Introduction
The first laparoscopic pancreatoduodenectomy (LPD) was reported over 20 years ago by Gagner and Pomp. Thus far, several comparative studies, published in recent years, have confirmed the feasibility of laparoscopy in cases of pancreatic head malignancy [
1,
2]. Although minimally invasive oncologic surgery has become an accepted approach for many abdominal operations, it is still used to a limited extent in pancreatic surgery [
3]. It is generally accepted that the laparoscopic approach should follow the same principles as open surgery. Yet, in more difficult cases, a concern arises about the oncologic quality of the operation. Forced attempts to finish the procedure minimally invasively may lead to some compromises in the technique or adjustments to difficult operative conditions and the use of atypical surgical instruments. LPD is, without a doubt, one of the most complex abdominal procedures, involving recognition of difficult anatomy, meticulous vascular dissection and multiple gastrointestinal tract reconstructions. For this reason, the risk of the course of surgery not following the principles of classical pancreatic head surgery is particularly high.
Quite recently, some authors have suggested that the so-called artery first approach (meaning superior mesenteric artery (SMA) dissection in the early phase of resection, before any irreversible step is taken) has potential advantages, such as early determination of resectability and decreased R1 resection rate [
4‐
6]. So far, this approach has been described mostly in open pancreatoduodenectomy. The data on artery first LPD are sparse [
7]. Therefore, the aim of this paper was to discuss the feasibility of laparoscopic ‘uncinate process first’ pancreatoduodenectomy.
Results
Patients in Group 1 were comparable to patients in Group 2 in regard to demographic parameters, except for ASA grade (Table
1).
Table 1
Demographic characteristics of patient groups
Number of patients, n
| 12 | 19 | – |
Females, n (%) | 6 (50%) | 7 (36.8%) | 0.47 |
Males, n (%) | 6 (50%) | 12 (63.2%) |
Mean age, years ± SD (median) | 58.0 ± 12.4 (56) | 62.3 ± 8.5 (62) | 0.09 |
BMI, kg/m2 ± SD (median) | 24.5 ± 3.1 (24.5) | 25.6 ± 2.8 (25.6) | 0.38 |
ASA 2, n (%) | 12 (100%) | 13 (68.4%) | 0.01 |
ASA 3, n (%) | – | 6 (31.6%) |
Any comorbidity | 8 (66.7%) | 10 (52.6%) | 0.44 |
Cardiovascular | 4 (33.3%) | 4 (21.1%) | 0.45 |
Hypertension | 2 (16.7%) | 8 (42.1%) | 0.12 |
Diabetes | 2 (16.7%) | 3 (15.8%) | 0.95 |
Pulmonary disease | 2 (16.7%) | 1 (5.3%) | 0.30 |
There were no significant differences in the stage of cancer between groups. The LN yield in Group 1 was 19.3 and in Group 2 it was 13.9 (
p = 0.03). R0 resection rates did not differ (66.7 vs. 63.2%,
p = 0.84). The involved margins are presented in Table
2. Although involvement of the superior mesenteric artery margin and uncinate process margin seemed lower in Group 1, we did not perform the statistical analysis due to small sample sizes.
Table 2
Pathologic results
AJCC | | | 0.78 |
AJCC II A, n (%) | 7 (58.3%) | 9 (47.4%) | |
AJCC II B, n (%) | 3 (25.0%) | 7 (36.8%) | |
AJCC III, n (%) | 2 (16.7%) | 3 (15.8%) | |
T category | | | 0.93 |
pT2 | 3 (25.0%) | 5 (26.3%) | |
pT3 | 9 (75.0%) | 14 (73.7%) | |
N category | | | 0.55 |
pN0 | 7 (58.3%) | 9 (47.4%) | |
pN1 | 5 (41.7%) | 10 (52.6%) | |
Lymph nodes, n (%) | 19.3 ± 8.2 (16) | 13.9 ± 9.4 (13) | 0.03 |
Resection | | | 0.84 |
Resection R0, n (%) | 8 (66.7%) | 12 (63.2%) | |
Resection R1, n (%) | 4 (33.3%) | 7 (36.8%) | |
Resection margin involvement |
SMAM | 2 | 4 | |
SMVM | 1 | 1 | |
PUPM | 2 | 4 | |
PNM | – | – | |
BDM | – | – | |
Total operative time (467 vs. 425 min,
p = 0.13) as well as resection time (221 vs. 232 min,
p = 0.34) were similar in both groups. The estimated blood loss in Group 1 was 408 ml, while in Group 2 it was 392 ml (
p = 0.33). Complication rates were 66.7 and 63.2%, respectively (
p = 0.84). Median length of stay was 9 days in both groups (
p = 0.36) (Tables
3 and
4).
Table 3
Postoperative characteristics of patient groups
Mean total operative time, min. ± SD (median) | 466.7 ± 53.8 (445) | 425.0 ± 85.1 (420) | 0.13 |
Mean resection time, min. ± SD (median) | 220.7 ± 47.8 (230) | 232.3 ± 51.8 (245) | 0.34 |
Mean intraoperative blood loss, ml ± SD (median) | 408.3 ± 166.3 (300) | 391.7 ± 180.7 (250) | 0.33 |
Median length of hospital stay (IQR) | 9 (8–12) | 9 (6–12) | 0.36 |
Patients with complications, n (%) | 8 (66.7%) | 12 (63.2%) | 0.84 |
Clavien-Dindo 1, n (%) | 2 (16.7%) | 4 (21.1%) | 0.34 |
Clavien-Dindo 2, n (%) | 4 (33.3%) | 7 (36.8%) |
Clavien-Dindo 3, n (%) | 2 (16.7%) | – |
Clavien-Dindo 5, n (%) | – | 1 (5.3%) |
Readmission, n (%) | – | 1 (5.3%) | - |
Table 4
Types of complications according to Clavien-Dindo classification
I | Chyle leak | 1 | 1 |
Pancreatic fistula grade A | 1 | 2 |
Surgical site infection | – | 1 |
II | Urinary tract infection | – | 1 |
Delayed gastric emptying (requiring TPN) | 3 | 4 |
Pancreatic fistula grade B | 1 | 1 |
Surgical site infection (requiring antibiotics) | – | 1 |
III | Biliary anastomotic leakage (reoperation) | 1 | – |
Postoperative bleeding (reoperation) | 1 | – |
V | Death (anastomotic leakage, massive bleeding) | – | 1 |
Discussion
In this study, we have confirmed the feasibility of laparoscopic uncinate process first approach for pancreatic head malignancy. The change in the operative technique did not have any negative influence on the operative time, blood loss and complications. In addition, pathologic specimen quality was comparable.
Currently, there is a lot of evidence that laparoscopic surgery can be successfully implemented in most of gastrointestinal cancer cases leading to reduced postoperative morbidity without compromising long-term survival. It is also significant that, nowadays, patients prefer minimally invasive access for various reasons [
11‐
13]. However, the surgeons’ acceptance for laparoscopic surgery in pancreatic head malignancy is still low. This is due to the limited number of studies showing the clinical advantages of laparoscopy over open surgery [
3]. LPD still remains an extremely difficult operation, with a long learning curve and prolonged operative time [
14].
Only recently, a modified artery first approach was described in pancreatic head surgery, which allows early determination of SMA involvement [
15]. There are several potential advantages of the artery first approach. They include the following: better resection of mesopancreas with a more adequate lymphadenectomy, reduced blood loss and an easier identification of aberrant right hepatic artery [
16]. Although nowadays, triphasic computed tomography or endoscopic ultrasound allows for the delineation of resectable, borderline resectable or non-resectable tumours, they may be insufficient in selected cases. Therefore, the use of laparoscopy for early assessment of SMA infiltration may have potential benefits, one of them being a better selection of candidates for vascular resection or patients who would benefit from neoadjuvant chemotherapy and a second-look operation. Moreover, faster recovery and better general status after exploratory laparoscopy may allow introducing chemotherapy earlier.
The number of resected lymph nodes in our series is comparable with that of previous reports in both laparoscopic and open surgery [
17,
18]. Although there is a difference in the number of harvested lymph nodes between groups, we are not certain whether this is due to the change in the operative approach. This aspect has to be confirmed in a larger trial comparing the classical approach with the artery first approach. In addition, there are more accurate prognostic factors related to lymph nodes, such as the number of positive nodes or lymph node ratio [
19]. In addition, there are no advantages of extended lymphadenectomy on survival [
20]. Moreover, pancreatic cancer spread is characterized not only by lymphatic metastases, but also by perineural invasion that may potentially lead to lymphatic spread of cancer [
21]. Therefore, a simple lymphadenectomy, without the resection of peripancreatic soft tissues and extrapancreatic nerve plexus, is considered oncologically insufficient. For this reason, meticulous skeletonization of the mesenteric vessels, with regional lymphadenectomy and perivascular neural and soft tissue removal, is highly recommended [
22,
23]. According to some authors, the artery first approach is superior over the classical in terms of better mesopancreas dissection [
15]. This term was first used by Gockel et al. in 2007 [
24]. Mesopancreas does not contain any surrounding fascia and it is defined as an anatomical space bounded by the pancreatic neck (anteriorly), pancreaticoduodenal fascia (posteriorly) and superior mesenteric vessels (medially). It does, however, contain lymph nodes, nerves and smaller vessels [
25]. It has been suggested that novel approach to mesenteric vessels may facilitate mesopancreas resection, thus lowering the incidence of R1 resection [
21]. Although this sounds logical, it still has to be investigated with well-designed trials to fully answer the question whether it lowers the recurrence rate and survival. Besides, there are no standardized protocols of mesopancreas assessment as those used in rectal cancer surgery, for instance.
According to the review by Sanjay et al., there are six different approaches to SMA that may be considered as the artery first approach [
15]. In our series, we adapted the so-called uncinate process first approach described in open surgery by Hackert et al. [
10]. In this technique, the resection is performed in a retrograde way starting from the jejunum, whereas the transection of the pancreas is the last step of the resection phase. We agree with Hackert’s observations that in this way it is possible to safely and completely dissect the uncinate process from the retroperitoneum and the superior mesenteric vein under visual control of the vein and the artery.
In our material, we did not find any differences in R0/R1 rates. The R1 rate in our group is relatively high, which is due to fact we use the classification proposed by Esposito et al. (R1 resection margin is defined as positive margin within <1 mm of the tumour) [
9]. Lack of difference in R0 rates, on the one hand, confirms that the artery first approach enables comparable quality of resection. On the other hand, we must admit that both groups were relatively small, therefore we interpret these results cautiously.
The laparoscopic uncinate process first approach allowed us to achieve comparable, if not better quality of the specimen. Another important observation from our study was that it was not associated with a worse intraoperative and perioperative course. The resection time, intraoperative blood loss and postoperative complications were not different between groups.
However, one of the greatest limitations of the study is that the learning curve in the artery first approach has not been completed yet. Although most parts of the procedure do not differ regardless of the used approach and practically only one particular step is different from what we have learnt, this may still introduce bias. Besides, this single-centre analysis includes relatively small groups of patients. Nevertheless, we clearly showed the feasibility of this technique in laparoscopic setting.