LDP vs ODP for pancreatic adenocarcinoma: a case matched study from a single-institution

In the last few decades, with the development of laparoscopic instruments and skills, Laparoscopic distal pancreatectomy (LDP) has become widely accepted by surgeons for benign and low-grade tumors of the pancreas. Recent reviews and meta-analysis showed that LDP has the advantage of less blood loss and fewer hospital stay days as well as fewer postoperative complications compared with open distal pancreatectomy [1‐3]. However, application of laparoscopic approach has been restricted for malignant pancreatic lesions due to concerns over oncologic safety [4]. Unlike other gastrointestinal regions, such as stomach and colon, until now, only a few pioneer studies reported direct comparisons of oncologic outcomes between LDP and open distal pancreatectomy (ODP) for pancreatic ductal adenocarcinoma (PDAC) [5‐9]. In this study, we designed a 1:2 case-matched retrospective study from a single institution and analysed the short-term and long-term outcomes of the patients undergoing either LDP or ODP for PDAC.

In recent years, LDP has been gradually accepted as standard approach to treat benign or low grade lesions located in the body or tail of the pancreas. The technical feasibility, safety and clinical benefit has been well confirmed by various matched studies compared with open distal pancreatectomies [16‐18]. However, application of Laparoscopic distal pancreatectomies for PDAC was still limited due to the concern of oncologic outcome and surgical quality [4]. But reports emerging from some experienced centers are encouraging. Compared with conventional open approach, they demonstrated the advantages of less blood loss, shorter hospital stay and early return to normal activity with a similar morbidity, POPF, short oncology outcome (R0 resection rate, the number of harvested lymph nodes) and the overall survival rate [5‐9, 19, 20]. In this case-matched study we compared the short-term and long-term outcomes of patients undergoing distal pancreatectomy and the results were consistent with these reports.

Surgery remains the only opportunity for long-term survival for patients with resectable PDAC [21]. R0 resection is the most crucial prognostic factor [22]. In a multicenter analysis Kooby et al. [5] reported that the R0 resection rate of LDP and ODP for PDAC was 73.9 % (17/23) and 65.7 % (46/70). A multivariate analysis was conducted in the whole cohort and only blood >500 ml was associated with R1 resection while the method of resection (LDP or ODP) wasn’t correlated. Shin et al. [20] reported the largest single-institution study of LDP for PDAC (n = 70), the R0 resection rate was 75.7 % (53/70) for LDP while 83.8 % (67/80) for ODP. Lee et al. and Hu et al. [8, 9] reported in their series that patients included in LDP group were relatively in early stage and the R0 resection rate was 100 %. These case–control retrospective studies showed no significant difference of R0 resection between the LDP and ODP groups. In the present study, the R0 resections for LDP and ODP were 94.1 % and 85.3 % (P = 0.650) which was in accordance with those former studies. Recently, Sharpe et al. [19] reported outcomes for 769 patients of which 144 in the LDP group for PDAC through the National Cancer Data Base. In this retrospective survey, the LDP group had a decrease in margin positivity rate but the tumor size was smaller compared with the ODP group and LDP was more likely to be performed at academic/research institutions. The results were satisfactory for laparoscopic procedure although the heterogeneity might exist due to the type of study design or selection bias. Besides the essentiality of frozen section, extended resections were required in some cases in order for a definitive margin-negative surgery because of the aggressive nature of the disease [23]. Extended resections are feasible procedures with increased postoperative morbidity and better survival compared with palliative bypass procedures [24]. Although laparoscopic extended resection of the pancreas is technically demanding, its application is increasing in specialized centers. Croome et al. [25] reported data from Mayo clinic of 31 patients undergoing total laparoscopic pancreaticoduodenectomy with major vascular resection, there was no significant difference of the total complications comparable with open group and with less mean operative blood, less hospital stay. We previously reported the first laparoscopic hepatopancreatoduodenectomy case with favorable perioperative outcome and showed no sign of recurrence over a year [26]. The data of LDP combined with extended resections is rare. Shin et al. [20] reported 6 (8.6 %) cases of concurrent resections for PDAC by laparoscopic procedure including 5 left colectomies and 1 gastrectomy. After propensity score-matched (including age, BMI, tumor size, concurrent resection) analysis, the overall survival was similar between the LDP group and ODP group while concurrent resection ration were balanced between the two groups. Ricci et al. [27] reported 6 (18.7 %) extended resections including resection of liver wedge, stomach, left adrenal gland and colon among 32 LDP. In our study, we had 5 cases (29.4 %) of extended resections in LDP group with 1 R1 resection while 7 cases (20.6 %) in ODP group with 5 R1 resections. We abolished the laparoscopic procedure of two cases because of invasion to superior mesenteric artery (SMA). Despite the sample data was too small to make any persuasive conclusion, it may achieve R0 resection of locally advanced PDAC in selected patients through laparoscopic procedure by skilled surgeons. Completion of the learning curve, a fixed surgical group and suitable selection criteria were efficacious to carry out these complex goals [27, 28] and we insist on using 5 trocars strategy in order for the cooperation of the main surgeon and the first assistant. Until now, there is no standard indication of LDP for PDAC. As reported from previous studies and meta-analysis, surgeons are mostly inclined to conduct LDP for smaller tumor size [2, 6‐8]. Although Kooby et al. [5] reported tumor size (>4 cm) was not associated with positive resection margin, a huge tumor would be an obvious obstacle for exposure of the operation field. So, patients forwith tumor size >5 cm in body and >10 cm in tail of the pancreas were reserved for open procedure and were exluded in this study. The median survival was both 14 months in LDP and ODP groups in this series. Kooby et al. [5] reported median survival 16 months both for LDP and ODP groups and Magge et al. [6] reported 19 months for the entire cohort. Lee et al. [8] reported a median follow-up 39 months for the minimally invasive surgery group (including 4 robotic cases) using their inclusion criteria (Yonsei criteria) which mainly consisted of early stage pancreatic cancer. Compared with previous studies, the survival data in this study was not fulfilling. In Cox proportional hazards analysis extended resection, perineural invasion were strong factors for worse survival. The high ration of extended resection (23.5 %) and perineural invasion (72.5 %) of the whole cohort indicated the cases enrolled in this study were relatively in advanced stage due to lack of early diagnosis of the disease probably. The median survival for no extended resection group was 15.0 months and was consistent with the previous case-matched studies.

This study has several critical limitations, including the retrospective design and low number of patients enrolled in the study. Adjuvant treatment is believed to prolong overall survival [29], but in this study the Cox proportional hazards analysis showed no association with overall survival (P = 0.380). The poor differentiation was not associated with overall survival but the moderate differentiation showed association. The small sample of this study might be the reason and didn’t have sufficient statistical power to evaluate the outcome. The study span lasted 11 years and only 1.3 LDP cases per year were performed. The surgical technique was not standardized between the laparoscopic and open approach. Also the follow-up time was short especially for the LDP group and it was difficult to calculate the 5-year survival. Until now, the oncologic safety and long-term survival were not tested by any randomized controlled study between LDP and ODP for PDAC, so it is not sufficient enough to make a conclusion that LDP is oncologic equivalence to ODP [3, 4]. As Kooby and Kang commented it was difficult to conduct an RCT because of the infrequence of diagnosis and opportunity for operation of PDAC in the pancreatic body and tail [4, 5]. The result of this study could provide valuable evidence to support use of LDP for PDAC even in relatively advanced stage.

In conclusion, the results in our study validated that LDP was technically feasible and safe for PDAC in selected patients and the short-term oncologic outcomes were not inferior to ODP in this small sample study However the long-term oncologic safety of LDP for PDAC has to be further evaluated by multicenter or randomized controlled trials.LDP with extended resection for PDAC is better performed in highly specialized centers and with suitable selection criteria.