World Journal of Surgery

, Volume 32, Issue 5, pp 904–917

Is Laparoscopic Resection Adequate in Patients with Neuroendocrine Pancreatic Tumors?

Authors

    • Surgical Department, ICMD Hospital Clínic de BarcelonaUniversity of Barcelona
  • Laia Blanco
    • Surgical Department, ICMD Hospital Clínic de BarcelonaUniversity of Barcelona
  • Rebeca Cosa
    • Surgical Department, ICMD Hospital Clínic de BarcelonaUniversity of Barcelona
  • Héctor Rendón
    • Surgical Department, ICMD Hospital Clínic de BarcelonaUniversity of Barcelona
Article

DOI: 10.1007/s00268-008-9467-2

Cite this article as:
Fernández-Cruz, L., Blanco, L., Cosa, R. et al. World J Surg (2008) 32: 904. doi:10.1007/s00268-008-9467-2
  • 512 Views

Abstract

Since the first reports with laparoscopic resection of islet cell tumors in 1996, the experience worldwide is still limited, with only short-term outcomes available. Some have suggested that a malignant tumor is a contraindication to laparoscopic resection.

Aim The aim of this study was to evaluate the feasibility, safety, and long-term outcome of the laparoscopic approach in patients with functioning, nonfunctioning, or overt malignant pancreatic neuroendocrine tumor (PNT). To our knowledge this is the largest single-institution series on this subject to date.

Patients and methods A total of 49 consecutive patients (43 women, 6 men; mean age 58 years, range 22–83 years) underwent laparoscopic pancreatic surgery (LPS) from April 1998 to June 2007. Preoperative localization was done by computed tomography, magnetic resonance imaging, endoscopic ultrasonography, and Octreoscan imaging. Other than 9 PNTs localized in the head of the pancreas, all tumors were located in the left pancreas. Malignancy was diagnosed based on the presence of lymph nodes or liver metastasis. There were 33 patients with functioning tumors: 4 with gastrinomas (mean size 1.2 cm), 1 with a glucagonoma (4 cm), 3 with vipomas (3.2 cm), 2 with carcinoids (5.2 cm), 20 with sporadic insulinomas (1.4 cm), 2 with insulinoma/multiple endocrine neoplasia type 1 (MEN-1) (4.4 cm), and 1 with a malignant insulinoma (13 cm). Sixteen patients had a nonfunctioning tumor (mean size 5 cm). The following techniques were performed: laparoscopic spleen-preserving distal pancreatectomy (Lap SPDP), laparoscopic distal pancreatectomy with splenectomy (Lap SxDP) and laparoscopic enucleation (Lap En)/laparoscopic excision (Lap E). Lymph node dissection was performed when malignancy was suspected (Strasberg´s technique). Evaluation criteria included operative and postoperative factors, pathologic data including R0 or R1 resection (the pancreatic transection margin and all transection margins on the specimen were inked). Long-term outcomes were analyzed by tumor recurrence and patient survival.

Results Four cases (8.2%) were converted to open surgery. Overall, Lap SPDP, Lap SxDP, and Lap En/Lap E were performed in 15 (33.3%), 8 (17.8%), and 22 (48.9%) patients, respectively. The operative time and blood loss was significantly lower in the Lap En group compared with the other laparoscopic techniques. The group of patients with malignant tumors undergoing Lap SxDP had a longer operating time and greater blood loss compared with the other distal pancreatectomy (Lap DP) techniques. Overall, the postoperative complications were significantly higher in the Lap En group (42.8%) than in the Lap DP (Lap SPDP + Lap SxDP) group (22%). These complications were mainly pancreatic fistula: 8.7% after Lap DP and 38% after Lap En. The overall morbidity was significantly higher after Lap SPDP (26.7%) than after Lap SxDP (12.5%) owing to the occurrence of splenic complications in the Lap SPDP group without splenic vessel preservation two of seven (28.5%). The means and ranges of hospital stay after Lap SPDP, Lap SxDP, and Lap En/Lap E were 5.9 (5–14), 7.5 (5–12), and 5.5 (5–7) days, respectively (NS). Pathology examination of the specimen showed R0 resection in all patients with malignant PNT. The mean time to resumption of previous activities for patients undergoing Lap DP or Lap En was 3 weeks. There were no postoperative (30 days) or hospital deaths.

Conclusions This series demonstrates that LPS is feasible and safe in benign-appearing and malignant neuroendocrine pancreatic tumors (NEPTs). The benefits of minimally invasive surgery were manifest in the short hospital stay and acceptable pancreas-related complications in high-risk patients. LPS can achieve negative tangential margins in a high percentage of patients with malignant tumors. Although surgical cure is rare in malignant NEPTs, significant long-term palliation can be achieved in a large proportion of patients with an aggressive surgical approach.

Introduction

Pancreatic neuroendocrine tumors (PNTs) comprise a rare group of neoplasms with complex patterns of behavior requiring detailed specialist management. Surgery remains the only curative modality currently available for resectable PNTs [1]. Complete surgical resection may be possible for those tumors that are localized at presentation. The surgical management varies with tumor type, location, and size. Benign PNTs <2–3 cm in diameter may be treated with enucleation. However, malignant PNTs or tumors >3 cm diameter should be managed with either a Whipple resection or distal pancreatectomy, depending on the location of the tumor, with local lymph node resection. Furthermore, up to 15% of patients with PNT and metastatic disease to the liver have disease confined to one hepatic lobe, which may be resectable [2, 3].

The surgical treatment of pancreatic endocrine tumors in the setting of multiple endocrine neoplasia type 1 (MEN-1) is slightly different [4, 5]. Tumors are frequently multiple, and the disease in patients with multiple tumors and MEN-1 may run a more indolent course. Currently, surgical intervention is advocated when the tumors exceed 2.5 cm in diameter, as tumor size is associated with a higher likelihood of liver metastasis, which ultimately affects survival [6].

Laparoscopic pancreatic surgery (LPS) was first introduced in 1994 by Gagner and Pomp [7] and Cushieri [8] for patients with chronic pancreatitis. In 1996, Gagner et al. [9] reported the early experience with laparoscopic resection of islet cell tumors. Since then, the experience worldwide is still limited, and most reports on LPS are often based on limited experience with short-term outcomes. Assalia and Gagner [10] reported a review of the world literature dealing with laparoscopy of PNTs using the MEDLINE database covering the period January 1966 to October 2003. They found a total of 93 reported cases. The largest experience from the same institution comprised no more than 10 patients. Insulinoma was the most prevalent diagnosis, comprising 87% of all cases. Other PNTs were rare, with three gastrinomas (two of them metastatic and converted), one malignant vipoma (also converted), and seven unspecified “other” and nonfunctioning PNTs [10]. The laparoscopic procedures performed were equally divided between laparoscopic distal pancreatectomy (Lap DP) and laparoscopic enucleation (Lap En) (39 cases each), with 15 converted cases (16.1%). Recently, a multiinstitutional European study (25 European centers), comprising 127 patients (with a variety of pancreatic disorders), demonstrated that laparoscopic pancreatic resection is feasible and safe in selected groups of presumed benign pancreatic lesions requiring enucleation procedures or left pancreatic resections [11]. A point of criticism of this study was that only four centers (16%) reported more than 10 patients.

Some authors have suggested that malignant pancreatic neoplasms are a contraindication to laparoscopic resection because of concerns about the radicality of the resection and the oncologic outcomes [12].

The aim of this study was to evaluate the feasibility, safety, and long-term outcome of the laparoscopic approach in patients with functioning, nonfunctioning, and overtly malignant PNTs. To our knowledge, this is the largest single-institution series on this subject to date.

Patients and methods

Laparoscopic pancreatic surgery was carried out at our institution, Hospital Clínic de Barcelona, in 49 consecutive patients with PNTs between April 1998 and June 2007. All patients had lesions suitable for Lap DP or Lap En. Tumor size, mitotic count, differentiation grade, angioinvasion, and immunohistochemical staining of hormones, precursor amines, and peptides were evaluated to classify tumors according to the new World Health Organization (WHO) classification [13] (Table 1).
Table 1

Neuroendocrine pancreatic tumors

Type of tumor

Patients (no.)

Age (years)a

Sex

Tumor size (cm)a

Functioning

  Gastrinoma

4

46.3 (22–61)

3 M/1 F

1.2 (0.8–2.0)

  Glucagonoma

1

66

—/1 F

4

  Vipoma

3

70.5 (69–72)

—/3F

3.25 (3.0–3.5)

  Carcinoid

2

68.5 (64–73)

—/2 F

5.25 (3.5–7.0)

  Sporadic insulinoma

20

40 (22–66)

2 M/18 F

1.4 (0.6–2.0)

  Insulinoma MEN-1

2

35 (28–42)

—/2 F

4.4 (1.8–7.0) multiple

  Malignant insulinoma

1

83

—/1 F

13 × 9

Nonfunctioning

16

61 (44–78)

1 M/15 F

5 (2.5–11.0)

Total

49

58.8 (22–83)

6 M/43 F

4.7 (2.5–11.0)

MEN-1, multiple endocrine neoplasia type 1

aMean and range

Functioning tumors

Four patients (22, 55, 56, and 61 years of age, respectively) had symptoms compatible with sporadic Zollinger-Ellison syndrome (ZES); the basal gastrin level (>200 pg/ml; normal <100 pg/ml) and stimulated (with secretin) gastrin level (>1,000 pg/ml) were markedly elevated. Computed tomography (CT) showed a 0.8-mm hypervascular lesion in one patient (22 years old) in the area between the duodenum and vena cava and a 1-cm lesion in the uncinate process in another patient; these lesions were confirmed by Octrescan imaging (OcI). In another patient (55 years old) an 8-mm duodenal gastrinoma was diagnosed by endoscopic ultrasonography (EUS) and positive OcI. One patient (56 years old) with elevated gastrin levels had MEN-1 syndrome. This patient had had episodes of abdominal pain and hyperamylasemia for 1 year. CT scans showed two lesions (2.0 and 1.5 cm, respectively) in the body/tail of the pancreas and dilatation of the duct of Wirsung in the tail of the pancreas. OcI was positive at the midpancreas. In another patient (61 years old) with ZES, magnetic resonance imaging (MRI) showed an 11-mm tumor at the uncinate process of the pancreas with positive OcI.

Three women (ages 38, 69, and 72 years, respectively) had profuse watery diarrhea, hypokalemia, and hypotension for more than 1 year in one patient and more than 6 months in two patients. The diagnosis was based on elevated vasoactive intestinal polypeptide (VIP) levels >200 pg/ml; CT scan showed 2.5-, 3.0-, and 3.5-cm tumors, respectively, in the body/tail of the pancreas. One of these patients (38 years old) had gross diffuse liver metastatic disease with excellent general health and was a candidate for liver transplantation after primary tumor resection.

One woman, 66 years old had abdominal pain and elevated plasma glucagon levels >500 pg/ml. CT scans showed a 4-cm tumor in the body of the pancreas.

There were 23 patients (20 women, 3 men) with organic hyperinsulinism (11 patients with sporadic insulinoma and 2 MEN-1 patients were previously reported). The mean age of all patients was 43 years (range 16–83 years). In 20 patients with benign sporadic insulinoma, CT and EUS were performed in all patients and localized the tumor in 14 (70%) and 19 (95%) patients, respectively. MRI was performed in 15 patients and localized the tumor in 12 (80%). The 20 presumed benign sporadic insulinomas were localized to the head of the pancreas (3), the neck of the pancreas (6), the body of the pancreas (6), and the tail of the pancreas (5). The mean size was 14 mm (range 0.6–20.0 mm). One 83-year-old woman had neuroglycopenic syndrome for 3 years and was misdiagnosed with cardiovascular and neuropsychiatric disorders; CT scans showed a 13 × 9 cm tumor in the body/tail of the pancreas with multiple bilobar liver metastases. Two MEN-1 patients with insulinomas had multiple tumors (mean size 4.4 cm, range 1.8–7.0 cm) localized in the body/tail of the pancreas.

Two women, aged 64 and 73 years, respectively, had abdominal pain and diarrhea. Plasma serotonin plasma levels were 1.250 and 1.527 mg/ml (normal 0–200 mg/ml), respectively; and urinary 5-hydroxyindoleacetic acid (5-HIAA) levels were 480.0 and 193.4 mg/24 h, respectively (normal 0–10 mg/24 h). CT scans showed 3.5- and 7.0-cm tumors in the body and tail of the pancreas, respectively; in addition, one patient had two liver metastases, 3 and 5 cm in size, in segments VI–VIII; and the other patient had 4.0- and 11.5-cm tumors in segments V–VI and VIII and another 3-cm tumor in segment III. OcI showed positive spots in the pancreas and in the liver in both patients. EUS and fine-needle aspiration (FNA) cytology confirmed the suspicion of carcinoid tumors.

Nonfunctioning tumors

A total of 16 patients (15 women, 1 man) had nonfunctioning tumors. The mean age of all of these patients was 58.9 years (range 44–78 years). Eleven patients had abdominal pain, and the other five were asymptomatic. Abdominal ultrasonography and CT scans showed well defined tumors with a mean size of 5 cm (range 2.5–11.0). One patient had three gross liver metastases (>3 cm) in the right liver. EUS and FNA were used in five asymptomatic patients to confirm the diagnosis of PNTs.

Techniques of laparoscopic surgery

Laparoscopic distal pancreatectomy was performed with or without spleen preservation. When performing spleen-preserving distal pancreatectomy (Lap SPDP), this operation was performed with [14] or without splenic vessel preservation [15] (Warshaw’s technique). We have previously published the technical details of these operations, which were performed in patients with presumed benign pancreatic lesions [16, 17].

En bloc laparoscopic distal pancreatectomy with splenectomy (Lap SxDP) was performed in patients with suspected pancreatic malignancy. The principles of this operation follow the technique described by Strasberg et al. [18] in 2003 called radical antegrade modular pancreatosplenectomy (RAMPS). We have recently reported the laparoscopic technique of this operation in patients with ductal adenocarcinoma of the body/tail of the pancreas [19].

Laparoscopic tumor enucleation (Lap En) was performed, when indicated, for small neuroendocrine tumors (mainly insulinomas) and in selected cases of nonfunctioning tumor. We have published the technical details of this operation elsewhere [20, 21].

The use of intraoperative laparoscopic ultrasonography (LapUS) was an integral part of the laparoscopic procedure [20, 22]. It was necessary for defining the safe area (2 cm from the tumor) for pancreatic transection in cases of LapDP and to perform safe tumor enucleation in cases of LapEn, avoiding injury to the duct of Wirsung or the splenic vessels.

Clinical data analysis

Data on operative, intraoperative, and postoperative care were prospectively collected. Preoperative parameters included patient demographics (age, sex); and intraoperative parameters included total operating time, blood loss, and blood transfusion. Postoperative events were recorded using the following definitions.

Pancreatic fistula: According to the International Study Group on Pancreatic Fistula (ISGPF) [23], this entity was deemed present if there was any measurable drainage from an operatively placed drain on or after postoperative day 3, with amylase levels greater than three times the upper limit of normal serum amylase levels. All patients below this threshold were considered to have no biochemical evidence of fistula. The patients with a fistula were then classified into three grades of severity according to the ISGPF. Grade A fistulas: transient asymptomatic fistulas, evident only by elevated drainage amylase levels. Drains are removed within 3 weeks, almost always within the first 7 days after the operation. Grade B fistulas: symptomatic, clinically apparent fistulas that require diagnostic evaluation and therapeutic management. Operatively placed drains may remain in situ at the time of discharge and are frequently required for management longer than 3 weeks. Grade C fistulas: severe, clinically significant fistulas that require major intervention during clinical management.
  • Splenic complications: splenic infarct, focal or massive, detected by color Doppler ultrasonography (CD-US); perisplenic fluid collection with pain in the left upper quadrant of the abdomen.

  • Abscess: culture-positive purulent drainage from an intraabdominal fluid collection obtained percutaneously or operatively; and/or radiographically confirmed fluid collection with systemic or localized signs of infection.

  • Pneumonia: presence of a new infiltrate on chest radiography and the following: body temperature >38°C; abnormal elevation of white blood count; or positive Gram stain or culture, requiring intravenous antibiotic treatment.

  • Length of stay (LHS): days from the initial operation to hospital discharge.

  • Conversion: need for an abdominal incision to deal with any intraoperative complication; the need to make an abdominal incision prematurely when safe dissection is not possible on oncologic grounds.

  • Completeness of resection: assessed by reviewing the operative and pathology reports. The pancreatic transection margin and all tangential margins on the specimen were inked. An R0 resection was considered to have been performed if the primary tumor was removed with negative margins. Patients with microscopically positive margins or grossly positive margins were classified as having had an R1 or R2 resection, respectively.

  • Recovery: measured by recording the patient’s time to return to activities of daily living.

Statistical analysis

Perioperative data, postoperative complications, and resection modalities were compared using chi-squared statistics, analysis of variance (ANOVA), and Student’s t-tests. Statistical significance was accepted at p < 0.05. All statistical computations were performed using the SPSS 11.5 program for Windows.

Follow-up

All patients underwent regular follow-up examination postoperatively at 6 months and annually thereafter.

Results

Among the group of four patients with gastrinoma, a duodenal gastrinoma was found in one patient on Lap US in close proximity to the papilla of Vater; it was enucleated after conversion to open surgery. The second patient, with a tumor between the common bile duct and vena cava, underwent laparoscopic tumor excision; the tumor proved to be a primary lymph node gastrinoma. In the third patient, with a suspected gastrinoma in the uncinate process of the pancreas, an 11-mm nodule outside the pancreatic capsule of the uncinate was resected laparoscopically; frozen-section pathology examination confirmed a neuroendocrine nodule. This patient was converted to open surgery and underwent pylorus-preserving pancreaticoduodenectomy (PPPD). On pathology examination a 0.4-mm pancreatic tumor (gastrinoma) was found in the pancreatic head in close proximity to the second duodenal wall; the extracapsular pancreatic nodule was a metastatic lymph node. Lap SPDP was performed in the fourth, MEN-1, patient. Lap US was performed in the area of the head of the pancreas and duodenum to confirm the absence of other gastrinomas. Pathology examination of the specimen showed gastrinomas of 2.0 and 1.5 cm in the body/tail of the pancreas; the 2.0-cm tumor obstructed the duct of Wirsung. All four patients were cured, with resultant normal basal and stimulated (with secretin) gastrin levels at a mean follow-up of 16 months (range 2–24 months).

In the group of three patients with vipoma, one patient was converted to open surgery for widespread peripancreatic lymph node metastases and 3 years after pancreatosplenectomy developed diffuse liver metastases and was lost of follow-up. One patient is still alive 5 years after Lap SPDP. The third patient (38 years old), with gross, diffuse liver metastasis, underwent Lap SxDP; pathology examination of the specimen showed a 2-cm tumor in the body of the pancreas and four metastatic lymph nodes. This patient is awaiting liver transplantation.

Lap SPDP was performed in the patient with a glucagonoma. The tumor was diagnosed as a high-grade malignant carcinoma. At 2.5 years after surgery, the patient remains asymptomatic, although OcI showed positive imaging in the area of the celiac trunk. This patient underwent a reoperation using the laparoscopic approach, and a lymphadenectomy was performed around the celiac artery and branches. All lymph nodes removed were invaded by tumor. One year later OcI identified multiple spots in the liver. This patient is alive 4 years after surgery with diffuse liver metastases.

In two patients with pancreatic carcinoids and liver metastasis, Lap SxDP with a modified RAMPS procedure was performed. In one patient, concomitant extended right hemihepatectomy and partial liver resection was carried out through a separate right J abdominal incision. In the other patient a right hepatectomy was performed 8 weeks after primary tumor resection using an open right J abdominal incision. Both patients are clinically and biochemically cured 6 and 12 months after surgery, respectively.

In 20 patients with presumed benign sporadic insulinoma at the time of laparoscopic exploration of the pancreas, the tumor was visible in eight patients (48%) and LapUS confirmed the localization of preoperative imaging modalities in 19 patients (95%) (Fig. 1). Lap SPDP was performed in four patients (20%) and Lap En in 15 patients (75%). One patient was converted to open surgery because the tumor was not found after LapUS. This particular patient had a pedunculated insulinoma at the inferior border of the body of the pancreas managed with enucleation. All patients are cured, with a mean follow-up of 36 months (range 1–54 months). Lap SPDP in one with splenic vessel preservation and in another without splenic vessels preservation were performed in two MEN-1 patients with multiple tumors, who remain asymptomatic 40 and 44 months after surgery, respectively. Lap SxDP and a modified RAMPS procedure were performed in the patient with malignant insulinoma, who is free of clinical symptoms 15 months after surgery despite having had bilobar liver metastases.
https://static-content.springer.com/image/art%3A10.1007%2Fs00268-008-9467-2/MediaObjects/268_2008_9467_Fig1_HTML.gif
Fig. 1

Distribution of 20 sporadic insulinomas. Each tumor is depicted according to size (mm, diameter), location in the pancreas, whether resection (R) or enucleation (E), and if visible (V) during laparoscopic inspection

In the group of 16 patients with nonfunctioning tumors, Lap En was performed in six patients with a mean tumor size of 3 cm (range 2.5–3.2 cm); Lap SPDP was performed in six patients and Lap SxDP in four patients. Among the latter four patients, concomitant right hepatectomy was performed in one patient after Lap SxDP using an open right J abdominal incision. The overall incidence of malignancy among the 16 patients was 25% (four patients). Two tumors were classified as low-grade carcinoma and two other as high-grade carcinoma. It was noteworthy that all of the tumors <3 cm were benign (6/6 patients) but malignancy was found in tumors ≥5 cm (4/10 patients). One patient with a 5-cm tumor (high-grade carcinoma) presented with liver metastases in the left lobe and was managed with left hepatectomy; the patient developed diffuse liver metastases 4 years later and is still alive. The patient after Lap SxDP and concurrent right hepatectomy developed liver metastasis 2 years after surgery and was managed with radiofrequency ablation and; diffuse liver metastases appeared 4 years later, but the patient is still alive. Two patients with low-grade malignant PNTs are disease-free 3 and 5 years after surgery.

Overall analysis of perioperative data

The operating time, blood loss, length of hospital stay, and medical and surgical complications were compared with respect to the laparoscopic technique used (four cases converted to open surgery were excluded): Lap En/Lap E (22 cases), Lap SPDP with splenic vessel preservation (seven cases) and without splenic vessel preservation (eight cases), and en bloc pancreatosplenectomy (eight cases) (Table 2). Overall, Lap SPDP, Lap SxDP, and Lap En/Lap E were performed in 15 (33.3%), 8 (17.8%), and 22 (48.9%) of the 45 patients, respectively. In our series, only four patients (8.2%) were converted to open surgery for the following reasons: (1) Inability to localize a pedunculated insulinoma at the inferior border of the pancreas. (2) Two cases of gastrinoma, one localized in the posterior duodenal wall in close contact to the ampulla of Vater and another in a patient in whom laparoscopy localized the tumor to the uncinate process of the pancreas but in an extracapsular position (previously diagnosed by MRI and OcI). The operation was converted to a pylorus-preserving pancreaticoduodenectomy (open) because of the uncertainty of dealing with lymph node metastases from an intraduodenal or intrapancreatic gastrinoma. (3) One patient with a vipoma with oncologic features that precluded a safe laparoscopic approach (Table 2).
Table 2

Neuroendocrine pancreatic tumors: techniques of laparoscopic surgery

Type of tumor

No.

Lap SPDPa

Lap SxDP

Lap En

Lap E

Conversion

Functioning

  Gastrinoma

4

1 (25.0%)

1 (25.0%)

2 (50.0%)

  Glucagonoma

1

1 (100%)

  Vipoma

3

1 (33.3%)

1 (33.3%)

1 (33.3%)

  Carcinoid

2

2 (100%)

  Sporadic insulinoma

21

4 (19.0%)

1 (4.8%)

15 (71.4%)

1 (4.8%)

  Insulinoma MEN-1

2

2 (100%)

Nonfunctioning

16

6 (37.5%)

4 (25%)

6 (37.5%)

Total

49

15 (30.6%)

8 (16.3%)

21 (42.9%)

1 (2%)

4 (8.2%)

Lap SPDP, laparoscopic spleen-preserving distal pancreatectomy; Lap SxDP, laparoscopic distal pancreatectomy with splenectomy; Lap En, laparoscopic tumor enucleation; Lap E, laparoscopic tumor excision

aLap SPDP with splenic vessels preservation (SVP) eight cases (53.3%) and without SVP seven cases (46.7%)

The operating time (mean 120 min) and blood loss (<220 ml) were significantly lower in the Lap En group compared with the other laparoscopic techniques (p < 0.01). The group of patients with malignant tumors undergoing Lap SxDP had a longer operating time and greater blood loss (mean 310 min; mean 720 ml) compared with the other techniques of distal pancreatectomy performed in the other groups included in this study (p < 0.05). Overall postoperative complications were significantly higher (p < 0.001) in the Lap En group (42.8%) than in the Lap DP (Lap SPDP + Lap SxDP) (22%). The complications were mainly pancreatic fistula: 8.7% after Lap DP (Lap SPDP + Lap SxDP) and 38.0% after Lap En (p < 0.001) (Table 3). Biochemical pancreatic fistula (grade A) was observed in 8.7% after Lap DP and in 19.0% after Lap En (p < 0.001). The severity of the pancreatic fistula was also significantly higher in the Lap En group (grade B 14.3%, grade C 4.8%) compared to Lap DP (0%). The overall morbidity was significantly higher after Lap SPDP (26.7%) than after Lap SxDP (12.5%) owing to the occurrence of splenic complications in the Lap SPDP group without splenic vessel preservation (2/7, 28.5%). Two patients had abdominal pain in the left upper quadrant of the abdomen; CD-US showed a focal splenic infarcts 3 and 4 cm, respectively. One of these patients developed a splenic abscess that was successfully treated with radiologic intervention. Reoperation was necessary 36 h after Lap En in a patient with an insulinoma because of bleeding from a gastroepiploic vessel. The means and ranges of the hospital stay after Lap SPDP, Lap SxDP, and Lap En/Lap E were 5.9 (5–14) days, 7.5 (5–12) days, and 5.5 (5–7) days, respectively (NS) (Table 3). One patient (with a hospital stay of 5 days) was rehospitalized with abdominal pain 10 days after Lap En for a nonfunctioning PNT. This patient developed an intraabdominal collection (9 cm on CD-US) that was drained with radiologic intervention. A pancreatic fistula of 100–150 cc was drained for 2 weeks. Pathology examination of the specimen showed R0 resection for all malignant PNTs. The mean time to resumption of previous activities for patients undergoing Lap DP or Lap En was 3 weeks. There were no postoperative (30 days) or hospital deaths.
Table 3

Neuroendocrine pancreatic tumors: complications of laparoscopic surgery

Complication

Spleen-preserving distal pancreatectomy

En -bloc -spleno pancreatectomy

Enucleation

Excision

No. of patients

15

8

21

1

Pancreatic fistula

  Overall

1 (6.7%)

1 (12.5%)

8 (38%)

  ISGPF grade A

1 (6.7%)

1 (12.5%)

4 (19%)

  ISGPF grade B

3 (14.3%)

  ISGPF grade C

1 (4.8%)

Splenic infarct

  Overall

2 (13.3%)

   

  Asymptomatic

   

  Symptomatic

2 (13.3%)

   

  Abscess

1 (6.7%)

   

  Postoperative bleeding

  

1 (4.8%)

 

Total

4 (26.7%)

1 (12.5%)

9 (42.8)

Length hospital stay (days) mean and range

5.9 (5–14)

7.5 (5–12)

5.5 (5–7)

 

ISGPF: International Study Group on Pancreative Fistula

Discussion

According to Assalia and Gagner’s [10] review of the literature, insulinomas are the most prevalent, mostly benign, and single and are curable with surgical excision, they comprise most of that experience. Hence, there is no consensus among surgeons about the indications for and limits of the laparoscopic approach in patients with PNT. Moreover, some have suggested that malignant pancreatic neoplasms are a contraindication to laparoscopic resection [12]. In the current report, LPS was feasible in 100% of 49 PTN patients, both with benign tumors (77.5%) and malignant tumors (22.4%). To our knowledge, this report is the first to present data on the short- and long-term outcome of laparoscopic pancreatic resection in patients with malignant PNTs.

Also, LPS was performed in three patients with MEN-1: two patients with insulinomas and one patient with gastrinoma. Functioning tumors (e.g., gastrinomas, vipomas, glucagonomas, carcinoids, insulinomas) and nonfunctioning tumors comprised 67.3 and 32.6% in this series, respectively.

Nonfunctioning PNTs are located most commonly (approximately 60%) in the pancreatic head but can be found anywhere within the pancreas. Generally, the diagnosis is often delayed (>5 years), and tumors become clinically apparent when they are already inoperable and/or metastatic. In patients with localized, nonmetastatic disease, complete surgical resection of the primary tumor is the treatment of choice [2426]. The median survival has been reported as 7.1 years [27]. However, only 48% of the patients with localized, nonmetastatic disease who underwent resection of the primary tumor were alive and without evidence of recurrent disease at a median follow-up of 2.7 years (range 1–8 years) from diagnosis [27]. The authors emphasized that is inappropriate to assume that complete resection of the primary tumor in the absence of metastatic disease represents long-term cure. Phan et al. [28] have reported that the 5-year survival rate following resection approximates 50%.

In patients with metastasic disease, whether resection of the primary tumor is appropriate should be based on the presence of clinical symptoms and the location of the tumor. In the absence of a symptomatic primary tumor with unresectable extrapancreatic metastatic disease, distal pancreatectomy is probably not indicated. In selected patients with localized liver metastases, combined resection of the primary tumor and liver metastases should be attempted [2933]. The 5-year survival for patients who were cleared of pancreatic and liver disease was 65% in recent series [29]. However, the tumor recurred in three-fourths of patients who underwent resection with curative intent; most of these recurrences were detected within 2 years.

In the current series, the mean size of the nonfunctioning PNT was 5 cm. Patients with tumors ≤3 cm (six patients) were managed with Lap En and local lymph node dissection to exclude metastatic lymph node involvement; benign tumors were the final diagnoses in this group. Overall, malignancy of nonfunctioning PNTs was found in 25% of cases; however, in tumors >5 cm, malignancy was found in 40% of cases. R0 laparoscopic resection was achieved in 100% of patients with malignant nonfunctioning PNTs. Based on these results, nonfunctioning PNTs ≤3 cm may be managed with an organ-preserving operation such as LapEn. In patients with tumors >5 cm, laparoscopic pancreatic resection should be an oncologic resection (modified Strasberg’s operation) including radical lymph node dissection of the peripancreatic, portal, hepatic, and superior mesenteric areas (Figs. 13). At a mean 37 months’ follow-up (range 2–89 months), no patients with a benign or low-grade malignancy had tumor recurrence; these preliminary data are yet to be confirmed. However, high-grade malignancies tend to recur. In a high-volume surgical series, the overall survival for PNT was 41 months; half of the patients developed metachronous liver disease during the observation period [27].
https://static-content.springer.com/image/art%3A10.1007%2Fs00268-008-9467-2/MediaObjects/268_2008_9467_Fig2_HTML.gif
Fig. 2

Modified Strasberg operation for malignant pancreatic neuroendocrine tumor. (1) Lymph node dissection in the area of common hepatic artery and celiac trunk. (2) The splenic artery is clipped at the origin. (3) The splenic vein is clipped at the junction with the mesenteric vein. (4) Laparoscopic linear stapling. (5) Lymphadenectomy on the left side of the mesenteric superior artery. (6) Lymphadenectomy on the area of the superior border of the pancreas

https://static-content.springer.com/image/art%3A10.1007%2Fs00268-008-9467-2/MediaObjects/268_2008_9467_Fig3_HTML.gif
Fig. 3

Surgical steps described in Fig. 2 are completed. Here, the spleen is mobilized inferiorly and superiorly (after coagulation of the short gastric vessels). Pancreatosplenectomy is performed

Gastrinomas, as other PNTs, may occur sporadically or as part of MEN-1. Gastrinomas are more common in men than in women. The mean age at diagnosis is 45–50 years, and approximately 20% have MEN-1. Approximately 65–90% of all gastrinomas found at surgery occur in the pancreatic head/duodenal area [34]. The percentages of gastrinomas that behave in a benign or a malignant fashion are not well established; malignancy is usually diagnosed by the presence of metastases. In one study, only 5% of patients who had duodenal gastrinomas had liver metastasis compared with 52% of patients who had pancreatic gastrinomas [34]. Surgical treatment results in an excellent prognosis and should be offered to good-risk candidates who have ZES [3537].

The role of laparoscopy in the management of patients with gastrinoma is controversial. Recently, Norton and Jensen [38] have given four reasons for recommending against a laparoscopic approach: “First, gastrinomas are 3 to 10 times more common in the duodenum than the pancreas. Second, many gastrinomas are not localized preoperatively; especially duodenal gastrinomas, and this will likely decrease the success rate. Third, many gastrinomas are associated with adjacent lymph node metastases. This will make a laparoscopic approach more difficult, prolong its duration, and may limit its success. Fourth, greater than 75% of gastrinomas are in the pancreatic head region in the so-called gastrinoma triangle, complicating the laparoscopic approach.” We believe these criticisms should be kept in mind so as to apply the laparoscopic approach only in selected cases. The MEN-1 patient after Lap SPDP is disease-free 1 year after surgery. The patient with primary lymph node gastrinoma 18 months after laparoscopic tumor excision fits the definition [39]: “as occurring in a patient whose only extrahepatic tumor resected was in a lymph node and who was disease-free post-resection” (normal fasting serum gastrin levels, negative secretin tests, and no tumor on imaging studies).

Vipomas account for fewer than 10% of NPTs. They are much more common in women (with a female/male ratio of 3:1), and most frequently occur at around the fourth decade of life [40]. Up to 90% of vipomas originate in the pancreas, and they are usually solitary tumors. Most of the lesions are located in the distal pancreas. More than 60% of pancreatic vipomas are malignant, and by the time of diagnosis up to 60% have metastasized to lymph nodes, liver, kidneys, or bone. Analysis of the Mayo Clinic data reported by Smith et al. showed in only 44% of patients were the vipomas resectable; and of those tumors, only 28% were resectable for cure. Long-term survivors are not uncommon; in the Mayo Clinic series the mean survival was 3.6 years, and the longest overall survival was 15 years [40]. Of the patients with resectable disease, the laparoscopic approach may be indicated when the tumor is localized in the body/tail of the pancreas. In the current series, two patients are still alive 3 and 5 years after surgery, and one patient after Lap SxDP is awaiting liver transplantation [41, 42].

Glucagonomas are less than half as common as vipomas, with an annual incidence of 0.01–0.1 new cases per million. They are slightly more common in women (55%) and usually occur after 45 years of age [1]. Most glucagonomas are large solitary tumors that are almost exclusively found in the body or tail of the pancreas. The diagnosis is established by measuring the serum concentration of glucagon, with levels of >500–1,000 pg/ml. Most glucagonomas are metastatic at the time of diagnosis [43]. However, due to the slow-growing nature of these tumors and the alleviation of symptoms with resection, it is generally recommended that primary tumors deemed resectable on preoperative imaging should undergo surgery with removal of the primary tumor and resection of accessible hepatic metastases.

The Mayo Clinic published a 15-year experience with 21 patients with glucagonoma. Despite the presence of metastatic disease in all 21 patients at the time of diagnosis, only 9 died of their disease over the subsequent 5 years of follow-up [43, 44]. Our glucagonoma patient, 4 years after successful Lap SPDP, developed diffuse liver metastases but remains asymptomatic.

Carcinoid tumors are defined as tumors with neuroendocrine histologic features and evidence of serotonin production. The overall incidence of these tumors is approximately one to two cases per 100,000 population. Pancreatic carcinoids have a poorer outcome than other foregut carcinoids, perhaps because of their location; and they tend to produce relatively vague, nonspecific symptoms resulting in larger, more advanced tumors at presentation. The pancreatic tumors were universally metastatic at diagnosis [4547]. Pancreatic primaries had the poorest prognosis, with only 10% of the patients alive at 10 years. In the current series, two patients after Lap SxDP and liver resection for metastatic disease are disease-free. However, the follow-ups for these patients are relatively short.

The role of surgery in patients with insulinomas associated with MEN-1 has been defined by Demeure et al. [5] in a review of the literature comprising 60 patients. These authors suggested that patients with insulinomas associated with MEN-1 require a different surgical approach than that for patients with sporadic insulinomas. In most reports, enucleation or limited resection did not result in the development of recurrent hyperinsulinism for up to 15 years. However, others have reported recurrence rates of 40% at 10 years after enucleation [4, 48, 49]. Enucleation of an insulinoma alone in patients with MEN-1 would likely lead to missed tumors and a failed operation [49]. More than 75% of patients with insulinoma and MEN-1 had multiple pancreatic tumors. It seems that subtotal distal pancreatectomy, preserving the spleen, combined with enucleation of any tumors identified in the pancreatic head should be the standard operation. We believe that patients with MEN-1 insulinomas may benefit from the choice of the laparoscopic approach according to the principles developed during the past two decades from the standard open approach [50, 51]. During surgery, intraoperative LapUS may identify tumors not seen during preoperative localization studies. In addition, LapUS identifies the demarcation between normal pancreas and macroscopic disease and is useful for determining the optimal site of transection. In this series, two MEN-1 patients with hyperinsulinism underwent Lap SPDP and remain asymptomatic and normoglycemic at 40 and 44 months, respectively.

The role of surgery for MEN-1 ZES patients is debatable [5057]. In patients who have MEN-1 with ZES, 70–95% of primary tumors arise in the duodenum and 30–25% in the pancreas. One of our patients with ZES had MEN-1 with two tumors (2.0 and 1.5 cm, respectively) localized in the body/tail of the pancreas. This particular patient had, in addition, abdominal pain and dilatation of >5 mm of the duct of Wirsung in the tail of the pancreas. The patient underwent Lap SPDP, and clear obstruction of the duct of Wirsung by one the tumors was seen in the pathologic specimen. This patient is clinically cured 1 year after surgery.

One patient (83 years old) had severe hypoglycemia that was misdiagnosed as a cardiovascular disorder. A tumor mass 13 cm in diameter was found in the body/tail of the pancreas with diffuse bilobar liver metastases. Lap SxDP and peripancreatic lymphadenectomy cured the clinical and biochemical syndrome (15 months of follow-up) despite the presence of liver metastasis. The absence of hepatic metastasis is a major predictor of survival at 3 years [58]. For localized liver metastatic disease, concurrent resection of the primary tumor and synchronous hepatic metastasis is recommended [59]. Improved outcome has been demonstrated in patients with metastatic islet cell cancer managed aggressively with debulking surgery; their overall, progression-free, and symptom-free survival rates are 71% (median 76 months), 5% (median 21 months), and 24% (median 26 months), respectively, at 5 years [3]. Danforth et al. [58] reported a recurrence rate of 63%. Palliative resection was associated with a median survival of 4 years, compared with 11 months after biopsy of the tumor only. Aggressive secondary therapies such as chemoembolization or radiofrequency ablation should be used only to control hypoglycemia [59]. According with Hirshberg et al. [60] metastatic insulinoma may have a variable natural history. After the initial surgical resection, the biology of the tumor, rather than any treatment modality, was most likely the major determinant of long-term survival.

Sporadic insulinomas, which represent up to 70–80% of clinically symptomatic PNTs, occur in all age groups. Most are solitary lesions, but 10% are multiple. Most insulinomas are benign; only 10% have any evidence of malignancy. The lesions are small with a diameter of <2 cm in 90% and less than 1.3 cm in 50% of patients [61]. They are distributed equally within the head, body, and tail of the pancreas. EUS is the most sensitive modality for detecting insulinomas, with preoperative detection rates of 86–93%. In recent years, spiral CT scanning has become more successful in localizing insulinomas and may also provide additional information regarding suspected malignancy. The combination of biphasic thin-section helical CT and EUS resulted in an overall diagnostic sensitivity of 100%. Recently, MRI [62] has shown a detection rate of 85% and has improved in delineating the anatomic relation of the tumor with the pancreatic duct.

Surgical resection is the treatment of choice and offers the only chance of cure. The surgical strategy in patients with sporadic insulinoma should be restricted to removal of the solitary tumor in about 90% of patients. The use of enucleation or resection depends on the location of the tumor in the pancreas and the findings from intraoperative ultrasonography (IOUS). Laparoscopy and LapUS provide information similar to that obtained by means of open IOUS and can identify lesions that are undetectable by preoperative imaging techniques. Despite the advantages of LapUS, it is still worthwhile to attempt preoperative imaging, as it provides useful information for patient positioning and port placement. LapUS also facilitates operative decision making. Ultrasonographically, islet cell tumors are typically hypoechoic and easy to differentiate from the surrounding pancreatic parenchyma. This information helps the surgeon choose between enucleation (in an attempt to preserve healthy pancreatic parenchyma) or resection. LapUS is also essential in Lap En for deciding whether an anterior or a posterior surgical approach should be performed (Figs. 46). LapUS is of great help in guiding the enucleation procedure by avoiding injury to the pancreatic duct or large blood vessels. In cases of Lap DP, LapUS identifies the demarcation between normal pancreas and the tumor and is useful for determining the optimal site of transection.
https://static-content.springer.com/image/art%3A10.1007%2Fs00268-008-9467-2/MediaObjects/268_2008_9467_Fig4_HTML.gif
Fig. 4

Laparoscopic tumor enucleation in a surgically difficult area (I), the head of the pancreas. Laparoscopic ultrasonography (Lap US) is essential for deciding between using an anterior or a posterior surgical approach. The danger is possible injury of the Wirsung duct. Lap US may demonstrate the location of the duct of Wirsung, which permits its protection during the enucleation

https://static-content.springer.com/image/art%3A10.1007%2Fs00268-008-9467-2/MediaObjects/268_2008_9467_Fig5_HTML.gif
Fig. 5

Laparoscopic tumor enucleation in a surgically difficult area (II), the neck of the pancreas. Lap US provides information on the proximity of the tumor on the anterior or posterior wall of the neck of the pancreas. In this example, the tumor lies posterior to the Wirsung duct

https://static-content.springer.com/image/art%3A10.1007%2Fs00268-008-9467-2/MediaObjects/268_2008_9467_Fig6_HTML.gif
Fig. 6

In the situation described in Fig. 4, the inferior border of the pancreatic neck is widely separated from the mesenteric-portal vein axis, allowing safe tumor enucleation

Overall cure rates of 75–98% are reported after surgery. Recently, Grant [63] has reported a Mayo Clinic experience that includes 242 patients (216 primary and 31 reoperations on 233 sporadic and 14 MEN-1 patients). The tumors were enucleated in 59%, whereas distal pancreatic resection was necessary in 36%; only eight patients (4%) required a Whipple procedure. Cure was achieved in 98% of patients.

With advances in laparoscopic techniques, both laparoscopic enucleation and resection of pancreatic insulinoma have been performed successfully. Enucleation or limited pancreatic resection for solitary, small, benign insulinomas is better achieved using the laparoscopic approach in terms of parietal damage to the abdomen. We have recently reported guidelines for laparoscopic surgical strategies in the management of insulinomas [21]. The reported success for laparoscopic resection of insulinoma ranges from 60 to 100% [20, 6471]. In the current series, 20 patients had a benign insulinoma, with 9 localized in the head/neck of the pancreas and 11 in the body/tail. One patient was converted to open surgery and underwent enucleation. Among the other 19 patients, Lap En and Lap SPDP were performed in 15 (79%) and 4 (21%), respectively. The percentage of enucleation in this series compares favorably with other large series using the open approach [63].

Nevertheless, enucleation of a PNTs is not a simple procedure, whether performed with an open procedure or laparoscopically. It offers clear advantages to patients: complete tumor removal without loss of pancreatic parenchyma, avoiding possible diabetes and splenectomy when distal pancreatectomy is chosen for tumors localized in the left pancreas or a Whipple operation for tumors in the pancreatic head. However, pancreas-related complications may occur. In an international review of benign insulinomas treated by open surgery, Rothmund et al. [72] reported a complication rate of 32%. Hellman et al. [73] evaluated the outcome of 65 patients operated on for organic hyperinsulinism using the open approach. Procedures included 37 enucleations, 19 distal pancreatectomies, four combinations of enucleation from the head and distal pancreatectomy, two subtotal left pancreatectomies, and 3 pancreaticoduodenectomies. Pancreatic fistula occurred after enucleation from the head in 50%, after enucleation from the body/tail in 24%, and after distal pancreatectomy in 26%. It appears that enucleation from the head is associated with a high risk of pancreatic fistulas. Park et al. [74] reported experience with surgical treatment (using the open approach) in patients with islet cell tumors arising in the head of the pancreas (70% sporadic insulinomas). In 22 patients with simple enucleation, the mean operating time was 4 hours, the mean blood loss 360 ml, the median time for the patients to have a regular diet 7 days, and the median time to remove drains 22 days. In this experience 67% of patients had no complications; however, pancreatic fistula occurred in 15% of patients, requiring intravenous alimentation and a prolonged hospital stay (we interpret this as grade C pancreatic fistula according to the ISGPF). Furthermore, 12 patients (44%) had some amount of amylase-rich fluid drainage (biochemical grade A fistula). Other complications observed were pancreatitis (4%), intraabdominal abscess (4%), postoperative bleeding (4%), and delayed gastric emptying (4%).

In the literature the incidence of postoperative morbidity (mainly pancreatic fistula) after Lap En is 20–30% [6470]. Most of these fistulas were successfully managed conservatively with drainage, combination of drainage and endoscopic retrograde cholangiopancreatography with pancreatic duct stenting, or reoperation (open) for salvage. In our current series 38% of patients (15 insulinomas and six nonfunctioning tumors) developed a pancreatic fistula after Lap En (19% grade A, 14.3% grade B, 4.8% grade C) and were managed conservatively. The occurrence of the pancreatic fistula was more frequent after Lap En of tumors localized in the head/neck of the pancreas (5/9, 55.5%) compared with Lap En of tumors localized in the body/tail of the pancreas (3/10, 30.0%).

In contrast, the frequency of pancreatic fistula was significantly less in the group of patients managed with LDP (Lap SPDP + Lap SxDP) (8.7%), being biochemical grade A fistulas in most (no impact on the clinical outcome). Automatic stapling was used in all cases for transection of the pancreas. Recently, a meta-analysis of technique for closure of the pancreatic remnant after distal pancreatectomy (open surgery) supported a statistically nonsignificant, but possibly clinically relevant, trend toward the superiority of staple closure [75]. However, even this favored technique resulted in a fistula rate of 22.9% in the meta-analysis. This percentage is higher than the 8.7% observed in the current laparoscopic series.

We recommend Lap SPDP (33.3% in the current series) to prevent the potential long- and short-term complications associated with splenectomy: 53.3% with splenic vessel preservation and 46.6% without splenic vessel preservation. The overall complication rate was higher in the Lap SPDP group than in the Lap SxDP group because of the occurrence of splenic complications (13.3%) in the Lap SPDP group (without preservation the splenic vessels). CD-US should be performed routinely in all patients in the latter group to diagnose areas of focal infarct promptly and to prevent abscess formation with antibiotic administration.

In the literature, conversion rates range from 20% to 33% [6471, 76]. In this report 4 of 49 (8.2%) patients were converted to open surgery. None of the patients with nonfunctioning tumors were converted. Among the functioning PNTs the highest conversion rate was found in patients with gastrinomas (50%) and the lowest in patients with sporadic insulinomas (4.8%).

Conclusions

Surgery remains the only curative modality currently available for resectable PNTs. Laparoscopic surgery varies with tumor type, location, and size. Small PNTs, such as insulinomas and nonfunctioning PNTs <3 cm, may be treated with Lap En. Lap SPDP with and without splenic vessel preservation is feasible and can be achieved in most cases of benign PNTs. Other functioning and nonfunctioning PNTs (suspected malignant or overtly malignant) should be managed with laparoscopic pancreatic resection, including peripancreatic lymph node dissection (modified laparoscopic Strasberg’s technique). Laparoscopic pancreatic resection achieved R0 resection for all malignant PNTs. In addition, some patients presenting at the time of diagnosis with liver metastasis were treated with concomitant primary tumor and liver resection. Other patients, some years after Lap DP, presented with liver metastases that were treated with liver resection.

Recovery is an important outcome and can be measured by the patient’s time to return to activities of daily living. In the current series, the mean time for the patients who underwent Lap DP or Lap En to return to previous activities was 3 weeks.

In summary, the results of this prospective study indicate that LPS performed in patients with PNTs, both benign and malignant lesions, is associated with no mortality, morbidity rates that are lower than reported for open surgery, and all the advantages of a minimally invasive approach with reference to pain, cosmetic results, reduced hospital stay, and faster postoperative recovery. The promising results presented in this report makes LPS a valid option in selected patients with PNT.

Acknowledgment

We thank Tom Gorey, Professor of Surgery, Dublin (Ireland), for article preparation.

Copyright information

© Société Internationale de Chirurgie 2008