Neuroendocrine Tumors of the Gastrointestinal Tract

EpidemiologyEtiologyPathologyNatural History and DiagnosisTreatmentReferences

Neuroendocrine tumors manifest in the gastrointestinal tract mainly as carcinoid and pancreatic islet-cell tumors. They comprise an interesting group of rare neoplasms that are derived from neuroendocrine cells interspersed within the gastrointestinal system amd throughout the body. Neuroendocrine tumors are well known for producing various hormonal syndromes and for their indolent clinical course in most patients, although some of these tumors do not produce hormones of clinical significance. Patients may have symptoms for many years before the diagnosis is suspected and confirmed. Therefore, the index of suspicion must be high in order to diagnose these tumors in a timely fashion.

Symptoms are caused by hormonal excess, local tumor growth, or metastatic spread. Therapy should be dictated by the severity of symptoms and the pace of the disease. Surgical extirpation remains the only curative modality for localized disease, but palliation of hormone-related symptoms can be achieved with the somatostatin analog octreotide (Sandostatin) in a significant proportion of patients. Combination chemotherapy, biologic response modifiers, vascular occlusion treatment, and newer somatostatin analogs are being used for the treatment of metastatic disease.

Epidemiology

Neuroendocrine tumors constitute approximately 2% of all malignant tumors of the gastrointestinal system [1]. These tumors are particularly rare in pediatric patients. It is estimated that fewer than 2,000 new cases are diagnosed yearly in the United States.

Carcinoid tumors are more common than pancreatic islet-cell carcinomas (65:35 incidence). Although approximately 600 new cases of small-intestine carcinoids are seen yearly in the United States, autopsy data from the Mayo Clinic indicate that the incidence of these tumors may be approximately 650 cases per 100,000 people [1]. Of the noncarcinoid tumors, insulinomas and gastrinomas have the highest incidence, whereas other types are extremely rare. A center in Belfast has established a registry for tumors of the gastroenteropancreatic endocrine system [2]; these investigators have reported incidences of 0.9 and 0.4 cases per million individuals per year for insulinomas and gastrinomas, respectively.

Etiology

The precise etiology of neuroendocrine tumors is not well understood. Some insight into the molecular biology of these tumors can be gained by studying a subset of tumors that occurs as part of the multiple endocrine neoplasia type I (MEN I) syndrome. In 1954, Wermer recognized that a neoplastic disorder involving the anterior pituitary gland, parathyroids, and pancreatic islet cells was familial and transmitted in an autosomal dominant fashion [3].

More recently, Larsson et al reported linkage of the MEN I gene to the muscle phosphorylase locus on chromosome 11q13 [4]. Using another gene known to be localized to 11q13 (INT2), Bale et al found similar linkage of the MEN I gene with this gene locus [5]. Radford et al investigated DNA isolated from tumors and somatic tissues in 12 patients with MEN I and found loss of heterozygosity markers mapped to chromosome band 11q13 in 9 (82%) of 11 informative tumors [6]. There was no allelic loss from other chromosomes. Such a high incidence of chromosomal deletion involving 11q13 suggests that this region is important in the oncogenesis of neuroendocrine tumors. The same deletion on chromosome 11q13 (PYGM locus) was subsequently demonstrated in sporadic islet-cell tumors, pituitary tumors, and parathyroid tumors [7,8]. Preliminary data suggest that the candidate MEN I gene may be a tumor-suppressor gene and that deletions or mutations of this gene may play a major role in the development of neuroendocrine tumors [9].

Pathology

The Neuroendocrine Concept

Advances in the fields of histochemistry, electron microscopy, and immunocytochemistry have influenced which cells can be classified as neuroendocrine. Neuroendocrine tumors originate from cells that are capable of amine precursor (such as dopa and 5-hydroxytryptophan) uptake and decarboxylation (APUD cells)[10]. As a result, these tumors have high intracellular levels of carboxyl groups and nonspecific esterase. The latter is used as a neuroendocrine marker.

As the name suggests, neuroendocrine cells were previously considered to be derived from the neural crest, from which they subsequently migrate throughout the body. However, extensive embryologic investigations have shown that while certain APUD cells are, indeed, derived from the neural crest, the neuroendocrine cells of the digestive tract cannot be traced back to the neural ectoderm [11]. It is now clear that similar cell phenotypes can arise from different germ layers, and the search for a common origin of neuroendocrine cells has become less important as more phenotypic characteristics (markers) of these cells are identified by new immunocytochemical methods [11]. Regardless of embryonic origin, these cells have a common genetic program for the expression of several biochemical markers of neuroendocrine function [12].

Thus, the term “neuroendocrine” is used to define cells by their secretory products and cytoplasmic proteins rather than by their localization and embryologic derivation. The neuroendocrine cell system, therefore, includes all neuronal and endocrine cells that share a common phenotypic program characterized by simultaneous expression of certain marker proteins (general neuroendocrine markers) and cell type–specific hormonal products. These peptides are contained within membrane-bound vesicles from which they are released by a process of regulated exocytosis in response to external stimuli.

Neuroendocrine cells differ from neurons in that axons and specialized nerve terminals are absent in the former, and consequently, their mode of transmission is endocrine or paracrine rather than synaptic. The neuroendocrine cells normally form either small organs, distinct cell clusters within other tissues, or a network of cells dispersed in the lung and gut mucosa [11,13]. In addition to carcinoids and islet-cell neoplasms, other tumors that arise from these cells include small-cell carcinoma (of pulmonary and extrapulmonary origin), medullary carcinoma of the thyroid gland, neuroblastoma, and Merkel-cell tumors of the skin.

Diagnostic Pathologic Features

Gross Histology: Carcinoid tumors may show any of a variety of architectural patterns that have been referred to as insular, trabecular, glandular, mixed, or undifferentiated [14,15]. Typically, there is extensive necrosis of both sheets of cells and individual cells. In some tumors, such as midgut carcinoids, there is extensive stromal fibrosis surrounding tumor islets. However, other aspects of tumor cell appearance are uniform: a pale pink or granular cytoplasm and a round to oval nucleus with stippled chromatin. There is minimal mitotic activity, cytologic atypia, or nuclear pleomorphism, except in undifferentiated tumors, and secretory granules can be demonstrated under the electron microscope.

Such histologic features are inadequate to distinguish pancreatic endocrine tumors from carcinoids and malignant from benign neuroendocrine tumors. The actual clinical behavior, such as infiltration of adjacent organs or structures and metastases to regional lymph nodes or the liver, is what usually substantiates the malignant nature of the tumor. However, multivariate analysis has demonstrated that histologic type and primary site have independent prognostic significance. (Patients with glandular and undifferentiated histology and foregut or hindgut tumor location have a worse outcome [16,17].)

Further Diagnostic Techniques

For tumors that show the classic morphologic properties described above, use of special techniques merely confirms the diagnosis. Frequently, however, a pathologist will need to confirm the morphologic impression with standard histochemical, electron microscopic, or immunohistochemical analysis.

Standard Histochemical Analysis: This relatively old staining procedure relies on the ability of well-differentiated neuroendocrine tumors to stain positively with silver in the presence of endogenous cellular reducing agents (argentaffin cells) or with the addition of exogenous reducing reagents (argyrophilic cells). This silver stain rarely is positive in poorly differentiated lesions.

Electron microscopy reveals the neurosecretory or dense core granules of the neuroendocrine cells. In this setting, however, use of this method is uncommon, as it is limited by substantial cost, the extensive tissue preparation required, and the small number of cells that can be examined.

Immunohistochemical Markers: During the past 15 years, identification of a number of components of neuronal and neuroendocrine cells by immunohistochemical methods has markedly affected the classification of tumors as neuroendocrine. These methods provide reliable information at low cost, require relatively simple tissue preparation, and have a rapid turnaround time. Several general neuroendocrine markers associated with cytoplasmic proteins, small secretory vesicles, or dense-core secretory granules have thus been established [18-20].

Cytoplasmic Constituents: Neuron-specific enolase, a glycolytic enzyme found in the cytosol, is the best known marker of cells with neuroendocrine differentiation. However, this marker is nonspecific, as it stains positive on fibroadenomas of the breast, renal-cell carcinoma, and certain malignant lymphomas. Its positivity is therefore not considered to be diagnostic, and consequently, this reagent is also known as nonspecific esterase [13].

Secretory Vesicle Membrane Constituents: Synaptophysin is an integral membrane glycoprotein that is involved in calcium binding and occurs in presynaptic vesicles of neurons and small vesicles of normal and neoplastic neuroendocrine cells [19].

Granule Contents: Chromogranins A, B, and C are acidic proteins that serve as powerful universal markers for neuroendocrine tissues and tumors. Chromogranins are a family of soluble proteins located in large (dense-core) secretory granules. The most frequently used antibody is against chromogranin A [20].

Plasma membrane constituents include receptors for peptides or neurotransmitters (somatostatin, glutamate, and gamma-aminobutyric acid), and neural cell adhesion molecules (NCAMs), the most important of which appear to be NCAM and L-1 [11]. Somatostatin receptors are present in 82% of carcinoid tumors and in 67% to 100% of islet-cell tumors [21]. Moreover, most metastases of primary somatostatin receptor–positive tumors are also positive for this peptide. Somatostatin inhibits peptide hormone secretion of most neuroendocrine cells by a mechanism that involves the suppression of secetory pathways that are dependent on cyclic adenosine monophosphate and the disruption of the second messenger function of intracellular calcium [22].

Somatostatin receptor status correlates highly with the ability of long-acting somatostatin analogs, such as octreotide, to inhibit in vivo hormone secretion [23]. The presence of these receptors enables in vivo imaging of tumors using indium-111-labeled octreotide. Somatostatin analogs are thus used in both imaging and treatment of neuroendocrine tumors.

Growth Factors and Nuclear Antigens: The expression of growth factors and the presence of nuclear antigens, although not unique to neuroendocrine tumors, are of particular interest. Ki-67 is a monoclonal antibody against a nuclear antigen present in proliferating cells [24]. Patients who have tumors with a high index for Ki-67 were found to have a significantly shorter survival than those whose tumors are low in Ki-67 content [25].

Various growth factors have been studied, including platelet-derived growth factors, transforming growth factors-alpha and -beta (TGF-alpha and -beta), fibroblast growth factors, and epidermal growth factors, and the data suggest that platelet-derived growth factors may be involved in the autocrine stimulation of neuroendocrine tumor cells and stimulation of stromal cell growth through paracrine or autocrine mechanisms [26,27]. Moreover, the TGF family known for its fibroblast-stimulating ability (TGF-beta) may play a role in the proliferation of fibroblasts and their production of matrix in the fibrotic lesions associated with carcinoids [28]. In endocrine pancreatic tumors, expression of CD44 correlates with the tumor's ability to give rise to lymph node metastases, and may play a vital role in determining the fate of metastasizing cells [29].

Different types of neuroendocrine cells share many specific properties and express several proteins in common, but the expression of any one marker protein is not an absolute criterion. Thus, there is no “universal” marker. In addition to the above substances that serve as general markers, immunoperoxidase staining has shown that tumors can synthesize numerous biogenic amines and peptides. These include 5-hydroxytryptamine, 5-hydroxytryptophan, serotonin, insulin, glucagon, somatostatin, vasoactive intestinal polypeptide, growth hormone, corticotropin melanocyte-stimulating hormone, gastrin, pancreatic polypeptide, calcitonin, and substance P. Clinical signs of hormone hyperfunction may therefore be correlated to their precise source [30]. Tumors are named primarily according to the predominant peptide secreted that can be related to the clinical features.

Natural History and Diagnosis

Neuroendocrine tumor cells can synthesize and secrete a variety of physiologically active peptides that generate disabling symptoms. However, some tumors produce peptides only after a prolonged period of inactivity [1]. In the case of islet-cell carcinoma, the dominant hormone and hormonal syndrome can also change [31].

Patients with nonsecreting tumors usually present with a large tumor burden but may lack any of the cancer-associated cachexia or morbidity that may be associated with smaller, peptide-secreting tumors. Thus, some patients with nonsecreting tumors are diagnosed with a large pancreatic mass or enlarged liver during a routine physical examination, whereas others with small but hormonally active tumors can have significant debilitating symptoms that require prompt diagnostic evaluation. In a study of 154 patients with gastrointestinal carcinoid tumor, 60% of tumors found at surgery were asymptomatic and 40% were symptomatic [32].

Because of their rarity, a specific staging system for neuroendocrine tumors of the gastrointestinal system is not required, but as is the case with many other tumors, involvement of lymph nodes and size of the primary tumor are the two most important prognostic factors.

Carcinoid Tumors

Carcinoid tumors are the most common neuroendocrine tumors. They arise from neuroendocrine cells located primarily in the submucosa of the intestine but can also arise in the main bronchi. Between 1981 and 1989, a shift in the anatomic origin of these tumors was noted, with an increase in the percentage of carcinoids of the bronchus from 10% to 30% and a decrease in the number of tumors of the jejunoileum and rectum [33]. Moertel has reported that the majority of carcinoid tumors arise from the appendix (45%), small intestine (30%), and rectum (15%) [1].

In 1 of every 200 to 300 appendectomies, usually in young adults, a carcinoid tumor is found incidentally. In patients with appendiceal tumors 1 cm in diameter, surgical cure is possible in 95% of cases, whereas patients with tumors between 1 and 2 cm in diameter remain at more than 50% risk of having established metastases that eventually manifest. A right hemicolectomy should be considered for patients with tumors > 2 cm in diameter, younger patients, and those with vascular invasion, as these patients have nearly a 95% risk of developing overt metastatic disease [34].

Rectal carcinoids occur more frequently in middle-aged adults. These tumors are found incidentally in approximately 1 in 2,500 proctoscopies as a small yellow-gray submucosal nodule in the anterior and lateral walls of the rectum. They seldom bleed [1]. Unlike other carcinoid tumors, rectal carcinoids may not take up silver stain and histochemically may not show evidence of serotonin production. The majority of rectal carcinoids are < 1 cm in diameter and do not metastasize; fulguration can be adequate treatment for these small lesions. Tumors larger than 2 cm almost always will metastasize. Patients with aneuploid rectal carcinoids appear to have a poor prognosis, as do those with other carcinoid tumors of the gastrointestinal tract [35].

Small intestinal carcinoid tumors are the carcinoids most frequently associated with clinical symptoms, are usually found in the distal ileum within 60 cm of the ileocecal valve, and tend to be multiple. The likelihood of spread is associated with the size and location of the primary tumor; primaries in the small bowel are more likely to metastasize than are those in the rectum or the appendix [15].

A fibroblastic reaction can be stimulated in the mesentery and retroperitoneum, leading to complications (recurrent, intermittent partial or acute complete intestinal obstruction), with intermittent obstruction and abdominal pain being the most common clinical presentation. Recently, it was shown that platelet-derived growth factor is involved in this fibroblastic reaction by stimulating stromal cell growth through paracrine and possibly autocrine mechanisms [26]. Metastases from mesenteric and iliac lymph nodes can result, though rarely, in progressive encasement of the mesenteric artery, followed by bowel ischemia and infarction.

Carcinoid metastatic deposits in the liver tend to grow slowly. Thus, a patient with large masses in a grossly enlarged liver can be fully active and productive, with minimal symptoms and normal or nearly normal liver function test results.

Carcinoid Syndrome: The most common systemic syndrome caused by carcinoid tumors is the carcinoid syndrome. It occurs when hormonal tumor products reach the systemic circulation. This usually implies the presence of disease that has venous drainage in the systemic circulation in such a way as to circumvent the liver and its “first-pass” effect. Such is the case with metastatic disease in the liver itself or primary disease in the bronchi. During the “first pass,” the liver is able to remove from the bloodstream even large amounts of a primary tumor's hormonal products before they reach the systemic circulation. Hepatic metastasis is the most frequently associated condition in patients with carcinoid syndrome. Because tumors of the jejunum, ileum, appendix, and ascending colon are the most common and frequently metastasize, they account for about 80% of the carcinoids that cause the carcinoid syndrome [35]. The syndrome also occurs in association with gastric, pancreatic, appendiceal, cecal, colonic, and, in rare instances, rectal carcinoids.

Carcinoid syndrome is characterized by paroxysmal flushing, watery diarrhea, abdominal cramping, telangiectasia of the face, episodes of asthma or wheezing, endocardial fibrosis and valvular heart disease, and pellagra-like lesions of the skin and oral mucosa [15]. The distinction between the carcinoid syndrome and malignant carcinoid syndrome (or carcinoid crisis) is based on the severity of symptoms [36]. Progressive, relentless symptoms (namely, symptomatic carcinoid heart disease, severe diarrhea, and episodes of hypotension or hypertension) constitute the malignant version of the syndrome. This may be precipitated by stressful interventions, such as anesthesia induction or hepatic arterial occlusion.

Several peptides have been implicated as a cause of the carcinoid syndrome. The principal agent appears to be serotonin, although other peptides, such as substance P, histamine, bradykinin, kallikrein, and perhaps prostaglandins, may play a minor contributory role. In the tumor, tryptophan is converted to 5-hydroxytryptophan, which is then converted to serotonin by l-dopa decarboxylase and is stored in the neurosecretory tumor granules or is released in the systemic circulation. Most of the serotonin is converted by monoamine oxidase to 5-hydroxyindole acetic acid (5-HIAA), which appears in increased amounts in the urine.

The severity of the syndrome appears to be related to tumor bulk, the level of urinary excretion of 5-HIAA, and the degree of liver dysfunction caused by metastatic disease [37]. Hepatic metastases frequently reach sizes several times that of the primary tumor and can produce large amounts of serotonin and 5-hydroxytryptophan. Urinary 5-HIAA can serve as a useful marker in monitoring the course of the disease and the effectiveness of treatment [1].

The exact etiology of flushing remains unexplained. Histamine, kinins, and prostaglandins, but not serotonin, have been implicated in various situations. Flushing is present in as few as 25% of patients initially but in as many as 75% at some time during the disease course [38]. It consists of a sudden onset of violaceous erythema of the face and neck and is often associated with diarrhea and an unpleasant feeling of warmth. Sometimes it is accompanied by itching, palpitations, and facial edema.

Serotonin is considered responsible for diarrhea, which is seen in up to 84% of patients during the course of their disease [38], and is manifested by frequent, profuse, watery stools often associated with abdominal cramps.

About one third of patients with carcinoid syndrome die of carcinoid heart disease and not of tumor growth [38]. Carcinoid heart disease is not due to direct involvement of the heart by the tumor, but rather, to fibrosis that primarily affects the endocardium of the right heart and is most commonly seen on the ventricular aspect of the tricuspid valve. This leads to tricuspid regurgitation; evidence of heart failure is seen in 80% of patients with such heart lesions [39]. Transforming growth factor-beta likely plays a major role in fibroblast proliferation and matrix production in carcinoid heart lesions, as well as in other locations of stromal fibrosis associated with carcinoids [28].

The laboratory diagnostic test most frequently used to detect carcinoid syndrome is the quantitative measurement of 5-HIAA levels in a 24-hour urine sample (normal range, 2 to 8 mg/24 h). False-positive results occur with consumption of serotonin-rich foods and medications that affect urinary 5-HIAA levels, such as guaifenesin, acetaminophen, and salicylates. The test may be false-negative in patients with carcinoids of the stomach, pancreas, and proximal duodenum, as these often do not produce significant levels of serotonin (and therefore of 5-HIAA). Urinary 5-HIAA levels correlate well with tumor mass and can be used as a marker for the extent of disease [1].

Imaging Studies: Rectal, colonic, and gastric carcinoids are most frequently localized by endoscopic methods. Some tumors can be localized by angiography, computed tomography (CT), or magnetic resonance imaging (MRI). In a recent study of patients with pancreatic carcinoid tumors, the primary tumor was detected on CT and MRI scans in 79% and 88% of cases, respectively. While CT is the standard imaging procedure for staging the disease, angiography is the most sensitive method for detecting liver metastases and for defining vascular supply to the tumor before surgery or embolization.

Somatostatin receptor scintigraphy has markedly improved the visualization of neuroendocrine tumors. Indium 111-labeled diethylenetriamine penta-acetic acid (DTPA) octreotide currently is widely available, and indium 111 is the most commonly applied isotope for this test. The sensitivity of a scintiscan is 80% to 90%. This method successfully revealed additional metastases that had not appeared on conventional imaging scans in about one third of patients with various neuroendocrine tumors [40]. In addition, scintiscans revealed carcinoids in about one third of patients in whom there was a strong clinical suspicion of insulinoma or gastrinoma but in whom conventional scans were negative [40].

One of the most recent imaging developments is based on the fact that in serotonin-producing tumors, the precursor 5-hydroxytryptophan can be labeled with carbon 11 and subsequently traced by means of positron emission tomography (PET). With this method, tumors can be localized and their metabolic activity (and thus, response to treatment) followed with repeat scans [41].

Survival Rates: The most common sites of metastases from carcinoids include regional lymph nodes, liver, lung, bones, and the peritoneal cavity. Frequent causes of death among individuals with these tumors include liver failure (most common) and cardiac failure (10% to 30% of patients), although up to 50% of patients may have cardiac involvement [36,38]. Rates of 5-year survival correlate with the site and extent of disease. Reported rates are highest (92% to 99%) in patients with carcinoids of the appendix and vary between 76% and 100% in patients with rectal carcinoids and between 42% and 71% in those with carcinoids of the small intestines [35]. In one study [42], 10-year disease-specific survival rates were correlated with Dukes' classification and were found to be 100% for Dukes' stage A, 80% for stage B, 55% for stage C, and 10% for stage D carcinoid tumors. The presence of carcinoid syndrome is associated with decreased survival, with median survival durations ranging between 2 and 8.5 years from the onset of symptoms [38,41,43].

Gastrinomas

The presence of jejunal ulceration is almost pathognomonic of gastrinoma. Recurrent peptic ulceration after adequate surgery or multiple ulcers should raise the suspicion of a gastrinoma. Patients with both peptic ulcers and diarrhea, patients whose ulceration persists with use of H2-receptor blockers or recurs frequently after their discontinuation, and patients in whom endoscopy and barium meals reveal hypertrophied gastric mucosal folds also warrant investigation for the presence of a gastrin-secreting tumor.

The Zollinger-Ellison syndrome, which occurs in 1 of 1,000 patients with peptic ulcer disease, is the classic syndrome associated with gastrinoma. Its hallmark is the hypersecretion of gastrin, which results in significant hyperacidity, abdominal pain, diarrhea, and multiple recurrent peptic ulcers [44]. This syndrome is more common in males (60%) than in females; the mean age at diagnosis is 45 to 50 years [44,45].

Gastrinomas are malignant in 60% of cases, and 50% of patients with these tumors have established metastases at diagnosis [46,47]. Most gastrinomas are located in the gastrinoma triangle, defined by the junction of the second and third portions of the duodenum and the junction of the neck and body of the pancreas [48]. Multiple tumors have been reported in approximately half of patients [49].

Most gastrinomas are indolent, and long-term survival, even with metastatic disease, is not uncommon. Median survival in patients with gastrinomas is between 3 and 6 years [50]. In patients in whom no tumor was found at laparotomy (tumor microscopic in size) or in whom complete resection was possible, 5- and 10-year survival rates were 70% to 100% [51-53]. It appears that duodenal tumors have a higher recurrence rate and metastatic potential than pancreatic tumors [35,52,54].

Approximately 20% of cases of Zollinger-Ellison syndrome occur as part of the MEN I syndrome. A patient with both duodenal ulcer and hypercalcemia should be suspected of having MEN I with a gastrinoma. The most commonly associated tumor is a parathyroid adenoma. There also appears to be an increased incidence of gastric argyrophil carcinoidosis in these patients [55]. Patients in this subset have a much higher incidence of multiple gastrinomas than patients with the sporadic Zollinger-Ellison syndrome. The survival rate of patients with MEN I is not adversely affected by the presence of adenomas in multiple organs [50]. The presence of MEN I in patients with Zollinger-Ellison syndrome is associated with improved survival [52].

An elevated concentration of gastrin in a blood sample from a fasting patient and increased basal gastric acid output (> 15 mEq/h in patients who have not undergone prior acid-reducing surgery) suggest the presence of gastrinoma. However, gastrin levels may also rise as a result of hypochlorhydria, chronic renal failure, vagotomy, short gut syndrome, retained gastric antrum, pyloric stenosis, and antral G-cell hyperplasia. A gastrin-provocative test is required to differentiate gastrinomas from these other causes of gastrin elevation. The secretin test has become the provocative test of choice because it is easy to perform, highly sensitive and specific, and not associated with side effects.

Localization Techniques: Gastrinomas are small submucosal tumors and can easily be missed during routine upper gastrointestinal endoscopy even in areas such as the duodenum. In 40% to 60% of patients, no gastrinoma is identified during surgery [56,57], and as stated above, 50% of patients with gastrinomas have metastases at the time of diagnosis. Localization of gastrinomas by imaging studies or other techniques is imperative, as it could avert surgery in metastatic disease or increase the number of tumors identified and removed by carefully planned surgery. In addition to external ultrasonography, CT, and MRI, currently used localization studies include intraoperative ultrasonography (used mostly to identify pancreatic gastrinomas), selective injection of secretin into abdominal arteries with collection of venous samples from the hepatic arteries and assessment of gastrin levels, indium-111-octreotide scan (as a single diagnostic test, this method identified 77% of all gastrinomas in one study [40]),and angiography.

Insulinomas

Insulinomas are the most common type of islet-cell tumor. Generally, these tumors are benign (90%), intrapancreatic (nearly 100%), solitary, and small (< 2 cm). The peak incidence of insulinomas occurs in patients between 30 and 60 years of age, more frequently women. Because of hyperinsulinism, most patients are overweight. Children younger than 15 years old are rarely affected. Islet-cell dysplasia in children may lead to hypoglycemic episodes because of the autonomous secretion of insulin. In one report, DNA ploidy analysis was not a useful prognostic factor for patients with insulinomas [58]. About 5% of these tumors are associated with the MEN I syndrome; consequently, family members of index cases need to be screened periodically [59].

Patients with insulinomas may present with neurologic or psychiatric disturbances. An important feature is that symptoms tend to occur at night or early in the morning. The diagnosis is made by demonstrating inappropriately high concentrations of both insulin and C-peptide in the blood and a blood glucose level of less than 50 mg/dL together with the clinical picture of a hypoglycemic episode (confusion, altered consciousness, visual disturbances, weakness, sweating, tremulousness, and less commonly, seizures). Provocative testing with a 48-hour fast can be done in a hospital setting. Frequent measurements of the blood glucose concentration should be made during the fast. Only when a low value is recorded should a sample be taken to provide serum for insulin and C-peptide assay.

Computed tomography, ultrasonography, and selective arteriography fail to localize an insulinoma in about 40% of cases. Moreover, because these tumors have no predilection for a particular area within the pancreas, blind partial pancreatectomy is not warranted. As most insulinomas are solitary and benign, surgical resection is curative in the majority of cases. Thus, it is important to use additional tumor localization procedures (portal venous sampling, indium-111-octreotide scanning) in cases in which the tumor cannot be visualized by conventional methods. If these additional studies are negative, the patient should undergo surgical exploration and intraoperative ultrasonography.

Recent reports demonstrated intraoperative ultrasonography to be a highly sensitive and specific procedure for localization of pancreatic endocrine tumors [60,61]. Rosch et al studied 37 patients in whom transabdominal ultrasonography and CT scans were negative but who were later shown to have 39 endocrine tumors of the pancreas [60]. Using endoscopic ultrasonography, these investigators were able to localize 32 of the 39 tumors; no tumor was incorrectly localized. Among the 22 patients who underwent both angiography and endoscopic ultrasonography, the latter was significantly more sensitive for tumor localization. Among 19 control patients without pancreatic endocrine tumors, endoscopic ultrasonography was negative in 18. These authors recommend this modality for preoperative localization of pancreatic endocrine tumors once the clinical and laboratory diagnosis has been established [60].

Glucagonoma

Glucagonomas are rare alpha-cell tumors of the pancreas that occur in people between 50 and 70 years old. These tumors are mostly malignant, with metastases present at diagnosis in 50% to 80% of patients [12]. The primary tumor is located within the pancreas, and the liver is the most common site of metastasis [62]. Symptoms may not appear until the tumor is quite large. The tumors are usually > 5 cm in diameter, penetrate the pancreatic capsule, and invade the regional lymph nodes. Glucagon levels are usually quite elevated (> 1,000 pg/mL; normal range, 150 to 200 pg/mL) at diagnosis.

Mild glucose intolerance, a characteristic rash called necrolytic erythema migrans, psychiatric disturbances, anemia, and thromboembolic disease with venous thrombosis or pulmonary emboli are the salient clinical features in patients with a glucagonoma [63]. Mild glucose intolerance is the most common feature and is usually the first symptom to be recognized in the majority (> 90%) of patients. Plasma concentrations of amino acids, particularly the glucogenic amino acids, are usually severely depressed in patients with this tumor. Plasma amino acid concentrations vary with the extent of disease and rise to normal levels following tumor resection.

Necrolytic migratory erythema is probably the most characteristic clinical feature of the glucagonoma syndrome. This skin lesion may precede the diagnosis by at least 5 years [12]. The initial lesion consists of red papules or pale brown macules on the face, abdomen, groin, perineum, or extremities. The erythematous areas form superficial bullae that eventually break down and become encrusted. The pathogenesis of the rash is unknown. Hyperglucagonemia, zinc deficiency, and hypoaminoacidemia have been proposed as contributory causes.

Somatostatinoma

Approximately 50 somatostatin-secreting tumors have been reported to date, the majority of which are found in the pancreas or duodenum [64,65]. These tumors are generally malignant. Patients with a somatostinoma present with diabetes mellitus, cholelithiasis, diarrhea, steatorrhea, hypochlorhydria, anemia, and weight loss. Disease is usually diagnosed late in its course, and metastases to lymph nodes, liver, and bone are found at diagnosis.

Vipomas

In 1958, Verner and Morrison described a patient with watery diarrhea, hypokalemia, and achlorhydria [66]. These symptoms appeared to be mediated by a hormone called vasoactive intestinal polypeptide (VIP) and other peptides secreted by malignant islet-cell tumors in adults and by benign ganglioneuroblastomas in children [12]. In adults, vipomas are located in the pancreas and are usually large and solitary, whereas in children, these tumors are extrapancreatic. The tumor is usually solitary and large and is already metastatic in 80% of patients at diagnosis [12].

The major presenting abnormality of patients with a vipoma is a large volume of secretory diarrhea, generally more than 3 L/d. The stool is essentially isotonic, and the diarrhea persists even during fasting with nasogastric secretion. Large amounts of potassium and bicarbonate are lost in the stool, leading to hypokalemia and metabolic acidosis. Diagnosis rests on the typical clinical presentation, the findings of a pancreatic mass, and elevation of plasma VIP levels.

Pancreatic Polypeptideoma

In the normal adult, pancreatic polypeptide (PP) is located primarily within the pancreas, where it is synthesized and released from PP cells. Little is known about the physiologic actions of PP. Fasting plasma levels increase with age, prolonged fasting, exercise, chronic renal failure, and pancreatic disorders [12]. A recent review found 21 patients with pancreatic polypeptideomas reported in the literature [64,67]. From 50% to 75% of nonfunctioning endocrine tumors can be classified as pancreatic polypeptideomas because they are associated with elevated fasting plasma PP concentrations.

Patients with pancreatic polypeptideoma range in age from 20 to 74 years (mean, 51 years) [12]. The clinical picture of pancreatic polypeptideoma is complicated by the finding that many functioning islet-cell tumors are multihormonal. Patients with high plasma PP levels have been reported as having watery diarrhea, chronic duodenal ulcer, PP cell hyperplasia, nesidioblastosis, and multiple endocrine tumors. Other symptoms seen in these patients include diarrhea, hypochlorhydria, and weight loss. Pancreatic polypeptideomas are often found unexpectedly in patients with symptoms produced by metastases to the liver.

Treatment

Whenever possible, a localized tumor should be completely resected, as this is the only treatment modality that is potentially curative [52]. Surgical resection can cure more than 90% of patients with carcinoid tumors up to 1 cm in diameter and approximately 20% to 25% of patients with lymph node metastases [1]. In symptomatic patients, partial resection of tumors has a definite and occasionally sustained palliative effect [68].

Patients with gastrinoma who undergo complete resection have an excellent prognosis. Gastric acid hypersecretion and its complications are adequately managed by medical treatment, but it is the tumor and its potential malignant behavior that determine survival. For this reason, all patients with the Zollinger-Ellison syndrome should be considered for surgical resection.

In patients with advanced islet-cell tumors, debulking surgery should be considered when all other treatment options are exhausted, as it may significantly improve the degree of symptoms and success of medical therapy by decreasing tumor load [69]. Data also suggest that debulking may prolong survival [53]. Surgery can be particularly challenging if multiple islet-cell tumors or extrapancreatic (usually duodenal) gastrinomas are present. Both preoperative and intraoperative tumor localization strategies are thus important for a successful outcome. If surgery is not possible or feasible, symptomatic localized disease may be treated with radiotherapy in some cases [70]. However, unresectable cancer remains incurable.

Knowledge of the natural history of neuroendocrine tumors should be the basis for the treatment of unresectable (advanced) neuroendocrine tumors [36]. In view of the indolent course of these tumors, it is prudent to withhold specific therapy in asymptomatic or mildly symptomatic patients (in whom symptoms do not interfere with daily routine activities). This group of patients should be monitored every 3 months. Only symptomatic patients or those with impending organ failure should be treated. Treatment usually entails a multidisciplinary effort aimed at addressing problems related to hormone hypersecretion and tumor expansion.

Management of Hormonal Excess

Octreotide is a long-acting somatostatin analog that can provide effective amelioration of carcinoid syndrome and diarrhea associated with vipomas [71-74]. The parent hormone, native somatostatin, contributes to the inhibition of the secretion of virtually every known endocrine and exocrine hormone [75]. It affects autonomic processes, gut motility, mucosal-cell proliferation, vascular smooth muscle tone, and intestinal absorption of nutrients. Native somatostatin effectively reduces the symptoms of carcinoid syndrome, but its general use is severely limited by a short half-life (1 to 2 minutes). The synthetic analog has a longer half-life (90 to 120 minutes), allowing for more practical management of patients.

The efficacy of octreotide, like that of the parent compound, is based on its ability to inhibit the synthesis and release of various peptides by neuroendocrine cells. Response to octreotide correlates largely with the presence of somatostatin receptors in the tumor tissue [76,77]. Using radiolabeled octreotide, Lamberts et al showed that a positive scan can predict that octreotide will control symptoms of hormonal hypersecretion [76,78].

Kvols et al studied the effects of octreotide (500 µg given subcutaneously three times daily) in 25 patients with metastatic carcinoid tumors and carcinoid syndrome [71]. Of these patients, 72% achieved a significant biochemical response (greater than 50% reduction in 5-HIAA levels). Seven patients obtained complete relief of flushing, and four had complete relief of diarrhea. After a median treatment period of 5 months, the severity of symptoms seemed to reach pretreatment levels in five patients, suggesting that the initial effectiveness of octreotide is excellent. Overall, the median duration of symptomatic improvement was approximately 12 months.

The recommended starting dose of octreotide is 50 µg injected subcutaneously every 8 hours. Patients need to be assessed every 2 to 3 weeks to determine whether the dose must be increased. The dose is titrated to a level that achieves the most ideal symptom control without producing major steatorrhea. In some patients, doses up to 500 µg every 8 hours may be required to obtain maximum effects. The estimated median duration of response for all responding patients is more than 1 year, with more than one third estimated to remain responsive for 2 years or more.

Although octreotide is well tolerated for prolonged periods, it may be associated with side effects, including steatorrhea and fat malabsorption, glucose intolerance, nausea and vomiting, pain at the injection site, and fluid retention. Rarely, long-term use of this compound can result in biliary sludge or cholelithiasis.

Anthony et al studied the dose-limiting toxic effects and potential antitumor efficacy of octreotide in 12 patients with the carcinoid syndrome [79]. The highest dose escalation was 2,000 µg every 8 hours. They observed two partial responses by CT scans; generally, 5-HIAA excretion was maximally suppressed at 150 µg every 8 hours and did not decrease further with increasing octreotide doses. In seven patients, symptoms such as flushing were better controlled at doses above 150 µg every 8 hours. Toxic effects did not increase as the dose was escalated above 150 µg every 8 hours. The investigators concluded that octreotide is well tolerated at higher doses, and the maximum tolerated dose has yet to be achieved.

Octreotide is also indicated for tumors that produce VIP, which causes watery diarrhea, hypokalemia, hypochlorhydria, hypophosphatemia, and, sometimes, hypercalcemia [80]. Most patients respond to this drug with a reduction in diarrhea. Other investigators have demonstrated a beneficial effect of octreotide in patients with gastrinomas; in such cases, the drug can lower the gastrin concentration and gastric acid secretion [73]. However, the greater proportion of patients with islet-cell tumors have only transitory responses that are frequently not of substantive benefit; the median duration of response is 2.5 months [1].

The efficacy of octreotide correlates with the presence of somatostatin receptors on tumor cells. Receptors are found in great number on vipomas and gastrinomas but less frequently on insulinoma cells. Still, an overall improvement in hypoglycemic symptoms has been observed in about 50% of patients with insulinoma in one study [81]. However, this treatment also may reduce compensating factors, such as growth hormone and glucagon, relatively more than insulin and thus worsen hypoglycemia. For this reason, patients given octreotide should be monitored in a hospital setting during the initiation of treatment.

The use of octreotide truly represents a dramatic development in the management of endocrine pancreatic tumors and carcinoid tumors. The hypersecretion of hormones, such as VIP, gastrin, and glucagon, as well as the secretory products of carcinoid tumors (eg, 5-hydroxytryptamine and tachykinins) and their clinical effects may be blocked successfully [73]. This drug provides excellent palliation of such tumors and often affords patients a better quality of life. Eventually, a decrease in (and finally, an absence of) clinical effectiveness occurs, however, despite the reintroduction of other treatment modalities [82].

Omeprazole (Prilosec) is another significant new drug for the management of gastric hypersecretion in patients with the Zollinger-Ellison syndrome [83]. This benzimidazole analog blocks the hydrogen-ion pump of the parietal cell and can practically abolish acid secretion. The drug has a long duration of action and is easily administered (on a once-daily basis). A reasonable starting dosage is 60 mg/d, which may be increased to 120 mg/d.

Omeprazole is more effective than H2-receptor antagonists in providing symptomatic relief and mucosal healing and does not cause significant toxic effects [84,85]. There are reports of gastric carcinoids developing in patients who had received omeprazole, but these patients also had MEN I syndrome, suggesting that the presence of the latter condition per se may be important in the pathogenesis of gastric carcinoids [55,86].

Other Therapies: Other symptoms produced by hormonal excess can be managed appropriately [12]. Recurrent hypoglycemia may be prevented by frequent carbohydrate-rich meals and diazoxide, a drug that inhibits the release of stored insulin in both normal and tumor islet cells. Patients with the glucagonoma syndrome may benefit from the empiric use of a high-protein diet, oral and topical zinc preparations, and amino acid infusions. Insulin may be administered to counteract the catabolic effects of glucagon. Oral anticoagulation protects against the development of venous thrombosis in patients with glucagon excess. Loperamide and codeine can supplement the antidiarrheal management in patients with vipomas.

Biologic Therapy

Interferons are produced by T-cells in response to various stimuli, such as viruses, bacteria, toxic drugs, and certain cytokines [87]. These substances mediate a wide range of biologic responses, including antiviral effects, antiproliferative effects, immunomodulation, gene activation, and differentiation. A possible mechanism of action of interferon is a direct inhibitory action on tumor-cell proliferation and hormone synthesis. Interferon may also inhibit tumor-derived hormones that serve as autocrine growth factors [88].

berg et al were the first to demonstrate the favorable biochemical effects of interferon in patients with the carcinoid syndrome [89]. Among seven patients treated with human leukocyte inteferon, the majority had a significant biochemical response (ie, a decline in tumor marker levels). Subsequently, in a larger number of patients, berg et al presented further evidence that various interferon preparations yield antitumor as well as biochemical responses in patients with carcinoids [90].

Moertel et al treated 27 patients who had carcinoid tumors with recombinant leukocyte interferon at a daily dose between 6 and 24 million U/m² given three times a week [91]. In many patients, severe toxic effects resulting from high starting doses of interferon led to frequent dose reductions. Of these patients, 20% experienced an objective tumor response and 39% had a biochemical response. However, the durations of responses and treatments were brief.

A recent report also questioned the efficacy of interferon in the treatment of patients with metastatic carcinoids [92]. The combination of octreotide and alpha interferon (IFN-alfa) was used to treat patients with carcinoid malignancies resistant to either octreotide or interferon alone. Biochemical responses (complete in 4 of 22 patients) were observed in 77% of patients, with a median response duration of 15 months [42].

Interferon also has been investigated in patients with advanced islet-cell tumors. In one study [88], 22 patients were given human leukocyte inteferon at an intravenous dose of 3 million to 6 million U/m²/d. Partial responses were noted in 16 of 20 assessable patients. The median duration of response was 8.5 months, and the therapy was tolerated by the patients. These results need to be confirmed.

Chemotherapy and Biochemotherapy

Carcinoid Tumors: Chemotherapy may be palliative in patients with carcinoid syndrome who are suffering from dominant symptomatic metastases and who are no longer responsive to octreotide therapy. In general, carcinoids are relatively resistant to chemotherapy [36]. Because of the general refractoriness of the tumor to cytotoxic drugs, there is no standard systemic chemotherapy for patients with metastatic carcinoid tumors. Fluorouracil and doxorubicin (Adriamycin, Rubex) are considered the most active chemotherapeutic agents [93]. In contrast, streptozotocin (Zanosar) and dacarbazine are relatively inactive against carcinoid tumors [93-95].

In general, single-agent response rates have varied between 10% and 25%; durations of responses have been short (usually less than 6 months), and complete remissions have been rare [94]. Doxorubicin-based combinations have resulted in response rates of up to 35% [96]. The combination of streptozotocin and fluorouracil has been reported to have a response rate of 33% [1]. The current status of combination chemotherapy for patients with carcinoids is characterized by low response rates (less than 35%), short durations of response (< 9 months), and rare complete remissions [36]. Anaplastic variants of carcinoids may be more responsive to combined etoposide (VePesid) and cisplatin (Platinol)[97].

On the basis of preclinical data suggesting that synergistic cytolytic activity occurs when doxorubicin is combined with IFNalfa, Ajani et al studied the effect of the combination of doxorubicin (40 mg/m² given over 72 hours) and IFN-alfa (5 million U/m² given subcutaneously on days 1 through 14, with the courses repeated every 28 days)[98]. Among 12 patients with carcinoids, one complete and two partial responses were observed. The biochemical response of the tumors was also significant.

Islet-cell carcinomas are relatively more sensitive to chemotherapy than are carcinoid tumors. Active drugs include streptozotocin, fluorouracil, chlorozotocin, doxorubicin, and dacarbazine [99,100]. Streptozotocin is a glucosamine- nitrosourea compound originally isolated from a culture of Streptomyces achromogenes. Broder and Carter reported an objective tumor regression in 50% of 52 patients with metastatic islet-cell tumors treated with this drug [101]. The median duration of response was more than 12 months.

The major toxic effects of streptozotocin are nausea, vomiting, myelosuppression, and renal impairment. Moertel et al studied 84 patients with islet-cell carcinomas in a randomized prospective trial of streptozotocin vs fluorouracil plus streptozotocin. Of 42 patients who received streptozotocin alone, 14 (33%) responded, as compared with 36% of those receiving the combination [102]. Complete remission rates were 12% in the study arm that received streptozotocin alone and 33% in the arm that received the combination. Patient survival was significantly greater among those who received combination chemotherapy than in those treated with streptozotocin alone. This study was the basis for recommending the combination of fluorouracil and streptozotocin as standard chemotherapy for patients with advanced symptomatic islet-cell tumors.

However, more recently, Moertel et al reported the results of a prospective, multicenter trial of 106 patients with advanced islet-cell tumors randomized to three treatment arms: chlorozotocin alone, streptozotocin plus doxorubicin, and fluorouracil plus streptozotocin (the standard arm)[103]. Chlorozotocin was given as a single intravenous injection (150 mg/m²), which was repeated every 7 weeks. For the combination regimens, streptozotocin was given by intravenous injection at a dosage of 500 mg/m²/d for 5 consecutive days every 6 weeks. Fluorouracil was given by intravenous injection at a dosage of 400 mg/m²/d for 5 days concurrently with streptozotocin. Doxorubicin was given along with streptozotocin by intravenous injection at a dose of 50 mg/m² on days 1 and 22 of each 6-week treatment, with a maximum total dose of 500 mg/m².

The combination of streptozotocin plus doxorubicin was superior to streptozotocin plus fluorouracil in terms of the rate of tumor regression (69% vs 45%, P = .05) and the length of time to tumor progression (median, 20 vs 6.9 months; P = .001). Streptozotocin plus doxorubicin also had a significant advantage in terms of survival (median, 2.2 vs 1.4 years; P = .004). Monotherapy with chlorozotocin, a relatively new drug that is structurally similar to streptozotocin, produced a response rate similar to that of standard therapy. The major toxic effect of chlorozotocin was hematologic depression. Compared with streptozotocin, the new drug caused less nausea and vomiting. The investigators concluded that the combination of streptozotocin and doxorubicin is superior to standard therapy and that chlorozotocin needs to be studied further as a constituent of combination drug regimens.

Despite these advances, curative chemotherapy has not been developed for neuroendocrine tumors. Novel agents and combinations of biologic and cytotoxic therapy still need to be identified. Preclinical data from studies using neuroendocrine tumor-cell lines may be helpful in this regard [104].

Local-Regional Therapy With Hepatic Arterial Embolization

Metastatic neuroendocrine tumors are hypervascular, and vascular occlusion can result in significant deprivation of the blood supply to the liver. Hepatic metastases receive most of their vascular supply from the hepatic arterial circulation, whereas normal liver parenchyma receives the majority of its blood supply from the portal vein and only 30% from the hepatic artery. Hepatic arterial embolization is an invasive procedure that involves the use of inert material, such as Ivalon particles or gel foam, to occlude the blood vessels. If feasible, this procedure is recommended for cases of carcinoid syndrome that are unresponsive or partially responsive to octreotide therapy and in patients with advanced islet-cell tumors suffering from dominant symptomatic metastases; the technique can provide effective palliation in these patients. Hepatic arterial embolization can also deliver chemotherapeutic agents intra-arterially, thereby delivering a large amount of drug into the tumor [105-107].

Carrasco et al reported the results of microembolization treatment in 25 patients with carcinoid syndrome [108]. The majority of patients achieved both a biochemical and objective response to therapy. The average duration of effective palliation was more than 11 months. Ajani et al also reported effective palliation using this technique in 22 symptomatic patients with islet-cell tumors [109]. Of 20 evaluable patients, 12 had a partial remission. The patients developed marked abnormalities in hepatic enzyme levels, which returned to baseline a few weeks later during the recovery period.

After the microembolization procedure, patients frequently experience abdominal pain, nausea, vomiting, and fever. Complications can be severe and may include death from hepatic failure or hepatorenal syndrome. Careful patient selection is therefore mandatory [108]. Patients who have jaundice, poor performance status (Zubrod 3 or 4), or a liver largely replaced by tumor (> 75%) are ineligible for this therapy. This modality is available only in major cancer centers and requires a team effort and skilled interventional radiologists. Its impact on the survival of patients with metastatic disease remains to be determined.

Future Directions

New agents are desperately needed for patients with neuroendocrine tumors. However, due either to the absence of new exciting agents or a lack of initiative on the part of sponsors to pursue investigation in rare diseases, no new agents have been demonstrated to be active in these diseases for some time. Moreover, research efforts focused on these relatively rare tumors are relatively few compared with endeavors dealing with more common malignancies. Nevertheless, efforts are being made in clinical and preclinical areas.

Various somatostatin analogs are now under investigation to assess whether they can control carcinoid syndrome more effectively than does octreotide and can produce a higher rate of objective regression. In addition, encapsulation of these agents for intramuscular injection results in sustained therapeutic blood levels for approximately 6 weeks. If found effective in controlling carcinoid syndrome, this approach would result in a dramatic advancement for patients who must now receive subcutaneous injections of the drug two to three times per day. In addition, the new analogs purportedly have a higher affinity for somatostatin receptors. If this is true, one can expect better imaging techniques and higher efficacy in therapeutic targeting than is possible today.

Investigational therapeutic approaches using somatostatin-receptor antibodies attached to radioactive isotopes such as indium 111 make use of the overexpression of somatostatin receptors in neuroendocrine tumors. Such approaches could selectively deliver relatively higher doses of radioactivity to tumor tissue and potentially ablate the tumor.

In the arena of vascular occlusion, encapsulation of various drugs in capsules in the range of 100 µm is being pursued. In addition, these capsules may be more effectively targeted by coating the surface with octreotide analogs. Such ideas are preliminary and must await confirmation in the preclinical setting. Because metastases from neuroendocerine tumors are often highly vascular, it may be useful to study angiogenesis inhibitors in patients with these diseases.

References:

References

1. Moertel CG: An odyssey in the land of small tumors. J Clin Oncol 5:1503–1522, 1987.

2. Buchanan KD, Johnston CF, O'Hare MMT: Neuroendocrine tumors. Am J Med 81:14, 1986.

3. Wermer P: Genetic aspects of adenomatosis of endocrine glands. Am J Med 116:363–371, 1954.

4. Larsson C, Shogseid B, Öberg K, et al: MEN-1 gene maps to chromosome 11 and is lost in insulinoma. Nature 332:85–87, 1988.

5. Bale S, Bale AE, Stewart S: Linkage analysis of multiple endocrine neoplasia type 1 with INT2 and other markers on chromosome 11. Genomics 4:320–322, 1989.

6. Radford DM, Ashley SW, Wells JA, et al: Loss of heterozygosity of markers on chromosome 11 in tumors from patients with multiple endocrine neoplasia syndrome type 1. Cancer Res 50:6529–6533, 1990.

7. Friedman EL, DeMarco PV, Gejman PV, et al: Allelic loss from chromosome 11 in parathyroid tumors. Cancer Res 52:6804–6809, 1992.

8. Thakker RV, Bouloux C, Wooding C, et al: Association of parathyroid tumors in multiple endocrine neoplasia type 1 with loss of alleles on chromosome 11. N Engl J Med 321:218–224, 1989.

9. Öberg K: Expressions of growth factors and their receptors in neuroendocrine gut and pancreatic tumors, and prognostic factors for survival. Ann NY Acad Sci 46–55, 1994.

10. Pearse AGE: The diffuse neuroendocrine system and the APUD concept: Related endocrine peptides in brain, intestine, pituitary, placenta and anuran cutaneous glands. Med Biol 55:115–125, 1977.

11. Langley K: The neuroendocrine concept today. Ann NY Acad Sci 733:1–17, 1994.

12. Gower WR, Fabri PJ: Endocrine neoplasms (nongastrin) of the pancreas. Semin Surg Oncol 6:98–109, 1990.

13. Klöppel G, Heitz P: Classification of normal and neoplastic neuroendocrine cells. Ann NY Acad Sci 733:18–23, 1994.

14. Wilander E: Diagnostic pathology of gastrointestinal pancreatic neuroendocrine tumors. Acta Oncol 288:363–369, 1989.

15. Cotran RS, Kumar V, Robbins SL: Robbins' Pathologic Basis of Disease, pp 818–820. Philadelphia, WB Saunders, 1994.

16. Moertel CG, Hanley JA: Combination chemotherapy trials in metastatic carcinoid and malignant carcinoid syndrome. Cancer Clin Trials 2:327, 1979.

17. Johnson LA, Lavin PT, Moertel CG, et al: Carcinoids: The prognostic effect of primary site histologic type variations. J Surg Oncol 33:81–83, 1986.

18. Polak JM: Diagnostic histopathology of neuroendocrine tumours, pp 1–39. Edinburgh, Churchill Livingstone, 1993.

19. Wiedenmann B, Huttner W: Synaptophysin and chromogranins/ secretograninsd: Widespread constituents of distinct types of neuroendocrine vesicles and new tools in tumor diagnosis. Virchows Arch B Cell Pathol 58:95–121, 1989.

20. Klöppel G, Veld PI: Neural and endocrine markers as diagnostic tools in pancreatic and gastrointestinal endocrine tumors. Acta Histochem XXXVIII(suppl):93–98, 1990.

21. Reubi JC, Laissue J, Waser B, et al: Expression of somatostatin receptors in normal, inflamed and neoplastic human gastrointestinal tissues. Ann NY Acad Sci 733:122–137, 1994.

22. Scherubl H, Hescheler J, Riecken EO: Molecular mechanisms of somatostatin's inhibition of hormone release: Participation of voltage-gated calcium channels and G-proteins. Horm Metab Res 27(suppl):1–4, 1993.

23. Reubi JC, Kvols LK, Waser B, et al: Detection of somatostatin receptors in surgical and percutaneous needle biopsy samples of carcinoids and islet cell carcinomas. Cancer Res 50:5969–5977, 1990.

24. Gerdes J, Schwab H, Lenke H, et al: Production of mouse monoclonal antibody reacting with a human nuclear antigen associated with proliferation. Int J Cancer 31:13–20, 1983.

25. Chaudhry A, Öberg K, Wilander E: A study of biological behavior based on the expression of a proliferating antigen in neuroendocrine tumors of the digestive system. Tumor Biol 13:27–35, 1992.

26. Chaudhry A, Funa K, Öberg K, et al: Expression of growth factor peptides and their receptors in neuroendocrine tumors of the digestive system. Acta Oncol 32:107–114, 1993.

27. Chaudhry A, Papanicolau V, Öberg K, et al: Expression of platelet-derived growth factor and its receptors in neuroendocrine tumors of the digestive system. Cancer Res 52:1006–1012, 1992.

28. Waltenberger J, Lundin L, Öberg K, et al: Involvement of transforming growth factor-beta in the formation of fibrotic lesions in carcinoid heart disease. Am J Pathol 142:71–78, 1993.

29. Chaudhry A, Gobl A, Eriksson B, et al: Different splice variants of CD44 are expressed in gastrinomas but not in other subtypes of endocrine pancreatic tumors. Cancer Res 54:981–986, 1994.

30. Solcia E, Capella C, Fiocca R, et al: The gastro-pancreatic endocrine system and related tumors. Gastroenterol Clin North Am 4:671–693, 1989.

31. Wynick D, Williams SJ, Bloom SR: Symptomatic secondary hormone syndromes in patients with established malignant pancreatic endrocrine tumors. N Engl J Med 319:605–607, 1988.

32. Thompson GB, van Heerden JA, Martin JK, et al: Carcinoid tumors of the gastroinestinal tract: Presentation, management and prognosis. Surgery 98:1054, 1985.

33. Vinik AI, McLeod MK, Fig LM, et al: Clinical features, diagnosis, and localization of carcinoid tumors and their management. Gastroenterol Clin North Am 18:865, 1989.

34. Moertel CG, Weiland LH, Nagorney DM, et al: Carcinoid tumor of the appendix: Treatment and prognosis. N Engl J Med 317:1699–1701, 1987.

35. Norton J, Levin B, Jensen R: Cancer of the endocrine system, in DeVita VT, Hellman S, Rosenberg SA (eds): Cancer: Principles and Practice of Oncology, pp 1371–1435. Philadelphia, JB Lippincott, 1993.

36. Ajani JA, Carrasco H, Samaan NA, et al: Therapeutic options in patients with advanced islet cell and carcinoid tumors. Reg Cancer Treat 3:235–242, 1991.

37. Moertel CG: Treatment of carcinoid tumors and the malignant carcinoid syndrome. J Clin Oncol 1:727–740, 1983.

38. Norheim I, Öberg K, Theodorsson-Norheim E, et al: Malignant carcinoid tumors. Ann Surg 206:115, 1987.

39. Lundin L: Carcinoid heart disease. Acta Oncol 30:499, 1991.

40. Krenning EP, Kwekkeboom DJ, Oei HY, et al: Somatostatin-receptor scintigraphy in gastroenteropancreatic tumors. Ann NY Acad Sci 733:416–424, 1994.

41. Öberg K: Treatment of neuroendocrine tumors. Cancer Treat Rev 20:331–355, 1994.

42. Agranovich AL, Anderson GH, Manji M, et al: Carcinoid tumor of the gastrointestinal tract: Prognostic factors and disease outcome. J Surg Oncol 47(1):45–52, 1991.

43. Davis Z, Moertel CG, McIlrath DC: The malignant carcinoid syndrome. Surg Gynecol Obstet 137:637, 1973.

44. Zollinger RM, Ellison EH: Primary peptic ulcerations of the jejunum associated with islet cell tumors of the pancreas. Ann Surg 142:709–129, 1955.

45. Jensen RT, Gardner JD: Zollinger-Ellison syndrome: Clinical presentation, pathology, diagnosis and treatment, in Dannenberg A, Zakim D (eds): Peptic Ulcer and Other Acid-Related Diseases, p 117. New York, Academic Research Association, 1991.

46. Townsend CM, Lewis BG, Gourley WK, et al: Gastrinoma. Curr Prob Cancer 7:1–33, 1982.

47. Townsend CM, Thompson JC: Gastrinoma. Semin Surg Oncol 6:91–97, 1990.

48. Stabile BE, Morrow DJ, Passaro E: The gastrinoma triangle: Operative implications. Am J Surg 147:25–31, 1984.

49. Fox PS, Hofman JW, Wilson SD, et al: Surgical management of the Zollinger Ellison syndrome. Surg Clin North Am 54:394, 1974.

50. Thompson JC, Terris BG, Wiener I, et al: The role of surgery in the Zollinger-Ellison syndrome. Ann Surg 197:590–607, 1983.

51. Malagelada JR, Edis AJ, Adson MA, et al: Medical and surgical options in the management of patients with gastrinoma. Gastroenterology 84:1524–1532, 1983.

52. Norton JA, Dopppman JL, Jensen RT: Curative resection in Zollinger-Ellison syndrome: Results of a 10 year prospective study. Ann Surg 215:8, 1992.

53. Ellison EC, Carey LC, Sparks J, et al: Early surgical treatment of gastrinoma. Am J Med 82:17, 1987.

54. Thom AK, Norton JA, Axiotis CA, et al: Location, incidence and malignant potential of duodenal gastrinomas. Surgery 110:1086, 1991.

55. Solcia E, Capella C, Fiocca R, et al: Gastric argyrophil carcinoidosis in patients with Zollinger-Ellison syndrome due to type 1 MEN: A newly recognized association. Am J Surg Pathol 14:503–508, 1990.

56. Deveney CW, Deveney K E, Stark D, et al: Resection of gastrinomas. Ann Surg 198:546, 1983.

57. McCarthy DM: The place of surgery in the Zollinger-Ellison syndrome. N Engl J Med 302:1344, 1980.

58. Graeme CF, Bell DA, Flotte TJ: Aneuploidy in pancreatic insulinomas does not predict malignancy. Cancer 66:2365–2368, 1990.

59. Jadoul M, Koppenshaar HP, Box MA, et al: Insulinomas in MEN-1 patients early detection and treatment of insulinomas in patients with MEN-1 syndrome. Neth J Med 37:95–102, 1990.

60. Rosch T, Lighdale CJ, Botit JF, et al: Localization of pancreatic endocrine tumors by endoscopic ultrasonography. N Engl J Med 326:1721–1726, 1992.

61. Doppmann J: Pancreatic endocrine tumors-the search goes on (editorial). N Engl J Med 326:1770–1772, 1992.

62. Stacpoole PW: The glucagonoma syndrome: Clinical features, diagnosis, and treatment. Endocrin Rev 2:347, 1981.

63. Bloom SR, Polak JM: Glucagonoma syndrome. Am J Med 82:25–36, 1987.

64. Vinik AI, Strodel WE, Eckhauser FE, et al: Somatostatinomas, PPomas, neurotensinomas. Semin Oncol 14:263–281, 1987.

65. Krejs GJ, Orci L, Conlon M, et al: Somatostatinoma syndrome (biochemical, morphological and clinical features). N Engl J Med 301:289–292, 1979.

66. Verner JV, Morrison AB: Islet cell tumor and a syndrome of refractory watery diarrhea and hypokalemia. Am J Med 25:375–380, 1958.

67. Adrian TE, Uttenthal LO, Williams SJ, et al: Secretion of pancreatic polypeptide in patients with pancreatic endocrine tumors. N Engl J Med 315:287–291, 1986.

68. Wilson H: Carcinoid syndrome. Curr Prob Surg 11:36–41, 1970.

69. Thompson NW, Eckhauser FE: Malignant islet cell tumors of the pancreas. World J Surg 8:940–951, 1984.

70. Rich TA: Radiation therapy for pancreatic cancer: Eleven year experience in the JCRT. Int J Radiat Oncol Biol Phys 11:759–763, 1985

71. Kvols LK, Moertel CG, O'Connell MJ, et al: Treatment of the malignant carcinoid syndrome: Evaluation of a long-acting somatostatin analogue. N Engl J Med 315:663–666, 1986.

72. Garden P, Comi RJ, Maton PN, et al: NIH conference: Somatostatin and somatostatin analogue (SMS 201-995) in treatment of hormone-secreting tumors of the pituitary and gastrointestinal tract and non-neoplastic disease of the gut. Ann Intern Med 110:350–354, 1989.

73. Wynick D, Bloom SR: Clinical review: The use of the long-acting somatostatin analog octreotide in the treatment of gut neuroendocrine tumors. J Clin Endocrinol Metab 73:1–3, 1991.

74. Öberg K, Norheim I, Theodorsson E: Treatment of malignant midgut carcinoid tumors with a long-acting somatostatin analogue octreotide. Acta Oncol 30:503–507, 1991.

75. Reichlin S: Somatostatin. N Engl J Med 309:1495–1501, 1983.

76. Lamberts SW, Bakker WA, Renbi JC, et al: Treatment with Sandostatin and in-vivo localization of tumors with radiolabeled somatostatin analogues. Metabolism 39:1525, 1990.

77. Reubi JC, Koals LK, Waser B, et al: Detection of somatostatin receptors in surgical and percutaneous needle biopsy samples of carcinoids and islet cell carcinomas. Cancer Res 50:5969–5972, 1990.

78. Kvols L, Reubi JC, Moertel C, et al: Somatostatin receptors may predict responsiveness of malignant neuroendocrine syndromes to therapy with somatostatin analogue (SMS 201-995, Sandostatin) (abstract 389). Proc Am Soc Clin Oncol 8:101, 1989.

79. Anthony LB, Winn SD, Krozely MG, et al: Relationship of octreotide dose to its efficacy and toxicity in carcinoid syndrome (abstract). Proc Am Soc Clin Oncol 10:387, 1991.

80. Kvols LK, Buck M, Moertel CG, et al: Treatment of metastatic islet cell carcinoma with a somatostatin analogue (SMS 201 995). Ann Intern Med 107:162–168, 1987.

81. Dunne MJ, Elton R, Fletcher T, et al: Sandostatin and gastroenteropancreatic endocrine tumors-therapeutic characteristics, in O'Dorisio TM (ed): Somatostatin in the Treatment of GEP Endocrine Tumors, pp 93–113. Berlin, Springer Verlag, 1989.

82. Öberg K, Erikson B: Medical treatment of neuroendocrine gut and pancreatic tumors. Acta Oncol 28:425–431, 1989.

83. Maton PN: Role of acid suppressants in patients with Zollinger-Ellison syndrome. Aliment Pharmacol Ther 5:25–35, 1991.

84. Solvell L: The clinical safety of omeprazole. Digestion 47:59–63, 1990.

85. Frucht H, Maton PN, Jensen RT: The use of omeprazole in patients with Zollinger-Ellison syndrome. Dig Dis Sci 36:394–404, 1991.

86. Maton PN, Lack EE, Collin MJ, et al: The effect of ZES and omeprazole therapy on gastric oxyntic endocrine cells. Gastroenterology 94:943–50, 1990.

87. Rosenberg SA, Longo DL, Litze MT: Principles and applications of biologic therapy in cancer, in DeVita VT, Hellman S, Rosenberg SA (eds): Cancer: Principles and Practice of Oncology, 3rd Ed. Philadelphia, JB Lippincott, 1989.

88. Eriksson B, Öberg K, Alm G: Treatment of malignant endocrine pancreatic tumors with human leukocyte interferon. Lancet 2:1307–1308, 1986.

89. Öberg K, Funa K, Alm G: Effects of leukocyte interferon on clinical symptoms and hormonal levels in patients with midgut carcinoid tumors and the carcinoid syndromes. N Engl J Med 309:129–133, 1983.

90. Öberg K, Norheim I, Lind E, et al: Treatment of malignant carcinoid tumors with human leukocyte interferon. Long term results. Cancer Treat Rep 70:1297–1304, 1986.

91. Moertel CG, Rubin J Kvols L: Therapy of metastatic carcinoid tumor and the malignant carcinoid syndrome with recombinant leukocyte A interferon. J Clin Oncol 7:865–868, 1989.

92. Valimaki M, Jarvinen H, Salmella P, et al: Is the treatment of metastatic carcinoid tumor with interferon not as successful as suggested? Cancer 67:547–549, 1991.

93. Legha SS, Valdivieso M, Nelson RS, et al: Chemotherapy for metastatic carcinoid tumors: Experience with 32 patients and a review of the literature. Cancer Treat Rep 61:1699–1703, 1977.

94. Stolinsky DC, Sadoff L, Braunwald J, et al: Streptozotocin in the treatment of cancer: Phase II study. 30:61–67, 1972.

95. Kissinger A, Foley FJ, Lemon HJ: Use of DTIC (dacarbazine) in the malignant carcinoid tumors. Cancer Treat Rep 61:101–102, 1977.

96. Ajani JA, Legha SS, Karlin DA, et al: Combination chemotherapy of metastatic carcinoid tumors with 5-FU, adriamycin, and cytoxan (FAC) and 5-FU, adriamycin, mitomycin-C, and methyl CCNU (FAMMe) (abstract). Proc Am Soc Clin Oncol 2:124, 1983.

97. Moertel CG, Kvols LK, O'Connell MJ, et al: Treatment of neuroendocrine carcinomas with combined etoposide and cisplatin: Evidence of major therapeutic activity in the anaplastic variants of these neoplasms. Cancer 68:227–232, 1991.

98. Ajani JA, Kavanagh J, Patt Y, et al: Roferon and doxorubicin combinations active against advanced islet cell or carcinoid tumors. Proc Am Assoc Cancer Res 30:293, 1989.

99. Haller I, Schutt A, Sayal Y, et al: Chemotherapy for metastatic carcinoid tumors: An ECOG phase II-III trial (abstract). Proc Am Soc Clin Oncol 9:395, 1990.

100. Hahn RG, Cnaan A, Kissinger A, et al: A phase II study of DTIC in the treatment of non-resectable islet cell carcinoma: An ECOG treatment protocol (abstract). Proc Am Soc Clin Oncol 9:417, 1990.

101. Broder LE, Carter SK: Pancreatic islet cell tumors with streptozotocin. Ann Intern Med 79:108, 1973.

102. Moertel CG, Hahnley JA, Johnsson LA: Streptozocin alone compared with streptozocin plus fluorouracil in the treatment of advanced islet cell carcinoma. N Engl J Med 303:1189, 1980.

103. Moertel CG, Lefkopoulos M, Lipsitz S, et al: Streptozocin-doxorubicin, streptozocin-fluorouracil or chlorozotocin in the treatment of advanced islet cell carcinoma. N Engl J Med 326:519–523, 1992.

104. Evers BM, Halbut SC, Tyring SK, et al: Novel therapy for the treatment of human carcinoid. Ann Surg 213:411–416, 1991.

105. Vallete J, Souquet JC: Pancreatic islet cell tumors metastatic to the liver: Treatment by hepatic arterial chemo-embolization. Horm Res 32:77–79, 1989.

106. Mavligit G, Pollack R: Islet cell carcinoma of the pancreas: Prospective (neoadjuvant) chemoembolization infusion in patients with concomitant primary lesions and liver metastases (meeting abstract). 3rd International Congress on Neo-Adjuvant Chemotherapy. Paris, France, February 1991.

107. Venook A, Stagg R, Frye J, et al: Chemoembolization of patient with liver metastases from carcinoid and islet cell tumors (abstract). Proc Am Soc Clin Oncol 10:386, 1991.

108. Carrasco CH, Charnsangavej C, Ajani JA: The carcinoid syndrome: Palliation by hepatic artery embolization. Am J Roentgenol 147:149–154, 1986.

109. Ajani JA, Carrasco CH, Charnsangavej C, et al: Islet cell tumors metastatic to liver: Effective palliation by sequential hepatic artery embolization. Ann Intern Med 108:340–344, 1988.