10 patients underwent LPD.

Demographic and preoperative data

Among the 10 patients who successfully completed LPD, 6 were men; 6 and 4 had ASA grade II and III, respectively. The median age of the patients was 54 (range: 39–70) years, and the median BMI was 23.6 (range: 19.6–27.2) kg/m². Four patients underwent PBD: 1 for cholangitis, 2 for bilirubin levels >20 mg/dL, and 1 because the patient deferred his surgery and a short metallic stent was placed in the waiting period. All the patients were operated for malignancy.

Intraoperative and postoperative data

A major proportion of the patients had a soft pancreatic texture with a small duct (90%), and PG was performed in all these patients except for one patient with a firm pancreas and a dilated duct in whom PJ was performed. Table 1 lists the details of intraoperative and postoperative outcomes.

Table 1: Intraoperative and postoperative data.

The overall complication rate was 40% with most of the events occurring in one patient with grade C POPF and was the only mortality in the series. A total of 2 reoperation were required: one due to a GJ leak that was managed laparoscopically and one due to PPH that was managed through the laparotomy suture ligation of the GDA stump.

Histopathology and early oncological outcomes

All the patients had malignancy. A total of 100% R0 resections were noted, and a satisfactory lymph node yield was achieved in 90% of the patients (Table 2).

Table 2: Histopathology and short-term oncological outcomes

Discussion

The first LPD was reported by Gagner and Pomp in 1994 [18]. Gagner reported their first series of 10 cases in 1997 in which they questioned the need for performing PD laparoscopically because of its high complication rate, long operative time, and no clinical benefits [19]. A literature review published in 2009 indicated that only 146 cases of LPD had been reported in the 15 years since the first case [20]. Subsequently, a large clinical series published by the Palanivelu group in 2007, followed by another one by the Kendricks group, have demonstrated satisfactory outcomes [21,22]. This led to increased interest in LPD in the last decade globally, and more than 20 studies compared LPD with open PD.

The goal of the application of laparoscopy to PD is not to reduce pancreas-specific complications. By contrast, the aim is to extend the benefits of MAS to PD in terms of blood loss, transfusion requirements, pain scores, intensive care unit and hospital stay, wound-related (SSI, wound dehiscence, and ventral hernia in the long term) and pulmonary complications, and functional recovery while demonstrating a non-inferiority in terms of pancreas-specific complications and oncological outcomes.

Palanivelu et al. (PLOT trial) and Poves et al. (PDAULAP trial) have reported a significantly shorter hospital stay (7 vs 13 days and 13.5 vs 17 days, respectively) [11,12]. The PADULAP trial indicated lower overall morbidity rates in the LPD arm and similar pancreas-specific complications and oncological outcomes in both the arms. The PLOT trial reported significantly lower blood loss in the LPD arm and similar outcomes in terms of complications and oncological safety in both the arms. The third RCT, the LEOPARD-2/3 trial [13], was prematurely stopped due to a higher mortality rate in the LPD arm (10% vs 2%). Although this trial had a robust methodology, it had flaws, especially with a 20% conversion rate and 22% of surgeries being graded below average in a video review by a panel of experts. Furthermore, two of the five deaths reported in the LPD arm were due to bowel ischemia following injury to the superior mesenteric vessels. The same group reported a mortality of 3.5% in 114 LPDs done prior to the start of this trial. A meta-analysis of these RCTs [10] indicated that LPD resulted in lower blood loss and comparable outcomes in terms of complications and hospital stay although the surgery duration was longer.

Conroy et al examined data from the National Cancer Database and reported that centers performing <20 LPD/year should be considered a LVC [23]. They observed that 82% of LPDs were performed at LVCs and higher mortality rates at such centres. In this series from a LVC, we noted that LPD is feasible and can be safely performed with diligent planning and audit (Table 3) compares our outcomes with those of studies conducted in high-volume centers reporting their experience of respective initial LPD case series. Most of these studies did not report their path to start LPD and do not indicate whether their initial cases were included in the analysis. We have included all patients who underwent LPD. Our conversion rate was 23%, and the most common reason for conversion was bleeding, which is also the common reason reported in the literature. In these patients, we completed surgery by making a small extension of the 5 cm incision that was required to extract the specimen in LPD.

Table 3: Initial LPD series (Minimum 10 completed LPD in the series).

Some institutions have a well-designed program for progression to MAS in PD. One such program is the one being followed at University of Pittsburgh Medical Centre (UPMC) for their robotic PD that is based on a five-step process involving simulator drills, biotissue drills, video-based learning, proctorship, and final review of the videos of robotic PD performed by an individual surgeon [27]. Zureikat et al. [25] described their experience of the initial 14 cases of LPD from the same institution; the surgeons practiced LDP on fresh frozen cadavers followed by some cases for whom they had a planned conversion before performing LPD. Surgeons at the Duke University hospital used a staged modular approach starting with the hybrid procedure and eventually performing LPD, which initially always involves two attending surgeons with a team approach [28]. One attending surgeon had expertise in minimally invasive foregut surgery, whereas the other was trained in open pancreatic surgery. We suggest the hybrid PD approach for LPD (our results of the same will be published elsewhere). We performed resection laparoscopically, and the more difficult reconstruction part was performed through mini-laparotomy of approximately 8 cm. The surgeon could decide whether to progress to LPDs or continue with the hybrid protocol depending on the experience and skill levels of the surgeon. Furthermore, if one can complete the resection phase in 240-300 minutes, a laparoscopic reconstruction may be attempted. However, if the resection takes a longer time, then it would be safer to perform the reconstruction through a mini-laparotomy. Some studies have suggested using the hybrid laparoscopy–robotic approach. We have taken a slightly different path to start performing LPDs. We extensively learned by watching videos. Furthermore, we split surgery into many small steps; these steps were essentially part of other laparoscopic surgeries including hemicolectomy, radical gastrectomy, radical cholecystectomy, choledochal cyst excision, and distal pancreatectomy. Thus, we utilized our experience gained while performing these surgeries and eventually applied it to LPD.

The strength of this study is that we included all patients who underwent LPD; thus, this study provides a true picture of our initial experience with LPD. In addition, we did not exclude any patient except if they had any suspicion of vascular involvement on imaging; thus, our study was not subjected to selection bias. This study has some limitations. Because of its retrospective nature, we do not have follow-up data on the time to functional recovery and start of adjuvant treatment. In addition, we did not perform a cost analysis that could have been difficult to calculate in our set up.