Neuroendocrine neoplasms (NENs) arise from the disseminated system of neuroendocrine cells and can occur in various parts of the body. However, they are most often found in the gastrointestinal tract and lungs. The term NENs includes both well-differentiated neuroendocrine tumours and neuroendocrine carcinomas (NECs), which account for 10–20% of all NENs. The following characteristics of NENs should be considered in the diagnostic and therapeutic process: proliferative activity, presence of somatostatin receptors (SSTRs), tumour growth rate, and extent of the neoplastic disease [1].

1. Incidence and epidemiology

Gastro-entero-pancreatic neuroendocrine neoplasms (GEP-NENs) are a rare, heterogeneous group of neoplasms that may produce hormones and biogenic (hormonally active) amines, but a significant majority of the tumours do not produce these substances in amounts that cause clinical symptoms. They constitute approx. 70% of all NENs and approx. 2% of all cancers of the gastrointestinal tract. The most common NEN locations in the gastrointestinal tract are the small intestine and the pancreas [1, 2]. The incidence of NENs in the years 1994–2009 increased from 2.48 to 5.86 per 100,000 persons/year [2]. Based on epidemiological studies conducted in the United States [the Surveillance, Epidemiology, and End Results (SEER) Program] and Norway [the Norwegian Registry of Cancer (NRC)], the most common gastrointestinal neuroendocrine neoplasms (GI NENs) originate in the small intestine or rectum (incidence of approx. 1.2/100,000 persons/year), pancreas (approx. 0.8/100,000 persons/year), and stomach or appendix (approx. 0.4/100,000 persons/year) [3–6]. Currently, the overall incidence rate for these neoplasms is 35/100,000 persons/year on average.

Between 1997 and 2012 there was a 6-fold increase in the prevalence of GEP-NENs. Local and regional NENs are diagnosed more frequently than those with distant metastases. The detectability of NENs is increasing, e.g. from 1973 to 2004 the incidence of NENs increased from 2.1 to 5.25 new cases per 100,000 persons/year, with the primary lesion most frequently found in the small intestine (37.4%). Since 2000, rectal NENs have been diagnosed more frequently than small intestine NENs [2, 6–8]. The incidence of GEP-NETs in the USA, based on SEER data, was 3.56/100,000 persons/year. In Europe, the incidence of GEP-NETs is also increasing — from 1.33 to 2.33/100,000 persons/year; however, these data come from different registries and are mainly retrospective. Higher incidence is observed among men (5.35/100,000 persons/year) compared to women (4.76/100,000 persons/year) [1, 5–7]. The vast majority of NENs are sporadic, well-differentiated tumours. However, GEP-NETs originating from the pancreas, duodenum, stomach, and much less frequently, from the thymus and lungs sometimes constitute an element of multiple endocrine neoplasia type 1 (MEN-1) syndrome. Pancreatic neuroendocrine tumours (PanNETs) may also be associated with von Hippel-Lindau (VHL) syndrome, tuberous sclerosis complex (TSC), and neurofibromatosis (NF). In these congenital diseases, NETs can be multifocal and occur 10–20 years earlier than in sporadic cases. The frequency of the hereditary causes (MEN-1, VHL) is estimated at about 5%. Genome studies revealed the presence of germline mutations in, e.g., MUTYH, CHEK2, and BRCA2 and a propensity to PanNETs in approximately 17% of the studied population [1].

2. Diagnostics

2.1. Biochemical diagnostics

Biochemical diagnostics of NENs involves the following:

A. Non-specific markers

The most frequently used diagnostic method is determination of the chromogranin A (CgA) concentration in the serum (less frequently in the plasma) [1, 8, 9]. CgA is a relatively stable protein in blood. However, there are two different methods for determining the concentration of CgA: radioimmunoassay (RIA) and enzyme-linked immunosorbent assay (ELISA) in the serum or plasma [9]. Unfortunately, there are no international CgA standards, and the differences between the available tests are significant. To monitor the course of the disease, it is advisable to determine CgA using the same method [7]. In their current recommendations, experts from the European Neuroendocrine Tumour Society (ENETS) emphasise that CgA can be helpful in diagnosing the disease, assessing response to treatment, and detecting early progression and recurrence. However, they indicate an urgent need for standardisation of CgA determination [2].

Another non-specific NEN marker is neuron-specific enolase (NSE). It is less sensitive and specific (30–50%) in the diagnosis of GEP-NENs compared to CgA [9]. Increased concentration of this marker is observed in neuroendocrine carcinomas. Its sensitivity is 63% in large cell NEC (LCNEC) and 62% in small cell NEC (SCNEC). The NSE concentration is also an independent prognostic factor in NEC. Its sensitivity is much lower in NETs G1 and G2, amounting to 19% and 54%, respectively. Simultaneous determination of CgA and NSE demonstrates greater sensitivity and specificity in the diagnosis of NENs [10, 11].

Pancreatic polypeptide (PP) may be a useful marker of non-functioning pancreatic NENs, especially those occurring in MEN-1 [12].

Great hopes have been pinned on new molecular markers, of which NETest is the most promising. It is used to analyse the expression profile of selected gene transcripts that are characteristic of NENs. Performing this test is justified both in the case of NEN diagnosis and to monitor the course of the disease response to treatment, and for early detection of progression [13–15].

In addition, circulating microRNAs are potential biomarkers of NEN due to, among others, the presence and stability in body fluids and specificity for a given tumour [16–18].

B. Specific markers

The choice of the specific GEP-NEN markers to be determined depends on the clinical picture and the type of suspected cancer [19–22].

Carcinoid syndrome is the most common set of clinical symptoms associated with the hormonal function of NETs. The clinical picture is discussed in the recommendations for small intestine NENs [23]. The classic form depends mainly on the excessive secretion of serotonin. The atypical form is observed in lung carcinoid tumours and gastric NENs, and it depends on the excessive secretion of serotonin, 5-hydroxytryptamine (5-HT; serotonin precursor), and/or histamine.

The screening test for carcinoid syndrome consists of a two-fold assessment of the daily excretion of the serotonin metabolite 5-hydroxyindoleacetic acid (5-HIAA) in the urine, while maintaining an appropriate diet. It should also be kept in mind that urine must be acidified during collection [7, 20, 24, 25] and that false-positive and false-negative results are possible [4].

Considering that about 20% of patients with poorly controlled carcinoid syndrome develop carcinoid heart disease [26], it seems reasonable to assess the most useful biomarker of carcinoid heart disease to date: the N-terminal pro B-type natriuretic peptide (NT-proBNP), which is both a diagnostic and prognostic biomarker of heart involvement [27].

Gastrointestinal neuroendocrine tumours (mainly of the pancreas) can cause ectopic production of adrenocorticotropic hormone (ACTH), growth hormone releasing hormone (GHRH), vasopressin, and parathyroid hormone-related protein (PTHrP). The diagnosis of ectopic syndromes depends on the clinical symptoms [7].

All patients with foregut NENs, and in particular those with thymic, duodenal (gastrinoma), and pancreatic NENs, should be tested for MEN-1 syndrome [28]. Basic screening tests in MEN-1 include ionised or total calcium, parathyroid hormone (intact PTH), gastrin, prolactin, and insulin-like growth factor 1 (IGF-1). In patients with suspected MEN-1, molecular testing should be considered to identify mutations in the menin-encoding MEN-1 gene [7, 11].

NETest is a new blood biomarker test for neuroendocrine neoplasms with potential application in the diagnosis and monitoring of disease activity progression [1, 29–31]. The NETest measures the neuroendocrine transcripts circulating in the blood [32]. The test result is obtained by algorithmic evaluation and expressed as an index of disease activity in the range from 0 to 100%. A negative test result (< 20%) excludes NEN or indicates a history of radical removal of the tumour; a positive test result (≥ 20%) correlates with the diagnosis of neuroendocrine neoplasm without identifying the primary tumour site [29, 30]. Clinical use of NETest also includes the identification of microscopic disease [33, 34], assessment of the radicality of surgical treatment [15, 35], somatostatin analogues [13], radioisotope [36], and disease progression [37]. A test score between 20 and 40% indicates disease stabilisation, while a score between 41 and 100% indicates disease progression [38, 39]. Based on the experience with NETest, a new indicator, the so-called PPQ (PRRT predictive quotient) has been developed, which predicts the response to radioisotope treatment with 95% accuracy [40, 41]. A meta-analysis evaluating the parameters of NETest as a clinical biomarker demonstrated the diagnostic effectiveness of the test to be 95–96%, with 84.5–85.5% accuracy in differentiating between disease stabilisation and progression [42].

Studies on the clinical use of NETest in the Polish population have confirmed the earlier reports regarding its usefulness in the diagnosis of NENs of the gastrointestinal tract [33, 43], respiratory system [44-46], pheochromocytoma and paraganglioma [29, 30, 47], as well as correlations with imaging tests [43] and disease activity (progression) [13, 43, 46, 48]. A comparison of the clinical utility of NETest and circulating chromogranin A demonstrated a significant advantage of the molecular biomarker in the diagnosis and monitoring of NENs [48–51].

The presented literature data indicate the potential clinical utility of NETest. The use of NETest in everyday clinical practice will enable the optimal inclusion of the test in the management algorithms in the Polish population of patients with NEN.

Detailed biochemical diagnostics is discussed in the recommendations regarding gastroduodenal [52], small intestine [23], colorectal [53], and pancreatic NENs [54].

Minimal consensus statement on biochemical tests in NENs

2.2. Pathomorphological diagnosis

2.2.1. Histopathological classification of NENs

In 2000, the unified histological classification of the World Health Organization (WHO) and the ENETS was introduced for gastrointestinal neuroendocrine tumours. GEP-NENs are derived from 15 types of diffuse endocrine system (DES) cells located in the gastrointestinal organs (oesophagus, stomach, duodenum, jejunum, ileum, appendix, colon, rectum) and the pancreas, liver, and bile ducts. The histological classification of GEP-NENs was updated in 2010. Two microscopic features: the degree of differentiation and the degree of histological maturity (G, Grade), constituted the basis for dividing the tumours into three categories. The key criterion for this division was the tumour grade assessed on the basis of the Ki-67 proliferation index and the number of figures of division in 10 large fields of view (LFOV). Considering the above-described features of neoplasms, GEP-NENs were divided into two main categories with the following subtypes:

1. Well-differentiated neuroendocrine tumours (WD-NETs):

2. Poorly differentiated neuroendocrine neoplasms (PD-NENs):

The 2010 update of the WHO/ENETS NEN classification was published in 2017 for endocrine organs. The category of pancreatic neuroendocrine neoplasms (PanNENs) with well-differentiated morphology turned out to be heterogeneous in terms of maturity/malignancy, and they were divided into three subtypes: low, intermediate, and high grade (NET G1, G2, G3). The poorly differentiated neoplasms were classified as neuroendocrine carcinomas (NECs) including both small- and large-cell lesions. The distinguished histological types of non-functioning pancreatic neuroendocrine neoplasms (NF-PanNENs) became the basis for the next update of the classification of all NENs of the gastrointestinal tract in 2019 [55]. In 2019, the GEP-NETs classification was ultimately revised and approved by the International Agency for Research on Cancer (IARC) of the World Health Organization in the form of a new fifth edition of the WHO classification published as the “WHO Blue Book” [56]. In 2019, the “WHO Blue Book” included GEP-NETs as a separate group of cancers in individual organs of the gastrointestinal tract for the first time. Table 1 presents the classification and grading of GEP-NENs according to the WHO 2019 criteria.

The assessment of the NEN grade is made on the basis of two features: the number of figures of division calculated per 2 mm² and the Ki-67 proliferation index. According to the criteria prior to 2019, the number of figures of division was counted in 10 large fields of view. However, due to the differences in field of view between microscopes, the principle of assessing mitosis per 2 mm² was adopted. The criteria for this evaluation are presented in Table 2.

2.2.2. Mandatory and conditional methods of NEN pathomorphological assessment

The guidelines developed by ENETS present the principles of material testing depending on its type. Fine-needle aspiration biopsy is not recommended as a diagnostic method in the absence of a primary tumour diagnosis. It can be used in the case of application of the cell block method or to confirm metastasis with a known primary site [57–59].

The rules for the processing of biopsy material from a primary tumour or metastases as well as the surgical material were presented in the previous edition of these recommendations [4].

Minimal consensus statement on pathomorphological examinations in NENs

Detailed pathomorphological diagnostics including the pTNM stage of pathological advancement according to ENETS and/or TNM AJCC/UICC criteria is discussed in the recommendations for gastroduodenal [52], small intestine [23], colorectal [53], and pancreatic [54] NENs.

2.3. Imaging diagnostics

Imaging diagnostics of NENs is associated with many difficulties resulting from small lesion size, often atypical location, and non-specific clinical symptoms. Hence, it is necessary to use various imaging methods, both anatomical and functional.

The following anatomical imaging techniques are used:

Classic imaging methods (CT, US, MRI) are useful primarily in assessing the stage of the disease and monitoring the response to treatment. They play a fundamental role in planning surgical treatment of the primary tumour. They also facilitate the performance a fine or core needle biopsy.

Endoscopic techniques have developed significantly in recent years. There is also better access to them. These methods enable not only diagnostic but also therapeutic procedures to be performed.

An important achievement in NEN diagnostics was the introduction of a scintigraphic test method of somatostatin receptor imaging (SRI). They are functional tests determining the density of the somatostatin receptor that allow the characterisation of lesions at the molecular level [60, 61].

Currently, tests with 68Ga-labelled somatostatin receptor agonists are used in clinical practice: [68Ga]Ga-DOTA-TATE and [68Ga]Ga-DOTA-TOC or 99mTc: [99mTc]Tc-HYNIC-TATE and [99mTc]Tc-HYNIC-TOC [62].

Tests using the above-mentioned methods are characterised by greater sensitivity in the diagnosis of primary disease sites and metastases to bones and lungs.

The joint use of morphological and functional imaging techniques has enabled increased sensitivity and specificity of the diagnostic methods used in NEN [63–66].

Both morphological and functional tests are used in the following: assessing the extent of the disease, determining the location of the primary lesion, selecting the surgical procedure, assessing the response to treatment, and qualifying for radioisotope therapy.

2.3.1. Ultrasonography

2.3.1.1. Abdominal ultrasonography

Ultrasound examination, due to its wide availability and low costs, is usually the first imaging examination performed. The sensitivity of the examination depends on the location of the lesion, the experience of the physician performing the examination, and anatomical and technical conditions [4].

In clinical practice, ultrasonography is used primarily in the initial diagnostics of pancreatic endocrine tumours and metastatic changes in the liver. Due to technical limitations, ultrasonography is of little use in the assessment of the remaining parts of the gastrointestinal tract [4, 67].

The ultrasound image of NEN is non-characteristic. The tumour is often clearly demarcated, hypoechoic, sometimes with a hyperechoic capsule, necrotic foci, and calcifications. However, the tumour may also be hyper- or isoechoic. Most lesions demonstrate rich vascularisation in Doppler examination [4, 67].

The sensitivity of abdominal ultrasonography in the diagnostics of metastatic foci in the liver is 82–88%, and its specificity is 92–95% [67, 68]. The sensitivity of the method in the diagnostics of pancreatic tumours is much lower lower and reaches only 39% (17–79%) [57, 69, 70]. Contrast-enhanced ultrasonography (CEUS) is useful in the ultrasound diagnostics of NEN; 78–86% of the lesions show contrast enhancement in the arterial phase. Due to the usually rich vascularisation of neuroendocrine tumours, patients diagnosed with NET are good candidates for CEUS. This method allows for the prognosis of tumour response to treatment and can be a valuable addition to multiphase CT examination, which reduces the radiation dose received by the patient during the diagnostic process and therapy [71, 72]. The sensitivity of CEUS in the diagnostics of liver metastases increases to 99% [73].

2.3.1.2. Endoscopic ultrasonography

Currently, the EUS test is used in the diagnostics of NENs of the pancreas, stomach, duodenum, and the last section of the large intestine [67]. The small distance between the ultrasound probe and the examined object allows the use of an ultrasonic wave with a higher frequency than in a conventional ultrasound machine. As a result, much better image resolution is achieved [4].

The approved indications for EUS include the assessment of the local advancement of neoplastic lesions of the gastrointestinal tract, and diagnostics of submucosal lesions and pancreatic and biliary diseases. The examination allows the visualisation of small diameter lesions and the assessment of the surrounding lymph nodes. The method also enables the precise determination of anatomical relations (the position of the tumour in relation to the bile ducts and main vessels) and the depth of infiltration into the gastrointestinal tract wall [4].

EUS is particularly useful in the diagnostics of pancreatic NENs (due to their usually small size). In general, the method has a sensitivity of 82–93% and specificity of 86–95% [74]. These parameters depend on the location of the lesion: for tumours located within the head and body of the pancreas it is about 90% (77–100%), and for peripherally located tumours — 75–80% [66, 69, 70]. The specificity of the method is estimated at 98% [75]. In the diagnostics of pancreatic lesions in high risk population, EUS is a more sensitive method than CT [76].

Endoscopic transrectal ultrasonography is the most sensitive method in preoperative assessment of local progression of rectal tumours. The sensitivity of the method in the assessment of the tumour and rectal wall infiltration is 76–93%, and in the case of metastases to the surrounding lymph nodes — 61–88% [77, 78].

2.3.1.3. Intraoperative ultrasonography

Intraoperative ultrasonography (IOUS) is used primarily to identify focal lesions in the pancreas. The sensitivity of this technique is 90% (74–96%), especially when combined with intraoperative palpation [79–81].

2.3.1.4. Intraductal sonography

Mini-probes can be inserted through the endoscope biopsy channel into the pancreatic or bile ducts. Intraductal sonography (IDUS) enables the assessment of the inside of the duct and its wall. Compared to EUS, it allows for better visualisation of pancreatic NENs in the immediate vicinity of the pancreatic duct and changes in lumens. The sensitivity of this examination is approximately 94% [82], which increases to almost 100% for lesions above 3 mm located in pancreatic duct [83, 84].

2.3.2. Endoscopic examinations

Endoscopic diagnostics is of fundamental importance in the diagnosis of neuroendocrine neoplasms originating from the stomach wall, duodenum, and large intestine [85, 86].

The use of colonoscopy as a tool for the screening of colorectal cancer has made it possible to detect GEP-NEN lesions not only in the rectum and colon, but also in the distal part of the small intestine. It should be emphasised that these lesions are most often found accidentally in tests performed due to non-specific symptoms, such as dyspepsia, anaemia, or during screening tests. Neuroendocrine tumours usually appear as an elevated polypoid lesions, and often only histopathological examination allows for proper diagnosis [2, 87]. The size of the lesion, the degree of gastrointestinal wall infiltration, and the possible presence of locoregional metastases influence the treatment strategy and can be assessed using EUS. This examination also makes it possible to obtain material for histopathological evaluation [2, 88].

Upper gastrointestinal endoscopy (oesophagogastroduodenoscopy) and colonoscopy with ileoscopy are often the first examinations performed in patients with suspected or diagnosed NENs of unknown primary site after finding metastases in the lymph nodes or the liver [4, 87]. The above is common in neuroendocrine neoplasms originating especially from the small intestine, and the determination of the primary tumour location may be important for the selection of the optimal method of management despite an earlier diagnosis based on histopathological and immunohistochemical examinations of the material collected during targeted metastatic biopsy. In case of doubt, it is recommended that an examination of the upper gastrointestinal tract be performed with a side-viewing endoscope because it allows for a better assessment of the area of the major duodenal papilla (papilla of Vater). Correctly conducted endoscopic examinations in search of the primary site in patients with liver metastases enable the detection of nearly 100% of primary lesions located in the stomach and 86% of lesions in the large intestine [5, 89, 90].

Examination of the small intestine is now possible with capsule endoscopy and enteroscopy. Video capsule endoscopy (VCE) is a non-invasive examination of the small intestine performed with a disposable, wireless capsule. After being swallowed by the patient, the capsule passively passes through the gastrointestinal tract, allowing the assessment of the mucosa of the small intestine along its entire length. This examination is not a substitute for upper gastrointestinal endoscopy or colonoscopy. The currently used capsules are not steerable, the selected fragment of the intestine cannot be re-assessed, and material cannot be collected for histopathological examination [5, 91, 92]. The limitation of capsule endoscopy is the capsule battery life. For this reason, in some patients with impaired peristalsis, the last section of the ileum may remain unexplored. The most common complication of capsule endoscopy (0.75% of all examinations) is capsule entrapment in a narrowing of the small intestine caused either by the use of non-steroidal anti-inflammatory drugs, other diseases (e.g. Crohn’s disease), or the neuroendocrine neoplasm itself. It should be kept in mind that NENs of the small intestine secrete growth factors causing desmoplastic reactions in the mesentery, which may lead to significant narrowing of the intestine [4, 93].

Current reports indicate a relatively low sensitivity of the examination with the use of an endoscopic capsule in the diagnosis of midgut lesions. The sensitivity of the examination, especially in the detection of submucosal and eccentrically growing lesions, is about 45%. Tumours of the small intestine are most often diagnosed accidentally, for example during the examination for gastrointestinal bleeding [4, 93–95].

A diagnostic method that allows for the assessment of the small intestine with the possibility of collecting material for histopathological examination and the possible use of endoscopic therapy is balloon-assisted (single or double) enteroscopy or spiral enteroscopy [96, 97]. During the examination, it is possible to simultaneously use EUS with miniature heads of 2 mm or 2.6 mm in diameter, introduced through the enteroscope biopsy channel [4, 96]. Capsule endoscopy and enteroscopy are complementary methods. Non-invasive capsule endoscopy allows for the initial determination of the location of the disease, while enteroscopy enables the collection of material for histopathological examination and therapeutic procedures [4, 97, 98].

Complete assessment of the small intestine during enteroscopy is obtained in approximately 80% of patients, and the diagnostic effectiveness of the examination is approximately 55% [99–101].

2.3.3. Computed tomography

CT examination by means of multidetector computed tomography (MDCT) is currently the standard method in assessing the location of disease foci and determining the grade of NENs. Computed tomography is also used to monitor the effects of treatment. However, this examination is characterised by relatively low sensitivity in locating the primary tumour if the patient is examined without proper preparation, using an inappropriate protocol. There are distinct examination protocols for the following locations, which also require different methods of patient preparation: CT of the pancreas is performed using a multi-phase protocol including, in addition to the native arterial phase, parenchymal phase (late arterial phase), portal phase, and in some cases, late phase. Another procedure involves the examination of the stomach, duodenum, and small intestine, which consists of stretching the gastrointestinal tract by filling its lumen with fluid of lower absorption than the intestinal wall [4, 102, 103]. Similar preparation for the examination can be used to assess the colon, although it is usually examined using CT colonography. In the case of gastrointestinal examination, it is recommended that between 900 and 2000 mL of fluid be administered, which should be moderately hyperosmotic to properly stretch the gastrointestinal tract. One of the more common errors leading to false positive diagnoses is a poorly distended intestine and the description of intestinal spasm as a pathology [103]. In turn, inadequately low electric current parameters during the examination lead to false negative results [104].

Depending on the method of filling the lumen of the gastrointestinal tract, the examination is called either “CT enterography” (if the patient drinks low-absorption contrast) or “CT enteroclysis” (if it is administered using a probe inserted into the small intestine). After the gastrointestinal tract is properly filled, a CT examination is performed before and after the intravenous administration of the contrast agent. Scanning after intravenous contrast administration should be performed in two phases — arterial and portal, involving the entire intestines and liver — in order to reveal possible metastases [4]. CT enteroclysis demonstrates sensitivity of 92.8% and specificity of 99.2% in detecting small intestinal tumours [105]. Most commonly, the morphological indicator of a neuroendocrine tumour in the small intestine or stomach is a hypervascularised mass that invades the lumen of the gastrointestinal tract (requiring differentiation from gastrointestinal stromal tumor [GIST]). Less commonly, only wall thickening is visible [106]. In addition to the tumour directly visible in the intestine, it is possible to visualise parenteral lesions such as metastasis to the mesenteric adipose tissue with or without desmoplastic reaction, with or without calcification, metastatic local lymph nodes, liver metastases, and ascites [106, 107].

The presence of a neuroendocrine neoplasm in the small bowel stimulates a desmoplastic reaction in the intestinal fat tissue, which may lead to impaired blood supply in the desmoplastic vessel basin as well as small intestine obstruction [105]. The symptoms indicating the malignant nature of a tumour include the following: large volume, necrosis, and features of infiltration of adjacent tissues (which occur in about 20% of patients). In the arterial phase, hyperdense lesions are most often observed, and less often they are hypovascularised or cystic lesions. In the portal phase, NET tumours are most often hypodense lesions because the contrast is quickly washed away from them [4].

In the diagnosis of pancreatic tumours, the sensitivity of the CT scan is 73% (63–82%) and the specificity is 96% (83–100%) [4, 108-110]. The sensitivity of the test in the assessment of metastatic lesions in the liver is 82% (78–100%) and the specificity — 92% (83–100%) [4, 58, 111–113]. In the diagnosis of extrahepatic metastases, the sensitivity of CT is 75% (63–90%) and its specificity is 99% (98–100%) [4, 57]. Due to the introduction of dual-energy CT in diagnostics and the normalised iodine uptake (NIU) assessment, it was possible to increase the sensitivity to 100% and specificity to 90.2% in the differential diagnosis of hypervascularised liver metastases [114].

Due to the insufficient sensitivity and specificity of individual examinations, it is recommended that anatomical and functional methods are used together when monitoring the response to treatment [115].

During active treatment, follow-up examinations are performed depending on the dynamics of the disease and the treatment method — not more frequently than once every three months, usually every six months. During clinical follow-up, the frequency of examinations is determined by the patient’s clinical condition and the results of biochemical tests. The morphological assessment should be conducted at least once a year. Due to its good availability and reproducibility, CT is the basic imaging method, while in the case of contraindications to CT or inconsistency of the CT image with the patient’s clinical condition, MRI is recommended.

2.3.3.1. CT colonography

In the assessment of colon pathology, classic colonoscopy is preferred because it enables histopathological verification of findings. However, if necessary, virtual colonoscopy (VC) examination by means of computed tomography can be performed. This method allows for three-dimensional mapping of the walls and contents of the large intestine. To obtain perfect 3D reconstructions, it is necessary to perform the examination with a sub-millimetre layer [4].

The examined person requires appropriate preparation, similarly to traditional colonoscopy. The preparation consists of completely emptying the large intestine of faecal masses and fluid (the residual faecal masses may cause false-positive results) [4, 116].

The complete assessment includes the analysis of toposcans and images in axial sections (treated as reference images), and the analysis of multiplanar and three-dimensional reconstructions (including 3D navigation algorithms). A new method that improves the effectiveness of result interpretation is computer-aided diagnosis (CAD) [4].

CT colonography is a safe and well-tolerated diagnostic method. The sensitivity and specificity of the method is comparable to that of classic colonoscopy.

The sensitivity of CT and colonoscopy is similar; according to various authors, it is 90% for lesions > 10 mm and 85% for lesions > 6 mm; the sensitivity and specificity in the diagnosis of malignant neoplasms is 88–100%, and of benign neoplasms — about 86% [117, 118]. The method has low sensitivity in detecting flat and small lesions, which should always be kept in mind. It is rarely used due to its low availability and the clear practical superiority of classic colonoscopy.

The quality of the obtained images depends on the patient’s cooperation and preparation [116]. This method, which has been used in imaging diagnostics for over 20 years, is currently recognised as a better, faster, cheaper, and safer method for assessing the interior of the large intestine compared to classic colonoscopy, which enables its use in the event of technical difficulties in performing an endoscopic colon examination [119].

2.3.4. Magnetic resonance imaging

The sensitivity and specificity of MRI are similar to those of CT in the diagnosis of both the primary NEN and its metastases [96]. The examination protocol includes obtaining the following images/sequences:

Endocrine tumours demonstrate a hypointense signal on T1-weighted images and hyperintense signal on T2-weighted images (rarely hypointense — if they contain a large component of fibrous tissue); they are markedly enhanced after administration of a contrast agent. Open-ring enhancement is observed in the case of cystic or necrotic tumours. Metastatic foci in MRI in 75% are characterised by a hypointense signal on T1-weighted images, and most of them are strongly enhanced following the administration of a contrast agent. MRI colonography is also possible. The advantages and disadvantages of this method are similar to those of CT colonography [4].

MRI examination — using the optimal protocol — enables the diagnosis of 80–95% of metastatic lesions in the liver [111, 118, 121, 122] and 73–93% of pancreatic NENs [111, 123]. In the diagnostics of extrapancreatic and extrahepatic foci, the sensitivity of the examination is much lower, ranging from 68 to 89% [124, 125].

Whole-body magnetic resonance imaging is considered a second-line examination in the assessment of liver metastases less than 10 mm in size and in the assessment of non-specific enhancement foci in the CT scan. It is also recommended in patients hypersensitive to the iodine-based contrast agents used in computed tomography [4]. The DWI sequence allows for the differentiation of benign and malignant lesions — both primary and metastatic [126].

There are also studies on the use of the DWI sequence to examine the whole body in patients with neuroendocrine tumours because it is particularly sensitive in the detection of metastases [127, 128].

2.3.5. CT/MRI enteroclysis/enterography

Currently, CT/MRI enterography/enteroclysis is used to assess the small intestine (see above). These methods increase the sensitivity of CT to 100% [102, 129, 130]. Compared to routine abdominal and pelvic computed tomography, CT enterography is twice as effective in diagnosing small intestine pathologies [74].

These methods allow the identification of even small, segmental thickenings of the intestinal wall, small intramural nodules, and segmental strictures of the lumen. MRI is characterised by better tissue resolution than CT; it enables the assessment of the layers of the intestinal wall and the degree of its infiltration by the neoplasm. The examination should include the area from the diaphragm to the pubic symphysis [4].

In CT/MRI enteroclysis, the contrast is administered by means of a probe introduced under fluoroscopic guidance behind the duodenojejunal flexure. In MRI enterography/enteroclysis, the use of biphasic contrasts is recommended to stretch the intestinal lumen [74].

The anti-reflux balloon prevents the contrast agent from flowing back into the duodenum. After the contrast medium is administered, an MRI scan of the abdominal cavity is performed using a surface coil. Fast T1- and T2-weighted sequences (e.g. HASTE, FIESTA), T2-weighted sequences with fat saturation, and post-i.v. administration of a contrast agent (T1-weighted images) should be used with a layer thickness of 3–5 mm. The patient should be in a supine position (the prone position is uncomfortable, and the patient cannot maintain it for a long time, so it is used less frequently, e.g. if artifacts are present). In the case of CT enteroclysis, after filling the intestinal lumen with a negative contrast agent (e.g. with an aqueous lactulose solution), i.v. contrast agent is administered in the volume of 1.5–2 mL/kg, at a rate of 3–4 mL/s. The examination is performed in the arterial phase activated using the smart prep function and with a delay of 30–60 s (after 45 s — intestinal phase). The layer thickness should be between 1 mm and 3 mm [4, 102].

In CT/MRI enterography, the contrast is administered p.o. one hour before the examination. In the MRI examination, i.v. administration of drugs inhibiting gastrointestinal motility is recommended (e.g. one ampule of Buscolysin 20 mg/mL, either i.v. or intramuscular (i.m.), if the patient has no contraindications) [4].

MRI enteroclysis provides a higher degree of intestinal loop distension and the possibility of assessing peristalsis; however, the examination is not as well tolerated by patients as is enterography. The sensitivity of CT enterography and enteroclysis is comparable. However, due to the long data acquisition time in magnetic resonance imaging, enteroclysis [4, 102] is recommended. Current studies report 86–94% sensitivity and 95–97% specificity of MRI enteroclysis in the detection of small tumours of the small intestine [129].

The examination time of CT enteroclysis is shorter than that of MRI enteroclysis. As a result, the quality of the examination is less dependent on cooperation with the patient, and the obtained images are usually of very good quality; therefore, the method is considered the gold standard in small intestine imaging, including in the diagnosis of neuroendocrine neoplasms [74]. However, it is associated with the patient’s exposure to ionising radiation [4].

Minimal consensus statement on imaging and endoscopic examinations in NENs

The choice of imaging method depends on the primary tumour location and its grade: US, CT, MRI, endoscopy [V, 2a]#.

Details of specific organ examinations are discussed in the recommendations regarding gastroduodenal [52], small intestine [23], colorectal [53], and pancreatic NENs [54].

2.4. Radioisotope diagnostics

2.4.1. Isotope diagnostics with the use of radiolabelled somatostatin analogues

Currently, in the imaging diagnostics of NETs, the most sensitive method is with the use of radiolabelled somatostatin analogues. Somatostatin receptor imaging (SRI) is performed using scintigraphy (planar, SPECT, or SPECT/CT) or positron emission tomography (PET/CT). The sensitivity of SRI depends on the technique (it is highest in PET/CT examination) and amounts to 54–100% for most GEP-NET types and locations [1, 4, 131]. The exception is insulinoma, in which overexpression of somatostatin receptors is found in 50–60% of cases [132].

The clinical indications for the use of SRI include the following: determining the primary tumour location, assessing the neoplasm grade, monitoring the patient after radical surgical treatment, assessing the effectiveness of the treatment and qualifying patients for antiproliferative therapy with somatostatin analogues (SSA) and peptide receptor radionuclide therapy using radiolabelled somatostatin analogues (PRRT), currently proposed new nomenclature: targeted radioligand therapy (RLT) [131].

Because the CLARINET study confirmed the antiproliferative effect of lanreotide in the presence of somatostatin receptors, with uptake at least comparable to that of the liver, SRI should be performed before the inclusion of SSA to achieve an antiproliferative effect [133]. The experts have not reached a consensus regarding this issue [1].

In 2019, a new classification of neuroendocrine neoplasms was introduced, with division into NET G1, G2, G3, and NEC [55]. The data published so far mostly concern NET G1 and G2. Currently, the data on imaging of somatostatin receptors in the NET G3 group are scarce. In the case of NEC, imaging of somatostatin receptors is not used routinely but may be helpful before treatment decision.

2.4.1.1. Diagnostics with the use of technetium (99mTc)-labelled somatostatin analogues

In most centres, technetium 99m (99mTc)-labelled somatostatin analogues have completely replaced scintigraphy with the use of [111In]ln-pentetreotide [4, 131]. Various scientific reports have shown greater sensitivity of scintigraphic examination with the use of [99mTc]Tc-HYNIC-TOC compared to [111In]In-pentetreotide [134]. The physical properties of 99mTc contribute to much better imaging quality, shorter examination time, and a lower dose absorbed by the patient. An additional advantage of [99mTc]Tc-HYNIC-TOC is the possibility of examining patients in any nuclear medicine laboratory. Currently, the standard examination is single-photon emission computed tomography (WB-SPECT/CT) used for scatter correction and anatomical localisation by means of CT, with the examination scope similar to that of PET.

The development of SPECT/CT gamma cameras allows for the use of quantitative measurements. However, there are currently no scientific reports regarding this aspect in neuroendocrine tumours.

2.4.1.2. Diagnostics with the use of somatostatin analogues labelled with positron markers

Scintigraphy with the use of positron markers is the imaging method with the highest resolution among radioisotope tests. Literature data indicate a higher sensitivity of imaging with positron-labelled somatostatin analogues (68Ga) compared to testing with [111In]In-pentetreotide or [99mTc]Tc-HYNIC-TOC/TATE [4, 131, 135, 136]. Among the somatostatin analogues, DOTATATE, DOTATOC, and DOTANOC are currently used, which differ in their affinity for individual SSTRs. The sensitivity, specificity, and diagnostic accuracy in PET studies using 68Ga-labelled somatostatin analogues ([68Ga]Ga-DOTA-SSA) are 97%, 92%, and 96%, respectively [4, 131]. In a meta-analysis of 10 studies, the [68Ga]Ga-DOTA-TATE PET/CT sensitivity was 90.9% [95% confidence interval (CI): 81.4–96.4%] and the specificity was 90.6% (95% CI: 77.8–96.1%) [136]. The sensitivity of [68Ga]Ga-DOTA-SSA depends on the degree of differentiation and is 92–100% for NET G1/G2, 67–92% for NET G3 and only 40–50% for NEC [137]. Abdominal MRI and CT and PET/CT with [68Ga]Ga-DOTA-SSA are preferred in the diagnostics of patients with clinical/biochemical suspicion of NEN (not confirmed in histopathological examination) [131]. [68Ga]Ga-DOTA-SSA PET/CT is preferred to confirm the expression of somatostatin receptors in all patients, both in the initial and subsequent stages of the disease progression, in finding the primary site, and in the assessment of somatostatin receptor expression before scheduled PRRT [131]. PET/CT examination with [68Ga]Ga-DOTA-SSA is a more sensitive method than classic scintigraphy in detecting latent or clinically suspected metastases in bones, being the fourth most frequent metastatic site, or metastases to lymph nodes. The use of PET/CT with [68Ga]Ga-DOTA-SSA instead of scintigraphic examination significantly affects the change of disease management in 13–71% of patients [4, 131, 134–136]. Moreover, in the PET/CT examination with [68Ga]Ga-DOTA-SSA, it is possible to measure the maximum standardised uptake value (SUVmax). The SUVmax value is correlated with the density of somatostatin receptors present on the cell surface, which is extremely important in qualifying patients for PRRT [138].

[68Ga]Ga-DOTA-SSA is recommended in the imaging of all NET types (except for the adrenal pheochromocytoma due to physiological distribution in the adrenal glands), medullary thyroid carcinoma, mild insulinoma, neuroblastoma, and ventral paraganglioma (characterised by variable SSTR expression) [131]. In centres equipped with PET apparatus, the examination with [68Ga]Ga-DOTA-SSA should be the examination of choice in the diagnostics of NEN.

Recently, copper-labelled somatostatin analogues have also been used in diagnostics — 64 (64Cu) [139]. Thanks to their longer half-life, 64Cu somatostatin analogues make it possible obtain images as much as 24 hours after the marker administration. There are currently no clear data on the advantages of [64Cu]Cu-DOTA-TATE over the commonly used [68Ga]Ga-DOTA-SSA [139].

2.4.1.4. Intraoperative somatostatin receptor imaging

The use of an intraoperative scintillation probe makes intraoperative somatostatin receptor imaging possible. The examination is useful in imaging the primary lesion and looking for metastases in the surrounding lymph nodes, which greatly simplifies and shortens the surgical procedure. 99mTc- or 68Ga-labelled somatostatin analogues are used in the intraoperative examination [140, 141].

2.4.2. Isotope diagnostics with the use of radiolabelled fluorodeoxyglucose [18F]FDG

Until now, it was thought that PET/CT with the use of radiolabelled 18F-fluorodeoxyglucose ([18F]FDG) was of little use in the diagnosis of NET due to its low sensitivity. However, it has been demonstrated that the accumulation of [18F]FDG in neoplastic foci is a significant negative prognostic factor that allows for a more precise characterisation of the biological malignancy of the tumour [4, 131, 142, 143]. PET/CT with [18F]FDG is positive in approximately 30% of NETs G1, 60% of NETs G2, 80% of NETs G3, and 80–100% of NECs [144, 145].

Experts concluded that [18F]FDG PET/CT is necessary in patients with NET G3, NEC, and selected patients with NET G2 with Ki-67 10–20%. In patients with NET G1/G2, this examination should be performed in the absence of SSTR expression in lesions found using CT/MRI and in patients with rapid disease progression (despite potentially low-grade malignancy) [1].

Statistically significantly shorter progression-free survival (PFS) and overall survival (OS) was observed in PET/CT positive patients with [18F]FDG [142, 143]. Arbitrary SUVmax > 2.5 is more often associated with a more aggressive course of disease and should prompt faster implementation of therapy [142, 146]. The use of [68Ga]Ga-DOTA-SSA and [18F]FDG significantly influences the modification of disease management (in up to 80% of cases) [147].

In 2017, Chan et al. proposed a classification of NEN advancement based on double marker imaging with the use of [68Ga]Ga-DOTA-SSA and [18F]FDG PET/CT examination, the so-called “NETPET score” (with five categories of results: P0 — both examinations negative, P1 — only [68Ga]Ga-DOTA-SSA positive; P2–P4 — both [68Ga]Ga-DOTA-SSA and [18F]FDG PET/CT positive, P5 — only [18F]FDG PET/CT positive, which correlate with tumour progression and overall survival) [148].

On the other hand, Mapelli et al. analysed the results of molecular imaging ([68Ga]Ga-DOTATOC and [18F]FDG PET/CT) in patients with PanNENs who were candidates for surgery, in order to determine the indicators of image data evaluation for preoperative analysis of more aggressive phenotypes [149]. In their work, they analysed the semi-quantitatively standardised value of SUVmax and SUVmean for both [68Ga]Ga-DOTATOC and [18F]FDG PET/CT, somatostatin receptor density (SRD) and total lesion somatostatin receptor density (TLSRD) for [68Ga]Ga-DOTA-TOC PET/CT, and metabolic tumour volume (MTV) and total lesion glycolysis (TLG) for [18F]FDG PET/CT examination. SRD and TLSRD were the only markers significantly associated with tumour size (pT3 or pT4 vs. pT1 or pT2; with the cut-off point of 18.3 for SRD and 275.4 for TLSRD; sensitivity and specificity, respectively: 78.1 and 72.5% for SRD and 75 and 66.7% for TLSRD).

Due to the varied biology of NENs, the PET/CT examination with [18F]FDG and the assessment of somatostatin receptor expression in the SRI are necessary, especially in the diagnosis of NET G2 and NET G3 and for proper qualification for PRRT [4, 131, 143]. PET/CT examination with [18F]FDG is also a predictive factor in response to PRRT [150, 151]. In recent years, a growing number of studies with the use of radiomics have appeared. Radiomics is based on a thorough qualitative assessment of imaging data along with quantitative analysis using advanced diagnostic tools for functional imaging (DWI MRI, [68Ga]Ga-DOTA-SSA, [18F]FDG PET/CT) and texture analysis considering radiomedical and radiogenomic features. It may be helpful in staging and stratifying cancer risk, optimising therapeutic management, and improving the prediction of response to treatment in patients with NET [152].

2.4.3. Isotope diagnostics with the use of radiolabelled dihydroxyphenylalanine [18F]FDOPA

PET diagnostics using fluoride-18 dihydroxyphenylalanine (DOPA) ([18F]FDOPA) is a promising imaging method for NETs [4]. [18F]FDOPA PET/CT sensitivity of 65–96% was demonstrated in the diagnostics of NETs [153, 154]. However, the significance of this study is not clear [1, 4, 131, 155]. This examination seems to be useful in the case of pancreatic tumours with secretory function and in other GEP-NETs with a negative SRI result [131, 153].

The European Association of Nuclear Medicine (EANM) guidelines for the diagnostics of NETs of the small intestine recommend [18F]FDOPA as an additional radiopharmaceutical. However, [18F]FDOPA does not provide information for PRRT planning [155].

2.4.4. Diagnostics with the use of radiolabelled meta-iodobenzylguanidine [123I]/[131I]mIBG

Another marker used in diagnostics and therapy is the radiolabelled guanidine derivative meta-iodobenzylguanidine ([123I]/[131I]mIBG), which is accumulated in the cell by means of the vesicular monoamine transporter 1 (VMAT1) and 2 (VMAT2) mechanisms.

Imaging with [123I]/[131I]mIBG is used primarily in pheochromocytoma and neuroblastoma, less frequently in other neoplasms with neuroendocrine differentiation. The sensitivity of scintigraphy with [123I]/[131I]mIBG in NET is on average 50% (40–85%) and is lower than in the case of [111In]In-pentetreotide [4, 131, 142, 156]. The best results are obtained when [123I]mIBG is used to visualise liver metastases. However, also in this case, the sensitivity of receptor scintigraphy is higher [142]. Therefore, currently [123I]/[131I]mIBG scintigraphy is mainly used in qualification for isotope treatment ([131I]mIBG) in situations where SRI is negative [4].

2.4.5. Other radioisotope tracers

As in the case of somatostatin analogues, the new radiopharmaceuticals introduced to diagnostics and treatment more and more often use the same ligand, which, in combination with an appropriate isotope, becomes the preferred theragnostic tool.

Currently, numerous scientific studies are being carried out on the use of radiotracers, such as [11C]C-5-hydroxytryptophan ([11C]C-HTP) and other new receptor markers, such as glucagon-like peptide-1 (GLP1) analogues in tumours such as insulinoma, and gastrin, and bombesin analogues in medullary thyroid cancer [157–161].

GLP-1 analogues in the diagnostics of insulinomas are particularly promising. GLP-1 analogues labelled with 111In, 99mTc, and 68Ga [157, 159–161] were used in studies. The preliminary results of the studies indicate the lack of GLP-1 receptor expression in most malignant forms of insulinoma (here, SRI is more often positive), suggesting the usefulness of imaging with labelled GLP-1 analogues in the differentiation between benign and malignant forms of insulin tumours [157, 160].

Research is currently underway on somatostatin receptor antagonists, characterised by a 4–12-fold higher number of somatostatin receptor type 2 (SSTR2) binding sites, lack of internalisation, and high stability, which translates into higher imaging sensitivity. Imaging with labelled antagonists demonstrated higher sensitivity in detecting lesions in the liver and spleen, lower in bones, and comparable in metastatic lymph nodes and primary tumours compared to agonists [162]. However, due to their limited availability, these markers are not used in routine diagnostics [131, 163].

Another studied receptor is CXCR4 — a chemokine family receptor. It has been found that this receptor is expressed in many neoplasms, and its expression is associated with a more aggressive course of the disease, earlier metastasis, higher risk of recurrence, and shorter survival [164]. Inverse expression of SSTR2 and CXCR4 was found in NETs, from G1 to G3, with an increase in CXCR4 expression and a decrease in SSTR2 expression along with increasing grade. Thus [68Ga]Ga-pentixafor PET/CT can serve as a non-invasive tool to assess the applicability of PRRT targeting CXCR4 in advanced SSTR-negative tumours [165].

Minimal consensus statement on imaging and radioisotope examinations in NENs

3. Treatment

3.1. Surgical treatment

The treatment of choice for the majority of locally and loco-regionally advanced GEP-NETs G1/G2 is surgical procedure, the scope of which depends on the general condition of the patient, and the primary location and the specificity (biology) of the tumour [1, 4, 166].

Most symptomatic or hormonally functional neuroendocrine tumours constitute an indication for surgical removal, regardless of their size. In the case of non-functional and asymptomatic tumours, the tumour size is one of the criteria determining the treatment method [166]. Small G1 and G2 lesions (tumour size limits are determined for individual organs) with low Ki-67 can be managed conservatively, while larger tumours should undergo minimally invasive treatment (e.g. endoscopy) or surgery. Follow-up and conservative treatment are possible only when there are no features suggesting a malignant nature of the tumour and when its close and constant surveillance is possible by performing a specific, established panel of examinations.

Most NETs G3 with high Ki-67 and NECs are treated with the same oncological management principles as with other gastrointestinal malignancies. Apart from symptomatic tumours, NEC stage IV is a contraindication to surgical resection, but it is not an absolute contraindication for surgical treatment in NET G3 [1].

In advanced metastatic NECs, resection, cytoreduction, or ablation of liver metastases are not recommended [4]. These methods can be used in selected cases of NET G3 metastasis management.

In the case of stage IV GEP-NET G1/G2 with the M1 feature in the liver, radical resection of the primary tumour and liver metastases is the best therapeutic option [1, 4]. Cytoreduction of hepatic metastases should be considered individually, especially in patients with functional tumours and uncontrolled symptoms as well as in patients with non-functional tumours and related symptoms. Cytoreduction is indicated if the tumour mass can be reduced by > 70% [1]. When metastases cannot be resected, ablative techniques and embolization (chemoembolization, radioembolization) are used [1].

Cytoreductive therapy consisting of the reduction of the primary tumour mass to reduce the symptoms, and obtain better systemic treatment efficacy is indicated if the tumour mass can be reduced by > 90% [1, 4].

The radical resection of stage IV primary G1/G2 tumours with unresectable liver metastases can also be considered [1, 167].

3.1.1. Indications for liver transplantation in GEP-NETs

Liver transplantation in patients with GEP-NETs should be considered in cases of unresectable neoplastic lesions in the liver parenchyma, both primary and metastatic. This method of treatment is also indicated in patients with recurrent hepatic neoplastic disease, after previous liver resection, ablation, or systemic treatment due to GEP-NET [4, 168]. Liver transplantation can be performed in cases of both symptomatic and asymptomatic tumours [4, 169, 170].

The condition for qualification for liver transplantation is the confirmation of liver-only location of GEP-NET metastases or primary tumour (no extrahepatic metastases in imaging examinations, PET/CT, or laparoscopy/diagnostic laparotomy) and G1 or G2 histopathological differentiation according to the WHO classification [169].

It is also recommended that the primary tumour be removed prior to liver transplant surgery, and it is emphasised that the most favourable prognosis is associated with a Ki-67 level below 10%, and according to some authors –– Ki-67 < 20%[169, 171].

Transplantologists commonly accept the so-called Milan criteria proposed by Vincenzo Mazzaferro for liver transplantation in patients with NET liver metastases:

The case of each potential recipient should be discussed at a multidisciplinary meeting of oncologists and transplantologists held to consider the dynamics of the disease, patient’s age and psychological attitude to the disease, and the proposed treatment strategy.

Given the shortage of organs from deceased donors, donation after circulatory death (DCD) or transplantation of a liver fragment from a living donor may be considered if such strategies are available [172].

Immunosuppressive therapy should be carried out with regimens involving calcineurin inhibitors (tacrolimus or cyclosporine) and glucocorticosteroids.

There are no convincing data on the efficacy of adjuvant chemotherapy, although the literature provides isolated reports on the use of a combination of cisplatin, doxorubicin, 5-FU with vincristine, and 5-FU with leucovorin or doxorubicin [173].

The infiltration of large hepatic vessels is assumed to be an unfavourable prognostic factor for survival, although this is not an absolute contraindication, as well as simultaneous resection of the tumour outside the liver [171].

Minimal consensus statement on surgical treatment in NENs:

3.2. Endoscopic treatment

The main goal of GEP-NEN treatment is the radical removal of the tumour, and in the case of functional tumours — also the control of clinical symptoms related to the production of specific hormones [4].

Although the basic method of radical treatment is surgical resection, technological progress in the field of endoscopic equipment and the development of new therapeutic endoscopy techniques make it possible to use this type of treatment in some cases. It has become possible mainly due to the introduction of diagnostic methods such as endoscopic ultrasonography, which enables precise assessment of the gastrointestinal wall and its individual layers along with the surrounding structures, and proper qualification of patients for surgical or endoscopic procedures.

The therapeutic management of GEP-NENs located in the upper gastrointestinal tract and large intestine, often detected during diagnostic endoscopy, depends on the size of the tumour, the depth of invasion, and the presence of metastases at diagnosis. Endoscopic GEP-NEN resection can be used as a treatment method only in the case of well-differentiated G1 and T1 tumours according to the TNM classification. Before deciding on endoscopic treatment, endosonographic examination is necessary to determine the size of the lesion and the depth of infiltration into the gastrointestinal wall with the assessment of the surrounding lymph nodes. The examination may be complemented by a fine-needle biopsy of the primary lesion and lymph nodes. Only lesions limited to the mucosa and submucosa are suitable for endoscopic removal, while in all other cases local or radical surgical excision is recommended, possibly with supportive treatment or chemotherapy [4, 85]. It is estimated that approximately 20% of gastric NENs, 10% of duodenal NENs, and as many as 80–90% of rectal tumours are eligible for endoscopic removal [174, 175]. Simple polypectomy performed with loop electrocautery is not recommended as a therapeutic method in the treatment of GEP-NENs due to the frequent presence of a positive surgical margin after the procedure.

The following endoscopic methods are currently recommended in the treatment of gastrointestinal GEP-NENs: a) modified endoscopic mucosal resection (mEMR), b) endoscopic submucosal dissection (ESD), and c) endoscopic full-thickness resection (EFTR) using a dedicated endoscopic cap [4, 176–184].

It should be emphasised that the classic EMR (injection of fluid under the lesion and using a loop) is no longer recommended, especially in the case of rectal lesions — as was the case until recently — due to the low percentage not exceeding 75% of complete resection. Three modifications of the classic EMR are available: cap-assisted endoscopic mucosal resection (EMR-C), endoscopic mucosal resection with ligation (EMR-L), and endoscopic mucosal resection with precutting of the mucosa around the lesion (EMR-precut). EMR modifications provide the opportunity for complete resection in over 93% of cases.

The most common complications of mucosal resection and endoscopic submucosal dissection include bleeding (up to 7%) and perforation (5%). They are more common in duodenal and gastric lesion removal than in the case of rectal lesions. Compared to mucosal resection, submucosal dissection is characterised by a higher rate of lesion removal in one piece regardless of tumor size [odd ratio (OR) 13.87] and a lower rate of local recurrence (OR 0.09). However, this method is time consuming and has higher complication rate (bleeding OR 2.2; perforation OR 4.09) [185].

The following histopathological criteria ensure the oncological safety of the endoscopic procedure: complete removal of the lesion (negative margin), no angioinvasion, as well as low mitotic activity and low proliferation index.

In most cases, after endoscopic treatment, further follow-up is recommended. The time intervals depend on the location and severity of the lesion at diagnosis [186].

Endoscopic methods are also applicable in the palliative treatment of NENs in the following cases:

Minimal consensus statement on endoscopic treatment in NENs

3.3. Systemic treatment

The aim of a systemic therapy is to relieve symptoms accompanying the disease and tumour growth control.

A. Symptomatic treatment

3.3.1. Somatostatin analogues

Somatostatin analogues (SSAs), octreotide and lanreotide, are the standard first-line treatment in NETs. These drugs reduce the secretion of hormones and biologically active substances, and in 70–80% of patients they reduce the most common symptoms (diarrhoea, flush) and significantly improve the quality of life [1, 4]. Long-acting SSAs constitute the treatment of choice in the case of the symptoms of the following: carcinoid syndrome, glucagonoma, VIPoma, and uncontrolled Zollinger-Ellison syndrome [1] (symptomatic treatment in these syndromes is included in the recommendations regarding pancreatic [54] and small intestine NENs [23]).

Generally, because SSAs are well tolerated, gastrointestinal adverse effects (abdominal discomfort, flatulence, nausea, diarrhoea) are usually transient. Other adverse effects include impaired glucose tolerance and cholelithiasis, occurring in 20–50% of patients (rarely symptomatic).

In cases of worsening of the patient’s condition, e.g. exacerbation of symptoms of carcinoid syndrome, administration of higher than standard doses of SSA may be considered (octreotide LAR 30 mg 1 × 4 weeks, lanreotide autogel 120 mg 1 × 4 weeks). The CLARINET FORTE study demonstrated that increasing the frequency of lanreotide administration to a dose of 120 mg every 14 days resulted in an increase in progression-free survival by 8.3 months in patients with intestinal NETs and by 5.6 months in patients with pancreatic NETs (in particular with Ki-67 ≤ 10%) [189]. The presented results of the study show that in patients with disease progression, it may be worth considering a higher frequency of lanreotide autogel dosing before switching to an alternative, but more toxic treatment [189].

SSAs are not the first-line treatment in the case of insulinoma (diazoxide is initially used to control hypoglycaemia) or gastrinoma [high-dose proton pump inhibitors (PPIs) are used to treat peptic ulcer disease]. In the case of malignant forms of insulinoma and gastrinoma, the use of SSA as a second-line treatment may be effective in relieving the disease symptoms [190].

Short-acting SSAs (octreotide, solution for injection 100 µg/amp.) are used when it is necessary to quickly control the clinical symptoms of GEP-NETs (e.g. carcinoid crisis), in the perioperative period or in selected cases, before starting treatment with long-acting analogues, in order to assess drug tolerance [4].

Pasireotide can be used off-label in situations where the previous measures are unsuccessful.

3.3.2. Telotristat

It is an oral tryptophan hydroxylase inhibitor that inhibits the rate of serotonin synthesis. In the Telestar study conducted on 135 patients with refractory carcinoid syndrome, the use of this drug demonstrated a significant reduction in the number of bowel movements. A sustained effect of telotristat (defined as a 30% improvement in bowel movements for over 50% of the 12-week study period) was noted in 44% and 42% of those treated with 250 mg or 500 mg doses three times a day, respectively. Adverse effects include mild elevation of blood transaminase levels, and cases of nausea and depressive behaviour have been observed at higher doses of this drug. Patients using telotristat showed a significant improvement in quality of life. The drug is approved for the treatment of diarrhoea associated with carcinoid syndrome in situations when SSA treatment is not efficient [1].

3.3.3. Other

The use of everolimus may also be considered in the treatment of uncontrolled symptoms of the disease, especially in the case of metastatic insulinoma or refractory carcinoid syndrome [although it is not approved by the European Medicines Agency (EMA) and the US Food and Drug Administration (FDA)]. Diazoxide remains a valuable drug, especially in malignant insulinoma, due to the inhibition of insulin secretion by tumour cells (bearing in mind that SSAs may increase hypoglycaemia). The symptoms of gastrinoma can be effectively managed with the use of PPIs. In progressive disease, the use of radioligand therapy may also be used to alleviate symptoms [1].

B. Antiproliferative treatment in patients with GEP-NEN

The choice of antiproliferative treatment depends on the primary tumour location, pathomorphological and clinical features, disease dynamics, and the presence of SSTRs. Treatment options include systemic and molecular targeted therapy. So-called biotherapy, with SSA and INFalfa, has the longest history. Other drugs include a selective m-TOR (mammalian target of rapamycin) pathway inhibitor — everolimus and a receptor tyrosine kinase inhibitor (TKI) — sunitinib [191–193]. These drugs have been registered, among others, in the treatment of inoperable and/or metastatic highly or moderately differentiated pancreatic neuroendocrine neoplasms in adults with progressive disease [194, 195]. Their role is discussed in the recommendations for pancreatic NENs [54].

It should be noted that none of the available treatment options provides full effectiveness, but rather stabilisation of varying duration.

SSAs constitute an established antiproliferative treatment in metastatic GEP-NETs, the efficacy of which has been confirmed in clinical trials — PROMID and CLARINET [196, 197]. The extended CLARINET trial demonstrated the efficacy of SSAs in patients with progressive GEP-NETs and good long-term tolerance of SSAs [198].

Based on the results of these studies, ENETS experts stated in their 2016 recommendations that SSAs can be used in stable or progressive disease or in patients with NEN the course of which has not been determined. SSA preparations are recommended as first-line therapy in midgut and pancreatic NETs. Octreotide is recommended for the control of midgut NET G1 with little hepatic involvement, and lanreotide is recommended in midgut and pancreatic NETs G1 and G2 (Ki-67 up to 10%) control, regardless of the degree of liver involvement (level of evidence 1) [199]. In patients with non-functional NETs G1 indolent disease (no symptoms, liver involvement < 10%, and stable disease), a watch-and-wait strategy is possible. This strategy is used less frequently with PanNET.

In turn, according to the 2020 European Society of Medical Oncology (ESMO) recommendations, SSAs are also recommended as the first-line treatment to control tumour growth (antiproliferative therapy) in advanced or metastatic, slowly growing, well-differentiated NETs G1/G2, SSTR (+) with Ki-67 up to 10% [1, 200]. Knowing the SSTR (+) status is generally required when deciding to use SSA, but it is not predictive of response to treatment. SSA may also be recommended for patients with an unknown SSTR status [200].

It is recommended that SSA be discontinued for four weeks prior to the planned SPECT or PET/CT examination in the case of long-acting preparations and 24–48 hours for short-acting ones. Treatment with SSA should be discontinued before the scheduled administration of PRRT; a five-week washout period is recommended for long-acting forms and a 24-hour withdrawal for short-acting ones.

Minimal consensus statement on somatostatin analogue treatment in NENs

3.3.4. Interferon alpha

The indications of interferon alpha (INF-a) are similar to those of SSAs [4]. Due to the greater number of adverse effects, it is more commonly a second-line drug used in the control of clinical symptoms of functional tumours.

IFN-a can also be used to relieve symptoms (3–5 million IU s.c. 3 × week). Its effectiveness is similar to that of SSAs, but it is used as a second-line drug due to adverse effects (fatigue, weight loss, fever, less commonly depression) [1].

Based on many years of experience and the most recent results of a large randomised trial (including 35% midgut NETs, PFS of 15.4 months for IFN-alpha and octreotide LAR), antiproliferative therapy with IFN-a may be considered if other treatment options have already been used or cannot be used (e.g. due to the lack of SSTR), especially in midgut tumours, where there are fewer therapeutic options compared to PanNET [201].

So far, in Poland, we do not have any experience with the use of INF-a in GEP-NETs because the drug is not available.

3.3.5. Chemotherapy

Chemotherapy (ChT) is one of many treatment options for GI NENs. Indications for chemotherapy depend on the histological characteristics of the neoplasm (grade, Ki-67), its primary location, disease dynamics, as well as the general condition of the patient and the comorbidities [4]. The role of chemotherapy is limited in well-differentiated G1 and G2 neuroendocrine neoplasms. It is one of the treatment options for NETs G3 and the mainstay of treatment for highly aggressive NECs.

The decision to use cytostatics depends on the following:

No specific predictors of response to specific ChT regimens have been identified and used in clinical practice so far, but benefits from the use of ChT can be expected in the following:

Neoadjuvant chemotherapy

Neoadjuvant chemotherapy may be considered prior to non-metastatic NEC resection. Usually 3–4 cycles of etoposide + cisplatin (or carboplatin) are recommended, especially when surgery cannot be performed within the optimal time of about four weeks or when there is a probability of not achieving radical resection (R1/R2) [202, 203]. In selected patients with locally advanced, unresectable primary pancreatic NETs G2 with Ki-67 of 5–20% and in those with more aggressive course of disease (NETs G3), systemic induction or neoadjuvant treatment may be considered in individual cases in order to reduce the extent of the disease and enable radical or cytoreductive treatment at a later stage [1].

Adjuvant chemotherapy

In NETs G1/G2 located in the gastrointestinal tract after radical surgical treatment, there are no indications for adjuvant therapy.

It should be emphasised that there is no clear evidence from controlled clinical trials on the effectiveness of adjuvant therapy in NET G3, only evidence from individual centres on which the experts base their opinion, and each case should be considered individually.

In the case of NEC, given the high recurrence rate after radical surgical treatment, adjuvant treatment with platinum based and etoposide regimens should be considered, usually 3–4 courses, and in some cases also in combination with radiotherapy (especially in such locations as oesophagus, duodenum, head of the pancreas, rectum and anal canal), although there is no clear evidence of patients benefiting from this procedure, and each case should be treated individually [1, 202, 203].

Palliative chemotherapy

In patients with well-differentiated neoplasms (NETs G1/G2/G3), ChT can be considered only as palliative management in the case of tumour dissemination (metastatic disease), inability to perform radical surgical treatment (locally advanced process), or lack of radicality (after cytoreductive treatment), as well as in the case of relapse after radical treatment with massive dissemination. The most important qualification criteria for palliative chemotherapy is the symptomatic nature of the disease and/or its dynamics, and the good general condition of the patient (WHO 0–2).

Palliative systemic treatment of NETs G1/G2/G3 is considered in the following cases:

It should be emphasised that chemotherapy is generally a treatment of moderate effectiveness in NETs G1/G2. In each case of advanced, highly/intermediately differentiated GEP-NET, before making a decision to administer chemotherapy, the possibility of using local palliative treatment of the primary lesion and/or metastases should be considered: excision (of the primary lesion and/or specific resectable metastases), removal of metastases (thermoablation, radiofrequency ablation RFA, NanoKnife), and local palliative techniques (CT-guided radioembolization and brachytherapy) or less toxic systemic methods (cold SSAs biotherapy, PRRT, molecular targeted therapy). The above considerations are based on expert opinions, global guidelines, and recommendations for the treatment of GEP-NETs.

The effectiveness of chemotherapy in the case of well-differentiated GEP-NETs G1/G2/G3 is much higher in neoplasms of the pancreas than in other locations [203, 205].

An indirect comparison of the results of clinical trials conducted in patients with GEP-NENs indicates a higher probability of obtaining a 43–70% response in patients treated for pancreatic NENs compared to GEP-NENs with a different primary lesion location, at about 5–15% [206, 207], although the interpretations of the results of clinical studies are hampered by different eligibility criteria, different patient populations, and different response assessment criteria [208]. The use of multi-drug regimens with streptozocin (STZ) and 5-fluorouracil (5-FU) or doxorubicin (DOX) makes it possible to obtain an objective response rate (ORR) of approximately 39–69% and overall survival (OS) of 15–30 months. Adding cisplatin to the above regimens significantly increases the rate of complications, and therefore triple regimens are not recommended [209-211]. However, the use of doxorubicin is limited by a cumulative dose of 500 mg/m2 due to the risk of cardiotoxicity [212]. Adverse effects typical of this regimen, in addition to haematological toxicity, are the following: nausea/vomiting, fatigue syndrome, symptoms of mucositis, diarrhoea, paraesthesia, impaired renal function, and cardiotoxicity [209]. In pancreatic NENs, STZ with 5-FU is still considered a standard of care, but in Poland STZ is not registered and is currently difficult to obtain.

Treatment of pancreatic NENs with CAPTEM regimen using capecitabine (CAP) and temozolomide (TEM) can be conisdered according to various recommendations (ENETS, ESMO, NCCN) as an alternative treatment depending on the availability of STZ/5-FU. The results of small prospective and retrospective studies using temozolomide in combination with anti-angiogenic drugs or capecitabine indicate that response rates (RR) range from 15% to 70% [207, 213, 214]. The value of temozolomide administered alone or in combination with capecitabine or anti-angiogenic drugs is still being evaluated in prospective clinical trials.

The first reports of the prospective phase II CAPTEM vs. TEM study in 145 patients with progressive pancreatic NETs confirm the effectiveness of temozolomide-based chemotherapy and suggest superiority of the combined regimen (CAPTEM) compared to TEM alone in terms of longer PFS [22.7 months compared to 14.4 months, respectively; hazard ratio (HR): 0.58, p = 0.023]. However, the objective response rate was not statistically different, ORR 33.3% vs. 27.8%. The median OS for temozolomide monotherapy was 38 months, and for CAPTEM the median value was not achieved (HR: 0.41; p = 0.01) [215].

It seems that a higher percentage of objective responses to alkylating agents, including temozolomide in pancreatic NENs, may correlate with the expression of DNA repair enzyme MGMT (O6-methylguanine-DNA methyltransferase) the deficiency of which is observed more often (about 50%) in PanNENs than in SINENs [216–219]. However, determination of MGMT expression or methylation status is not currently recommended as a criterion for selecting chemotherapy due to the lack of a standardised MGMT determination method, and the results of studies conducted remain controversial. Results from prospective clinical trials regarding this issue are required [202, 216, 220]. In patients with pancreatic NENs after the failure of first-line chemotherapy, there are alternative systemic treatment options in patients who are in good general condition: in the case of progression in the course of STZ therapy, the second-line treatment is temozolomide ± capecitabine (CAPTEM); in the case of progression when using the CAPTEM scheme as the first-line therapy (due to the lack of STZ availability), chemotherapy with oxaliplatin + 5-FU or capecitabine is recommended.

The effectiveness of the oxaliplatin regimen in the available studies ranged from 17% to 30%, and stabilisation was achieved in 50–67% of patients. However, these studies were conducted in very small groups of patients and require confirmation [221].

Assessing the actual value of chemotherapy in patients with NETs G1/G2 located outside the pancreas is even more difficult due to the small number of studies and their conflicting results. Because of the low ORR, some scientific societies do not mention chemotherapy as a therapeutic option in this indication. Systemic chemotherapy is not recommended for NETs of non-pancreatic origin, except in biologically aggressive tumours after exhausting other treatments. If this treatment is to be considered, two-drug regimens, analogous to those used in pancreatic neuroendocrine neoplasms, are applied. In clinical practice, two-drug regimens with 5-FU and doxorubicin or streptozocin/dacarbazine [221, 222], or the full oral temozolomide and capecitabine regimen (CAPTEM) are usually opted for [223]. Some reports also indicate the efficacy of the combination of oxaliplatin and 5-FU derivatives [224] or irinotecan with 5-FU [225].

G3 neuroendocrine tumours (NETs G3) and neuroendocrine carcinomas (NECs)

NET G3

The 2019 WHO classification identifies well-differentiated neuroendocrine tumours with a high proliferation index > 20% — NETs G3. Their biological diversity is noteworthy; they are more commonly characterised with lower chemosensitivity, lower dynamics and lower aggressiveness, most of them are associated with DAXX/ATRX gene mutations, and 60–70% of these neoplasms originate from the pancreas, probably as a result of cell dedifferentiation and selection of aggressive cell clones [1, 203]. In metastatic NET G3, the efficacy of platinum-based chemotherapy appears to be limited, with response rates ranging from 0% to 10%. In two studies no objective response rate to platinum-based chemotherapy was obtained in patients with NETs G3 [226, 227]. In a study by Heetfeld et al. the response rate to platinum-based chemotherapy was 2% in patients with NETs G3 and 39% in patients with NECs [228]. In another study, in 16 patients with PanNETs G3, the response rate to platinum-based chemotherapy was 10% [229]. Patients with PanNENs participating in the NORDIC study demonstrated higher response to the platinum-based regimen if their Ki-67 index was > 55% (42% vs. 15%), suggesting that aggressive lesions are more responsive to this regimen [230]. On the other hand, the results of studies using chemotherapy with alkylating agents in patients with NET G3 are more encouraging. In 16 patients with PanNET G3, the response rate to chemotherapy based on alkylating agents was 50% [229]. In addition, a multicentre study of patients with NEN G2 and G3 (including 11 patients with NET G3) evaluated the effects of capecitabine and temozolomide (CAPTEM) in 22% of patients receiving this regimen as the first-line treatment. There was a trend towards median PFS improvement in patients with NET G3 and Ki-67 < 55% (15 vs. 4 months, p = 0.117) and in patients who received CAPTEM as the first-line therapy (17 months vs. 8 months, p = 0.3) [231]. In a retrospective analysis of studies conducted in Polish centres among 32 patients with G3 neuroendocrine tumours treated with capecitabine and temozolomide, the disease control rate was twice as high in the group of patients with NET G3 compared to NEC (70% vs. 30%). Progression-free survival in NET G3 patients was 15.3 months (95% CI: 3.9–30.4), and in the case of patients with NEC it was 3.3 months (95% CI: 2.5–7.1). The median overall survival was 22 months (95% CI: 11.8–22.0) and 4.6 months (95% CI: 2.2–5.9), respectively [232]. The toxicity associated with the use of CAPTEM is acceptable, and its safety was further confirmed in a large retrospective study in a group of 426 patients, in whom severe thrombocytopaenia and G4 neutropaenia were more frequent in women than in men [231, 233, 234]. Recommendations indicate the possibility of using other regimens in this group of neoplasms; apart from capecitabine and temozolomide (CAPTEM), there is oxaliplatin + capecitabine or 5-fluorouracil (CapOX/XELOX/FOLFOX). However, these are only retrospective reports based on the experience of individual centres [230, 235, 236].

NEC

Chemotherapy is the basic method of palliative treatment of patients with advanced GEP-NECs, which are characterised by an aggressive course and a high proliferation index. They account for more than 10% of all gastrointestinal neuroendocrine neoplasms. Their rapid growth makes them more sensitive to cytotoxic treatment than well-differentiated neoplasms, but their prognosis is much worse [230]. In the case of NEC, the treatment of choice is the use of ChT based on regimens containing cisplatin or carboplatin and etoposide, which allows for 42–70% response (including complete response in 20–25% of patients), with the duration of response rarely exceeding 10 months and with the median overall survival ranging from 15 to 19 months [214, 215]. According to the results of the NORDIC NEC study, the effectiveness of cisplatin is comparable to that of carboplatin, and the choice is based on a different toxicity profile [230]. In the NORDIC NEC study, cancer patients with Ki-67 above 55% had a higher response rate (ORR 42% and 15%, respectively) but shorter survival (OS 10 months and 14 months, respectively) than patients with Ki-67 below 55%. The results of the analysis of the 305 study participants demonstrated that the negative prognostic factors included the following: poor physical condition [WHO/ECOG (Eastern Cooperative Oncology Group) performance status > 2], primary tumours located in the large intestine and rectum, and an increase in the number of platelets or lactate dehydrogenase (LDH) concentration. All of these factors were associated with a reduction in overall survival [230].

The effectiveness of chemotherapy should be assessed every 2–3 cycles. First-line therapy should be administered for a total of six cycles followed by a break in therapy in patients without disease progression. No maintenance treatment is recommended. In the case of achieving good response to first-line chemotherapy, maintained for at least three months after the completion of treatment, and in the absence of treatment toxicity (e.g. neurotoxicity, ototoxicity, renal failure), re-induction with platinum derivatives may be considered using a PE/CE regimen (cisplatin + etoposide/carboplatin + etoposide).

If tumour progression occurs within the first 4–6 months, second-line chemotherapy should be considered. The use of second-line chemotherapy can be considered individually only in patients with a good performance status [1, 230].

Topotecan turned out to be ineffective in the treatment of NEC [230]. There is no therapeutic standard for second-line treatment. Inclusion in clinical trials should be considered whenever possible.

In the available studies, chemotherapy according to the FOLFIRI regimen achieved an objective response rate of 24%, and a median PFS of about three months [237, 238], while the FOLFOX regimen can achieve and objective response rate of about 21–29% and a median PFS between 2.5 and 4.5 months [238, 239]. A combination of cytostatics and alkylating agents (temozolomide or dacarbazine) may be considered, although there is no evidence from prospective studies, and significantly better results are achieved in pancreatic NETs G3 and NECs with Ki-67 < 55% [232, 236].

Reports on the addition of the anti-angiogenic drug bevacizumab to chemotherapy using the FOLFOX or FOLFIRI regimen are promising [240]. However, this requires further confirmation; therefore, the use of bevacizumab should be limited to scientific studies [241]. Taxanes also show some efficacy in NEC [193].

MiNEN

MiNENs are often treated in a way analogous to their non-neuroendocrine (nNE) or neuroendocrine (NE) component, but there are no prospective data on the optimal therapeutic strategy [199].

Minimal consensus statement on chemotherapy in NENs

3.3.6. Targeted therapy

Molecularly targeted therapies

In patients with NETs of the gastrointestinal system, molecularly targeted drugs have proven to be an effective and safe therapy. The mechanism of action of targeted drugs includes blocking the activity of many receptors associated with the processes of neoangiogenesis, neoplastic cell proliferation, and inhibition of metastasis. The effectiveness of targeted therapies was first confirmed in pancreatic NETs G1/G2 — in the advanced stage of the disease, two drugs with anti-angiogenic properties were used: a selective mammalian target of the rapamycin (m-TOR) inhibitor — everolimus, and an inhibitor of numerous receptor tyrosine kinases (TKI) — sunitinib [191, 192, 242]. In Poland and in the other countries of the European Union, both drugs have been approved for the treatment of inoperable and/or metastatic highly or moderately differentiated pancreatic neuroendocrine tumours in adult patients with progressive disease [194, 195]. Their role is discussed in the recommendations for pancreatic NENs [54].

Additionally, on the basis of the results of the RADIANT-4 study, everolimus has been registered both in the USA and in Europe in the treatment of advanced NETS G1/G2 (hormonally nonfunctional) originating from parts of the gastrointestinal tract other than the pancreas, as well as from the lungs. A total of 302 patients participated in this prospective, placebo-controlled, randomised, phase III study, 24% of whom were patients with small intestine tumours, 13% were patients with rectal NENs, and approximately 30% were patients with lung NENs. The effects of the drug were also assessed in NEN patients with unknown primary site. There was a significant delay in time to disease progression in the everolimus group compared to the placebo group (11 months vs. 3.9 months), with a more than a two-fold reduction in the risk of progression or death (HR: 0.48; 95% CI: 0.35–0.67; p < 0.00001). The majority of patients had disease stabilisation (81% in the everolimus group vs. 64% in the placebo group), while objective responses were sporadic [243]. Currently, everolimus therapy in non-pancreatic locations is still not reimbursed in Poland.

m-TOR inhibitors may be considered as first- or second-line treatment options after chemotherapy or after SSA therapy, both “cold” and “hot” (PRRT), in locally advanced inoperable or well-differentiated metastatic NETs (G1 and G2) of the gastrointestinal tract. According to the consensus reached, targeted therapies should not be widely used as the first-line treatment due to the potential risk of complications and the lack of reliable research results [1, 199]. There is also no evidence to establish the exact sequence of using the various therapeutic options in the treatment of NENs [244]. In the event of sequential use of therapy, potential toxicity should be taken into account, as indicated by the observations of an Italian, retrospective multicentre study [244, 245] conducted with the participation of 169 patients, which demonstrated a significant increase in toxicity when everolimus was used in patients previously treated with PRRT and/or chemotherapy. Another, smaller, retrospective study conducted in Holland with 24 patients demonstrated that prior use of PRRT did not affect the safety of everolimus therapy [246]. The position of everolimus in the treatment algorithm of progressive NETs is being further investigated. We are currently waiting for the results of two prospective studies: the first study comparing everolimus vs. PRRT (COMPETE) and the second one comparing everolimus vs. streptozotocin-based chemotherapy in PanNETs (SEQTOR) (NCT03049189, NCT02246127).

In choosing the targeted therapy with either everolimus or sunitinib in the case of pancreatic NETs, knowledge about the adverse effects of treatment and the profile of the patient’s comorbidities is helpful.

Primary adverse events of grade I and II according to CTCAE (Common Terminology Criteria for Adverse Events v. 3.0) [247] associated with everolimus therapy are as follows: mucosal changes (64%), rash (49%), diarrhoea (34%), fatigue syndrome (31%), and infections (23%), and in the case of grade III and IV: anaemia (6%) and hyperglycaemia (5%). Whereas the treatment with sunitinib was associated with such CTCAE grade I and II complications as the following: diarrhoea (54%), nausea (45%), weakness (34%), and fatigue (32%), and serious, grade III and IV complications — neutropaenia (12%) and hypertension (10%) [191, 192]. Both targeted drugs are oral preparations taken in continuous therapy until disease progression or unacceptable toxicity (everolimus 10 mg/day, sunitinib 37.5 mg/day). The use of everolimus may be limited by uncontrolled diabetes or lung disease (non-infectious pneumonia is a class effect of rapamycin derivatives, including everolimus) [228], and in the case of sunitinib, by severe cardiovascular disease [229]. For both targeted drugs, no significant improvement in quality of life was demonstrated based on FACT-G (Functional Assessment of Cancer Therapy — General Questionnaire) quality of life forms.

The use of targeted therapies in combination with SSA in the treatment of hormonally functional NETs is standard practice [1]. The effectiveness of everolimus has been demonstrated in the treatment of hormonally functional pancreatic insulinomas in the scope of controlling symptoms of hypoglycaemia [247, 248] as well as in controlling symptoms of carcinoid syndrome (RADIANT-2) [249]. A similar hypoglycaemic effect was demonstrated in individual cases of pancreatic NETs with sunitinib [250]. The effectiveness of the combination of molecularly targeted drugs with SSA (everolimus + octreotide) in first-line hormonally non-functional NETs with somatostatin receptor overexpression is based on the results of the phase II RADIANT 1 study conducted in patients treated for PanNETs. In this study, PFS was achieved at 16.6 months in the group receiving everolimus in combination with Octreotide LAR 30 mg, compared with 9.7 months in the group receiving everolimus monotherapy [251]. There is still insufficient evidence of the advantage of the targeted drug — SSA combination over the targeted drug monotherapy (> 90% benefit in terms of disease control, no effect on PFS) [252]. The 2020 ESMO recommendations do not include maintaining the therapy with an SSA for non-functioning GEP-NENs based on the results of the COOPERATE-2 study, where the addition of pasireotide to everolimus did not bring benefits compared to everolimus monotherapy in progressive pancreatic NETs [1, 253]. There are currently insufficient data to support the use of other targeted therapies, including bevacizumab, sorafenib, pazopanib or axitinib, and cabozantinib, in the treatment of gastrointestinal NENs [254–257]. Studies are ongoing, and preliminary results regarding the use of lenvatinib or cabozantinib are promising.

It is worth mentioning that the recent publications on the activity of surufatinib in the Chinese population are encouraging. A phase III study (SANET) in both pancreatic NETs and non-pancreatic neuroendocrine tumours after previous use of molecularly targeted drugs evaluated the effect of surufatinib, a drug with both anti-angiogenic (inhibiting VEGFR1, VEGFR 2, VEGFR3 and FGFR1) and immunomodulating (CSF1R) potential. The median time to progression in patients receiving surufatinib (9.2 months) was significantly longer than that obtained in the placebo group (3.8 months) [258, 259]. Because the drug is not registered and there are only preliminary results of a phase II study in non-Asian population, the possibility of using its potential in clinical practise is limited [260].

Molecularly targeted drugs can be used as treatment options for first-line or subsequent lines of treatment following chemotherapy or administration of SSA, or SSA combined with PRRT.

The use of targeted treatment in NET G3 or NEC is currently not justified in medical practice due to the lack of evidence from controlled, prospective clinical trials. Reports on the efficacy of molecular targeted drugs show some efficacy in the NET G3 group. Phase II studies are currently underway to assess the effects of everolimus in NET G3 and NEC (NCT02113800, NCT02248012).

Molecularly targeted therapies — adjuvant treatment

There is no scientific evidence to justify the use of targeted therapies in adjuvant management in NETs G1/G2 and NETs G3 as well as NECs of the gastrointestinal system.

Minimal consensus statement on targeted treatment in NENs

3.3.5 Immunotherapy

Immunotherapy, which looks promising in many cancers, still does not play a significant role in the treatment of neuroendocrine neoplasms (NENs). Research to date shows some potential of immunotherapy in treating high-grade NENs.

A prospective, open-label, multicentre, phase II clinical trial, DART (Dual Anti-CTLA-4 and Anti-PD-1 Blockade in Rare Tumours, still ongoing, NCT02834013) using the combination of ipilimumab and nivolumab in various types of cancer, included 33 patients with NENs other than pancreatic, including 19 (56%) with NEC, 10 (31%) with NET G2, and 4 (12%) with NET G1. Primary lesions were located in the lungs (6 patients), small intestine (6), stomach (2), rectum (4), cecum (1), CUP (cancer of unknow primary) (5), and other sites.

For the entire cohort, the ORR was 25%. A response was only observed in patients with high-grade cancer [261]. When the group of patients with NEC was assessed, the overall response rate was 26% (regardless of the primary lesion site). The clinical benefit rate (response or stable disease for more than 6 months) was 42% in patients with NEC. The six-month progression-free survival rate was 32%, and the median overall survival was over 8.7 months. The treatment regimen was well tolerated. The most common symptoms were fatigue (32%) and rash (26%). No cases of fatal toxicity were reported [262]. In another phase II study involving 29 patients with advanced NETs the objective ORR was 24% and for pancreatic localisation the ORR was 43% [263]. The KEYNOTE-028 non-randomised, phase 1b study assessed the safety and efficacy of the programmed cell death 1 (PD-1) inhibitor pembrolizumab as monotherapy in a large cohort of patients with advanced solid tumours positive for programmed death ligand 1 (PD-L1) [264]. The study included 16 patients with pancreatic NEC, achieving stable disease (SD) in 14 and partial response (PR) in one patient. The six-month PFS was 40%, the 12-month PFS was –27%, and the 12-month OS rate was 87%. In a subsequent phase II study, KEYNOTE-158, including GEP-NET, three patients with PanNET and one patient with rectal NET had a partial response after a median follow-up of 24 months [265].

Spartalizumab

Spartalizumab is a humanised anti-PD-1 monoclonal antibody evaluated in a phase II, single-arm, multi-centre study of patients with well-differentiated metastatic NETs G1/G2 (32 GI-NETs; 33 PanNETs) and GEP-NEC. Higher expression of PD-L1 was observed in cells of the immune system in 21 patients with GEP-NEC than in patients with GEP-NET. These data did not reach the primary endpoint, defined as ORR ≥ 10% [266]. In the GEP-NEC group, ORR was 4.8% (95% CI: 0.1–23.8), and 12-month overall survival was 19.1%. Interestingly, the ORR was higher in patients with higher PD-L1 expression or CD8+ infiltration at baseline.

Toripalimab

Toripalimab is a humanised IgG4 antibody with the human PD-1 receptor as a target. A phase 1b study assessed its efficacy in patients with relapsed or metastatic NENs after first-line treatment [267]. In a cohort of 40 patients, the ORR was 20% and the median durability of response (DOR) was 15.2 months. Interestingly, in tumours with PD-L1 expression ≥ 10% the ORR was 50%, while in tumours with PD-L1 < 10% the ORR was 10.7% (p = 0.019).

At present there are insufficient data on the use of immunotherapy in NENs.

Conclusion

3.4. Radioisotope therapy

NET radioisotope therapy uses radiolabelled somatostatin analogues, while metaiodobenzylguanidine derivatives ([131I]mIBG) labelled with 131I are used sporadically [4].

Patients with advanced, unresectable NETs G1 and G2 are eligible for PRRT with radiolabelled somatostatin analogues. Recent data also show the effectiveness of this treatment in patients with NETs G3 [268].

There are no indications for the use of PRRT therapy as adjuvant treatment after radical surgery [4].

3.4.1. Peptide receptor radionuclide therapy with radiolabelled somatostatin analogues (PRRT), recently called targeted radioligand therapy (RLT)

The experience with NET radioligand therapy to date includes primarily the use of DOTA-Tyr3-octreotide and DOTA-Tyr3-octreotate labelled with 177Lu, 90Y, or a mixture of these radioisotopes. Data from non-randomised clinical trials demonstrate that the response to PRRT in patients with NET G1 and G2 (complete and partial remissions) can be obtained in approximately 8–46% of patients, and the median PFS after treatment is 25–36 months [131, 269-276]. In a retrospective analysis of 149 patients with NETs G3, a CR of 1% and a PR of 41% were obtained. Survival parameters were dependent on Ki-67, and in patients with NETs G3, Ki-67 of 20–55%, PFS was 16 months and OS was 31 months, while in patients with NETs G3, Ki-67 > 55%, PFS was 6 months and OS was 9 months [268].

NETTER-1, a prospective, randomised, phase III trial evaluating the effect of treatment with [177Lu]Lu-DOTATATE + Octreotide LAR 30 mg (patients with symptoms of carcinoid syndrome) vs. 60 mg Octreotide LAR in midgut neuroendocrine neoplasms, demonstrated a 79% (95% CI: 83% to 64%) reduction in the risk of neoplastic disease progression in the group of patients subjected to PRRT. The PRRT response rate (complete and partial remissions) was 19% (95% CI: 11–26%), and the estimated PFS was 40 months [277].

The best candidates for PRRT are patients with intensive accumulation of the radiotracer (at least Krenning 2/3, preferably 3/4) in all neoplastic lesions and with its homogeneous distribution [131].

If not all neoplastic lesions show increased accumulation of the radiotracer, therapy may be considered as a palliative treatment to reduce the symptoms of neoplastic disease and prolong overall survival. In individual cases, the use of PRRT can also be considered as a neoadjuvant treatment to regress the tumour mass before the planned surgery [4, 131, 278].

Qualification for targeted radioligand therapy

Patients with well-differentiated NETs characterised by significant expression of the somatostatin receptor confirmed in the SRI examination are eligible for treatment with radiolabelled somatostatin analogues — with uptake intensity in the tumour/metastatic lesions at least equal to that in the liver, i.e. 2 in the Krenning score [279]. The reference organ used in this score is the liver, which is characterised by one of the lowest physiological accumulations [280]. The current data indicate that the best PRRT effect is achieved when the uptake in metastatic lesions is 2.2-fold greater than in the liver or when SUVmax > 13–16.4 [281, 282].

In EANM consensus the experts agreed that patients whose neoplastic lesions are characterised by the uptake of at least 3/4 should be eligible for treatment [131].

Exclusion criteria for PRRT/RLT:

a) haemoglobin (Hb) < 8 g/dL,

b) platelets < 80 × 103/µL,

c) white blood cells (WBC) < 2 × 103/µL*,

d) lymphocytes < 0.5 × 103/µL,

e) neutrophils < 1 × 103/µL,

Patients with leukopaenia < 3000, neutropaenia < 1500, thrombocytopaenia < 100,000, and creatinine clearance < 60 mg/mL, due to an increased risk of adverse effects, should be qualified for treatment with caution [4].

Data on the effectiveness of treatment in people under 18 years of age are scarce, and therefore the treatment in this age group should always be considered individually.

Qualifying examinations before implementing PRRT/RLT:

PET/CT with [18F]FDG is recommended as a qualifying examination for PRRT/RLT to assess the biological malignancy of neoplastic disease and to exclude “mismatch” [18F]FDG + /[68Ga]Ga-DOTA-SSA foci, especially in patients with NETs G2 and G3 and as a prognostic factor [4, 131, 142, 143, 145, 146, 150, 151, 272].

Radioisotope therapy regimen

Treatment is normally carried out in four cycles at 8–12-week intervals with the use of 90Y-, 177Lu-, or 90Y/177Lu-labelled somatostatin analogues. Currently, due to lower nephrotoxicity and data from the NETTER-1 study, the 177Lu radioisotope or the 90Y/177Lu tandem treatment is preferred in most centres. In the course of PRRT, an infusion of an amino acid solution is necessary for radioprotection of the kidneys [4]. There is no conclusive evidence that treatment with SSAs reduces the effectiveness of treatment with radiolabelled somatostatin analogues. If there are clinical indications for the use of SSAs, this treatment should not be discontinued during PRRT; however, the interval between the administration of long-acting analogue and PRRT should be at least four weeks. If it is necessary to continue treatment with SSAs before the implementation of PRRT, short-acting analogues are recommended, which should be discontinued the day before starting PRRT [8]. Literature data on SSAs biologic therapy after the completion of PRRT/RLT (in patients without carcinoid syndrome symptoms) are scarce [276, 283], and the decision should be made on an individual basis.

In the summary of the product characteristics of Lutathera, it is stated that the use of corticosteroids may reduce the expression of SST2 receptors [284]. It is recommended that the administration of high doses of glucocorticoids be avoided during PRRT. In patients with chronic corticosteroid use, the expression of somatostatin receptors should be carefully assessed.

Adverse effects of PRRT/RLT

Adverse effects should be monitored using oncological criteria, optimally on the basis of National Cancer Institute Common Terminology Criteria for Adverse Events (NCI CTCAE) rev.4.03 or later. They mainly concern the haematopoietic system and kidneys. The infusion of amino acids, L-lysine + arginine, before treatment reduces the risk of radiation damage to the kidneys. In patients with carcinoid syndrome, breakthrough symptoms may occur during therapy; in these patients, short-acting SSAs should be used before, during, and immediately after PRRT.

Retreatment with radiolabelled somatostatin analogues

In the case of progression after effective radioisotope therapy lasting for at least a year, a repeating PRRT may be considered [131, 273, 285–287]. However, in the case of repeated PRRT, a shorter PFS should be expected [287]. If a decision is made to repeat PRRT— due to the greater toxic effect of 90Y — it is recommended that 177Lu is used. Individual dosimetry measurements should also be considered.

3.4.2 Treatment with [131I]mIBG.

Treatment with [131I]mIBG should be considered in patients with a negative SRI examination result and preserved [131I]mIBG accumulation in the primary tumour and/or metastases. This treatment is primarily palliative and allows for the relief of clinical symptoms, including carcinoid syndrome, and other clinical symptoms associated with advanced neoplastic process [4, 131]. Contraindications to treatment include bone marrow depression (according to the above-described criteria). In patients with an active thyroid gland, it is necessary to block the uptake of free 131 iodine not bound to mIBG carrier (using Lugol’s iodine or sodium perchlorate).

Qualification for [131I]mIBG radioisotope therapy

Diagnostic scintigraphy with [131I]mIBG or [123I]mIBG is the basic examination used in the qualification for radioisotope therapy. Due to its physical properties, the iodine 123 isotope is preferred (its availability in Poland is limited because of high costs). The examinations should be performed using the whole-body SPECT/CT method with a careful analysis of the location of the radiotracer collection and its comparison with the NET foci visible in CT or MRI. The remaining laboratory tests necessary for the qualification are the same as those used during the qualification for PRRT [4, 131].

3.4.3. PRRT/RLT result assessment

Assessment of treatment results should include structural examinations (i.e. CT, MRI) and somatostatin receptor imaging (note: in order to assess the effectiveness of treatment, SRI must be performed using the same methods as in the case of the qualification for the therapy) approximately three months after the completion of treatment, and subsequently every six months for two years [4, 131]. Further follow-up depends on the clinical course of the disease. RECIST 1.1 is currently used to assess the objective response rate. In the case of assessment of the response to PRRT, the criteria for assessing the effects of treatment are still being discussed, including functional imaging [4, 131].

Sharma et al. used standard and modified Positron Emission Tomography Response Criteria in Solid Tumors (PERCIST) to assess the response to PRRT based on baseline and follow-up [68Ga]Ga-DOTATATE PET/CT scores. They correlated PFS with a single lesion baseline maximum standardised uptake value (SUVmax), the average maximum SUV (SUVmax-av) for up to five lesions, the tumour-to-spleen SUV ratio (SUVT/S) and the tumour-to-liver SUV ratio (SUVT/H), showing that baseline SUVmax and SUVmax-av predicted the response to PRRT ([177Lu]Lu-DOTATATE) [282]. A comparison of anatomical CT imaging (RECIST1.1 and Choi) and receptor imaging (PERCIST) analysis demonstrated a higher [68Ga]Ga-DOTATATE PET/CT value in predicting disease progression [288].

The role of PRRT/RLT in the treatment of advanced NENs

Experts recommend PRRT as the second line of treatment (after unlabelled analogues) for unresectable or disseminated GEP-NETs G1/G2/G3 with Krenning 3/4 uptake in all lesions [131]. During the expert discussion, it was concluded that PRRT may constitute the first-line treatment in patients with unresectable or disseminated NETs in selected cases depending on SSTR expression, symptoms, and primary site location [131]. However, in the recommendations of ENETS, PRRT is not considered as the first-line treatment. It may be considered in patients with advanced disease progression despite the use of another therapeutic method. In patients with midgut tumours, it is recommended as the second-line treatment after the failure of SSA and as an alternative to everolimus [85, 199].

According to ESMO guidelines, PRRT can be used as second-line treatment in patients with NET G1/G2 of the small intestine and Ki-67 < 10% with progression on SSA [1].

Undoubtedly, the progression of neoplastic disease is an indication for the implementation of cytotoxic therapy (chemotherapy/radiotherapy/targeted therapy); however, there are no studies assessing which of them constitutes the most effective first-line treatment. When selecting treatment, the location of the primary site, the dynamics of the neoplastic process, the degree of histological maturity, and the expression of somatostatin receptors should be taken into account.

The main indication for PRRT is disease progression after biologic therapy with SSAs (level 1 evidence in NETs of the small intestine); however, in the case of advanced disease that poses a risk of multiorgan failure, PRRT may be considered as first-line treatment. In other GEP-NEN locations, PRRT can be considered when disease progression is observed in imaging examinations, and individually in the case of advanced disease and high expression of somatostatin receptors in the SRI [4, 131].

Development directions

Research is currently underway to improve the effectiveness of PRRT. The studies involve the use of somatostatin receptor antagonists, which are characterised by much stronger binding to the membrane receptor. In preliminary studies, the accumulation of the [177Lu]Lu-DOTA-JR1 SSTR antagonist was higher than the [177Lu]Lu-DOTATATE SSTR agonist in metastatic NET lesions. Unfortunately, greater accumulation of the radiopharmaceutical was observed in critical organs — kidneys and bone marrow [289, 290].

Another direction of PRRT development is visible in the works on local administration to the hepatic artery [35, 36], and the use of the a-emitters 213Bi and 225Ac [291, 292].

One of the methods that increases the sensitivity of cells to ionising radiation is the use of compounds called radiosensitisers. For this purpose, attempts are being made to combine PRRT with capecitabine, temozolomide, or 5-FU [293–295].

Another attempt was the combination of PRRT with mTOR (mammalian target of rapamycin kinase, everolimus) inhibitor. However, in the preclinical study, combination therapy was less effective than PRRT alone [296].

Minimal consensus statement on radioligand therapy in NENs

3.5. Radiotherapy in neuroendocrine neoplasms

The interest in radiotherapy as an effective method of local treatment resulted from the implementation of sophisticated modern irradiation techniques. Retrospective analysis (data from the 1990s and the 2000s) of the results of treatment with conformal radiotherapy, often combined with simultaneous chemotherapy, demonstrated benefits of the therapy used as radical or adjuvant postoperative treatment, but the toxicity of the treatment limited the total dose to 50.4 Gy/28 fractions, which is suboptimal [297–299]. In contrast, brachytherapy has been shown to be highly effective in the treatment of NENs metastases to the liver due to the high fractional dose of radiation [300]. The use of stereotactic body radiotherapy (SBRT) enables precise administration of high fractional radiation doses into the tumour, which significantly increases the bioequivalent dose deposited in the tumour. Low proliferation neoplasms (Ki-67) are radio-resistant to conventional fractionation, but sensitive to higher fractional doses that are used in SBRT. Therefore, these techniques provide 70–100% local control, which is comparable to surgical resection. SBRT is used in unresectable tumours of the lung, rectum, pancreas, or liver [301–303].