Parenchyma-sparing resections for pancreatic neoplasms

Introduction

Pancreaticoduodenectomy (PD) and distal pancreatectomy (DP) with splenectomy are the standard resections performed in patients with malignant tumors of the pancreas. Nowadays these procedures are associated with low mortality and acceptable morbidity in high-volume centres 1-3. However, in recent years there has been a dramatic increase in the diagnosis of benign or low-grade malignant neoplasms of the pancreas [4, 5]. Thanks to high-resolution imaging these neoplasms are frequently detected when they are small in size and are diagnosed even in young or middle-aged patients. In this setting, standard pancreatectomies can be performed but they result in a significant loss of normal pancreatic tissue without a real oncological need and with subsequent impairment of exocrine and endocrine pancreatic function 6-8. This is of obvious interest, especially for young patients with benign tumors of the pancreas and a long life expectancy. Therefore, parenchyma-sparing resections of the pancreas have increasingly become an option to treat these tumors in order to preserve pancreatic parenchyma as much as possible [9]. Depending on the size, site and number of lesions, the procedures of choice are enucleation 10-24, middle pancreatectomy (MP) 25-41, and middle-preserving pancreatectomy [42]. On the other hand, these “limited” resections of the pancreas appear to be associated with a significant risk of postoperative complications, particularly pancreatic fistula, and their oncological safety is also debated [9, 10, 17, 22, 27, 28, 30, 42-46].

In the present article, the available experience of parenchyma-sparing resections of the pancreas is reviewed in order to assess their effectiveness by analyzing indications, postoperative complications and long-term functional and oncological results.

Pancreatic enucleation

First reported in 1898 by Ernesto Tricomi in Italy [47], pancreatic enucleation is nowadays considered the procedure of choice to treat insulinomas [45]. However, there is a wide spectrum of lesions amenable to enucleation including non-functioning pancreatic endocrine tumors (NF-PETs) [10-12, 46], serous and mucinous cystadenomas 13-17. Enucleation of solid pseudopapillary tumors (SPTs), small branch-duct intraductal papillary mucinous cystic neoplasms (BD-IPMNs) and other benign conditions such as cystic lymphangioma have been reported as well 18-20.

Beyond feasibility, a patient can be considered a candidate for enucleation if:

With regard to NF-PETs, only neoplasms with a low likelihood of malignancy (<2 cm in size, no enlarged locoregional lymph nodes, no distant metastases) can be treated with enucleation [44, 46]. Not all the authors agree to treating gastrinomas with an enucleation. In fact, despite the fact that gastrinomas are usually small, they are malignant with nodal involvement in up to 80–90% of cases and standard resections are associated with better disease-free survival than atypical resections [44, 48].

At the time of operation a careful exploration of the entire abdominal cavity must be performed in order to confirm the preoperative findings. Enlarged peripancreatic lymph nodes should be taken for intraoperative examination and, in case of metastatic disease, standard pancreatic resection with lymphadenectomy must be performed. Intraoperative ultrasound (IOUS) is helpful to clearly identify the lesion, its morphology, its proximity to the main pancreatic duct and the site [49]. In patients with multiple endocrine neoplasia type-1 (MEN1), IOUS is of paramount importance to rule out the presence of multifocal disease within the pancreatic gland [46]. Technically, surgery is performed by staying outside of the tumoral pseudo capsule to ensure complete resection. The dissection can be carried out in a classical way very carefully tying all the small vessels or using clips and selective ligation of small vessels, harmonic scalpel and tissue coagulation devices. In these latter cases the proximity of the lesion with the main pancreatic duct may represent a contraindication to their use for its possible thermal injury. A drain should be left in place near the resection bed. Intraoperative frozen-section examination may provide information on the nature of the lesion. The completeness of surgical excision and the presence of negative margins must be evaluated and confirmed at final pathology. In the setting of NF-PETs some peripancreatic nodes should be excised and evaluated at final pathology as well [50]. Reoperation must be considered in the event of incomplete resection, positive resection margins or presence of nodal metastases [10]. However, in these cases, the histology of the primary pancreatic tumor as well as the patients' age and willingness to undergo strict follow-up evaluation over time versus immediate re-resection must be taken into account.

Different series on pancreatic enucleation have been published. Park and colleagues [11] reported successful enucleation in 27 of 30 (90%) endocrine tumors of the head of the pancreas. The most frequent complication was pancreatic fistula (15%) with no local recurrence during long-term follow-up. Talamini et al. [14] described their experience with enucleation of mucinous cystadenoma of the pancreas. Of the 36 mucinous cystadenomas observed by these authors between 1990 and 1997, 10 underwent enucleation. Enucleation was associated with less blood loss compared to standard resections. Moreover, overall morbidity did not differ between the two groups, but pancreatic fistula was found in 50% of enucleation versus 12% of standard resections. All the fistulas resolved spontaneously and no recurrences were reported. Recently, our group reported the experience of 61 enucleations performed between 1990 and 2005 [10]. Endocrine tumors, including 22 insulinomas and 16 NF-PETs, were the most common indication for the procedure (62%), followed by epithelial cyst (8%), serous cystadenoma (8%), mucinous cystadenoma (5%) and SPTs (5%). Only a small BD-IPMNs with negative margins was enucleated without recurrence in the long-term follow-up. However, considering the biological behaviour of IPMNs and the need for clear negative margins, enucleation should carefully evaluated in these cases [51, 52]. Interestingly, in our personal experience IOUS helped in the decision to perform enucleation in 16 patients (26%) who were initially candidates for standard pancreatectomy. Overall morbidity was 43% with a pancreatic fistula rate of 38%. However, the rate of clinically significant pancreatic fistula was 23%. Reoperation was required in five patients (8%) but there was no mortality. After a median follow-up of 61 months no patient developed tumor recurrence.

With regard to long-term functional outcome, enucleation can well preserve pancreatic function, and several authors reported no exocrine/endocrine insufficiency [9, 11, 14]. In our series of 61 patients no patient developed exocrine pancreatic insufficiency, while two patients (3%) developed diabetes [10].

Recently, the combined experience of the San Raffaele Hospital in Milan, Italy and our institution was analyzed to evaluate the oncological effectiveness of atypical resections—-including enucleation and MP—-for insulinomas and NF-PETs [50]. Overall, 91 patients underwent enucleation to treat 65 insulinomas and 26 NF-PETs. Of these tumors, 65 were benign, 25 had an uncertain biological behaviour and one neoplasm was a well-differentiated non-functioning carcinoma with nodal metastases. After a median follow-up of 60 months, only one recurrence was reported. One patient with NF-PET with uncertain biological behaviour developed liver metastases 86 months after initial surgery; no lymph nodes were excised during enucleation in this case. Therefore, enucleation seems to be a safe end effective procedure to treat both insulinomas and well-selected NF-PETs. However, nodal sampling should be routinely performed during enucleation for NF-PETs to rule out metastatic disease and to increase the diagnostic effectiveness.

Finally, more recently, laparoscopic enucleation of the pancreas has been described, particularly to treat insulinomas and NF-PETs 21-24. Early experiences seem to be favourable, reporting less blood loss and shorter hospital stay after laparoscopic enucleation compared to open surgery, but the morbidity rates were similar. However, large series with long-term follow-up are lacking. Therefore, laparoscopic enucleation is feasible and safe, but the clear role for this approach still remains to be better defined. Development of laparoscopic intraoperative ultrasonography, as well as a hand-assisted approach for selected cases, may improve the number and the effectiveness of these procedures [9].

In conclusion, pancreatic enucleation is the procedure of choice for insulinomas and it has a role for well-selected patients with NE-PETs, serous and mucinous cystadenoma and other benign conditions. Although a high fistula rate as well as high morbidity rates have been reported compared to standard resection, recurrence is rare and long-term functional results are very good.

Middle pancreatectomy

Enucleation is not technically possible in all the cases. In particular, if the tumor is close to the main pancreatic duct, enucleation must be avoided [10]. For benign or low-grade neoplasms in the neck or proximal body of the pancreas that cannot be treated with enucleation, middle pancreatectomy (MP)—-also known as central or median pancreatectomy—- has been proposed as an alternative technique 25-41. This operation was described the first time by Guillemin and Bessot [25] in 1957 to treat a patient with chronic pancreatitis. However, the surgical technique nowadays applied worldwide was described by Dagradi and Serio [26] from our institution in 1984 to treat an insulinoma. Briefly, in MP, the anterior face of the pancreas is exposed, a Kocher manoeuvre is performed and the superior mesenteric, portal and splenic vessels are dissected free from the posterior aspect of the pancreas. As pointed out for enucleation, IOUS might be helpful to (1) clearly detect the tumor, (2) exclude additional pancreatic lesions and (3) evaluate the relationship between the tumor and peripancreatic vessels and the main pancreatic duct [28]. The central portion of the pancreatic gland with the lesion is resected, possibly with a margin of at least 1 cm to both cut pancreatic ends. Pancreatic transection can be performed with the scalpel or a linear stapler. The transacted head of the pancreas and main pancreatic duct are usually over-sewn. A Roux-en-Y loop is created and an end-to-side pancreatico-jejunostomy is performed to the distal stump [25-28, 30, 31]. A drain is left near the pancreatic anastomosis. As alternative techniques, a pancreatico-gastrostomy can be performed instead of a regular pancreatico-jejunostomy [37]; moreover, some authors have described a double pancreatico-jejunostomy of the two stumps on the same jejunal loop [29].

MP can be considered for different lesions including serous and mucinous cystadenomas, insulinomas, NF-PETs (size <2 cm, no vascular or peripancreatic infiltration, no lymph node or liver metastases), non-invasive intraductal papillary mucinous neoplasms, true cyst of the pancreas, cystic lymphangioma and small SPTs 27-36. As pointed out for enucleation there are important recommendations to follow:

One of the major deterrents to the widespread acceptance of MP among pancreatic surgeons is the high rate of complications associated with the procedure. Roggin et al. [27], in a review of 207 patients from 16 series who underwent MP, reported an overall morbidity rate of 33% and a fistula rate of 22%. Muller et al. [29] described a single center experience with 40 MPs, and compared their perioperative and long-term outcomes with those of 40 patients who underwent PD and other 40 who had a DP. In this series 23 patients had only focal chronic pancreatitis and the remaining 17 neoplastic lesions. Muller and co-workers [29] reported a morbidity rate of 27.5% with mortality of 2.5% (one patient died of postoperative intra-abdominal haemorrhage), a reoperation rate of 5% while only three patients (7.5%) developed a pancreatic fistula. Matched-pairs analysis with patients who underwent PD or DP revealed no major differences of perioperative parameters, morbidity or mortality.

The experience of the University of Verona and of the Massachusetts General Hospital was recently reported with 100 MP, comparing short- and long-term operative outcomes after MP taking as a control group 45 patients who underwent an extended distal pancreatectomy (EDP) for neoplasms in the middle pancreas [28]. The two most common indications for MP were endocrine tumors (33%) and serous cystadenoma (27%). Overall morbidity after MP was 58% with no mortality nor reoperation. Overall pancreatic fistula rate was 27% even if the rate of clinical significant pancreatic fistula was 17%. However, MP and EDP groups showed no significant differences in overall morbidity, abdominal complications or pancreatic fistula. Finally, in the aforementioned cohort of 100 MP, at final pathology seven patients (7%) had a malignant tumor and five (5%) had a positive resection margin. Interestingly, all five of these patients had an IPMN and no patient underwent re-resection soon after MP.

In the last years, some laparoscopically or laparoscopic-assisted MPs have been described [40, 41]. Rotellar and colleagues [41] reported a series of 9 laparoscopic MPs with a morbidity rate of 33%, no mortality and a median length of stay of 5 days. A laparoscopic duct-to-mucosa pancreatico-jejunostomy with interrupted 3/0 silk stitches was performed with fibrin glue around the anastomosis. The results of these preliminary experiences need to be confirmed in prospective multicentric randomized studies evaluating indication, postoperative and long-term outcomes, and costs after laparoscopic and open MP.

Considering long-term functional results, MP is an effective procedure to preserve pancreatic function. In our experience, after a median follow-up of 54 months, the incidences of endocrine and exocrine insufficiency were 4 and 5%, respectively, compared to 38 and 15.6% in patients who underwent EDP [28]. Roggin et al. [27] found in their review incidences of endocrine and exocrine impairment after MP of 3.6 and 3%, respectively. Similar results were reported by our group in a recent study where the development of pancreatic insufficiency was evaluated in 162 patients with benign tumors who underwent different pancreatic resections—-62 pancreaticoduodenectomies, 64 distal pancreatectomies, and 36 parenchyma-sparing resections including 21 MPs and 15 enucleations [6]. The long-term incidences of endocrine and exocrine insufficiency were significantly lower after parenchyma-sparing resections, with the probability of developing both endocrine and exocrine insufficiency being higher for PD and DP than for MP and enucleation (58, 29 and 3% at 5 years, respectively).

Considering oncological results, MP is not an adequate operation for malignant tumors, and it should be performed only for benign and low-grade neoplasms 27-37. In our experience, 3/7 patients with a malignancy and 5/5 patients with positive resection margin had an IPMN [28]. Of these patients, two with malignant main-duct IPMN and positive margins had a local recurrence. Thus, the indication to MP in IPMNs should be carefully evaluated only for small, benign branch-duct IPMNs of the middle pancreas, while main-duct IPMNs should always be treated with standard pancreatectomies [28, 51-53. Another important recommendation must be made for endocrine neoplasms. As for enucleation, it is strongly advisable to remove multiple peripancreatic nodes. In fact, in the combined experience between our institution and San Raffaele Hospital, where 36 MP for endocrine tumors including 12 insulinomas and 24 NF-PETs were performed, at final pathology two of the NF-PETs were found to be carcinomas [50]. Overall, three patients who underwent MP for NF-PETs, including the two well-differentiated carcinomas, showed tumor recurrence during follow-up. The lack of lymph node histological examination may “under-stage” some patients. Therefore, a good advice is that whenever parenchyma-preserving resections are performed for endocrine neoplasms, lymph node sampling of peripancreatic nodes should always be performed in order to increase the effectiveness of the procedures.

In conclusion, in experienced hands, MP is an appropriate operation for benign and low-grade malignant neoplasms with good long-term functional results despite a significant perioperative morbidity. Careful patient selection and appropriate intraoperative evaluation are of paramount importance in this setting.

Middle-preserving pancreatectomy

Multifocal diseases of the pancreas are uncommon entities that may occur in patients with MEN1 syndrome, IPMNs, and multiple metastases from renal cell carcinoma [42, 46, 52, 53]. In most of these cases a total pancreatectomy (TP) is required to remove all the lesions. However, the need for this major and challenging operation should be carefully balanced with patient's life expectancy, risk of tumor recurrence and patient's ability to manage exocrine/endocrine pancreatic insufficiency following TP. Recently, we have proposed an alternative operation, namely middle-preserving pancreatectomy (MPP), for patients affected by multifocal pancreatic lesions sparing the body of the gland [42].

In MPP, after a careful IOUS, the resection of the pancreatic tail with splenectomy is performed, with the preservation of the great pancreatic artery. The distal margin is sent for frozen-section examination to be sure that it is free from disease. The distal cut edge is then over-sewn with selective ligation of the main pancreatic duct. Then a regular PD with or without pylorus-preservation is carried out. The proximal margin of the body is then checked by a second frozen section and then anatomized to a jejunal loop.

Between 1999 and 2007, 5 MPPs were performed in our institution, two for multifocal endocrine neoplasm in the setting of MEN1 syndrome and the remaining three for IPMNs. Median operative time was 365 min. Only one patient developed pancreatic fistula, while the postoperative course was uneventful in the remaining patients who were discharged after a median of 10 days. After a median follow-up of 20 months, two elderly patients developed insulin dependent diabetes and exocrine insufficiency soon after surgery while another patient developed mild exocrine insufficiency. No recurrence was observed. MPP seems to be safe and feasible, but larger series with longer follow-up are needed to better evaluate the long-term outcomes and benefits of the procedure. However, actual data suggest that MPP could be an effective alternative to total pancreatectomy. In fact, it may decrease the risk of both exocrine and endocrine insufficiency, and if pancreatic insufficiency does occur after MPP, its management is definitively easier compared to the postoperative status after total pancreatectomy.

Conclusions

Parenchyma-sparing resections of the pancreas are safe and effective procedures for treating well-selected benign or low-grade malignant neoplasms of the pancreas. Although the associated morbidity rates are significant, long-term functional outcomes are very good and tumor recurrence is uncommon. Table 1 summarizes the long-term outcomes after different parenchyma-sparing resections of the pancreas of the largest series in the present literature. Moreover, careful patient selection and performance of these operations in high-volume centres is of paramount importance to achieve the best results. For NF-PETs, lymph node sampling and intraoperative evaluation is strongly recommended in order to avoid an inappropriate oncological procedure despite the preoperative selection.