Work up of Gastroenteropancreatic Neuroendocrine Tumors

SYNOPSIS

GEPNETs are rare causes of abdominal cancer. Their spectrum of presentation overlaps with other intra-abdominal neoplasms but can have unique features related to diverse hormonal secretions and desmoplasia. The workup of patients for possible GEPNETs involves initial recognition of unusual clinical features associated with the tumors, followed using complementary cross-sectional and functional imaging, analysis of blood or urine concentrations of appropriate tumor-secreted chemicals, and tissue biopsy. Functional imaging takes advantage of the NET-specific expression of somatostatin receptors. There are also characteristic features supporting the diagnosis on contrast-enhanced cross-sectional imaging. The use of tumor markers for biochemical diagnosis requires an understaging of the confounding variables affecting these assays. There are unique and specific immunohistochemical staining and grading requirements for appropriate diagnosis of these tumors.

INTRODUCTION

The workup of patients with neuroendocrine tumors of the gastrointestinal tract or pancreas (GEPNETs) generally involves initial recognition of the clinical features associated with the tumors, followed by appropriate use of cross-sectional and functional imaging, analysis of appropriate tumor markers, and tissue biopsy. Some patients present incidentally, and workup then involves imaging, biochemical testing and biopsy. This article will review these aspects of GEPNET management.

CLINICAL PRESENTATION

The presentation of GEPNETs may be incidental or based on symptoms. Symptoms may relate to mass effects, size and site of the primary tumor or metastatic disease, or to hormonal secretions of these tumors. Both hormonal and mass symptoms may differ based on the site of origin of the primary tumor.

INCIDENTAL PRESENTATION OF NETs

Screening colonoscopy may commonly reveal incidental findings of small rectal neuroendocrine tumors (NETs) ¹ or, very rarely, reveal terminal ileal NETs ². Appendicular NETs may be found incidentally in appendicitis specimens (where the small NET lesions at the appendiceal tip may not have played a role in the development of appendicitis ³). Up to 40% of pancreatic NETs may be noted incidentally ⁴. Gastric and duodenal lesions may also be noted incidentally on endoscopy for non-specific symptoms. Consistent with the increased rate of incidental diagnosis of NETs, the increased incidence of NETs is partially related to sooner diagnosis of early stage lesions ^5,6. The relatively slow growth of these tumors may enhance the likelihood of being diagnosed at some time during another presentation or ‘routine’ endoscopy.

SYMPTOMATIC PRESENTATION OF NETs

PRESENTATION RELATED TO SIZE AND SITE OF PRIMARY OR METASTATIC TUMORS

The specific presentations of NETs are partially dependent on their sites of origin, as follows:

Gastric NETs

These consist of 3 etiologically distinct groups ⁷: type 1 associated with chronic atrophic gastritis, low gastric acid and high gastrin; type 2 associated with functional gastrinoma, high gastric acid and gastrin; and type 3 with generally normal gastric acid and gastrin. Of these, type 3 may be associated with mass effect of the primary tumor and lymph nodal or hepatic metastatic disease. Type 1 lesions are generally small and multiple, and these patients present with non-specific dyspeptic symptoms of chronic atrophic gastritis, frequently in the setting of a personal or family history of other autoimmune disorders, e.g., thyroid disease ⁸. Type 2 lesions also tend to be small and multiple, with a low propensity for metastatic spread, and present with peptic ulceration or duodenal mass effect of the associated gastrinoma. However, Type 3 lesions generally present as larger (>2cm), higher grade, single lesions with a high metastatic risk. Their presentation may be like gastric adenocarcinoma with abdominal pain, weight loss and iron deficiency ^9,10. Type 3 gastric NETS may occasionally present with atypical flushing (patchy, cherry red, pruritic, prolonged, with chemosis) associated with histamine release.

Duodenal NETs

These are generally small unifocal lesions which may metastasize to local nodes. They may be associated with intestinal obstructive features, anemia, jaundice (if close to the ampulla of Vater), or hormonal features of Zollinger Ellison Syndrome (if gastrinoma) ¹¹.

Duodenal NETs may also present with features of type 1 neurofibromatosis ¹².

Pancreatic NETs (PNETs)

These are referred to as either functional PNETs or non-functional PNETs. The non-functional PNETs (approximately 80% of all PNETS) consist of those that are not associated with the production of any hormonal output, or the production of hormonal output lower than a threshold needed to produce symptoms, or the secretion of substances with no known associated symptoms (e.g., chromogranin A, pancreatic polypeptide). Non-functional tumors tend to be bigger than functional tumors on diagnosis ¹³. They may present with abdominal pain (from primary or related to bulky nodal or hepatic metastases), early satiety, variceal GI bleeding, or jaundice ¹⁴.

Jejuno-ileal NETs

These tumors may present with abdominal pain ^15,16 related to the primary tumor, metastatic nodal, peritoneal or hepatic lesions, or desmoplastic reaction. Pain may relate to acute or recurrent sub-acute small bowel obstruction. Gastrointestinal bleeding and anemia may occur from ulceration of the primary tumor, or varices from venous obstruction due to mass effect, venous infiltration or perivascular fibrosis of mesenteric vessels. There may be associated nausea, vomiting, weight loss, and early satiety.

Colorectal NETs

Seventy percent of rectal NETs are noted incidentally during screening colonoscopies and are localized in more than 80% of cases at diagnosis ¹⁷. Less frequent presentations are with altered bowel habit (13%), rectal bleeding (6%), abdominal or rectal pain (5%). Colonic NETS tend to present later, frequently with bulky metastatic nodal or liver disease, and symptoms of abdominal pain, gastrointestinal blood loss or obstruction ¹⁸.

GEPNET Metastases

In tertiary referral centers up to 77% of pancreatic and 91 % of intestinal NETS will present with distant metastases ¹⁹. With all GEPNETS the predominant sites of metastatic disease are liver, local and distant lymph nodes, peritoneum and mesentery, followed by bone, with lung and brain more rarely involved ²⁰. Metastatic disease can be symptomatic by bulky nodal, mesenteric, or liver lesions causing abdominal pain, early satiety, nausea, and weight loss. Nodal or mesenteric lesions, with disease encasing the mesenteric vasculature, may cause intestinal ischemia and post prandial abdominal pain. Hydronephrosis may occur because of retroperitoneal or mesenteric nodal disease, or related desmoplastic reaction, resulting in flank pain, sepsis and decreased renal function. Either the bulk of liver disease, or adenopathy in the porta hepatis, or adenopathy or mesenteric lesions around the superior mesenteric vein, can be associated with portal vein or mesenteric vein thrombosis, resulting in hypersplenism, varices, and/or ascites.

Desmoplastic Reaction

Jejuno-ileal NETS are particularly associated with desmoplastic reactions ²¹. Evidence of intra-abdominal fibrosis may be noted on cross-sectional imaging in up to 50% of those presenting with small bowel NET ²². The fibrosis frequently leads to the development of bowel ischemia, small bowel obstruction or hydronephrosis ²³. It can be difficult to determine if some of these symptoms are related to bulk of disease or desmoplastic reaction. The etiology of the fibrotic desmoplastic reaction in GEPNETs is still unclear. While multiple lines of evidence suggest that serotonin causes fibrotic reactions in the right sided heart valves and abdomen, other NET-secreted tachykinins and transforming growth factor (TGF) β may also play a role ²⁴.

PRESENTATION RELATED TO HORMONAL SECRETION OF PRIMARY OR METASTATIC TUMORS

Gastrointestinal NETs (GI-NETs)

The presentation of GI-NETs related to their hormonal output depends on the specific hormone(s) that are produced by the tumors and their metastases. The hormones produced vary by site of the primary GI-NET (Table 1). For example, gastric NETs can produce histamine from enterochromaffin-like (ECL) cells, mid gut NETs (jejunum to cecum) produce serotonin from enterochromaffin (EC) cells, whereas serotonin is not produced by colorectal NETs, which are derived from peptide YY- and enteroglucagon-expressing L cells.

Table 1.

Hormonal secretions by tissue

Tissue	Hormones	Symptoms / Syndrome
Gastric	Histamine, CGA	Atypical flush, wheeze, angioedema
Duodenal	CGA, Somatostatin, Gastrin	Cholelithiasis, steatorrhea, diabetes, ZE Syndrome (gastrinoma)
Jejuno-ileal, appendiceal, cecal	Serotonin, CGA, Pancreastatin	Carcinoid Syndrome
Colorectal	Pancreatic polypeptide	No hormonal symptoms
Pancreatic	See Table 2	See Table 2

CGA: chromogranin A; ZE: Zollinger Ellison.

The classical carcinoid syndrome describes a symptom complex which may include flushing (90%), diarrhea (70%), abdominal cramping (40%), valvular heart disease (30%), telangiectasia (25%), wheezing (15%), or pellagra (5%) ²⁵. While many tumor-secreted chemicals may play a role in the symptoms of carcinoid syndrome, serotonin is thought to play a dominant role in the diarrhea and pro-fibrotic effects (cardiac valvular disease and desmoplastic reaction). Histamine, prostaglandins, and substance P may play dominant roles in the carcinoid flushing reaction ²⁶. Flushing associated with small intestinal NETs is predominantly facial, lasting less than a minute, and has a cyanotic hue, whereas that from gastric NETs may be bright red, prolonged, more widely spread, and associated with chemosis and wheeze or other histaminergic features ²⁶. (Table 1)

In a recent analysis of the Surveillance, Epidemiology, and End Results (SEER) database, features of carcinoid syndrome were recorded most often in patients with small bowel and cecal NETs (32%), appendiceal (17%), colorectal (11%) and lung (8%) ²⁷ (patients with PNETs were not analyzed). Carcinoid syndrome data for colorectal tumors may have been overestimated in this study due to misclassification of cecal tumors, since descending colon and rectal tumors are generally derived from neoplastic L-cells, which do not secrete serotonin, whereas cecal (and small bowel) tumors are derived from serotonin-secreting EC cells.

The prevalence of carcinoid syndrome increases with the presence of liver metastatic disease 27. This is thought to relate to the metabolic degradation of serotonin and other carcinoid syndrome hormonal mediators delivered to the liver via the portal vein, whereas serotonin generated in liver metastases is delivered into the hepatic veins, entering the systemic circulation without further liver metabolism. Carcinoid syndrome may occasionally occur in individuals without evident liver metastatic disease. This may be due to unrecognized liver metastatic disease ^27,28, or to tumor with systemic rather than portal venous drainage (some bulky retroperitoneal or pelvic nodes, peritoneal implants, primary ovarian or lung NETs).

Pancreatic NETs

Functional PNETs (approximately 20% ²⁹) present with characteristic symptoms related to the predominant hormone secreted. The occurrence of carcinoid syndrome in patients with PNETs is unusual (<1%) and relates to the small proportion of patients with PNETs expressing serotonin (8%) ³⁰. Ultra-rare PNETs secreting renin, luteinizing hormone, insulin-like growth factor (IGF) 2, cholecystokinin (CCK), glucagon-like peptide (GLP) 1 have also been described, some associated with clinical syndromes ³¹. Specific hormones secreted by functional PNETs and the characteristic clinical presentation of each are noted in Table 2. Non-functional PNETs may be associated with elevated serum concentrations of chromogranin A, pancreatic polypeptide, or pancreastatin. (Table 2)

Table 2.

Hormonal secretions in functional PNETS

Hormone	Features	% of F-PNET	% Malignant	Additional Notes
Insulin	Recurrent hypoglycemia	40%	10%
Glucagon	Diarrhea, glossitis, necrolytic migratory erythema, weight loss, hyperglycemia, blood clots	~5%	80%
VIP	Diarrhea, hypokalemia, achlorhydria	<5%	80%	Also in colon, liver, adrenal tumors
ACTH	Cushingoid facies, weight gain, diabetes, hypertension	<1%	>80%	Co-existing ZE 35%, insulinoma 5%
GHRH	Acromegalic features, diabetes	<1%		Also in lung NETs, 75% have MEN1
PTHRP	Hypercalcemia	<1%		Also in multiple other cancers
Gastrin	Pain, Diarrhea (ZE Syndrome)	20%	>90%	Also in duodenum NETs, 25% have MEN1
Somatostatin	Diabetes, cholelithiasis, steatorrhea, weight loss	<5%	75%	Also in duodenum, jejunum, 50% have NF1
Serotonin	Flushing, diarrhea (Carcinoid syndrome)	<1%	>95%	8% with elevated urine 5HIAA without syndrome

ACTH: adrenocorticotrophic hormone; F-PNET: functional pancreatic neuroendocrine tumor; GHRH: growth hormone releasing hormone; 5HIAA: 5-hydroxyindoleacetic acid; MEN1: multiple endocrine neoplasia type 1; NF1: neurofibromatosis type 1; PTHRP: parathyroid hormone-related peptide; ZE: Zollinger Ellison; VIP: vasoactive intestinal polypeptide.

Data from Refs ^{14,30,32–34}.

Presentation of genetically driven GEPNETs

The majority of GEPNETs are sporadic but about 10 – 20% of patients with PNETs harbor genomic mutations, some of which produce wellcharacterized syndromes. These are generally autosomal dominant in inheritance pattern, although family history maybe lacking due to de novo gene mutation, variable penetrance of features, incomplete family history, etc. Pancreatic NETs are rarely the first manifestation of any of these syndromes and so other features should be evident on history. The possibility of an associated genetic syndrome should be considered in all patients with PNETs and further genetic testing pursued if there are features suggestive of genetic mutations, as the management of the NET may differ in the context of genetic disorders ³⁵. Table 3 lists the specific PNET-associated genetic disorders which need to be considered.

Table 3.

PNET-associated genetic disorders

Syndrome	Gene/ inheritance	Incidence of PNETs	Associated features	Differences from incidental
Multiple Endocrine Neoplasia Type 1	MEN1 / AD	~60%	Hyperparathyroidism 95%; pituitary adenoma 40%; duodenal NET, adrenal adenoma, lung & thymus NET, lipoma, angiofibroma	Multifocal microadenomas
Von Hippel-Lindau	VHL / AD	~15%	Cranial and retinal hemangioblastoma 70%; renal cell carcinoma 90%; pheochromocytoma 15%; endolymphatic cell tumor 10%.
Neurofibromatosis Type 1	NF1 /AD	<10%	Neurofibroma, optic glioma, café au lait spot, Lisch nodule	Somatostatinomas
Tuberous Sclerosis	TSC1 / AD	Rare	Multiorgan hamartoma, epilepsy, angiofibroma, mental retardation
Glucagon cell hyperplasia and neoplasia	GCGR / AR	Rare		Nesidioblastosis with α cell hyperplasia; High plasma glucagon without diarrhea, rash or diabetes 39

AD: autosomal dominant; GCGR: glucagon receptor gene; MEN1: multiple endocrine neoplasia type 1; NF1: neurofibromatosis type 1; TSC1: tuberous sclerosis complex subunit 1; VHL: Von Hippel-Lindau.

Duodenal NETs may also be associated with multiple endocrine neoplasia, type 1 (MEN1) or neurofibromatosis, type 1 (NF1). Familial occurrence of small bowel GI-NETs has rarely been described ³⁶. A genomic mutation in Inositol Polyphosphate Multikinase has been proven in one family, whose NETs presented in a manner typical of sporadic GI-NETS but with more tumor multifocality ³⁷. Recently, genomic mutations in MUTYH (encoding a DNA repair enzyme and also associated with Familial Adenomatosis Polyposis), and CHEK2 and BRCA2 (encoding tumor suppressor genes and also associated with breast, ovarian and prostate cancers), were described in some patients with PNETs ³⁸. (Table 3)

THE ROLE OF IMAGING IN WORKUP OF PATIENTS WITH GEPNETs

Introduction

For patients with NETS, both cross-sectional (CT, MR and US) and functional (⁶⁸Gallium-DOTATATE PET, Octreoscan®, ¹⁸FDG-PET) imaging are used to determine that the patient has neuroendocrine cancer, localize the primary tumor, stage the disease, achieve adequate diagnostic biopsy, determine amenability for surgery or other intervention, and follow response to therapy. Cross-sectional and functional imaging are used in a complementary manner to take advantages of the relative strengths of each modality. The relative role and reported accuracy of different modalities vary significantly based on availability of newer technologies in CT and MR over time, recent availability of ⁶⁸Ga-DOTATATE PET imaging, and variable local expertise with these modalities and with endoscopic US (EUS).

Functional Imaging

The role of functional imaging is further described elsewhere in this issue. 68 Gallium-DOTATATE PET and Octreoscan® functional imaging for GEPNETs are based on the unique feature of high-level somatostatin receptor expression in 85–100% of well-differentiated GEPNETs ⁴⁰. With the 2016 approval of ⁶⁸Ga-DOTATATE preparations for use in PET imaging, the ⁶⁸Ga-DOTATATE PET/CT (or MR) has replaced the Octreoscan® as the functional imaging agent of choice for well-differentiated NETs, due to far higher sensitivity ⁴¹, particularly for lesions less than 2cm ⁴², coupled with ease of use (less radiation exposure and 1–2 hour total duration rather than 2 days with Octreoscan®). Newer somatostatin receptor-based PET imaging agents are being developed, e.g., ⁶⁴Cu-DOTATATE, which promise to broaden the availability of this imaging technology further. Given its expense, professional societies have developed ‘appropriate use criteria’ for ⁶⁸Ga-DOTATATE PET imaging. Appropriate indications include: localization of the primary tumor, initial staging after histologic diagnosis, pre-operative staging, evaluation of masses not amenable to biopsy, evaluation of patients with biochemical evidence and symptoms of NET without evidence of tumor on conventional imaging or prior biopsy, and others ⁴³.

Functional imaging in patients with unknown primary NETs

Approximately 12% of patients with NETs have unknown primary lesions, despite standard cross-sectional imaging and biopsy ⁵. Functional imaging with ⁶⁸Ga-labelled somatostatin analogs and PET/CT can improve primary tumor identification for a further 38–59% of patients ⁴⁴. It is important to be aware of the potential for false positive localization of ⁶⁸Ga-DOTATATE in the uncinate process of the pancreas and in areas of inflammation ⁴⁵. Intrapancreatic accessory spleen can also be misdiagnosed as a small PNET on both CT and ⁶⁸Ga-DOTATATE PET and may require further nuclear imaging to resolve ⁴⁶.

Identifying the primary tumor is important because multiple retrospective studies suggest that resection of the primary NET, along with debulking of lymph nodes and hepatic metastases, may have a survival benefit for NET patients ⁴⁷. Identifying the primary tumor is also important because NETs from different sites vary in their prognosis ⁵ and their response to therapies ⁴⁸. Furthermore, many clinical trials recruit only patients with NETs of specific tissue origin.

¹⁸ FDG-PET imaging for high-grade and poorly-differentiated NETs

Expression levels for somatostatin receptors, and thus positivity on ⁶⁸Ga-DOTATATE imaging, declines with Grade 3 and poorly-differentiated lesions ⁴⁹. However, Grade 3 and poorly-differentiated GEPNET lesions may be positive on ¹⁸FDG-PET scans in up to 100% of cases. Thus, ¹⁸FDG-PET scanning is helpful in the staging of patents with high-grade tumors and identification of those at risk for more rapid progression ⁵⁰.

GEPNET features on cross-sectional imaging

Jejuno-ileal NETs are generally small (mean 2cm, range 0.4–6.6cm ⁵¹) and primary tumors are frequently not detected by cross-sectional imaging. Cross-sectional CT or MR imaging may suggest a small bowel localization of a primary tumor by demonstrating the consequences of desmoplastic reaction and mesenteric fibrosis, or by enlargement of lymph nodes (calcified in up to 70% ⁵²) in characteristic nodal fields along the superior mesenteric artery ⁵³, or by the demonstration of other specific sites of metastatic disease ⁵⁴. Oral contrast (e.g., water for upper intestinal lesions or dilute barium in sorbitol for distal small bowel lesions, with use of an anticholinergic to inhibit peristalsis) may enhance finding the primary intestinal lesion(s) by producing bowel distension and limiting motion artifact ⁵⁵. Other options for identifying small bowel primary lesions include capsule endoscopy, CT enteroclysis, and double balloon enteroscopy ⁵⁶, although the role of these investigations is unclear in comparison to careful intra-operative palpation of the complete small bowel ⁵¹. The role of CT or MR scans in gastric, duodenal, rectal, or colonic NETs is to adequately stage the disease, detecting local and regional adenopathy and distant metastases, and to assess the degree of local invasion and vascular compromise. Direct or EUS-aided visualization by endoscopy is p in these sites.

GEPNETs and their metastases generally enhance (appear hyperintense) during the arterial phase of intravenous contrast infusion on CT and MR, with low intensity on T1 and intermediate to high intensity on T2 weighted non-contrast images on MR. Hepatic metastases may be isointense on portal venous phase CT imaging, or may present as low attenuation lesions in this phase. Hence, the use of ‘triple phase’ CT imaging, incorporating pre-contrast, arterial phase, and portal venous phase imaging, can be crucial to diagnosing these lesions ⁵⁷. Magnetic resonance, especially using hepatic arterial phase images, fat-suppressed fast spinecho T2-weighted images, diffusion-weighted (DW) images, and hepatocyte-specific contrast media (e.g., gadoxetic acid), demonstrates significantly more hepatic metastatic lesions than CT imaging ⁵⁸, although even these protocols miss abundant micrometastases ⁵⁹. Magnetic resonance cholangiopancreatography (MRCP) imaging sequences can define specific obstruction and infiltration around the biliary and pancreatic ducts, although the better spatial resolution of CT can also provide clear pre-operative information on pancreatic and bile ducts patency and vascular encasement.

Endoscopic Ultrasound

While transabdominal US has a relatively limited diagnostic imaging role besides facilitating accurate tissue biopsy, EUS has an important role in imaging and biopsy of pancreatic NETs and in imaging of gastroduodenal and rectal lesions and local nodal screening. Endoscopic US in rectal and gastric lesions can define the depth of invasion and clarify the appropriate treatment strategy (endoscopic submucosal resection versus surgical resection). It offers the highest available imaging sensitivity for pancreatic NETs, although delineation and biopsy of pancreatic tail lesions may be less accurate. However, EUS is highly operator dependent and technically challenging.

Finally, patients with serotonin secreting NETs or elevated N-terminal pro-B-type natriuretic peptides (NT-proBNP) should have transthoracic echocardiography to evaluate for carcinoid valve disease ⁶⁰. Specific details of available imaging modalities for GEPNETs are noted in Tables 4 and 5.

Table 4.

Imaging modalities available for GEPNETs

Modality	Advantages	Disadvantages	Indications	Sensitivity (%)	Specificity (%)
CT Abdomen and Pelvis (Triple Phase)	Cost, wide availability, fast imaging, good spatial resolution, good liver views (triphasic)	Misses smaller or isodense liver lesions, or small nodes and bone lesions	Initial diagnosis / staging, biopsy, pre-op planning	82	86
MR with / with-out contrast; hepatocyte-specific contrast (e.g., gadoxetate)	Best for liver, pancreas, bone, brain; better differentiation of normal vs. tumor; no radiation (preferred for MEN1 screening)	Longer scan times – movement artifact in bowel / abdominal nodes, claustrophobia; less spatial resolution; cost (MR>CT)	Initial diagnosis / staging, pre-op planning	79 (PNETs) 75 (Liver metastases)	100 (PNETs) 98 (Liver metastases)
68 Ga- DOTATATE PET	Visualizes small primary GI-NETs, lesions in bone, small nodes and peritoneum, missed on CT or MR	Cost (PET>>MR >CT); availability; SSTR2 negative lesions	Grade 1, 2, WD Grade 3; specific indications43	92 (All NETs) 52 (primary unknown)	88 (All NETs)
18 FDG-PET	Visualizes higher grade 2 / grade 3, & PDNET	Cost, generally negative or poor uptake in WDNET	Grade 3, PDNET	52% (G1, G2 GEPNETS) 100% (PDNET) 61
EGD / Colonoscopy/ EUS	Direct visualization & biopsy of gastric, duodenal, colorectal, some terminal ileal NETs, visual-	Operator dependent, EUS expertise limited, difficulty with NET det- ection in tail of pancreas	Imaging and guided biopsy in stomach, duodenum, pancreas, colorectal (incidental	EUS 86 (PNETs)	EUS 92 (PNETs)
	ization of strictures, bleeding sites, wall thickness, adjacent nodes; high definition pancreas evaluation and biopsy		findings in terminal ileum)
Transabdom- inal US	Availability	Operator dependent, poor imaging quality	Guided biopsy	88 (Liver metastases)	95 (Liver metastases)
				39 (PNET detection rate)

SSTR2: somatostatin receptor 2; WD: well-differentiated; PD: poorly-differentiated.

Data from Sundin A, Arnold R, Baudin E, et al. ENETS Consensus Guidelines for the Standards of Care in Neuroendocrine Tumors: Radiological, Nuclear Medicine & Hybrid Imaging. Neuroendocrinology. 2017;105(3):212–244.

Table 5.

Indicated imaging by disease site

NET Location	Recommended imaging
Gastric Type 1, 2, 3	Recommend EGD in Types 1–3; consider EUS (for infiltration, nodal assessment); CT or MR staging with Type 3; consider 68Ga-DOTATATE PET for Types 2 and 3
Duodenal	Recommend EGD; consider EUS (for infiltration, nodal assessment); recommend CT or MR abdomen for staging; consider 68Ga-DOTATATE PET
Jejuno-Ileal, Appendix, Cecum	Recommend CT or MR abdomen/pelvis; consider specific enterography or enteroscopy protocols for jejuno-ileal or colonoscopy for cecal; recommend 68Ga-DOTATATE PET for staging; consider echocardiogram
Distal Colo-Rectal	Recommend CT or MR abdomen/pelvis; recommend colonoscopy; consider EUS for rectal lesions; recommend 68Ga-DOTATATE PET
Pancreatic	Recommend CT or MR for staging; consider EGD / EUS for staging / biopsy; recommend 68Ga-DOTATATE PET for staging
Poorly differentiated NE carcinoma (NEC)	Recommend CT chest/abdomen/pelvis; Consider 18FDG-PET, 68Ga-DOTATATE PET, Nuclear bone scan

EGD: esophagogastroduodenoscopy.

Data from Kunz PL, Reidy-Lagunes D, Anthony LB, et al. Consensus guidelines for the management and treatment of neuroendocrine tumors. Pancreas. 2013;42(4):557–577.

THE ROLE OF BIOCHEMICAL MARKERS IN WORKUP OF PATIENTS WITH GEPNETs

Since GEPNETS arise from secretory cells of the GI tract and pancreas, it is not surprising that multiple (>20) proteins, peptides, and amines have been documented as secretory products of these tumors ⁶³. Measurement of these chemicals in blood or urine has been used extensively as an aid to diagnosis and monitoring of patients with GEPNETs. Some of the chemicals secreted result in evident symptoms (e.g., serotonin-induced diarrhea, or insulin-induced hypoglycemia), while others are not associated with evident symptomatic effects (e.g., chromogranin A (CGA)). A minority of GEPNET patients have ‘functional tumors’ with 19% of newly diagnosed patients with GI-NETs presenting with carcinoid syndrome ²⁷ and 27% of patients with PNETs presenting with hormonal syndromes ¹⁵. There are commercially available assays for a subset of these secreted products. Some biochemical markers have also been suggested to have prognostic implications, e.g., neurokinin A ⁶⁴ or pancreastatin ⁶⁵, though such claims have not yet been sanctioned by consensus panels ⁶⁶.

The appropriate role of biochemical markers in diagnostic testing remains somewhat controversial because of poor sensitivity for early disease, and multiple influences on specificity. The hepatic clearance of serotonin can be up to 80% ⁶⁷ and thus blood serotonin or its excreted breakdown product, urine 5-hydroxyindoleacetic acid (5HIAA), may not be abnormal until GEPNETs are associated with (serotonin-secreting) hepatic metastases. Additionally, the commonly used chromogranin A (CGA) is elevated by proton pump inhibitors (PPI), chronic atrophic gastritis, and renal dysfunction, amongst many other conditions ⁶⁸. Chromogranin A falls with a half-life of 5 days on withdrawal of PPI ⁶⁹. All current biochemical markers have issues affecting their sensitivity or specificity including medications, diet, fasting state, technique of phlebotomy, handling of sample, lack of standardization between laboratories, and others ⁷⁰. Results of these tests must to be considered in the context of these confounding factors.

Additionally, it is clear that there are cases of functional NETs where the secreted substance(s) causing the syndrome are not assayed in standard, regularly performed, assays. A quarter of patients diagnosed with carcinoid syndrome diarrhea have normal 24-hour urinary 5HIAA ⁷¹. Serotonin accounts for only 50 % of the pro-diarrheal activity in small intestinal NETs ⁷² and plays little role in flushing. Tachykinins, a family of peptides secreted by small intestinal NETs, can cause both flushing and diarrhea, but are not routinely assayed ⁷³.

Professional society recommendations for use of tumor markers in GEPNETs are noted in Table 6. In general, laboratory evaluation of functional pancreatic tumors is guided by the specific syndrome exhibited (e.g., insulin and glucose testing for hypoglycemia with suspected insulinoma), and CGA and pancreatic polypeptide (PP) are considered in non-functional tumors. In jejuno-ileal, appendicular and cecal NETs (all derived from transformed EC cells), CGA and some assay of serotonin concentration are suggested (serum or blood serotonin, plasma or 24 hour urine 5HIAA ⁷⁴), with consideration of testing pancreastatin if there are factors present which influence the level of CGA (PPI use, decreased renal function etc). While professional societies have not endorsed other biochemical testing, studies support consideration of pancreastatin and neurokinin A as prognostic markers in jejunoileal NETs^75,76.

Table 6.

Recommendations for use of biomarkers in diagnosis and management of NETs

Society / Organization	NET Type
	Colorectal	Jejuno-ileal, appendiceal, cecal	Gastroduodenal	Pancreatic	HighGrade GEPNEC
NANETS 62, 78, 79	CGA (Rec) U5HIAA (Cons)	CGA, U5HIAA (Rec)	Gastrin (Rec); U5HIAA, CGA (Cons)	CGA (Rec) U5HIAA, gastrin, glucagon, insulin proinsulin, VIP, PP, PTHRP, GHRH, ACTH, (others as clinically indicated)	No Rec
ENETS80, 81, 31, 16, 10, 18	CGA (Rec)	Appendix: CGA (Rec), U5HIAA (Cons) J-I: CGA, U5HIAA (Rec)	Gastric type 1, 2: Gastrin, CGA (Rec) Gastric Type 3: CGA (Rec) Duodenal: CGA (Rec) others as indicated clinically or genetically	NF-PNET: CGA (Rec) F-PNET: based on specific symptoms	CGA, NSE (Cons)
NCCN82	CGA, Urine or Plasma 5HIAA*			CGA, PP (others as clinically indicated)*

Cons: consider; Rec: recommend; ACTH: adrenocorticotrophic hormone; CGA: chromogranin A;

PP: pancreatic polypeptide; PTHRP: parathyroid hormone-related peptide; F-PNET: functional PNET; GEPNEC: gastroenteropancreatic neuroendocrine carcinoma; GHRH: growth hormone releasing hormone; J-I: jejuno-ileal; NF-PNET: non-functional PNET; NSE: neuron specific enolase; U5HIAA: 24-hour urine 5-hydroxyindoleacetic acid; VIP: vasoactive intestinal polypeptide.

^*All NCCN recommendations are Category 3 (“Based upon any level of evidence, there is major NCCN disagreement that the intervention is appropriate”).

To improve the clinical utility of tumor markers, there has been significant interest in the use of multianalyte marker testing related to miRNA profiling, isolation of circulating tumor cells, and analysis of circulating gene transcripts over-expressed in neuroendocrine tumors. The only commercially available multianalyte test is the NETest®. This test involves quantitative reversetranscription and polymerase chain reaction (PCR) amplification of 51 genes whose expression is enhanced in NETs. Early data suggests high sensitivity (>95%) and specificity (>95%) in the detection of all GEPNETs ⁷⁷. There are no evident confounders for the NETest® of the type that interfere with biochemical markers. The NETest® has not yet been recommended by professional organizations. (Table 6)

THE ROLE OF BIOPSY IN WORKUP OF PATIENTS WITH GEPNETs

An adequate biopsy and pathological examination by an expert pathologist are crucial in the workup of patients with GEPNETs. Biopsies may be percutaneous (US, CT or MR-guided), endoscopic, or by endoscopic US. Minimum requirements for pathological reports have been published ⁸³,⁸⁴ and reports should conform to the most recent World Health Organization classification ⁸⁵. In general, core biopsies are preferred to cytology ⁸⁶, to allow for full diagnostic assessment. However, cytology is frequently sufficient for pre-operative diagnosis of PNETs and discrimination of poorly-differentiated Grade 3 from well-differentiated Grade 1/2 morphology ³¹,⁸³. The diagnostic accuracy of pancreatic fine needle aspiration (FNA) cytology is 97% and the concordance rate for grade assessment, between Ki67 on FNA versus postoperative histology, is 83% ⁸⁷.

The pathological assessment includes conventional histology of hematoxylin and eosin (H&E)stained slides to establish that the morphology is consistent with the diagnosis of NET and to review the differentiation state, an important component of grading. These studies may also reveal evidence of mixed neuroendocrine carcinoma (MANEC), where greater than 30% of cells are of non-endocrine origin. For gastric carcinoids, biopsy of fundic mucosa and assessment for chronic atrophic gastritis is appropriate.

The tumor grade (which influences therapy options, since chemotherapy is preferred over other options for metastatic high-grade lesions) is established from immunohistochemical studies using the MIB-1 antibody to label Ki67, a cellular proliferation-related protein, to determine the Ki67 Proliferation Index, and (optionally) counting the number of mitotic cells per 10 high power fields or 2 mm² on H&E stained slides. Tumor grade may increase between primary and metastatic lesions ⁸⁸, and thus biopsy of a metastatic lesion may be preferred to initially guide therapy options.

A panel of immunohistochemical studies establish the tumor as a neuroendocrine tumor. These must include chromogranin A and synaptophysin to confirm that the biopsied lesion is a NET. Other markers may also be considered, e.g., cytokeratin to differentiate GEPNETs from pheochromocytoma or paraganglioma NETs ⁶²_,⁸³.

The site of origin of the tumor (if biopsy is from a metastatic lesion) is an important influence on the surgical approach. Specific immunohistochemical markers (e.g., CDX2 (prominent in small bowel), and PAX6 or ISL1 (prominent in pancreas)) may help to differentiate the cell of origin for metastases from an unknown primary. Further immunohistochemistry can be performed to confirm the specific hormones secreted by the tumor ¹⁹. These studies are generally not necessary since functional tumors are defined by the associated clinical syndrome and blood biochemistry ⁸⁹. However, these studies may be of use in some patients, e.g., with MEN1 where most gastrinomas arise from the duodenum and any accompanying PNETs are either non-functional or are secreting hormones other than gastrin ⁸⁹.

KEY POINTS.

• Gastroenteropancreatic Neuroendocrine Tumors may present incidentally, or with features related to the size and site of the primary tumor or metastases, or with symptoms related to the secretions of these tumors.

• All biochemical tumor markers have potential for false-positive and false-negative results.

• Cross-sectional imaging (performed with specific protocols) and functional imaging have complementary roles in the diagnosis and management of Gastroenteropancreatic Neuroendocrine Tumors.

• Biopsy assessment must fulfill updated (WHO 2017) criteria, to optimally guide therapy choices.

SUMMARY.

Understanding the unique clinical, laboratory, imaging, and pathology features of GEPNETs is important in the pre-operative evaluation of these rare tumors.