Pancreatic endocrine tumors are uncommon tumors that belong to the family of amine precursor uptake and decarboxylation (APUD) neoplasms. They are called functional if the hormone released is associated with a clinical syndrome and nonfunctional if they are not associated with clinical symptoms. The incidence of functional pancreatic endocrine tumors is approximately 4 per million population per year.1 The most common functional pancreatic endocrine tumors are gastrinomas (20%) and insulinomas (50%).2 The number of nonfunctional pancreatic endocrine tumors diagnosed annually is 1 to 2 per million population.3 Pancreatic polypeptide secreting tumors comprise the majority of nonfunctional pancreatic endocrine tumors. Surgical resection is the only treatment that can possibly cure patients with pancreatic endocrine tumors. Preoperative workup includes establishing the diagnosis clinically and biochemically, evaluating for the presence of multiple endocrine neoplasia type 1 (MEN-1), and imaging to determine the location of the primary tumor or tumors and the presence of metastatic disease. The imaging studies needed to localize and stage pancreatic endocrine tumors depend on the tumor type. Because pancreatic endocrine tumors other than gastrinomas and insulinomas are typically large at diagnosis, noninvasive imaging such as somatostatin receptor scintigraphy, computed tomography (CT), and magnetic resonance imaging (MRI) suffices to detect and stage these tumors. In contrast, hormonally active gastrinomas and insulinomas produce dramatic clinical symptoms, so patients come to medical attention when their tumors are small to occult at the time of initial presentation. In fact, 90% of these tumors are detected when 2 cm or less.2 Gastrinomas cause the Zollinger-Ellison syndrome (ZES), which is characterized by multiple and unusually located peptic ulcers and diarrhea. In sporadic ZES, the apparent primary gastrinomas are located in the duodenum (40%-50%), pancreas (30%-40%), lymph node (10%-15%), and nonpancreaticoduodenal or nodal sites (≤5%).1 The percentage of gastrinomas that are malignant, as defined by metastatic disease, is 60% to 90%.1,3 In contrast to pancreatic gastrinomas, duodenal gastrinomas tend to be small (<1 cm), can be multiple, and are more associated with lymph node rather than hepatic metastases.4 About 20% of patients with ZES have MEN-1, which is an autosomal dominant trait characterized by hyperplasia or tumors of the parathyroids, pituitary, and enteropancreatic endocrine system.1 The tumors in these patients are invariably multiple and typically small; furthermore, the primary tumors are located in the duodenum (70%-95%) and pancreas (10%-25%).1 For almost all ZES patients, gastric acid hypersecretion can be controlled effectively over the long term with proton pump inhibitors such as omeprazole or lansoprazole. The main limitation on long-term survival in ZES patients is tumor growth.5 Clinically, there are two forms of gastrinomas. The aggressive form, which involves about 25% of patients with ZES, is more common in females and patients without MEN-1 and is associated with large pancreatic tumors, liver metastases, short disease duration, and a 10-year survival of 30%.4,5 The nonaggressive form, involving 75% of patients with ZES, is associated more with duodenal gastrinomas, absence of liver metastases, and a 10-year survival of 96%.4,5 For patients with sporadic and localized gastrinomas, surgical exploration with intent for a curative resection is the treatment of choice. The cure rate for these patients is 34% at 10 years.6 In contrast, the role of surgery in patients with ZES and MEN-1 is controversial.1 Because of multiple tumors in these patients, the cure rate with local resection is less than 10%,4 and a cure is not possible without a pancreaticoduodenectomy. Pancreaticoduodenectomy is not routinely indicated in patients with ZES, with or without MEN-1, because a survival benefit with this procedure has not been demonstrated, and the procedure itself can have substantial morbidity.4 However, some surgeons will resect primary tumors larger than 2.5 cm on preoperative imaging in patients with ZES and MEN-1 in an attempt to prevent development of metastatic disease, because development of hepatic metastases correlates with primary tumor size.1 The biochemical diagnosis of ZES is established with elevated fasting serum gastrin levels (>200 pg/mL), increased basal gastric acid secretion (≥15 mEq/h), and a positive secretin stimulation test (>110 pg/mL increase in serum gastrin over basal levels after 0.4 µg/kg is given intravenously).1 The role of preoperative imaging in patients with ZES is to identify the location and number of primary tumors and assess for metastases. The initial imaging study of choice is somatostatin receptor scintigraphy (SRS).7 Gastrinomas express somatostatin receptors to which the somatostatin analog octreotide binds. SRS is the most sensitive noninvasive imaging modality for detecting primary gastrinomas or liver metastases. It detects approximately 58% of primary tumors and 92% of liver metastases,8 but sensitivity depends on tumor size. SRS detects 30% of tumors 1.1 cm or less, 64% of tumors between 1.1 and 2 cm, and 96% of tumors larger than 2 cm.9 SRS misses mainly duodenal gastrinomas, which tend to be small (≤1 cm).9 Because SRS does not provide information on tumor size and gives only an approximate location, CT or MRI, also performed routinely, are useful for detecting larger primary tumors and distant metastases. Single-photon emission computed tomography (SPECT)-CT hybrid imaging and software fusion of SRS with CT or MRI may help localize areas of increased activity on SRS. If results of both SRS and CT or MRI are negative, an invasive localization study such as endoscopic ultrasonography (EUS) or arteriography with ASVS using secretin should be considered. The choice depends on local expertise. EUS has a sensitivity of about 85% in detecting pancreatic gastrinomas and about 43% in detecting duodenal gastrinomas.4,10 Limitations of EUS include low sensitivity for tumors in the duodenal wall, tumors in the pancreatic tail, and small (<0.5 cm) tumors and false-positive results with lymph nodes, splenules, and pancreatic nodules.10 In patients with sporadic ZES, whether EUS should be a routine preoperative study is controversial.4,11 In patients with ZES and MEN-1 who are surgical candidates, there is more agreement that EUS should be included in the preoperative workup to identify additional pancreatic endocrine tumors and metastatic lymph nodes.4 The advantage of ASVS over anatomic imaging is that ASVS is a functional study, so its sensitivity for detecting gastrinomas does not depend on tumor size. Presently the indications for arteriography with ASVS in patients with ZES are controversial. A relative indication is failure to localize a gastrinoma with noninvasive imaging studies. The sensitivity of all preoperative imaging techniques in detecting duodenal gastrinomas is low. The most sensitive technique for detecting duodenal gastrinomas is intraoperative duodenotomy.12 A stronger relative indication is to identify the tumor or tumors secreting gastrin in patients with ZES and MEN-1 who are surgical candidates, because these patients typically have multiple pancreatic endocrine tumors.11,13 Insulinomas cause symptoms of hypoglycemia, which can be classified as neurologic or adrenergic. Neurologic symptoms include visual disturbances, confusion, altered consciousness, and less commonly seizures. Adrenergic symptoms include sweating, tremulousness, weakness, and palpitations. Unlike gastrinomas, insulinomas are exclusively found in the pancreas and are distributed evenly throughout the organ. Five to 15% of insulinomas are malignant, as defined by metastatic disease.1 Most insulinomas are solitary. In the 2% to 10% of patients with multiple insulinomas or pancreatic endocrine tumors, one should evaluate for MEN-l.1 About 4% of patients with insulinomas have MEN-1.2 Because of the characteristic symptoms, insulinomas found at surgery are small: 90% less than 2 cm, 66% less than 1.5 cm, and 40% less than 1 cm.2 For insulinoma patients without metastases, surgical resection is curative. Preoperative workup includes establishing the diagnosis of insulinoma and obtaining noninvasive imaging. The diagnosis of insulinoma is confirmed by showing inappropriately high serum insulin levels with simultaneous symptomatic hypoglycemia during a supervised 72-hour fast and excluding other causes of hypoglycemia. Patients with insulinoma typically have serum insulin levels 6 µU/mL or higher.14 The most common definition of hypoglycemia is a blood sugar level less than 45 mg/dL during a fast.14 Insulinoma patients also should have an elevated plasma C-peptide level (≥200 pmol/L) and no sulfonylurea in the plasma. There is agreement in using cross-sectional imaging such as CT or MRI to assess for metastatic disease, but the necessity for preoperative imaging to localize insulinomas in cases of presumed benign solitary insulinoma is controversial.15,16 Because insulinomas are exclusively intrapancreatic, careful exploration and assessment with direct palpation and intraoperative ultrasound allow the surgeon to identify the vast majority of tumors. In fact, the sensitivity of intraoperative ultrasound in localizing insulinomas is 86% to 98% in experienced hands.14,17, Nevertheless, preoperative localization of insulinomas is useful for several reasons. First, it helps the surgeon decide whether an open or laparoscopic approach is feasible and whether enucleation or resection is required to remove the tumor.18–21 Second, if surgical exploration fails to localize an insulinoma, blind distal pancreatectomy is no longer recommended.22 Positive preoperative localization is a requisite for a reoperation for insulinoma. Third, in patients with MEN-1, functional localization with ASVS can identify the tumor or tumors among the multiple pancreatic endocrine tumors responsible for hyperinsulinism. Fourth, because data from ASVS are usually concordant with intraoperative ultrasonography, preoperative data from ASVS can be helpful when expertise with intraoperative ultrasonography is not available or when findings are particularly subtle even to experienced ultrasonographers. Finally, in the differential diagnosis of hypoglycemia with hyperinsulinism, there is an entity called noninsulinoma pancreatogenous hypoglycemia syndrome (NIPHS).23 Histologically, it is characterized by nesidioblastosis, which refers to islet cell hypertrophy and neoformation of islet cells from pancreatic exocrine duct cells.23 Clinically, NIPHS is characterized by postprandial rather than fasting hypoglycemia. In the Mayo Clinic series, NIPHS accounted for 9% of their hyperinsulinemic hypoglycemia patients.14 ASVS showing an insulin gradient in all pancreatic arterial distributions allows metabolic confirmation of this diagnosis. Preoperative localization begins with noninvasive imaging such as CT, MRI, and/or transabdominal ultrasonography. The sensitivities of these noninvasive modalities are highly variable in different series but are typically less than 50% to 65%.14,17 If noninvasive imaging is negative, one should consider invasive modalities such as EUS or arteriography with ASVS using calcium. Again, the choice depends on local expertise. The sensitivity of EUS in localizing insulinomas is about 80%.10 The sensitivity in the pancreatic tail is lower than that in the pancreatic head.
Arteriography and Arterial Stimulation with Venous Sampling for Localizing Pancreatic Endocrine Tumors
Indications
Gastrinoma
Insulinoma