The neuropeptide somatostatin was first identified in 1973 in ovine hypothalami (1) and has since been recognized as a major neurotransmitter and neuromodulator, with a mostly inhibitory capacity in that it regulates exocrine secretions, glandular secretions, neurotransmission (2), smooth muscle contractility, and absorption of nutrients. Of particular interest is the ability of somatostatin and somatostatin analogs to act as cytostatic agents in their effect on tumor cells (3), as evidenced in many animal tumor models and cultured tumor cell lines.

Human somatostatin, a 14-amino-acid peptide (somatostatin-14; Fig. 43.1A), appears in several organ systems, such as the central nervous system, the hypothalamopituitary system, the gastrointestinal tract, the exocrine and endocrine pancreas, and the immune system. The diverse actions of somatostatin are mediated through interaction with a family of different specific, high-affinity somatostatin receptors (SSTRs) located on the plasma membrane of the target cells. To date, five human SSTR subtypes (SSTR1 to SSTR5) have been cloned and partially characterized (4). All five SSTR subtypes bind the endogenous somatostatin-14 with high affinity (in the nanomolar range). However, the short biologic half-live of somatostatin in vivo (about 2 minutes) prevents its application in the clinic. To overcome this drawback, analogs of somatostatin-14 consisting of 8 amino acids were developed (5). These octapeptides (Fig. 43.1) exhibit a biological half-life on the order of several hours, and among them octreotide (Fig. 43.1B) is one of the clinically most relevant.

Although all five human SSTR subtypes bind somatostatin-14 with high affinity, there are differences in the binding affinities of the analogs; octreotide, for example, is bound with high affinity by the SSTR2 and SSTR5 receptor subtypes and with a modest affinity by SSTR3 but is not bound by subtypes SSTR1 and SSTR4.

Receptor-binding studies have identified SSTRs in the human brain as well as in numerous peripheral tissues, including the pituitary, pancreas, gut, and thyroid and the adrenal and immune systems, to name a few. The subtype most frequently expressed is usually SSTR2.

SSTRs, however, are not only expressed in the above tissues under physiological conditions. In the past decade, there has been increasing evidence for SSTR expression in various human cancers (6,7), in particular in gastroenteropancreatic neuroendocrine tumors (GEP-NETs).

Phenotypically, the cells of GEP-NETs belong to the system of disseminated neuroendocrine cells. This system is composed of cells that are scattered throughout the mucosa of the gastrointestinal tract or form the islets in the pancreas. These cells have many histologic similarities to neural cells such as secretory granules and the markers chromogranin A, synaptophysin, and neuron-specific enolase. These features led
to the designation “neuroendocrine cells.” They constitute the so-called diffuse endocrine system (8).

Misunderstanding was perpetuated for the clinician by the application of different classifications and terminologies to the many varied types of neuroendocrine tumors of the gastrointestinal tract. In the past, pathologists called these tumors carcinoids because their histology is quite similar without special staining. This practice still is continued sometimes. In 1963, these carcinoids were subclassified (9) according to their embryologic site of origin as foregut carcinoids (lung, stomach, duodenum, biliary system, and pancreas), midgut carcinoids (small intestine, appendix, cecum, and proximal colon), and hindgut carcinoids (distal colon and rectum).

This classification was the first to emphasize clinicopathological differences between the tumor groups composing the gastroenteropancreatic neuroendocrine tumors but was not generally accepted by clinicians, because it proved too imprecise in distinguishing between the (biologically) different GEP-NET entities. Therefore, instead of “carcinoid,” the neutral terms “neuroendocrine tumor” and “neuroendocrine carcinoma” were chosen for use in the new World Health Organization (WHO) classification of 2000 (10). It relates histopathology to biological behavior (i.e. prognosis). Three major categories of GEP-NETs have been defined: (a) well-differentiated neuroendocrine tumors (benign or low-grade malignancy), (b) well-differentiated neuroendocrine carcinomas, and (c) poorly differentiated neuroendocrine carcinomas. The criteria on which categorization of the tumors is based are, for example, size, presence of metastases, and proliferation index. The proliferation index is determined by the percentage of cells staining positively with a monoclonal antibody directed against a nuclear antigen in proliferating cells (Ki-67/MIB-1) (greater than 2% indicates increased degree of malignancy).

The term “carcinoid” is not completely abandoned. For gastroenteric NETs, it is still used synonymously with the term “well-differentiated neuroendocrine tumor.” The term “malignant carcinoid” is still used synonymously with the term “well-differentiated neuroendocrine carcinoma” (11).

GEP-NETs are among the tumors that most frequently express SSTRs with high density and homogeneous distribution (others include medullary thyroid cancer, small cell lung cancer, meningioma, medulloblastoma, glioma, breast cancer, lymphoma, and renal cell cancer). Pancreatic NETs (e.g., gastrinomas, glucagonomas, and VIPomas) and gut NETs (foregut, midgut, and hindgut NETs) usually express SSTRs in 80% to 100% of the cases. Insulinomas, however, have a lower incidence of SSTR expression (about 50%).

Among the five SSTR subtypes, the SSTR2 subtype is the most frequently expressed in GEP-NETs, while SSTR4 is rarely detected. It is well known that undifferentiated GEP-NETs express SSTRs less often (and, if expressed, in lower densities) than well-differentiated GEP-NETs.

Management of GEP-NETs is primarily based on surgical removal of the primary lesion and the metastases (12), as these slow-growing tumors are only minimally responsive to systemic chemotherapy.

Neuroendocrine tumors of the gastrointestinal tract and pancreas may present with hormone-related symptoms or without hormonal symptoms (so-called nonfunctioning tumors). Therefore, the treatment might have different aims, such as to control excessive hormonal secretion or retard tumor growth. Surgery is generally considered as the first-line treatment for patients with localized disease, because it can be curative. In patients with metastatic disease, surgery is palliative, directed toward a reduction of the tumor burden and a consequent decrease in hormone secretion.

Cytotoxic treatment has been applied to a number of GEP-NETs, particularly those with high proliferation capacity, with minor beneficial effect.

Interferon alfa has been used for the treatment of classic midgut carcinoids for almost 2 decades and has generated biochemical and symptomatic improvement in about 50% of patients but significant tumor reduction in only 10% to 15%. Stabilization of the disease has been obtained in 60% to 80% of patients up to 3 to 4 years (13,14).

Nonradiolabeled somatostatin analogs have been used for the management of clinical symptoms in patients with GEP-NETs, and significant biochemical and symptomatic improvement has been obtained in 50% to 60% of patients, but tumor reduction has been seen in only 3% to 5%. (14). Beta particle–labeled somatostatin analogs are well-established radionuclide therapy modalities for GEP-NETs (15). Yttrium 90 (⁹⁰Y)- and lutetium 177 (¹⁷⁷Lu)-radiolabeled somatostatin analogs have been used for the treatment of GEP-NETs. The physical characteristics of ⁹⁰Y and ¹⁷⁷Lu suggest that ⁹⁰Y-labeled analogs are more suitable for radiotherapy of larger tumors, ¹⁷⁷Lu-labeled analogs more suitable for smaller tumors (16,17).

In practice, imaging of GEP-NETs is routinely performed with standard imaging techniques such as ultrasonography, CT, and MRI. There is little difference in sensitivity between CT and MRI, although the former is probably superior for localizing the primary tumor and thoracic lesions, whereas the latter may be superior in characterizing liver lesions. However, it is occasionally difficult to visualize GEP-NETs with CT or MRI because such tumors are frequently small and can be disseminated over several organ systems. Visualization of SSTRs in GEP-NETs is therefore considered an alternative approach, provided that the density of SSTRs is high enough for this type of imaging.