6.3 Neuroendocrine (carcinoid) tumours and the carcinoid syndrome

Neuroendocrine tumours (NETs) are derived from cells of the diffuse neuroendocrine system, which are present in organs throughout the body. Originally, Pearse proposed that tumours develop from migration of cells from the neural crest; however, it is now thought that the tumour cells are derived from multipotent stem cells (1).

The term ‘karzinoide’ (meaning carcinoma like) was initially introduced by Siegfried Oberndorfer in 1907 (2). The term carcinoid tumour has historically been used; however, with advances in the understanding of the tumour biology, and the recent WHO classification, the term NET or endocrine tumour is considered more appropriate, and more details are given in the historical introduction in Chapter 6.1.

The reported incidence is 2.5–5 cases per 100 000 population (3), however, due to their rather indolent nature the prevalence of these tumours is much higher—approximately 35 per 100 000 population in the USA (4). The incidence of different NETs has risen over the last three decades, with the greatest increased in bronchial NETs (5), which account for 10–30% of all NETs. This increase in incidence of NETs is partly due to improved diagnostic techniques, both radiological and endoscopic.

Most cases are sporadic; however, some occur as part of genetic syndromes, including multiple endocrine neoplasia 1 (MEN 1), MEN 2, von Hippel–Lindau syndrome and neurofibromatosis 1 (6). The incidence of MEN 1 in NETs varies dependent on the site, from very rare in midgut NETs but occurring in up to 25–40% of gastrinomas (7). Approximately one-third of individuals with MEN 1 develop gastric carcinoids.

NETs of the gastrointestinal tract have been classified according to their embryological origin into foregut (bronchial, stomach, pancreas, gall bladder, and duodenum), midgut (jejunum, ileum, appendix, and colon, up to ascending colon), and hindgut (transverse and remaining colon, rectum). It is becoming apparent that tumours within each region can have markedly different clinical behaviour and, therefore, a shift towards categorization of tumours purely by anatomical location is being introduced (8). Additional tumours considered to be neuroendocrine include: thymic carcinoids, medullary thyroid cancer, phaeochromocytomas, and paragangliomas.

NETs can exhibit a diverse spectrum of pathology, from benign tumours to highly aggressive, poorly differentiated tumours (8). The WHO classification is used for describing tumours of gut and pancreas (9). Separate classifications systems are in use for bronchial, thymic, and thyroid NETs. The WHO classification for tumours is based on degree of differentiation and clinical behaviour; there are three types:

Bronchial carcinoid tumours are classified into four groups dependent on histological parameters, including mitotic activity and proliferation index. These groups are typical carcinoids, atypical carcinoids, large cell neuroendocrine carcinoma, and small cell lung carcinoma (10).

The European Neuroendocrine Tumour Society has proposed a TNM staging classification and this also includes a grading system of low, intermediate, or high-grade tumours dependent on their proliferation index, mitotic activity, and histological phenotype (11–13). Also, they stage tumours using the TMN classification. The classification so far has been published for GEP NETs but not other NETs. Further details are given in Chapter 6.1.

NETs can be separated into nonfunctioning and functioning tumours. The majority (approximately 60%) are nonfunctional tumours, i.e. with no symptoms attributable to secretion of metabolically active peptides. Functional tumours secrete substances that are metabolically active, which can lead to the development of specific clinical syndromes (Table 6.3.1). The most common functional syndrome is carcinoid syndrome, which is thought to be due to secretion of amines, kallikrein, and prostaglandins. Serotonin (5-hydroxytryptamine) is one of the main amines that is synthesized and secreted by these tumours.

Carcinoid syndrome occurs in 20–30% of patients with midgut carcinoid tumours and approximately 5% of bronchial carcinoids (14). Other foregut tumours (e.g. pancreatic neuroendocrine tumours) can cause carcinoid syndrome although this is uncommon (1%). Hind gut tumours are generally nonfunctional and rarely cause carcinoid syndrome. Carcinoid syndrome is usually seen in patients with liver metastases (in 95% patients), but excess tachykinins, serotonin production from retroperitoneal metastases, or ovarian tumours can bypass the liver to cause the syndrome.

Normally, serotonin is synthesized from tryptophan, and is subsequently metabolized by monoamine oxidase to 5-hydroxyindoleacetic acid (5-HIAA), which is subsequently secreted in the urine in healthy individuals. Approximately 99% of tryptophan is used for the synthesis of nicotinic acid and less than 1% converted to 5-hydroxytryptamine (5-HT). However, in patients with carcinoid tumours there is a shift towards the production of 5-HT. The increased production of 5-HT and other products and their direct release into the systemic circulation, due to liver metastases, leads to the development of carcinoid syndrome (15).

Patients often describe having symptoms for many months prior to presentation. The two most common symptoms are diarrhoea and flushing, whilst wheeze occurs less commonly. Often diarrhoea is associated with crampy abdominal pain and urgency, and can occur during both day and night. Flushing is characteristically described as a sudden onset of pink to red discoloration involving the face and upper trunk. This usually lasts a few minutes and can occur intermittently throughout the day. Triggers leading to flushing and diarrhoea include stress, tyramine-containing foods (chocolate, bananas, walnuts) and alcohol. In patients with atypical flushing, which may last for several hours, telangiectasia and hypertrophy of the face may be seen. Wheeze is caused by bronchial constriction mediated via tachykinins and bradykinins. This is more common in those with bronchial carcinoid tumours.

A raised jugular venous pressure and features of right heart failure may be present in patients with carcinoid heart disease related to carcinoid syndrome. Right-sided cardiac murmurs of tricuspid regurgitation and pulmonary stenosis may be heard on cardiovascular examination (16).

Other hormone-related manifestations include morphoea (subcutaneous thickening of the lower limbs) and a pellagra-type rash if nicotinic acid deficiency has been induced. With severe, long-standing hepatomegaly or local infiltration, inferior vena cava obstruction, or even lymphangiectasia leading to ascites may occur.

The prognosis of patients with carcinoid syndrome varies widely and although some patients may have rapidly progressive disease, in others survival for decades may occur.

Diagnosis of NETs requires biochemical, topographical imaging, and, importantly, histological diagnosis. Efforts should be made to identify the primary tumour site, which can be difficult since some primary lesions are small and not detected by conventional cross-sectional imaging.

Patients with suspected NETs should undergo biochemical testing, including fasting gut hormones (glucagon, vasoactive intestinal peptide, somatostatin (SST), and gastrin), chromogranin A, and pancreatic polypeptide. In addition to specialized blood tests, routine lab tests including full blood count, urea and electrolytes, liver function tests, carcino-embryonic antigen, α-fetoprotein, β-human chorionic gonadotropin, Ca 19-9, and ESR should be performed. Urinary 5-HIAA assay should also be performed in all patients with suspected carcinoid syndrome (17). Table 6.3.2. shows the different biochemical tests that are used for diagnosis of NETs.

Histology remains the gold standard for diagnosing NETs. Specimens should be immunostained with a panel of antibodies to general neuroendocrine markers. These include chromogranin A, synaptophysin, and PGP9.5. In addition, the tumour should be stained with an antibody to the Ki-67 protein, since the Ki-67 proliferation index is of benefit in grading tumours (11).

The histological characteristics of NETs vary according to the degree of differentiation. Low-grade NETs originating from the gut were previously termed ‘typical’ carcinoids; these tumours had classic histological architecture of trabecular, or ribbon-like cell clusters, with little or no cellular pleomorphism and occasional mitoses. The higher grade and poorly differentiated tumours had increased mitotic activity and evidence of necrosis. The WHO classification gives clear parameters for categorizing NETs into the three main categories described earlier (9).

Cross-sectional imaging is usually with contrast CT, including arterial phase enhancement, of the abdomen, chest, and pelvis. MRI is the most sensitive modality for liver metastases (18). Studies of CT in carcinoid tumours show an overall sensitivity of 80% in detecting lesions (19, 20). The sensitivity and specificity of CT and MRI alone are lower than the combination of ¹¹¹In-octreotide scan with CT or MRI (21).

Nuclear medicine imaging is important in staging of disease and determining suitability for therapy with SST analogues and peptide receptor. The two main nuclear medicine scans used in staging NETs are ¹¹¹In-octreotide and ¹²³I-metaiodobenzylguanidine (MIBG), with newer modalities, including PET scanning, being introduced.

There are five different SST receptor (SSTR) subtypes, all of which have strong affinity for SST (22). Octreotide is an SST analogue which has a strong affinity for SSTR-2 and to a lesser extent SSTR-5 receptors. NETs predominantly express SSTR-2. Synthetic radiolabelled SSTR analogues, such as ¹¹¹In-pentetreotide, enable SSTR scintigraphy to be performed (23).

SSTR scintigraphy is now established in localizing NET (24). Prospective studies have shown that inclusion of SSTR scintigraphy in the diagnostic work-up of patients alters management in up to 47% of cases (25). The sensitivity of SSTR scintigraphy for the detection of GEP NETs has been well studied. The sensitivity has been reported to be between 67 and 100%, with no significant difference in carcinoid tumours from foregut, midgut, or hindgut origin (26–28). With pancreatic NETs, sensitivity of SSTR scintigraphy is dependent on the type of functional tumour. Gastrinomas detection has a sensitivity between 56 and 80%, VIPoma is 60–70%, and insulinoma lower at 50% due to a lower expression of SSTR-2 (23). With phaeochromocytomas, SSTR scintigraphy is often negative and other imaging modalities, such as MIBG, should be used. Medullary thyroid cancer express SSTR-1 therefore may be negative on SSTR scintigraphy. False-positive scans can be seen in patients with chronic inflammation and granulomatous disease. SSTR scintigraphy detection is also affected by the size of NET and will often not detect lesions less than 1 cm (29).

MIBG has been used for two decades to visualize carcinoid tumours. The method was initially developed to detect phaeochromocytomas. MIBG shares the same method of uptake as noradrenaline and is not dependent on SSTR receptor expression. In phaeochromocytomas, MIBG has sensitivity of 87% and specificity of 99%; however, for carcinoid tumours it only has 50% sensitivity and specificity, whilst in pancreatic NETs uptake may be seen in less than 10% of cases (30). In general, ¹²³I-MIBG scintigraphy was shown to be less sensitive than ¹¹¹In-octreotide in identifying carcinoid tumours (30).

PET scanning in other malignancies is well established; however, its role for NETs is still evolving. [¹⁸F]2-fluoro-2-deoxy-glucose (FDG)-PET is only suitable for high-grade tumours and is of minimal use in low-grade tumours due to their slow glucose turnover. Experimental agents of interest include gallium-68 (Ga-68) DOTA-octreotide and Ga-68 DOTA-octreotate, 5-hydroxytryptophan (5-HTP) and 3,4-dihydroxyphenylalanine (DOPA). Studies with Ga-68 DOTA-octreotide had a greater sensitivity than conventional SSTR scintigraphy (31, 32) (Fig. 6.3.1).

If the primary site has not been identified by conventional imaging, it is worthwhile performing endoscopy of the upper and lower gastrointestinal tract. In addition, if patients are known to have a primary lesion in the gastrointestinal tract endoscopy will allow visualization of the lesion and the option of histological diagnosis. For detection of gastric, pancreatic, and duodenal lesions, endoscopic ultrasonography is a sensitive method for staging disease and providing information regarding depth of invasion and potential resectability of the lesions. In addition, biopsies can be performed to provide histological diagnosis. Endoscopic ultrasonography has an accuracy of 90% in staging of rectal carcinoids (33).

Capsule endoscopy can be used to diagnose small bowel carcinoid tumours, and appears to be at least as good as enteroscopy for identifying lesions. Obviously, the drawback is the inability to obtain a histological diagnosis. In small case series there appears to be advantage of capsule endoscopy over conventional small bowel investigations using CT and barium follow-through (34). To exclude the possibility of obstruction, a barium follow-through should be performed prior to capsule endoscopy.

For bronchial NETs, which commonly arise in large to midsize airways, bronchoscopy is of use in assessing the lesion and obtaining histological diagnosis (5).

Therapies for NETs incorporate those required for control of symptoms due to hormonal secretion from tumours, and also antiproliferative therapies. The management of NETs requires the use of a number of different therapies including: surgery, biotherapy, chemotherapy, peptide receptor targeted therapy, and tumour embolization. The best way to provide the most appropriate management plan for patients is through a multidisciplinary approach. Different therapies may be required at different clinical stages, and in patients with indolent disease and mild symptoms merely symptomatic relief may be all that is required for some years.

Surgery is the only method of cure and therefore should be considered and undertaken in all patients where feasible. In patients with localized tumours resection of the primary lesion should be performed, especially with bronchial tumours which are often localized. Debulking surgery should be considered in cases where increasing hormonal symptoms are present that cannot be controlled using medical therapy.

SST is a small polypeptide hormone, which occurs naturally in the human body and binds with a high affinity to the five recognized SSTRs. Activation of SSTRs leads to activation of common signalling pathways, such as inhibition of adenyl cyclase and modulation of mitogen activated protein kinase through G-protein dependent mechanisms (35). The effect of SST on tumour growth may be through the suppression of the synthesis and secretion of growth factors and growth promoting hormones. SST also appears to inhibit angiogenesis and cell proliferation in in vitro models. Its antiangiogenic effect appears to be through inhibition of angiogenic factors such as vascular endothelial growth factor, insulin-like growth factor-1, and platelet-derived growth factor (36, 37).

Short-acting octreotide, which needs to be administered three times a day. Long-acting octreotide-LAR and Lanreotide Autogel have a 28 day duration of action (38). Both are equally effective at controlling symptoms related to carcinoid syndrome, with improvement seen in approximately 85% of cases (39). Biochemical markers, such as chromogranin A and urinary 5-HIAA, are found to decrease by at least 50% in 60–80% of cases following therapy (40). In a study performed by Garland et al., of 27 patients with positive SSTR scintigraphy and commenced on octreotide-LAR, all had good symptom control initially; however, the majority of patients developed progressive disease and required further therapies for symptom control (41). Side effects include gastrointestinal disturbances, including pancreatic insufficiency which may require enzyme replacement therapy, gallstones, and glucose intolerance. Tolerance to SST analogues is a recognized phenomenon and there is a need for new biotherapy agents. Pasireotide, a new multiligand SST analogue, is currently being trialled. Recent studies have demonstrated that the majority of NETs coexpress dopamine and SSTRs, which has led to development of chimeric agents; these have shown promising results in NET cell lines (42, 43).

Interferon therapy has been used for symptomatic control in patients with NETs since 1982. It has been found to be beneficial in reducing symptoms of flushing and diarrhoea in patients with carcinoid syndrome in 50–60% of cases. Significant biochemical responses are reported in 40–50% of cases (44). Its mechanism is action is unclear though is thought to act through antisecretory and immunomodulatory functions. Its antitumour effect is not as pronounced as with SST analogues, with radiological evidence of tumour regression being less common. In a study of 111 patients treated with interferon-α, 15% demonstrated a greater than 50% reduction in tumour size (45).

Studies have shown that disease stabilization occurs in 40% of patients following combined therapy with SST analogues and interferon-α, which is similar to that of SST analogues alone (46). A randomized study with over 100 patients showed there was no significant survival benefit of SST analogues with interferon-α compared to SST analogues alone (47).

Chemotherapy has been widely used in the treatment of NETs for over three decades. Its precise role is not clearly defined; it is, however, often used as first-line therapy for unresectable, poorly differentiated NETs and pancreatic well-differentiated NETS, which are often chemosensitive. Studies have demonstrated wide variation in response rates with chemotherapy; this may, in part, be due to inclusion of different types and grades of NETs. The overall response rate for intestinal carcinoid is less than 30% (48–55).

Metastases from NETs are often isolated to the liver and therefore embolization of the liver can result in necrosis of tumour tissue and consequent decrease in hormonal secretion. Embolization is commonly performed radiologically and can be performed with particles or chemoembolization. Contraindications to performing hepatic artery embolization include: portal vein thrombosis, liver failure, and biliary reconstruction.

Symptomatic response is seen in 40–80% of cases, with a biochemical response (56) for hepatic embolization of 7–75%, and 12–75% for hepatic chemoembolization (57). In the latter study (57), Gupta et al. demonstrated no additional benefit of chemotherapy to transarterial hepatic embolization in metastatic midgut tumours. Complications postprocedure include ileus, portal vein thrombosis, hepatic abscess, hepatic fistula, encephalopathy, and renal insufficiency.

The overexpression of SSTR-2 has allowed for the development of targeted peptide receptor therapy. The mechanism of action appears to be that the radiopeptide binds to the SSTR-2 receptor and is internalized by the cell, thereby delivering radioactivity for a long period of time, with beta emitting radionuclides irradiating neighbouring tumour cells. Contraindications include bone marrow suppression, renal impairment, liver failure, very poor performance status, and inability to self care. A number of studies have been published using peptide receptor radionuclide therapy; however, the criteria for objective response has varied in studies (Table 6.3.3). The two radiopeptides that are currently in use are Yttrium-90 and Lutetium-177. Unfortunately, there are no randomized studies of peptide–receptor radionuclide therapy, thus evaluation of their true benefit and optimal radionuclide is difficult.

Kwekkeboom et al. recently published the largest series to date of over 500 patients treated with Lu177-DOTATATE (58). Of these patients, response data was available in 310: 2% had complete response, 28% partial response, and 16% had minor response. The median time to progression was 40 months and median overall survival from start of treatment was 46 months; median survival from diagnosis was 128 months. The overall survival for these patients seems much higher than historic controls, were survival was usually around 60 months.

The anatomical site and biology of NETs is important in determining management. With the wide variety of therapies and a number of trials underway, patients with NETs are best managed in a multidisciplinary team setting in a specialist centre. Further randomized control trials are needed to determine the optimal treatments for patients, which in view of the rarity of the cancers need to be performed in national and international studies.