A, Solitary, benign, intrapancreatic insulinoma. B, Simple enucleation of an insulinoma. C, Localization of an insulinoma with laparoscopic intraoperative ultrasonography.
Customize your JAMA Network experience by selecting one or more topics from the list below.
Jaroszewski DE, Schlinkert RT, Thompson GB, Schlinkert DK. Laparoscopic Localization and Resection of Insulinomas. Arch Surg. 2004;139(3):270–274. doi:10.1001/archsurg.139.3.270
Copyright 2004 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2004
Laparoscopic localization and resection of insulinomas are feasible and safe and may decrease morbidity and the length of hospitalization.
Case series identified through retrospective medical record review.
Mayo Clinic, Scottsdale, Ariz, and Rochester, Minn, a single-institution tertiary care medical center.
Nine patients (6 men and 3 women; mean age, 54 years) who underwent attempted laparoscopic insulinoma resection between September, 1997, and April, 2002.
Laparoscopic localization and resection of insulinoma tumors of the pancreas.
Main Outcome Measures
Intraoperative complications, conversion to open laparotomy, and length of hospitalization.
Preoperative localization was successful in 5 patients (56%). Four conversions to open exploration were required because of an inability to identify the tumor (n = 3) or to perform a safe laparoscopic resection owing to proximity to the portal vein (n = 1). Laparoscopic intraoperative ultrasonography facilitated identification of lesions (mean tumor size, 2.1 cm) in 4 (80%) of 5 patients; 3 patients underwent successful laparoscopic resection. Average hospital stay was shorter after laparoscopic resection (4.5 vs 7.0 days in uncomplicated cases). At follow-up (mean, 29 months; range, 3-57 months), 1 patient who underwent laparoscopic enucleation had recurrent hypoglycemia.
Laparoscopic localization and resection of insulinomas are feasible and safe. Laparoscopic ultrasound aids successful insulinoma localization and laparoscopic resection. Preoperative localization is not required, but previous knowledge of tumor location helps focus intraoperative ultrasound and limit pancreatic mobilization. Excluding patients with pancreatic leaks (observed in both groups), hospital stay and time to recovery may be shortened by using laparoscopic insulinoma resection.
Insulinoma is the most common functional neuroendocrine tumor of the pancreas,1,2 with an incidence of 4 per million person-years.3 Patients often initially have symptoms of neuroglycopenia (eg, confusion, somnolence, and visual disturbances) and hypoglycemia-induced catecholamine release (eg, tachycardia, sweating, and tremor). Diagnosis is confirmed biochemically during a supervised 72-hour fast. A diagnosis of insulinoma is made in symptomatic patients who demonstrate hypoglycemia, an inappropriate insulin level, and increased concentrations of C peptide.1,3 Plasma screening results for both classes of sulfonylureas must be negative during end-of-fast studies.
In most patients, the lesions are solitary, benign, and intrapancreatic (Figure 1, A). Excision of the insulinoma is the treatment of choice. Ninety percent of these tumors are benign and can be treated with simple enucleation (Figure 1B).1 At Mayo Clinic, 62% of nearly 200 insulinomas were enucleated.
The advantages of minimally invasive procedures are well known, and technical advances in laparoscopic surgery have led to its increasing use in the treatment of pathologic intra-abdominal conditions. Advanced laparoscopic techniques have been used to resect islet cell tumors in a few patients worldwide.4-10 Patients who undergo open exploration of the pancreas have considerable postoperative pain, resulting in extended hospitalization. Laparoscopic pancreatic procedures should shorten hospital stays and possibly decrease patient morbidity.
Frequently, preoperative localization of an insulinoma is not possible. In a few highly selected patients, intraoperative ultrasonography (IOUS) has been reported to be more sensitive than preoperative US and other intraoperative techniques for localizing insulinomas.11-13 In most such cases, IOUS can be used to localize insulinomas more precisely and, thus, to facilitate safe resection (Figure 1C).11-14
We describe our experience with laparoscopic localization and resection of insulinomas of the pancreas, and we review the reported success with this procedure at other institutions.
Nine patients underwent attempted laparoscopic resection of insulinomas between September 1997 and April 2002 at Mayo Clinic (6 patients were treated in Scottsdale and 3 were treated in Rochester, Minn). The diagnosis of insulinoma was made using a supervised 72-hour fast. All 9 procedures were initiated with the laparoscope, and laparoscopic IOUS (LIOUS) was conducted in 5 of the 9 patients. A review of the clinical records examined the variables of age; sex; preoperative localization studies; location, type, and site of the insulinoma; type of procedure; operative time; postoperative stay; morbidity; use of IOUS or LIOUS; and follow-up for recurrence. Additional follow-up was provided by telephone and written contact with patients.
Two surgical approaches were used. Three of the 9 patients underwent planned laparoscopic distal pancreatectomy. They were placed in the lateral decubitus position, and 3 trocars (10 mm) were used along the left subcostal region to facilitate the planned distal pancreatectomy. For distal pancreatectomy, the pancreas was dissected carefully away from the splenic vessels. The tail of the pancreas was freed and transected with a linear stapler. Six patients underwent planned pancreatic exploration and enucleation of the insulinoma if possible. They were placed in the dorsal lithotomy position, and 5 or 6 trocars were used. The lesser sac was opened by dividing the gastrocolic ligament with ultrasonic shears (Ethicon LCSC5; Ethicon Endosurgery, Cincinnati, Ohio). Attachments between the pancreas and the posterior aspect of the stomach were separated. When available, a 5-MHz laparoscopic ultrasonic probe (model 8566-S; B-K Medical Systems, Wilmington, Mass) was used to scan the pancreas. Enucleated tumors were removed in a bag through the 10-mm trocar site. One or 2 drains were left in the pancreatic vicinity.
Nine patients (6 men and 3 women) underwent attempted laparoscopic surgical treatment of insulinomas. Mean patient age was 54 years (range, 34-73 years). These tumors were located in the uncinate process (1 patient), the proximal body (2 patients), the junction of the tail and the body (2 patients), or the tail of the pancreas (4 patients). Mean tumor size was 2.1 cm (range, 1.5-3.0 cm).
Preoperative imaging was conducted in 8 patients (Table 1). Computed tomography (CT) in 7 of 9 patients showed evidence of a mass or "fullness" in the pancreas in 3 patients (43%). In 1 patient, an octreotide scan failed to show any areas of abnormal activity. Endoscopic US in 1 patient accurately identified the lesion. One patient had abdominal US findings that showed a mass; however, subsequent CT findings were negative.
Lesions were identified visually in 5 patients, and LIOUS was available to confirm visual findings for only 2 of these 5. Both patients underwent successful laparoscopic procedures (2 enucleations). Two other patients underwent planned laparoscopic distal pancreatectomy for tumor resection. The fifth patient had an apparent lesion at the superior border of the pancreas that could not be differentiated clearly from the splenic artery. Conversion to an open procedure showed that the lesion was an insulinoma resting on the portal vein.
Four patients did not have visible tumors. Laparoscopic IOUS was available for use with 3 of these patients and facilitated the identification of lesions in 2. One lesion was successfully enucleated from the tail of the pancreas. In the second patient, a lesion in the uncinate process adjacent to the superior mesenteric vein required conversion to celiotomy to facilitate safe dissection. The third patient had no lesion visualized, and findings with LIOUS were negative. At conversion to celiotomy, an insulinoma was palpated at the distal tail of the pancreas in the splenic hilum. Subsequent IOUS verified this lesion. For the fourth patient with no visible tumor, LIOUS was not available. Preoperative endoscopic US identified a lesion in the tail of the pancreas, and the patient underwent laparoscopic distal pancreatectomy. The specimen did not contain the lesion, and the operation was converted to celiotomy. The lesion was found proximal to the line of transection and was removed using further proximal resection.
The mean operative time was 215 minutes (range, 160-285 minutes) for the laparoscopic approach and 203 minutes (range, 179-225 minutes) for procedures converted to celiotomy. The average hospital stay was 9.4 days (range, 2-29 days) for patients whose lesions were resected laparoscopically (Table 2) and 11.8 days (range, 5-26 days) for patients whose surgical procedures were converted to celiotomy (Table 3). Excluding 1 patient in each group with a pancreatic leak, the average hospital stay was 4.5 days after laparoscopic surgery and 7.0 days after celiotomy.
Three patients had early postoperative complications, 2 of which were due to postoperative pancreatic leaks. One leak occurred after laparoscopic enucleation of a lesion, and the other occurred after open distal pancreatectomy. Both leaks resolved with conservative therapy. The third patient was hospitalized for 9 days because he experienced persistent fever of unknown cause and an elevated white blood cell count after open enucleation of a lesion.
Patient follow-up was 3 to 57 months (mean, 29 months). Eight patients remained asymptomatic. One patient who had laparoscopic enucleation was hyperglycemic in the immediate postoperative period and had recurrent hypoglycemia at 6-week follow-up. Biochemical testing results suggested recurrent disease, but CT findings did not reveal an insulinoma. No additional surgical treatment was undertaken. This patient's symptoms were minimal, and his blood glucose concentrations remained stable with use of diazoxide (50 mg twice daily).
Benign insulinomas of the pancreas are ideally suited for laparoscopic resection. They generally are singular, small, and well encapsulated.7 With advancing technology and increasing surgical experience, laparoscopic resection of the pancreas undoubtedly will contribute increasingly to the care of patients with benign insulinomas. Laparoscopic resection is technically demanding and, thus, is challenging even for experienced surgeons with advanced laparoscopic skills.
The lack of tactile sensation during laparoscopic surgical procedures for insulinoma is a concern. Detection of an insulinoma during a laparoscopic procedure depends solely on visualization or the use of LIOUS. In all but 1 of our patients, LIOUS helped localize and resect the tumors.
Similar results have been reported from other series. Berends et al5 found that LIOUS failed to localize insulinomas in 1 of 10 patients, and Gagner et al4 were unable to visualize 2 of the 5 insulinomas with LIOUS in their series. In our patients, the size of the tumor did not always correlate with successful intraoperative localization. The largest lesion (3.0 cm) was not identified with LIOUS but was identified by manual palpation after conversion.
Preoperative knowledge of tumor location helps focus US localization and limits mobilization and exposure of the pancreas.11,13,15 Yet, the use of preoperative imaging techniques for insulinomas is controversial. Many techniques to localize insulinomas have been described, and there is a wide range of reported accuracy with each one. Detection rates of transabdominal US and CT are lower than 50% to 60%.11,15-17 Spiral CT has improved this sensitivity considerably.17 Magnetic resonance imaging has identified 45% to 91% of insulinomas in several series.11,17-19 Endoscopic US seems to be even more sensitive, with preoperative detection rates of 86% to 93%.16,20-22 Invasive techniques such as percutaneous transhepatic venous sampling for insulin are successful in 70% to 95% of cases, and arterial stimulation and venous sampling for insulin identify insulinomas in 75% to 80% of patients.11,13 Laparoscopic IOUS complements preoperative localization techniques and generally yields excellent results.11-14
The patient benefits of minimally invasive surgical procedures are well known. Patients successfully treated in our series realized the advantages that laparoscopic surgery offers over upper abdominal incisions, and they visibly experienced less postoperative pain. By excluding patients with pancreatic leaks, the hospital stay for patients who had undergone laparoscopic resection was shorter than that for patients whose laparoscopic procedure was converted to open exploration (4.3 days vs 7.0 days). However, if patients with pancreatic leaks are included (9.4 days vs 11.8 days), the average stay for patients with laparoscopic procedures improved only slightly. Pancreatic leaks are a well-known complication of pancreatic operations. We noted a pancreatic leak in 1 patient in the laparoscopic group and in 1 in the open group. The patient whose procedure was converted had undergone a distal pancreatectomy, whereas the laparoscopic patient had undergone enucleation of a 1.7-cm tumor in the tail of the pancreas. Both pancreatic leaks were noted about 1 week postoperatively, and they resolved without further surgical intervention.
The published studies on laparoscopic resection of insulinomas include case reports and small series that indicate a similar incidence of complications and conversions. Berends et al5 successfully resected 6 of 10 insulinomas laparoscopically. Five of these were enucleations. They found 2 pancreatic fistulas in the laparoscopic group (total hospitalization, 3-21 days). Gagner et al4 published a series on patients with mixed histologic findings who had pancreatic tumors for which laparoscopic resection was attempted. Five of these were insulinomas, 3 of which were removed successfully by laparoscopy. This series did not report the type of procedures or the complications associated with them. The overall hospital stay for these laparoscopic patients was 4 to 7 days, so patients who had laparoscopic treatment for insulinomas may have been released with minimal complications within 7 days of the procedure. Other case reports4-10 noted minimal complications and hospital stays of 3 to 10 days (Table 4).
Laparoscopic ultrasound probes frequently do not provide the resolution of standard intraoperative probes. Improvements in technology should resolve this problem eventually.
In conclusion, as indicated by the results of this small preliminary series, laparoscopic localization and resection of insulinomas seem to be feasible and safe. With appropriate equipment and expertise, treatment of these tumors in most areas of the pancreas is possible by laparoscopic resection. Conversion to open exploration should be considered for tumors that cannot be identified accurately or for those located close to vascular structures that prevent safe resection. Failure of the laparoscopic procedure did not impede open localization or treatment and did not affect patients negatively. Overall, patients did not experience increased complications as a result of laparoscopic attempts, and, when pancreatic leaks are excluded, they had a shorter hospitalization and a faster return to usual activity. Our sample size is small, however, and 1 patient showed evidence of recurrent disease. Further study is required.
Corresponding author: Richard T. Schlinkert, MD, Division of General Surgery, Mayo Clinic, 13400 E Shea Blvd, Scottsdale, AZ 85259.
Accepted for publication August 5, 2003.