World J Surg (2009) 33:1966–1970 DOI 10.1007/s00268-009-0145-9
Pancreatic Insulinoma: A Surgical Experience Marı´a Nayvı´ Espan˜a-Go´mez Æ David Vela´zquez-Ferna´ndez Æ Paulina Bezaury Æ Mauricio Sierra Æ Juan Pablo Pantoja Æ Miguel F. Herrera
Published online: 23 July 2009 Ó Socie´te´ Internationale de Chirurgie 2009
Abstract Background Small size, high benignity rate, and sporadic nature make insulinomas suitable for laparoscopic resection. On the other hand, occult location or multicentricity mandate open surgery. This study was designed to analyze a series of patients who had pancreatic insulinomas and underwent initial treatment at our institution. Methods Clinical records of the 34 patients with pancreatic insulinomas who underwent surgical resection between 1995 and 2007 were reviewed. Main variables for analysis were cure of the disease and surgical complications. Results There were 20 women and 14 men with a mean age of 40 ± 13 years. Mean size of the tumors was 2.2 ± 1 cm. Laparoscopic resection was completed in 14 of 21 patients. Most tumors that were resected by laparoscopy were solitary, benign, and located in the body and tail of the pancreas. Open surgery was selected for 13 patients, including 7 sporadic (5 in the head), 4 related to the MEN syndrome, and 2 malignant tumors. Surgical morbidity occurred in 23 patients. The most common complication was pancreatic fistula (3/13 in open, 4/14 in laparoscopic, and 6/7 in conversions). One patient in the open group died 15 days after surgery from massive PTE.
M. N. Espan˜a-Go´mez D. Vela´zquez-Ferna´ndez M. Sierra J. P. Pantoja M. F. Herrera (&) Department of Surgery, Instituto Nacional de la Nutricio´n Salvador Zubira´n, Vasco de Quiroga 15, Tlalpan, Mexico City 14000, Mexico e-mail:
[email protected] P. Bezaury Department of Radiology, Instituto Nacional de la Nutricio´n Salvador Zubira´n, Mexico City, Mexico
123
Postoperative normoglycemia was achieved in all patients and persisted for a follow-up period of 4 ± 3.7 years. Conclusions Most insulinomas in our series were small and benign. Tumors that were located in the body and tail were more often amenable for laparoscopic resection. The cure rate was very high. Pancreatic fistula was the most frequent complication.
Introduction Insulinoma is the most common pancreatic neuroendocrine neoplasm [1]. Surgery is the treatment of choice and has an extremely high success rate [2]. Improvements in minimally invasive techniques have made possible laparoscopic resection of some tumors. Favorable features of insulinomas that make them amenable for laparoscopic resection are: (a) small size (tumors are typically \2 cm at the time of diagnosis); (b) high benignity rate (90%); (c) solitary and sporadic nature of most tumors (close to 10% of tumors are associated with the MEN 1 syndrome) [3]. The first laparoscopic resection of an insulinoma at our institution was successfully performed in 1997. Since then, we have attempted laparoscopic resection for most cases. This study was designed to analyze the results and complications of a series of 34 insulinomas who underwent initial treatment at our institution during a 12-year period.
Materials and methods Medical records of the 34 patients with pancreatic insulinomas who underwent initial surgical resection between
World J Surg (2009) 33:1966–1970
1967
1995 and 2007 at our institution were reviewed. The main variables for analysis were cure of the disease and surgical complications. Patients were divided into three groups according to the surgical technique: laparoscopic surgery, open surgery, and conversion from laparoscopic to open surgery. Results between groups were analyzed in a comparable fashion.
Table 2 Localizing studies showed different accuracy to identify tumors located in the various anatomic pancreatic areas Head and neck
Body and tail
CT (26)
2/8 (25%)
9/18 (50%)
MRI (14)
5/6 (83.3%)
Endoscopic ultrasound (14)
5/6 (83.3%)
6/8 (75%)
Angiography (21)
5/7 (71.4%)
14/14 (100%)
5/8 (62.5%)
Results There were 20 women and 14 men with a mean age of 40 ± 13 years. All patients presented with symptoms of neuroglycopenia and had a biochemical diagnosis of organic hypoglycemia. The fasting glucose levels were abnormal in all patients. Mean glucose value was 43.8 ± 10.84 mg/dl. Low levels of glucose after an overnight fasting period were found in 85% of patients. All patients underwent a supervised fasting test and developed symptomatic hypoglycemia (mean glucose value, 39.8 ± 7.8 mg/dl) combined with hyperinsulinemia (mean insulin value, 43.9 ± 28.7 lU/ml) within 48 h. Proinsulin was measured in seven patients and was abnormal ([5 pmol/l) in all seven. C-peptide was measured in 15 patients and it was abnormal in 10 (66.6%). Several localizing studies were used throughout the study period. CT scan was better at detecting tumors in body and tail, MRI for tumors in head, angiography for tumors in body and tail, and endoscopic ultrasound detected the same number of tumors in head and body. There was no significant differences. Because we do not analyze the population who ultimately did not have an insulinoma, only sensitivity and positive predictive values can be calculated (Table 1). The used localizing studies showed different accuracy to identify tumors located in the various anatomic pancreatic areas (Table 2). Our standard protocol for angiography includes the calcium stimulation test if the tumor is not easily visualized in the images. From the 20 patients who underwent
Table 1 Sensitivity and PPV of the localizing studies used throughout the study period Study
N
CT
26
45
MRI
14
91
84
EU
14
84
91
Angiography Ca stimulation
21 9
82 100
PPV positive predictive value
Sensitivity (%)
PPV (%) 100
77 88.8
angiography, the calcium stimulation test was performed in 9. In 4 patients the stimulation test confirmed a suspected lesion, and in 5 patients the results indicated regional hypersecretion of insulin. In 4 of these 5 patients a tumor was found in the anatomic region of insulin hypersecretion at surgery, and in 1 patient no tumor could be found. Intraoperative ultrasound was used as a part of our surgical exploration in 30 patients. In all but one case, surgical exploration with intraoperative ultrasound was able to find the disease. In all patients the histopathological study confirmed the presence of a neuroendocrine tumor. Immunohistochemistry was positive for insulin in 29 of 34 stained tumors. Three tumors were diagnosed as malignant (8.8%). Diagnosis of malignancy was based on capsular invasion in two cases and lymph node metastases in one. In one patient (2.9%) after extensive exploration, a tumor could not be found. Subtotal pancreatectomy was performed and the patient became normoglycemic after surgery. The pathological analysis along with the clinical results suggested nesidioblastosis. We had four patients with MEN-1, three of whom had multiple tumors. Immunohistochemical analysis confirmed multiple insulinomas in two patients and an insulinoma associated to six additional nonfunctioning endocrine tumors in one patient. The maximum median size of the tumors by pathology was 2 (range, 1.1–5.2) cm. The laparoscopic approach was initially selected in 21 patients (61.8%) and 13 cases (38.2%) were treated by open surgery. Conversion from laparoscopic to open was required in seven patients (33%), of whom five had tumors immersed in the pancreatic head, one had a malignant insulinoma, and in the last case we were unable to localize the tumor during the laparoscopic evaluation. Several surgical procedures were performed. Surgical techniques divided by approach and tumor location are shown in Tables 3 and 4. A total of 33 patients (97%) became normoglycemic after surgery and 1 patient developed diabetes. One morbidly obese patient died from pulmonary thromboembolism 15 days after surgery. The median hospital stay was 10 (range, 3–61) days.
123
1968 Table 3 Surgical procedures performed with the different approaches
World J Surg (2009) 33:1966–1970
Surgery
Laparoscopy (14)
Conversion (7)
Open (13)
Enucleation
9
4
4
Distal pancreatectomy/splenectomy
4
1
2
Splenic sparing distal pancreatectomy
1
1
2
Pancreaticoduodenectomy
0
1
2
Subtotal pancreatectomy
0
0
2
Central pancreatectomy
0
0
1
Table 4 Surgical approach related to tumor localization site
Enucleation Laparoscopic
Open
Conversion
Laparoscopic
Open
Conversion
Head
1
4
4
0
2
1
Neck
1
0
0
0
1
0
Body
4
0
0
3
3
0
Tail
3
0
0
2
3
2
Total
9
4
4
5
9
3
The complication rate was 67.6%. The most common complication was pancreatic fistula, of which there were 13 cases (38.2%): 3 of 13 in open, 4 of 14 in laparoscopic, and 6 of 7 in conversions. The next most common complication was peripancreatic collection, which occurred in four patients (11.8%): three in the open and one in the laparoscopic group. Two patients required early reoperation: one due to bleeding and one due to an uncontrolled pancreatic fistula. Table 5 shows patients’ characteristics and complications divided by type of surgery. One patient with a malignant insulinoma died from the cancer 16 months after surgery. Another patient with malignant insulinoma is currently asymptomatic after 4 years of follow-up, and the other patient was lost for follow-up 20 months after surgery. In a mean follow-up of 3.5 ± 4 years, all patients with benign tumors are alive and normoglycemic, including the four patients with MEN syndrome. Table 5 Comparison of variables according surgical approach
Pancreatic resection
Discussion Since the first laparoscopic cholecystectomy was performed in 1983, there have been important advances in the techniques and the technology, allowing the use of laparoscopy to treat a variety of surgical procedures, including tumor resection in solid organs, such as the liver, adrenals, and pancreas [4, 5]. Pancreatic surgery has evolved slowly but consistently throughout the time. General features of insulinomas, such as small size, high benignity rate, and sporadic nature, have made these tumors suitable for laparoscopic resection [6]. It has been recognized that pancreatic insulinomas have an even distribution in the head, body, and pancreatic tail [7]. Tumors located in the pancreatic head and the uncinate process often have a hidden location and require extensive mobilization of the pancreatic head and duodenum to be resected. Tumors in the body and tail, in theory, may be
Laparoscopic (14)
Open (13)
Conversion (7)
Age (year)
42.4 ± 13.6
39.1 ± 13.5
35.1 ± 8.2
Male/female ratio
4/10
7/6
3/4
MEN-1*
0
4
0
Malignancy, n
0
2
1
Blood lost (ml)
210.8 ± 317.4
461.5 ± 368.1
430.7 ± 357.7
Pancreatic fistula**
4
3
6
Fluid collections
1
3
0
Other
4
4
1
Surgical mortality, n
0
1
0
Mean hospital stay (days)
9.8 ± 11.6
17.8 ± 16.8
16.8 ± 8.8
Complications, n
* P \ 0.05, v2 test (selection bias); ** P = 0.02, v2 test (0.02)
123
World J Surg (2009) 33:1966–1970
1969
easily exposed and resected by enucleation or distal pancreatectomy. In our experience we have found that tumors in these locations can be more commonly treated by laparoscopic resection. In 1993 Gagner and Pomp [8] reported the first successful laparoscopic distal pancreatectomy and pancreatoduodenectomy. During the following 10 years, very few reports of laparoscopic pancreatic surgery were published. Table 6 summarizes the most representative reports of the laparoscopic management of pancreatic endocrine tumors during this time period. In this series, the most sensitive preoperative localization study was angiography. Unfortunately this is an invasive study with a potential for complications. Many centers have suggested that endoscopic ultrasound is the best localizing modality for insulinomas [9]. Endoscopic ultrasound has been incorporated to our diagnostic armamentarium during the last 5 years and has been used with an increasing frequency. We hope that we can get enough experience with this technique in the near future. The most common complications of laparoscopic pancreatic surgery are pancreatic leaks. Frequency of pancreatic fistulas seems to be higher in laparoscopic (23%) than in open surgery (15.3%) [10–12]. Because most of surgical morbidity is associated with fluid collections, different surgical alternatives have been proposed to avoid or diminish this complication. In a recent meta-analysis conducted by Knaebel [13], several techniques to prevent pancreatic leaks after distal pancreatectomy were evaluated. There were no significant differences between handsewn and stapler closure of the pancreatic stump. Fibrin glue applied to the cut end of the pancreatic remnant appeared to reduce the leak rate. We have used all of these different surgical strategies to avoid fluid collections. At the beginning of our series, the pancreatic bed was left open after ultrasonic energy had successfully controlled the hemostasis after enucleation. Distal pancreatectomy
was performed using staplers with no additional reinforcement. Subsequently we started closing the pancreatic bed after tumor enucleations and over sewing the stapler line, using interrupted sutures of nonabsorbable material. Fibrin glue on all transected tissue also was used in some cases. Although the numbers are very small for any comparison, it is our impression that the combination of closure of the pancreatic parenchyma with sutures and the use of fibrin glue has diminished the leak rate. We also have found that pancreatic fistulas were more common in the patients who needed conversion from laparoscopic to open surgery. This may be related to the fact that these patients required a more extensive dissection. To ensure good control of any fluid collection, we have adopted a more liberal use of drain tubes. In all cases, closed suction drains (Blake or Jackson Pratts) where placed close to the enucleation bed or the pancreatic stump. Intraoperative ultrasound has been very helpful in many aspects, such as the intraoperative identification of the tumor and the assessment of the relationship between the tumor and the pancreatic duct. Before the laparoscopy era, it was stated that careful intreoperative exploration, both visually and by palpation, aided with intraoperative ultrasound and was the best localizing modality for pancreatic insulinomas [9]. We and others feel that only well-localized tumors should be laparoscopically explored, but as more experience is gained in the laparoscopic management of these tumors, it may be a day where patients will be explored based on the biochemical diagnosis with no preoperative localization. As shown in this and other series, laparoscopic resection of pancreatic insulinomas is feasible. In the era of evidence-based medicine, a randomized study becomes mandatory to assess the potential value of minimally invasive surgery for the management of these tumors. The low frequency of the disease may be an important limitation to
Table 6 Summary of reports on laparoscopic management of pancreatic insulinomas Study
Year
N
Cure rate (%)
Complications
Conversions
Gagner et al. [14]
1996
10
100
1 bleeding
4/10
Fernandez Cruz et al. [15]
2002
10
Park and Heniford [16]
2002
9
2 fistulas
2/10
2 bleeding
1/9
1 wound infection. 1 fistula Grama´tica et al. [17]
2002
9
100
Lo et al. [18]
2004
10
100
1 fistula
Iihara et al. [19]
2001
7
100
4 fistulas
1/7
Brendes et al. [20]
2000
10
2 fistulas
4/10
1 fistula 1 pleural effusion
0
1 abscess 6/10
123
1970
get the necessary experience and to obtain the number of patients needed for a clear assessment. From this small experience, we have learned that insulinomas located in the body and tail of the pancreas can be resected laparoscopically with good long-term results and that pancreatic leaks are a common complication if the pancreatic bed or stump is not sealed closed.
References 1. Pasieka JL, McLeod MK, Thompson NW et al (1992) Surgical approach to insulinomas. Assessing the need for preoperative localization. Arch Surg 127:442–447 2. Alexakis N, Neoptolemos JP (2008) Pancreatic neuroendocrine tumours. Best Pract Res Clin Gastroenterol 22:183–205 3. Grant CS (2005) Insulinoma. Best Pract Res Clin Gastroenterol 19:783–798 4. Mu¨he E (1986) The first cholecystectomy through the laparoscope. Langenbecks Archiv fu¨r Chirurgie 369:804 5. Gagner M (2003) Pioneers in laparoscopic solid organ surgery. Surg Endosc 17:1853–1854 6. Finlayson E, Clark OH (2004) Surgical treatment of insulinomas. Surg Clin North Am 84:775–785 7. Service FJ, McMahon MM, O’Brien PC et al (1991) Functioning insulinomas-incidence, recurrence, and long-term survival of patients: a 60-year study. Mayo Clin Proc 66:711–719 8. Gagner M, Pomp A (1994) Laparoscopic pylorus-preserving pancreatoduodenectomy. Surg Endosc 8:408–410 9. Norton JA, Cromack DT, Shawken TH et al (1988) Intraoperative ultrasonographic localization of islet-cell tumors: a prospective comparison to palpation. Ann Surg 207:160–168
123
World J Surg (2009) 33:1966–1970 10. Rothmund M, Angelini L, Brunt LM et al (1990) Surgery for benign insulinoma: an international review. World J Surg 14:393–398 11. Fernandez-Cruz L, Saenz A, Astudillo E et al (2002) Outcome of laparoscopic pancreatic surgery: endocrine and nonendocrine tumors. World J Surg 26:1057–1065 12. Shimizu S, Tanaka M, Konomi H et al (2004) Laparoscopic pancreatic surgery: current indications and surgical results. Surg Endosc 18:402–406 13. Diener MK, Heukaufer C, Schwarzer G et al (2008) Pancreaticoduodenectomy (classic Whipple) versus pylorus-preserving pancreaticoduodenectomy (pp Whipple) for surgical treatment of periampullary and pancreatic carcinoma. Cochrane Database Syst Rev 2:CD006053 14. Gagner M, Pomp A, Herrera MF (1996) Early experience with laparoscopic resections of islet cell tumors. Surgery 120:1051– 1054 15. Fernandez-Cruz L, Saenz A, Astudillo E et al (2002) Laparoscopic pancreatic surgery in patients with chronic pancreatitis. Surg Endosc 16:996–1003 16. Park AE, Heniford BT (2002) Therapeutic laparoscopy of the pancreas. Ann Surg 236:149–158 17. Gramatica L Jr, Herrera MF, Mercado-Luna A et al (2002) Videolaparoscopic resection of insulinomas: experience in two institutions. World J Surg 26:1297–1300 18. Lo CY, Lo CM, Fan ST (2000) Role of laparoscopic ultrasonography in intraoperative localization of pancreatic insulinoma. Surg Endosc 14:1131–1135 19. Iihara M, Kanbe M (2001) Okamoto T et al. (2001) Laparoscopic ultrasonography for resection of insulinomas. Surgery 130:1086– 1091 20. Brendes FJ, Cuesta MA, Kazemier G et al (2000) Laparoscopic detection and resection of insulinomas. Surgery 128:386–391
