1. Epidemiology, definitions, incidence, classifications, prognosis/survival

The incidence of pancreatic neuroendocrine neoplasms (PanNENs), both functional pancreatic neuroendocrine neoplasms (F-PanNENs) and non-functional pancreatic neuroendocrine neoplasms (NF-PanNENs), is constantly increasing and amounts to around 8–10 million cases per year [1–3]. PanNENs represent approximately 30% of all gastro-entero-pancreatic neuroendocrine neoplasms (GEP-NENs). PanNENs represent 4% of all neuroendocrine neoplasms (NENs) originating from diffuse endocrine system (DES) cells present in the gastrointestinal tract and pancreas, as well as in the bronchi and other organs. PanNENs account for 1–2% of all pancreatic tumours [2–6].

NF-PanNENs constitute the majority (60–90%) of PanNENs. They are usually detected in more advanced stages because of their slow growth and often asymptomatic or oligosymptomatic course [7–13]. NF-PanNENs sometimes exhibit the ability to produce peptides such as hormones (including neurohormones) or neurotransmitters, for example pancreatic polypeptide (PP), chromogranin A (CgA), neuron specific enolase (NSE) subunit b chorionic gonadotropin (b-hCG), calcitonin, neurotensin, and synaptophysin, but their clinical symptoms are mute in this respect.

The clinical symptoms of F-PanNEN depends on the hormones or biogenic amines, and their metabolites, secreted by the tumour [12, 14–16]. The clinical manifestations of F-PanNEN may change during the natural course of the disease depending on the secreted biologically active substances. The most common F-PanNENs include [15, 17] insulinoma — inulin-producing tumour and gastrinoma — gastrin-producing tumour.

Rare F-PanNENs include the following: glucagonoma — glucagon-producing tumour, VIPoma — vasoactive intestinal peptide (VIP) producing tumour, GHRHoma — growth hormone-releasing hormone (GHRH)-producing tumour, ACTHoma — adrenocorticotropin hormone (ACTH)-producing tumour, F-PanNEN causing carcinoid syndrome — producing serotonin, tachykinins, and PTHrPoma — PTH-related peptide-producing tumour. The existence of somatostatinoma is being questioned [18].

Very rare F-PanNENs include the following: reninoma — renin-producing tumour, LHoma — luteinising hormone (LH)-producing tumour, cholecystokininoma — cholecystokinin (CCK)-producing tumour, as well as tumours secreting erythropoietin, insulin-like growth factor-2 (IGF-2), and glucagon-like peptide-1 (GLP-1). F-PanNEN can also occur in extra-pancreatic locations. The clinical picture reflects the activity of the above-mentioned biologically active substances produced by these tumours. In the case of very rare F-PanNENs, there are often doubts as to the interpretation of the occurring symptoms [17, 19].

Most PanNENs are sporadic (non-hereditary) tumours, although a variable proportion of different F-PanNENs occur as components of hereditary syndromes. Multiple endocrine neoplasia type 1 (MEN-1) remains the most significant inherited disorder, accounting for 20–30% of gastrinomas and less than 5% of insulinomas or rare F-PanNENs tumours [20, 21]. Hereditary PanNEN sometimes coexists with von Hippel-Lindau disease (VHL) (10–17% of VHL patients develop NF-PanNEN), neurofibromatosis type 1 (NF-1), and, very rarely, tuberous sclerosis [20].

The median survival in all patients with PanNENs is 136 months and drops to 77 months in patients with lymph node metastases. In 64% of patients with distant metastases, the median survival is only 24 months [2].

2. Clinical characteristics of PanNENs

2.1. Clinical symptoms of functional pancreatic neuroendocrine neoplasms (F-PanNENs)

Insulinoma — pancreatic neoplasm that produces insulin is the most common hormonally functional neuroendocrine tumour of this organ. Clinical characteristics of insulinoma are presented in Table 1.

Clinical symptoms (specified in Tab. 1) are related to insulin secretion and subsequent hypoglycaemia, and not to the tumour mass effect (its diameter usually does not exceed 2 cm).

The prognosis: in the case of benign tumours, it is very good; in over 95% of patients complete recovery is achieved following surgery. In patients with distant metastases, the median survival is less than two years. Tumour diameter greater than 2 cm, Ki-67 proliferation index over 2%, various molecular and chromosomal disorders, for example loss of the 3p or 6q fragment, are factors associated with shorter survival [12, 14, 19, 22, 23].

Gastrinoma is described in the first part of the recommendations — in the article Update of the diagnostic and therapeutic guidelines for gastro-entero-pancreatic neuroendocrine neoplasms [24], and in the article Gastroduodenal neuroendocrine neoplasms including gastrinoma — update of the diagnostic and therapeutic guidelines (recommended by the Polish Network of Neuroendocrine Tumours).

The remaining F-PanNENs included in the group of rare and very rare F-PanNENs are specified in Table 2. The prognosis in these cases depends on tumour size and the presence of distant metastases. Five-year survival in the advanced stage of the disease is estimated at 29–45%. Unfavourable prognosis is associated with Ki-67 above 2%, the occurrence of distant metastases, chromosomal disorders, and the presence of cytokeratin-19 [17, 19, 25].

The prognosis for patients with F-PanNENs occurring in the course of MEN-1 syndrome remains unclear; however, it is associated with the significant and ever-increasing effectiveness of treatment of F-PanNEN in this syndrome [20]. Compared to sporadic tumours, genetically determined PanNENs occur in younger patients; they are more often multifocal, more aggressive, and may be accompanied by cell proliferation.

2.2. Clinical symptoms of non-functional pancreatic neuroendocrine neoplasms (NF-PanNEN)

They do not cause characteristic symptoms related to the secretion of biogenic peptides or amines and their metabolites. In some NF-PanNENs, the presence of these substances is confirmed by immunohistochemical methods, but they are not secreted into the blood or are secreted in such small amounts that they do not cause any clinical symptoms. Most of the NF-PanNENs are well-differentiated neuroendocrine tumours (WD-NETs). They occur with the frequency of 1.8/1,000,000 persons/year in women and 2.6/1,000,000 persons/year in men. The frequency of their detection increases with age, and the peak incidence is observed in the 6th and 7th decade of life. In 3–53% (on average 19%) of cases, they coexist with MEN-1 syndrome (more often in the elderly) and in 13–17% of cases with VHL [12, 14].

Symptoms: NF-PanNENs are usually diagnosed late, when they have already reached a significant size, causing compression or infiltration of adjacent organs, or producing distant metastases [12, 14, 16]. The most common symptoms include abdominal pain (35–78%), weight loss (20–35%), and loss of appetite and vomiting (45%). Internal haemorrhage (4–20%), jaundice (17–50%), or palpable abdominal tumour (7–40%) are less common [12, 14, 16]. Hepatic metastases are observed in approximately 1/3 of patients with de novo NF-PanNEN. This percentage is lower than in previous studies (46–73%) [12, 14, 26].

Prognosis: in the available studies, the mean survival of patients with NF-PanNEN was determined at 38 months, with a five-year survival rate of 43% [16]. Patients with distant metastases had an average survival of about 23 months, compared with overal survival (OS) of 70–124 months .in the case of regional disease advancement only [12, 14, 16]. The degree of histological differentiation of the tumour is also an important factor influencing survival [16]. Other unfavourable prognostic factors include age over 40 years, high dynamics of the development of liver metastases (25% increase in their volume within 6–12 months), and the occurrence of bone metastases [12, 14, 16]. It has also been reported that the presence of calcifications in preoperative computed tomography (CT) in patients with PanNEN (observed in 16% of cases) is correlated with the degree of malignancy and the presence of lymph node metastases in highly differentiated PanNEN [27]. The extent of liver metastases affecting one or two lobes, or the presence of additional abdominal metastases were considered important predictors of survival, regardless of the tumor grade of histological maturity (G) [28]. It has been observed that most patients with advanced PanNEN experience progression over time, and the best predictor of progression proliferation index (Ki-67) [29]. Lymph node involvement and the number of affected lymph nodes in patients with PanNEN have significant prognostic value. On the other hand, the absence of symptoms is associated with a better prognosis, regardless of cancer stage, especially in NF-PanNEN. The prognosis in patients with PanNEN in the course of MEN-1 remains unclear [20]. NF-PanNEN in MEN-1 syndrome are often stable or slow-growing tumours, which, in the case of tumour size ≤ 2 cm, translates into watch-and-wait management [13].

3. PanNENs diagnostics

3.1. Biochemical diagnostics of PanNENs

It includes blood tests for the hormonal biochemical markers secreted by PanNENs. It is used in differential diagnosis, assessment of treatment effectiveness, and prognosis.

The characteristics of biochemical biomarkers specific and non-specific for PanNENs are included in the first part of the recommendations [24].

3.1.1. Biochemical diagnostics of functional pancreatic neuroendocrine neoplasms (F-PanNENs)

This diagnostics is based on the determination of increased concentrations of specific hormonal markers in the blood [e.g. gastrin in Zollinger-Ellison syndrome (ZES), insulin in insulinoma] in combination with changes observed in examinations and laboratory tests characteristic for a particular tumour type, such as excessive secretion of gastric juice in ZES or hypoglycaemia in insulinoma, etc. [12, 14, 18]. In sporadic PanNENs, the cell type may change, and the tumours may then produce additional peptides (in addition to the ones specific for a particular type of tumour) or biogenic amines and their metabolites. It is associated with worse prognosis, especially when the tumour secretes ACTH [12, 14].

Insulinomas are usually benign tumours with normal blood CgA levels, but they can be higher if the tumour is malignant and has metastasised. Unlike other PanNENs, CgA does not have very high diagnostic sensitivity in the case of insulinoma [30].

The diagnosis of insulinoma is facilitated by identification of the Whipple triad:

The diagnostic criteria for insulinoma are still evolving and differ by consensus, include endogenous hyperinsulinism (clinical symptoms and biochemical criteria): glucose < 55 mg/dL (3.0 mmol/L); insulin ≥ 3.0 μU/mL (18 pmol/L); C-peptide ≥ 0.6 ng/mL (0.2 nmol/l); and proinsulin ≥ 5.0 pmol/L.

The gold standard in the diagnosis of insulinoma is still the 72-hour fasting test, although some studies suggest that the 48-hour test may be sufficient. The fasting test is performed in a hospital setting with serial measurement of blood glucose. In patients with insulinoma, hypoglycaemia usually occurs within 24 hours [12, 14]. When a patient develops symptoms of hypoglycaemia and the blood glucose is ≤ 2.2 mmol/L (≤ 40 mg/dL), blood should be drawn to determine the levels of C-peptide, proinsulin, and insulin. The lack of adequate insulin suppression in the presence of hypoglycaemia confirms the presence of an autonomously secreting insulinoma [10, 12, 14].

The results of some studies confirm that in patients with MEN-1, insulin-secreting tumours occur earlier than gastrin-secreting tumours, and in 25% of patients, insulinoma appears before the age of 20 years [31]. Therefore, if insulinoma is diagnosed before the age of 20 years or there are multiple insulin tumours at any age, MEN-1 should be suspected and an appropriate genetic test should be performed.

The interpretation of the above diagnostic criteria for insulinoma takes into account the possibility of drug-induced hypoglycaemia (assessment of the amount of sulfonylurea and its metabolites in the blood and/or urine, and the concentration of C-peptide and insulin) [12, 14, 18], and postprandial, reactive, autoimmune hypoglycaemia.

The biochemical diagnostics of gastrinoma is discussed in the second part of the recommendations — in the article Gastroduodenal neuroendocrine neoplasms including gastrinoma — update of the diagnostic and therapeutic guidelines (recommended by the Polish Network of Neuroendocrine Tumours).

3.1.2. Biochemical diagnostics of rare PanNENs

This diagnostics includes confirmation of the presence of elevated concentrations of specific and nonspecific biochemical markers in the blood. Specific markers include, for example, the concentration of glucagon (when glucagonoma is suspected; concentration > 1000 pg/mL), VIP (concentration > 170 pg/mL) [12, 14]. A non-specific marker such as CgA may only indicate the presence of a neuroendocrine tumour and may be used to monitor the course of the disease but is not a basis for the diagnosis of hormonally active PanNEN syndrome. All biochemical tests should be performed at the first visit. Cushing syndrome, acromegaly, hypercalcaemic syndrome, and other syndromes require separate biochemical confirmation.

The assessment of specific NENs markers is useful in identifying and monitoring individual tumours, as shown in Table 3 [12, 14]. Indications for their determination depend on the clinical features of a patient with PanNEN.

The concentration of some peptides increases significantly after meals and may remain at an elevated level for over 6 hours after eating — blood samples for tests are collected only in the morning after overnight fasting [12, 14, 18]. In the case of CgA, this condition is not necessary; however, if the blood samples are not collected under fasting conditions, this information should be recorded. The concentrations of all PanNEN markers in the blood, except for insulin, are elevated in patients with renal failure; therefore, interpretation of the results in these patients is difficult. Among the numerous markers assessed in the blood, CgA has prognostic value for the majority of PanNENs [12, 14, 32, 33].

3.1.3. Biochemical diagnosis of non-functional pancreatic neuroendocrine neoplasms (NF-PanNENs)

In the diagnosis of PanNENs, it is recommended that blood CgA be measured, which is a marker of most NENs. The concentration of chromogranin B (CgB) is sometimes increased when the concentration of CgA is within the reference values [12, 14, 18].

NF-PanNENs can also secrete other substances, including PP; however, the percentage of patients with elevated PP levels is clearly lower than the percentage of patients with elevated CgA levels [18]. In the biochemical diagnostics of NF-PanNEN, NSE is also used, the sensitivity of which in NEN G1 and NEN G2 is 19% and 54%, respectively. NSE is mainly determined in pancreatic neuroendocrine carcinomas (PanNEC) if the CgA concentration is normal [12, 14].

3.1.4. Pancreatic neuroendocrine carcinomas

The determination of circulating CgA (serum, plasma) concentration and other markers of this group of PanNENs usually gives negative results. PanNECs lose their secretory function over time. NSE can be considered a marker of these tumours [12, 14]. Its sensitivity is approximately 62–63%, and it is an independent prognostic factor for PanNEC [12, 14, 34].

Minimal consensus statement on biochemical tests in PanNEN diagnostics:

3.2. Pathomorphological diagnosis

3.2.1. Histopathological classification

The World Health Organization (WHO) classification of NEN published in 2017 was the first to introduce the division of highly differentiated PanNETs into three grades of histological maturity (high/medium/low): PanNETG1/G2/G3 [35]. This division was the result of clinical observations which showed that the NEC, a category of neoplasms with the proliferation index (Ki-67) above 20%, is heterogeneous, including tumours with both highly and poorly differentiated histopathological tissue [36, 37]. Based on the results of the conducted research, these neoplasms were divided into two categories, highly differentiated ones, most often with the Ki-67 proliferation index between 20 and 50% (PanNET G3) and highly malignant NECs with the proliferation index above 20%, but usually over 50%. Neuroendocrine carcinomas have been defined as a separate group of poorly differentiated and low-grade neoplasms divided into large and small cell carcinomas [38, 39].

In 2019, the 5th edition of the WHO gastrointestinal system classification was issued [40, 41]. For the first time, it included the classification of NENs of individual organs of the gastrointestinal system. Its main feature, unique in terms of histopathological oncological terminology, is the standardised division of these neoplasms in each organ into WD-NET and poorly differentiated NEC. In addition, the NETs, like the PanNETs beforehand, were classified into three subtypes based on the grade of their histological maturity; NET G1 — high grade, NET G2 — intermediate grade, and NET G3 — low grade. The criteria of this classification are the number of figures of division and the proliferation index (Ki-67).

The WHO 2019 classification precisely defines the category of mixed neoplasms by introducing the term MiNEN (mixed neuroendocrine — non-neuroendocrine neoplasms). In pancreatic neoplasms, three categories of MiNEN have been distinguished, including acinar NET, ductal NET, and NEC.

The organ distinctiveness of pancreatic NEN compared to NEN developing in other organs of the gastrointestinal system is due to the cell type and function or the lack of endocrine function of tumours, which constitute the basis for their division into two categories. Regarding functional neoplasms, nomenclature based on their clinical symptoms is also used, such as insulinoma, gastrinoma, VIPoma, glucagonoma, somatostatinoma, ACTH-producing tumour, enterochromaffin-cell carcinoid, and serotonin-producing tumour.

An additional characteristic feature of pancreatic NET is the microscopic variants related to the type of neoplastic cells distinguished in the 2019 WHO classification: oncocytic, pleomorphic, and clear cell as well as the cystic variant. The WHO classification of non-functional and functional pancreatic neoplasms is presented in Table 4 [42].

3.2.2. Prognostic factors in non-functional pancreatic neuroendocrine tumours (PanNETs)

The assessment of prognostic factors in PanNENs plays a key role in disease course prediction and treatment planning. Among numerous prognostic parameters, the proliferation index is the most important microscopic feature that enables patient division into groups that constitute the basis for establishing prognosis and selecting the treatment method. Tumours with Ki-67 below 20% divided into PanNET G1 (up to 3%) and PanNET G2 (from 3 to 20%) are characterised by better prognosis than PanNET G3. It has been established that the five-year survival in the case of PanNET G2 is 62%, and in PanNET G3 — 29%. On the other hand, patients with small or large cell neuroendocrine carcinoma, with Ki-67 above 20%, usually close to 100%, have poor prognosis, and their five-year survival is 16%.

The size of the PanNET and the extent of the infiltration constitute an important prognostic factor. The tumour feature (pT) tumour feature according to the American Joint Committee on Cancer/Union for International Cancer Control (AJCC/UICC) pathological tumour–node–metastasis (pTNM) classification determines the cancer stage [43]. Microadenomas below 0.5 cm in diameter are considered benign and are not included in the staging system. However, it is impossible to predict if and when they may progress to PanNET. On the other hand, tumours of less than 2 cm in diameter, referred to as low-grade PanNET, usually have an indolent course. The risk of metastasis is associated with a diameter greater than 3 cm. Approximately 55–75% of PanNETs demonstrate an aggressive course with extra-organ infiltration, metastasis, and/or recurrence. After surgery, five-year survival is 65–86%, and 10-year survival is 45–68%. In contrast, the five-year survival of patients with metastases is 59%, and the 10-year survival is reduced to 36%. Table 5 presents the 8th edition of AJCC/UICC pTNM classification of 2017 [43].

Pathomorphological diagnosis includes the evaluation of microscopic and immunohistochemical factors, assessed obligatorily and conditionally. They are presented in Table 6.

Targeted molecular sequencing demonstrated the presence of mutations in the following genes: TP53, KRAS, PIK3CA/PTEN, and BRAF. The gene mutations found in PanNETs (MEN-1, DAXX, ATRX) and the mutations characteristic of ductal carcinomas (SMAD4/DPC4) are rarely found in PanNECs. Mutations in the TP53, BRAF, PIK3CA, PTEN, WNT, and CTNNB1 genes are used in the targeted cancer treatment [42].

Minimal consensus statement on pathomorphological examinations

3.3. Endoscopic diagnostics

Classic endoscopy of the gastrointestinal tract is of limited importance in the diagnosis of PanNETs [12–14].

The presence of multiple ulcers of atypical location (distal duodenum or jejunum) during the examination of the upper gastrointestinal tract, the coexistence of severe oesophagitis, and recurrence of ulcers after pharmacological or surgical treatment suggest that diagnostics towards gastrinoma is advisable.

In rare cases, PanNETs, both functional and non-functional, may cause stenosis of the biliary tree, and in such cases, after exhausting the noninvasive diagnostic methods, an attempt to obtain material for histopathological examination during biliary stenting as part of endoscopic retrograde pancreato-cholangiography may be done.

3.4. Ultrasound examinations

3.4.1. Transabdominal ultrasound

The sensitivity of conventional ultrasound (US), most often performed as first-line examinations in the detection of primary lesions and in the assessment of the stage of the disease, is low for small tumours. On average, US detects about 39% of primary lesions, such as insulinomas and gastrinomas [44]. The sensitivity of this method is higher in the detection of metastatic lesions to the liver, reaching 82–88%, and its specificity is 92–95%. For larger tumours, most often non-functional pancreatic tumours and late diagnosed glucagonoma, the sensitivity of transabdominal ultrasound is higher [12, 14, 45].

Currently, third-generation contrast agents are used in ultrasound diagnostics, which are gas microbubbles surrounded by phospholipids, with a long half-life in the bloodstream and perfusion-dependent greyscale enhancement. Contrast-enhanced ultrasound (CEUS) is particularly useful for the assessment of neuroendocrine tumours in both the pancreas and their metastases in the liver due to the very good vascularisation of these lesions, which makes it possible to distinguish PanNET from adenocarcinoma [46, 47]. Both PanNET lesions and PanNET metastases to the liver usually demonstrate a homogeneously hypervascular type of enhancement in the arterial phase of the examination in 90.7% and 83.9%, respectively, before treatment initiation [46]. As the contrast medium is washed out, echogenicity decreases rapidly during the venous phase of the examination [46, 48]. A correlation has been found between the homogeneously hypervascularised enhancement pattern and low proliferative activity and thus low tumour grade [46]. The CEUS examination is very well suited to monitoring the response to treatment aimed at reducing the degree of vascularisation of the lesion, which makes it possible to extend the time between follow-up CT examinations of the abdominal cavity [46, 49]. CEUS detects tumours of less than 2 cm in diameter with sensitivity similar to that of endoscopic ultrasound (EUS) (95%). The method sensitivity in PanNENs is 94%, specificity — 96%, positive predictive value — 75%, and negative predictive value — 99%. The CEUS examination is characterised by 82% sensitivity in the recognition and assessment of the morphology of focal lesions in the liver [46].

3.4.2. Endoscopic ultrasound

EUS – despite the fact that compared to computed tomography and magnetic resonance imaging (MRI) it is much more dependent on the operator’s experience – is currently considered to be the most sensitive method of detecting pancreatic neuroendocrine tumours, especially small ones with diameter < 2 cm (even 1–2 mm). The method sensitivity and specificity are estimated at 57–94% and 98.0%, respectively, with a likelihood ratio for a positive result of 11.1 [95% confidence interval (CI): 5.34–22.8] and 0.17 (95% CI: 0.13–024) for a negative result, which means that a normal result of EUS examination practically excludes the presence of a pancreatic tumour [50].

For comparison, computed tomography demonstrates sensitivity of 100% for lesions of > 20 mm in size, but it fails in 15% of lesions with a diameter of ≤ 20 mm and in 68.4% of < 10 mm in diameter, which means that the average sensitivity, similarly to magnetic resonance imaging, is 73% (39–94%) [51].

In small insulin-secreting tumours, the sensitivity of EUS is 94–100% [52]. Also, in pancreatic gastrinomas, the sensitivity is close to 100%, but it decreases in multifocal and extra-pancreatic lesions; in the case of gastrin-secreting tumours located in the duodenum and outside the pancreatic parenchyma the sensitivity of the examination is estimated at 50%.

In the diagnostics of pancreatic tumours, curved linear array echoendoscopes are preferred because they allow for visual and therapeutic diagnostic procedures as well. The optimal images of the isthmus, body, and tail of the pancreas as well as the celiac trunk, portal confluence, and splenic vessels are obtained through the gastric wall, head of the pancreas, portal vein, and bile ducts — through the wall of the duodenal bulb; and the uncinate process — from the descending part of the duodenum.

EUS can obtain a high-resolution image of the pancreatic-biliary area without disturbances caused by the presence of intestinal gases or subcutaneous fat, which affect the results of transabdominal ultrasound [53].

EUS enables the assessment of vascular invasion, i.e. the assessment of resectability and the assessment of the distance between the tumour and the main pancreatic duct, which affect, among others, the therapeutic decisions (enucleation vs. resection). Large lesions of the PanNET type may cause compression on the main pancreatic duct, causing abrupt-type stricture with upstream dilatation and a normal-calibre duct downstream of the stricture visible in the EUS (widening of the pancreatic duct and its infiltration indicate a NET G2/NET G3), but more often the stenosis results from fibrosis secondary to serotonin production [54].

The differentiation of pancreatic tumours, especially in the case of doubts regarding the differentiation from adenocarcinoma, but also, e.g., intra-pancreatic spleen, is facilitated by the colour Doppler EUS and contrast-enhanced harmonic EUS (CH-EUS).

PanNET, unlike ductal carcinoma, is characterised by rich vascularisation. Colour Doppler EUS enables assessment of the vascularisation of solid pancreatic tumours (although the method does not allow for the assessment of small vessels with slow blood flow), and the use of ultrasound contrast agents (SonoVue or Sonazoid), especially in the harmonic method, allows for their differentiation [55–57].

EUS elastography enables the assessment of tissue elasticity (“hardness”) and can be a good method complementary to fine-needle biopsy in differentiating malignant and benign focal pancreatic lesions. The colour-coded dynamic EUS elastography (where red means soft tissue, blue — hard, stiff tissue typical for malignant changes, and green — a structure of intermediate hardness) is 99% sensitive and 74% specific [58].

Another EUS-based diagnostic method is needle-based confocal laser endomicroscopy (nCLE-EUS) — mainly useful in differentiating cystic lesions with PanNET detection sensitivity of 100%, specificity of 80–100%, and negative predictive value of 100% [59].

Endoscopic ultrasound enables material acquisition for pathomorphological examination of solid lesions and the content of cystic lesions for serological evaluation and marker determination.

Material for pathomorphological examination under the control of EUS can be obtained by means of the following:

Complications related to the material acquisition under EUS control are rare, and no differences in the frequency of their occurrence have been demonstrated between EUS-FNA and EUS-FNB. The most common ones include temporary abdominal or chest pain; however, in 1% of patients, acute pancreatitis is observed after puncture of the pancreas. Bleeding into the lumen of the gastrointestinal tract or beyond it (0.69%), infection (0.44%), and perforation (0.21%) are less common. To minimise these complications, general rules apply regarding the modification of anticoagulant and antiplatelet therapy before the biopsy is performed (EUS biopsy is classified as an endoscopic procedure with a high risk of bleeding), and in the case of cystic lesion puncture — antibiotic prophylaxis [62].

Diagnostic EUS enables the following:

There are reports confirming the safety of watch-and-wait strategies in the case of small non-secreting PanNETs of < 2 cm in diameter [2, 63].

It is worth emphasising that if such a lesion is clearly visible in transabdominal ultrasound, there is no need for EUS monitoring.

EUS techniques include EUS-fine-needle tattooing (EUS-FNT) of lesions or EUS-guided fiducial implantation (EUS-F) into small lesions before a scheduled surgery, which facilitates quicker intraoperative location of the lesions. The use of this method is important especially in the case of laparoscopic procedures, which do not allow palpation control. In addition, the precise location of the lesion makes it possible to obtain an adequate resection margin and to maintain healthy pancreatic tissue. However, it should be emphasised that tattooing may cause acute pancreatitis [64, 65].

EUS is particularly indicated in MEN-1 and von Hippel-Lindau syndrome. The incidence of neuroendocrine pancreatic lesions in these patients is estimated at 40–80%. Hormonally functional tumours can be diagnosed earlier due to typical clinical and biochemical symptoms; however, in most patients, hormonally non-functional tumours (approx. 50% of the lesions) are diagnosed late, which is associated with poor prognosis. Despite the controversies regarding the optimal monitoring method and the intervals between examinations, EUS or MRI are recommended as methods of monitoring every 6–12 months, because early detection of pancreatic lesions allows the application of radical treatment [66–69].

3.4.3. Intraductal endoscopic ultrasound

Intraductal ultrasound (IDUS) can be more effective in detecting PanNENs than EUS. In this method, a 2 mm probe is inserted into the main pancreatic duct (Wirsung duct) through the duodenoscope channel [12, 14].

3.4.4. Intraoperative ultrasound

Intraoperative ultrasound (IOUS) has sensitivity similar to EUS in the detection of small PanNENs. The accuracy of the examination in combination with the intraoperative palpation assessment is up to 97%. In the case of gastrinoma, the precision of the examination within the pancreas is close to 100%, while in the case of extra-pancreatic locations, it drops to 58%. Intraoperative ultrasound also allows the detection of multifocal tumours and metastatic lesions within the liver and the assessment of the distance of the tumour, especially a small one, from the pancreatic duct, in order to properly qualify the patient for resection or enucleation of the lesion [70]. IOUS is also performed during laparoscopy [12, 14].

3.5. Computed tomography and magnetic resonance imaging

In accordance with the current guidelines, the diagnostics of pancreatic pathology uses multi-phase examinations with the application of multidetector computed tomography (MDCT) and MRI. These methods are primarily important in assessing the stage of neoplastic disease and in monitoring the response to therapy [12, 14]. They are also useful in assessing the anatomical location and resectability of the primary lesion [12, 14]. The sensitivity of individual imaging methods depends on the tumour type and location [12, 14].

3.5.1. Computed tomography

Computed tomography (CT) is the most popular method of imaging pancreatic pathologies including neuroendocrine neoplasms. This is due to its wide availability, short examination time, repeatability of examinations, and high spatial resolution of the method, allowing for the performance of the multidimensional reconstructions necessary for the assessment of resectability of the lesion and the planning of a surgical procedure [12, 14]. The protocol for multiphase computed tomography (CT) examination of the pancreas is presented in Table 7.

Before the administration of the contrast agent, F-PanNENs are usually isodense, and rarely hypodense, in relation to the remaining pancreatic parenchyma, and calcifications are perfectly visible. Most tumours are highly vascularised (80% in the case of insulinoma); therefore, they are intensively enhanced in MDCT in the arterial phase. Metastases behave similarly. Therefore, in the arterial phase, the MDCT examination should include both the pancreas and the liver. At this stage of the examination, it is also possible to assess the tumour-visceral artery relationship and to evaluate possible variants of arterial vascularisation of the pancreatic field, which may be of key importance in planning surgical treatment [2]. In the parenchymal phase, the assessment is limited to the pancreas and concerns the tumour morphology and the degree of contrast washout. The portal venous phase involves the pancreas again, along with the liver and the vessels of the portal system [2, 12, 14, 71]. Some researchers suggest also performing the test in the late phase, 150–180 seconds after contrast administration, to further assess the degree of contrast washout from the tumour [12, 14, 72]. Contrast stasis in PanNET lasting over 150 seconds is associated with a more aggressive course of the neoplastic disease, as well as mild enhancement in the arterial phase [73, 74]. If there is a clear image of rapid, intense enhancement in the arterial phase, the late phase may be omitted. In typical neuroendocrine tumours, the contrast enhancement should be at least 60 IU lower in the late phase compared to the arterial phase. Other types of enhancement in PanNENs include uneven contrast washout (with over or below 1/2 of the tumour mass), or slowly growing enhancement when the tumour is more clearly visible in the equilibrium phase, where the healthy pancreatic parenchyma reduces absorption. This is a typical behaviour of tumours containing a lot of connective tissue. Thus, in the parenchymal and late phases, neuroendocrine tumours are not always isodense and are therefore imperceptible on CT examination. Some tumours maintain their enhancement in these phases of the examination or are only just starting the process of contrast accumulation. Tumours that exhibit mild enhancement, especially in the arterial phase, are usually poorly differentiated, so the degree of enhancement correlates with the patient survival [12, 14, 73, 75–78]. Higher enhancement is associated with better developed vascularisation found in histopathological examination and is associated with a lesser stage of tumour advancement [74, 78, 79].

NF-PanNETs are characterised with lower enhancement after contrast agent administration and are heterogeneous due to areas of necrosis. In adenocarcinoma, calcification is very rare, while in hormonally functional and non-functional PanNEN it can be observed in at least 25% of cases. Larger tumours exhibit dilatation of the pancreatic duct and atrophy of the parenchyma, which is more common in NET G3 and NEC [78, 80]. The features that clearly differentiate malignant lesions from benign ones include infiltration of adjacent structures and distant metastases. Hepatic metastases are detected in the arterial phase of the examination as hyperdense areas, while in the portal phase the contrast is washed and they are hypodense [12, 14]. Metastases are more common in heterogeneous tumours with unequal enhancement and vein infiltration. Mild tumour enhancement in the venous phase is an unfavourable prognostic factor [73, 78, 80].

Due to the shorter scanning time, reduction of the number of motion artifacts, and obtaining thin (1–2 mm) sections, MDCT allows for multiplanar and spatial reconstructions, which facilitate the imaging of structures below 1 cm in size and enable full assessment of vessel infiltration by the tumour [76, 77]. The sensitivity of MDCT with contrast enhancement and the use of 1 mm layers in the diagnosis of insulinoma reaches 85–94%, while for various types of NEN the sensitivity of multi-row CT is 61–93% and the specificity is 71–100% [12-14, 16, 81]. Thin-layer CT also enables a very accurate assessment of the tumour morphology, including its margins — a poorly defined tumour border is an independent predictor that helps distinguish between NET G3 and NET G1/G2 [80].

In everyday practice, the role of CT in the assessment of PanNEN is to describe the tumour morphology with precise determination of its location. However, with regard to the infiltration exceeding the organ boundaries, the involvement of the adjacent adipose tissue, duodenum, common bile duct, stomach, spleen, intestinal loops, adrenal glands, and the extent of infiltration of arteries and veins should be assessed with information on the part of the vessel and the length of the segment affected by the infiltration. The description should also include information about enlarged local lymph nodes and the assessment of the liver for the presence of metastases. Based on the CT description, it should be possible to classify the tumour using the pTNM classification [12, 14, 15]. The description should also include the assessment of tumour resectability according to National Comprehensive Cancer Network (NCCN) criteria [12, 14, 82].

Work is currently underway on the use of artificial intelligence to assess various PanNET features in both CT and MRI, which would make it possible to better objectify lesion assessment and prognosis of the degree of lesion malignancy [79, 83]. The study of the role of the extracellular volume fraction (EVF) in the prognosis of the degree of malignancy in PanNETs is also an attempt to objectify the tumour description. It was found that there is a statistically significant difference in EVF values between NET G1 and NEC [84].

3.5.2. Magnetic resonance imaging

The sensitivity of MRI using the optimal examination protocol in PanNEN diagnostics is similar to that of CT, reaching 79% (54–100%) [13]. On the other hand, MRI is better than CT in terms of the diagnosis of liver metastases, the sensitivity of MRI is 75% (from 70 to 80%), and the maximum level of specificity approaches 98% [13]. Similarly, MRI exceeds CT in the assessment of metastasis to the brain and bone [13]. MRI offers higher tissue resolution in combination with multiplanar imaging. The method limitations include lower availability compared to CT, higher price, longer examination time, and the necessity of patient cooperation. The method is recommended especially in younger patients (because it does not require the use of ionising radiation), and in those for whom CT provided an ambiguous result. NETs are hypointense on T1-weighted images and hyperintense on T2-weighted images. Intravenous administration of contrast agent increases the method sensitivity [12, 14, 71]. In a multiphase examination, after intravenous administration of contrast, enhancement is observed in accordance with the above-mentioned CT image enhancement pattern.

The administration of hepatotropic contrast is recommended to assess metastatic lesions in the liver (Tab. 8) [2, 85].

To increase the spatial resolution after the administration of contrast medium (CM), data acquisition should be performed using thin layer (1–2 mm), 3D, T1 sequence, and fat saturation (FS). In MRI spectroscopy, which uses the chemical shift phenomenon, it is possible to determine the chemical composition of tissues. Relatively increased lipid content in NEN facilitates differentiation in doubtful situations.

In recent years, diffusion-weighted imaging (DWI) has been introduced. NETs, especially those with a high connective tissue content, limit the diffusion of water molecules, thus generating an intense signal in the DWI sequence with accompanying apparent diffusion coefficient (ADC) reduction [12, 14, 86].

It is promising to observe that better differentiated PanNETs represent a higher value on ADC maps than PanNECs [12, 14, 85]. The DWI sequence is a highly valuable tool in the assessment of possible metastases in the liver and lymph nodes and their differentiation [2].

3.6. Radioisotope diagnostics

Somatostatin receptor imaging (SRI) in PanNEN is used as the method of choice in the initial clinical staging, early relapse diagnosis, disease monitoring, and therapeutic decision-making. In a meta-analysis, Calabro et al. (18 publications) assessed the sensitivity and specificity of SRI in patients with PanNET at 79.6% (95% CI: 71–87%) and 95% (95% CI: 75–100%), respectively, and the primary site detection rate at 81% [87]. Mapelli et al. analysed the results of two-path molecular imaging ([68Ga]Ga-DOTA-DPhe1-Tyr3-Octreotide ([68Ga]Ga-DOTATOC) and 18F-fluorodeoxyglucose Positron Emission Tomography ([18F]FDG PET)) in PanNENs patients who were candidates for surgery, with the aim of determining possible indicators of image data evaluation for preoperative analysis of more aggressive phenotypes [88]. In their work, they performed semi-quantitative analysis of the maximum standardised uptake value (SUVmax) and the mean for both [68Ga]Ga-DOTATOC PET and [18F]FDG PET, the somatostatin receptor density (SRD) and the total lesion somatostatin receptor density (TLSRD) for [68Ga]Ga-DOTATOC PET scans, and metabolic tumour volume (MTV) and total tumour glycolysis (TLG) for [18F]FDG PET scans. In the study, SRI, SRD, and TLSRD were the only indicators significantly related to tumour size [pT3 or pT4 vs. pT1 or pT2; cut-off value 18.34 (p = 0.0002) for SRD and 275.41 (p = 0.0031) for TLSRD; sensitivity and specificity, respectively: 78.1 and 72.5% for SRD and 75 and 66.7% for TLSRD].

PET/CT imaging with [18F]FDG is most often performed in the case of NET G3 and NET with Ki-67 ≥ 10% — detailed recommendations in the case of PanNET do not differ from the general principles of GEP-NETs management. Many recently published studies emphasise the value of simultaneous imaging with [18F]FDG and SRI in pancreatic NET G3 before planned radioisotope therapy with isotope-labelled somatostatin analogues (PRRT — peptide receptor radionuclide therapy), emphasising the usefulness of [18F]FDG PET/CT in predicting treatment outcomes [89, 90]. When the [18F]FDG PET result is positive, it indicates the need for a more aggressive treatment regimen. Patients with PanNETs G3 should be examined using both SRI and [18F]FDG PET because this would indicate which patients could benefit from PRRT, even as first-line treatment [91].

Similarly to NECs in other locations, as well as in the case of rapidly progressing PanNETs or in tumours without somatostatin receptor expression, it is justified to extend the diagnostics to metabolic imaging with [18F]FDG [92].

When selecting the imaging method, especially in patients in whom the applied treatment methods do not bring the expected results, or in the case of unexpected disease progression in the course of currently used therapy, it should be kept in mind that some of the disease foci may undergo dedifferentiation, and it is also worth applying PET with labelled glucose in these clinical situations.

Small tumours, including insulinoma, as well as, inter alia, rare tumours, e.g. with ectopic ACTH secretion, still constitute a challenge in PanNET diagnosis. When available, mainly in specialised centres, molecular imaging with [18F] F-DOPA can be applied using labelled analogue for the GLP-1 receptor [11C]C-methionine [93]. Imaging with [18F] F-DOPA, in accordance with the 2017 European Association of Nuclear Medicine (EANM) guidelines, is not recommended in PanNET due to the potential risk of false positives in non-neuroendocrine pancreatic neoplasms, such as pancreatic pseudopapillary tumour, and pancreatic serous cystadenoma [94, 95].

In the case of difficulties in visualising the foci of a benign insulin-secreting tumour, depending on method availability, there are indications for imaging with the use of a labelled GLP-1 analogue [96].

Brand et al. developed a bimodal probe (PET/fluorescence) for imaging examination of the expression of GLP-1 receptor (GLP-1R) in the pancreas and in tumours originating from pancreatic islets. It may enable pre- and intraoperative assessment of GLP-1R expression in tumour tissues to facilitate differential diagnosis of noninsulinoma pancreatogenous hypoglycaemia (NIPHS) and intraoperative real-time detection of tumour margins, neoplastic infiltration, or residual neoplastic cells in the operating bed [97].

Minimal consensus statement on imaging and radioisotope examinations

4. Treatment

4.1. Surgical treatment

4.1.1. Surgical treatment of well-differentiated pancreatic neuroendocrine tumour G1/G2 (PanNET G1/PanNET G2)

A. General recomendations

The decision regarding PanNET treatment method should be made by a multidisciplinary team of physicians experienced in the treatment of neuroendocrine tumours [13].

Sporadic PanNET

Most PanNETs of ≤ 2 cm in diameter are benign and characterised by a moderate risk of malignancy. In selected cases, tumours ≤ 2 cm in diameter can be observed. This applies to asymptomatic tumours, NF-PanNETs G1, and NF-PanNETs G2 with a low proliferation index (Ki-67), especially those located in the head of the pancreas, in the absence of radiological suspicion of malignancy, and taking into account the individual characteristics of the patient (age, patient’s decision, concomitant diseases) [13, 18, 63, 98, 99]. Recommendations of some scientific societies and the results presented in some publications recommend following up tumours of up to 1 cm, leaving the decisions about 1–2 cm tumours to the therapeutic team and the patient [2, 100, 101].

Tumours > 2 cm in size require resection [13, 18, 98, 99].

Symptomatic and functional tumours, regardless of their size, most often require resection [2, 13, 98, 102].

PanNET concomitant with MEN-1 syndrome

Asymptomatic and NF-PanNETs concomitant with MEN-1 syndrome should be excised if they are > 2 cm in diameter, accompanied by metastases to lymph nodes (or have other radiological features of malignancy), and demonstrate rapid growth (annual size increase > 0.5 cm) [13, 18, 102].

NF-PanNETs G1 and NF-PanNETs G2 with low Ki-67 and smaller than 2 cm can be observed [13, 18, 102]. The presence of numerous nodules sometimes requires resection of the entire organ.

Functional tumours usually require resection, but this may differ in the case of gastrinoma [13, 103].

Surgical procedures

The method of surgical treatment of PanNETs depends on the degree of tumour malignancy, its size, location, infiltration of adjacent organs, the presence of distant metastases, the patient’s general condition, and the ability to control clinical symptoms. Patients are eligible for either radical or palliative treatment just to improve the patient’s quality of life. In the case of tumours located within the head of the pancreas, pancreatoduodenectomy is performed, while in the case of tumours located in the body and tail, it is distal resection with or without removal of the spleen. Segmental infiltration of the veins (portal vein and superior mesenteric vein) should not be treated as an absolute contraindication for PanNET resection. In a selected group of patients, procedures involving the resection and reconstruction of the vessel as well as multi-organ resection should be considered [2, 13, 102].

In selected cases of small and well-defined PanNETs, atypical resections can be performed including enucleation and resection of the middle segment, also with the use of the laparoscopic method [2]. Middle segment resection is most often performed in the case of small lesions located in the body of the pancreas. The prerequisite for enucleation is the benign nature of the lesion and the maintenance of > 2–3 mm distance between the tumour and the duct of Wirsung [2]. The size of the tumour is usually limited to 2 cm. Enucleation of the lesion carries the risk of damaging or closing the duct of Wirsung and the occurrence of related complications in the form of pancreatic fistula and acute postoperative pancreatitis. Performing resection of a large fragment of the pancreas may, in addition to the above-mentioned complications, cause symptoms of failure of the exocrine and endocrine function of the pancreas [104].

Because pancreatic tumours are often malignant, regional lymph nodes (12–15 nodes) must be removed during extensive resection [2, 105, 106]. In the case of enucleation and resection of the middle segment, it is recommended that one/several lymph nodes are collected for examination [2]. In the case of PanNET, intraoperative ultrasound is recommended.

Currently, pancreatic resections (especially distal resections and enucleation) are increasingly performed using the laparoscopic method or robot-assisted surgery; however, the decision of whether an open surgery or minimally invasive procedure is used should be made by a surgeon specialising in surgical treatment of the pancreas at a reference centre [13, 102, 107, 108].

In selected cases of PanNETs G1 and PanNETs G2, cytoreductive procedures are performed towards both the primary tumour and liver metastases [2]. It is believed that advanced PanNETs G3 and NECs should not be operated on if disseminated metastases are confirmed, in such cases cytoreductive procedures are not recommended either [2, 13].

B. Functional tumours

Insulinomas and gastrinomas are the most common hormonally functional tumours. Others are rare functioning pancreatic neuroendocrine tumours (RF-PanNETs). F-PanNETs require surgical resection regardless of their size. This differs in the case of gastrinoma in MEN-1 syndrome.

Gastrinoma, Zollinger-Ellison syndrome

Gastrinomas are most often located in the head of the pancreas, and 60–90% are malignant lesions associated with lymph node involvement [2]. Radical removal of these lesions is recommended. The range of procedures depends on the location and size of the tumour. They include enucleation, mid-segment resection, distal resection, and pancreatoduodenectomy with lymphadenectomy [2, 12, 14, 103].

The role of surgical treatment in patients diagnosed with ZES but having negative results of preoperative imaging examinations (lack of tumour identification and location) remains debatable [2, 18]. Exploratory surgery may be considered in patients with ZES, who do not have medical contraindications to the procedure. It should be performed by a surgeon experienced in the treatment of NET to search for the primary tumour [12, 14, 18]. Because gastrinomas are often multiple and occur in the duodenum, a thorough examination of the abdominal cavity, especially the pancreatic-duodenal area, using intraoperative ultrasound is recommended during surgery [2, 13, 18]. Currently, routine duodenotomy is not recommended [2].

Surgical treatment is not routinely recommended in ZES syndrome, a component of MEN-1 syndrome, in the case of tumour/tumours with a diameter ≤ 2 cm.

In patients with multiple small gastrinomas, only major surgery (e.g. pancreatoduodenectomy or total pancreatic resection) can ensure radical treatment. However, these procedures are not routinely recommended due to complications, good long-term prognosis in patients with tumours ≤ 2 cm in diameter, and the fact that excessive secretion of gastric juice can be efficiently controlled by pharmacological methods in patients with MEN-1/ZES [2, 13, 18].

Surgical treatment is recommended when the tumour is > 2 cm in size [2, 13, 103]. According to some authors, pancreatoduodenectomy is the recommended procedure in the case of tumours located in the head of the pancreas, because smaller-scale procedures are associated with the risk of recurrence [13].

Insulinoma

Insulinomas are benign in 85–95% of cases [2]. Due to the symptoms, surgical treatment is recommended. Tumour enucleation is possible in most cases. If it is not (anatomical conditions, tumour size, suspicion of a malignant lesion), a more extensive resection is required [2, 102, 109]. If possible, a lymph node or nodes should be collected during the procedure for diagnostic assessment. The procedure can be performed laparoscopically [2, 102]. In the absence of tumour identification and determination of its location through imaging, routine surgical exploration is not recommended [2].

In selected cases of insulinoma, in patients with a sporadic tumour or with MEN-1 syndrome, percutaneous and EUS-controlled ablation with ethanol administration or radiofrequency ablation (RFA) can be used [110–112].

Rare and very rare functional tumours

Radical treatment is recommended in this group of tumours, and the extent of resection and removal of lymph nodes corresponds to the procedures performed in the case of gastrinoma.

C. Non-functional tumours

Treatment of asymptomatic sporadic NF-PanNETs depends on the tumour size [13]. In the case of NETs G1 and NETs G2 ≤ 2 cm with low Ki-67 (< 10%), lesion follow-up or conservative treatment may be considered, especially in cases where, due to their location, they would require extensive resection [13, 18, 113]. The exceptions include tumours with invasive features (e.g. dilatation of the main pancreatic duct, suspicion of lymph node metastases). Because some tumours < 2 cm in size are characterised by lymph node involvement, it seems advisable to collect a node/several lymph nodes for examination, even using parenchyma-sparing procedures [13, 114].

The recommendations of some scientific societies and the results presented in some publications recommend following up tumours of up to 1 cm, leaving the decisions about 1–2 cm tumours to the therapeutic team and the patient [2, 100, 101].

In the case of NF-PanNETs in MEN-1 syndrome, tumours are usually stable and slow growing, therefore the “watch-and-wait” approach can be used when the tumour size is ≤ 2 cm [13]. In the case of PanNETs > 2 cm, tumour resection is recommended. Due to the high risk of nodal metastases, it is most often necessary to perform pancreatectomy (pancreatoduodenectomy or distal resection) with regional lymphadenectomy [13, 101, 102].

D. Advanced PanNETs/metastatic disease

In the case of advanced PanNET, the purpose of resection is to reduce the tumour mass, and in the case of F-PanNET, additionally, to reduce the severity of symptoms related to hormonal activity [13].

Cytoreduction of the tumour mass is being debated. It is considered in patients with NETs G1 and NETs G2 (Ki-67 < 10%) without metastases outside the abdominal cavity [13]. Cytoreduction is considered when more than 90% of the tumour mass can be removed, even in the presence of liver metastases [13, 115].

Cytoreductive surgery can be performed with resection and/or ablation of liver lesions [116]. RFA is applicable when there are less than 10 focal lesions in the liver, and the largest of them is less than 5 cm in diameter (optimally 3 cm). Cytoreduction of liver lesions is considered advisable when the tumour mass can be reduced by min. 70% (especially in patients with F-PanNET or carcinoid syndrome) [2, 116–118].

Radical excision of liver metastases is the gold standard in the treatment of advanced PanNET, so whenever resection is possible, it should be performed [116]. The method of resection depends on the patient’s general condition, size, location, and number of lesions, and includes enucleation, hepatic wedge resection, segmental resection, non-anatomical resection, or hemihepatectomy. In selected cases, the importance of simultaneous surgical resection of primary PanNET with resection/cytoreduction of liver metastases is emphasised [2, 116, 119].

In the case of unresectable metastatic lesions in the liver, cholecystectomy should be performed in order to avoid ischaemic complications in the gallbladder as a result of possible application of (chemo)embolization. Other methods of treating metastases are locoregional therapies (various ablation methods, embolization) and liver transplantation. It is assumed that transplantation is performed in a selected group of patients with severe symptoms related to the production of hormones. The eligibility criteria for transplantation are as follows: no extrahepatic disease, histological confirmation of a well-differentiated NET (NET G1/ NET G2, Ki-67 < 10%), earlier removal of the primary tumour, metastases involving < 50% of the liver volume, stable disease for at least 6 months, and age < 60 years [13, 102]. The general condition of the patient determines the qualification for procedure.

In patients for whom resection of liver metastases is not possible, palliative treatment is recommended in the form of selective hepatic artery embolization (HAE), transarterial chemoembolization (TACE), or radioisotope embolization [transarterial radioembolization (TARE), selective internal radiation therapy (SIRT)] [13, 116]. RFA, cryoablation, and microwave ablation (MWA) as well as irreversible electroporation (IRE) are used in tumours ≤ 5 cm in size [13, 116].

In exceptional cases, the possibility of resection of the stage IV primary tumour (with the presence of unresectable liver metastases) is considered [2, 102, 120, 121].

Due to the small number of studies, no clear recommendations have been issued for patients with PanNETs of unknown primary site and liver metastases, which would define the advisability of surgery in search of a primary tumour with simultaneous cytoreduction of metastases [122, 123].

There is no consensus on the resection of peritoneal metastases [124, 125]. Resection should be considered in a specialist centre with the use of multi-directional treatment [126]. It is recommended that the accessible intraperitoneal metastatic lesions be removed. Currently, there are no recommendations regarding the use of intraperitoneal chemotherapy and intraoperative hyperthermic intraperitoneal chemotherapy (HIPEC) in conjunction with surgery in the treatment of peritoneal metastases of PanNET [127].

Palliative surgical treatment

If tumour resection is impossible, palliative surgery is used, which may affect the quality of life. In the event of mechanical jaundice, it is recommended that biliary-intestinal anastomosis or bile duct drainage is performed. When an unresectable pancreatic tumour obstructs the passage of food through the duodenum, bypass is recommended, usually by means of gastrointestinal anastomosis.

4.1.2. Surgical treatment of pancreatic NETs G3 and NECs

PanNETs G3 and NECs are very rare neoplasms, accounting for less than 1% of gastrointestinal malignancies. At diagnosis, metastases are observed in up to 85% of cases [26, 128]. Most of the principles of oncological management regarding pancreatic adenocarcinomas apply to NETs G3 with high Ki-67 and NECs.

In patients with advanced and disseminated pancreatic NECs, due to the high aggressiveness of this cancer, surgery is not recommended [2, 13].

In patients with advanced and disseminated NET G3, resection should be considered as part of combination therapy when R0 resection is possible. However, cytoreduction of liver metastases is not recommended [2].

According to some authors, locoregional therapy of aggressive PanNET G3 with liver metastases, including thermal ablation methods and hepatic intraarterial therapies, improves treatment outcomes [129, 130].

Minimal consensus statement on surgical treatment:

4.2. Endoscopic treatment

PanNET treatment is in the domain of surgery, although in recent years we have observed the development of procedures performed under the control of EUS as a less invasive locoregional therapeutic method compared to surgical procedures [131, 132].

The effectiveness and the occurrence of complications depends primarily on the characteristics of the neuroendocrine tumour (secreting vs. non-secreting), its size, as well as the degree of differentiation and the degree of histological maturity (Grade), affecting the choice of the treatment method.

Endoscopic pancreatic PanNET ablation performed under EUS control is an alternative treatment for patients with hormonally functional (insulin- and gastrin-secreting) and non-functional NETs G1/NETs G2 with a diameter ≤ 2 cm, who are not candidates for surgery or who refuse surgery. Due to the often aggressive course, VIP-, glucagon-, and ACTH-secreting PanNETs should be carefully qualified for ablative treatment.

The preferred method of ablation in patients with both functional and non-functional Pan-NETs is EUS-guided radiofrequency ablation (EUS-RFA). The RFA method uses the phenomenon of thermoablation, i.e. damage to tissues caused by heat generated by the flow of an alternating current with a frequency of 300–500 kHz (radiofrequency). The method is relatively safe provided that prophylactic antibiotic therapy is used, along with acute pancreatitis prevention by rectal administration of diclofenac and fluid aspiration from the cystic lesion. Tumours located less than 2 mm from the main pancreatic duct should not be treated as eligible for treatment. The incidence of complications is estimated at 3.5% [110, 112, 133–137].

Currently, there are two devices available to perform RFA of pancreatic lesions: the Habib EUS-guided RFA probe (EndoHPB, EMcision UK, London) and the EUSRA EUS-RFA system (Taewoong Medical Co., Gimposi, Geyonggi-do, South Korea). The results of ablation are promising. At 6 and 12 months after treatment, complete regression was reported in 64.2% and 85.7%, respectively, and this delayed effect can be explained by the occurrence of immune response to RFA [110]. After the ablation of functional Pan-NETs, EUS follow-up is recommended to confirm the lack of blood flow in the lesion.

Today, EUS-fine needle injection (EUS-FNI) of alcohol is used less commonly because it is associated with a higher incidence of complications. It can be performed with the injection of 95% ethanol, which leads to relief of symptoms in 94.7% of patients with small hormonally functional lesions such as insulinoma that demonstrate homogeneous enhancement [138–140].

The use of a 1:1 solution of 99% alcohol with lipiodol enables regression or stabilisation of NF-NETs G1/G2 with a diameter ≤ 2 cm, although therapeutic success is observed in only 50–62.5% of patients [141].

Complications of ablation occur in 0–20% of patients and include abdominal pain, fever, bleeding, and the development of acute pancreatitis.

Classic endoscopy can be used in the symptomatic treatment of advanced PanNET lesions with the following:

Minimal consensus statement on endoscopic treatment

4.3. Systemic treatment

4.3.1. Somatostatin analogues

Systemic therapy should be used to control tumour-related clinical symptoms (symptomatic treatment) and to control tumour growth (antiproliferative therapy). Somatostatin analogues (SSAs) (octreotide, lanreotide) reduce the secretion of hormones and biologically active substances and therefore reduce the severity of clinical symptoms, e.g. in patients with glucagonoma and VIPoma [3, 13]. In turn, the choice of antiproliferative therapy, including SSAs (PROMID, CLARINET studies) depends on the type of tumour, its pathomorphological features, size, growth dynamics, clinical symptoms, and the result of SRI [3, 13 147].

Some general principles should be followed in choosing systemic treatment. SSAs are recommended as the first-line therapy in patients with F-PanNEN, although there is little evidence confirming their effectiveness [3]. In patients with F-PanNEN G1 and F-PanNEN G2 qualified for surgical treatment (of local or locoregional disease), symptomatic treatment with SSAs is recommended in preoperative management. Overall, SSAs are also recommended as the first-line treatment to control tumour growth (antiproliferative therapy) in advanced or metastatic, slowly growing tumours (well-differentiated PanNENs G1, PanNENs G2, somatostatin receptor (SSTR) [+]) (both F-PanNEN and NF-PanNEN) with a Ki-67 value of up to 10% [3, 13]. The use of SSAs may be considered in PanNENs G3 in patients with strong SSTR expression in the receptor or immunohistochemical examinations, with a low stage and slow disease progression [3]. 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 [13]. However, the treatment effects of patients with SSTR (–) PanNENs G1/G2 are uncertain. Due to the relatively mild adverse effect profile of SSAs, their use in small SSTR (–) PanNEN with slow clinical progression may be considered [3].

Short-acting SSAs are used when it is necessary to quickly control the clinical symptoms of F-NEN, in the perioperative period, or in selected cases, before starting treatment with long-acting SSA, in order to assess their tolerance in patients (discussed in the first part of the recommendations [24]).

SSAs do not constitute the first-line treatment of insulinoma (discussed in the first part of the recommendations [24]). SAAs reduce symptoms of hypoglycaemia in patients with advanced SSTR (+) insulinoma. Due to the risk of paradoxical hypoglycaemia, they should not be used in SSTR (–) cases. Instead of assessing receptor expression, short-acting octreotide can be administered to check the treatment efficiency [3]. Pasireotide may be considered in insulinoma due to its relatively high affinity to SSTR3 and SSTR5 and the fact that its use is associated with the occurrence of hyperglycaemia [3, 148]. In the case of malignant forms of insulinoma, the use of SSA as a second-line treatment may be effective in relieving the disease symptoms (discussed in the first part of the recommendations [24]).

4.3.2. Molecularly targeted treatment (targeted therapy)

Only two molecularly targeted drugs are available in Poland: the mammalian target of rapamycin (m-TOR)pathway inhibitor everolimus and the multi-tyrosine kinase inhibitor sunitinib. Both drugs are approved for the treatment of inoperable or metastatic well (G1) or moderately (G2) differentiated PanNENs. The optimal moment to start targeted treatment should be slow disease progression observed over the previous 12 months, radiologically assessed according to Response Evaluation Criteria in Solid Tumours 1.1 (RECIST 1.1) as progressive disease (PD) (the occurrence of new lesions or an increase in measurable lesions by at least 20%).

Based on the results of two placebo-controlled studies, the median progression-free survival (PFS) was significantly increased, by approximately 5.5–6 months, for both drugs, with remission rates of 5% for everolimus and approximately 10% in the group of patients treated with sunitinib [149, 150].

Everolimus

Everolimus is an oral drug recommended in a daily dose of 10 mg, the effectiveness of which has been demonstrated, inter alia, in clinical trials (RADIANT-1, RADANT-3) in patients with advanced, progressive PanNENs G1 and PanNENs G2. The time to progression was prolonged in the group of patients previously treated with both chemotherapy and somatostatin analogues (mPFS 9.7–16.6 months) [151–153].

Median overall survival (OS) showed a favourable trend in the everolimus vs. placebo group [44.0 months vs. 37.7 months; hazard ratio (HR): 0.94; p = 0.3] [152].

The effectiveness of everolimus has been demonstrated in the treatment of functional pancreatic insulinomas in the scope of controlling hypoglycaemia [154], as well as PanNEN with carcinoid syndrome (RADIANT-2) [155].

A similar hypoglycaemic effect was demonstrated in individual cases of PanNETs with sunitinib [156].

The use of targeted therapies in combination with SSAs in the treatment of PanNEN is standard practice (inhibition of tumour growth and improvement of symptom control).

The role of everolimus in the treatment algorithm is presented in Figure 1, but this is only a current proposal to organise the sequence of treatment that requires further confirmation by prospective clinical trials (e.g. the COMPETE study comparing everolimus to PRRT, the SEQTOR study comparing streptozocin (STZ) with everolimus, and others). Some clinical trials are still ongoing: NCT03049189, NCT02246127.

Figure 1. Treatment algorithm for locally advanced, inoperable, or metastatic pancreatic neuroendocrine neoplasms (NEN). SSA — somatostatin analogues; RLT (PRRT) — radioligand therapy (radioisotope therapy); CAPTEM — chemotherapy: capecitabine + temozolomide; STZ + 5-FU — chemotherapy: streptozocin + 5-fluorouracil; FOLFOX — chemotherapy: oxaliplatin + 5-FU; CAPOX — chemotherapy: capecitabine + oxaliplatin; FOLFIRI — chemotherapy: irinotecan + 5-FU; SRI — somatostatin receptor imaging; PE — chemotherapy: cisplatin + etoposide; KE — chemotherapy: carboplatin + etoposide; FOLFIRINOX — chemotherapy: oxaliplatin + irinotecan + 5-FU; CHTH — chemotherapy; Ki-67 — proliferation index; NEC — neuroendocrine carcinoma; NET — neuroendocrine tumour; ***disease restricted to the liver: metastasectomy (resection or cytoreduction of metastatic lesions), laser ablation, radiofrequency ablation (RFA), transarterial chemoembolization (TACE), and radioembolization, including selective internal radiation therapy (SIRT) and others

Currently, there is no evidence of the efficacy of everolimus therapy in PanNEC, while a small retrospective study demonstrated drug activity in a group of NETs G3 patients with a very short median PFS of 1.2 months [157].

Phase II prospective clinical trials of the use of everolimus in NETs G3 and NECs are currently underway (NCT02113800, NCT02248012).

The main side effects of everolimus were mucosal complications, hyperglycaemia, and rare cases of non-infectious interstitial pneumonia [158].

An increased risk of adverse events was demonstrated in a group of patients receiving PRRT or chemotherapy prior to everolimus [159].

Sunitinib is an orally administered drug with a daily dose of 37.5 mg, the effectiveness of which has been demonstrated in cases of PanNET compared to placebo in a phase III randomised multicentre study (PFS 11.4 months vs. 5.5 months, HR: 0.42; 95% CI: 0.26–0.66, p < 0.001) [150]; the overall response rate (ORR) was obtained in < 10% of patients. Sunitinib has been approved in the treatment of PanNET G1/PanNET G2 in treatment-naive or previously treated patients, but it is not recommended in NEC and NET G3, despite the promising results of a small phase II study [160].

Adverse effects associated with sunitinib include diarrhoea, fatigue syndrome, hypertension, congestive heart failure, lymphopaenia, and hair discoloration [158].

Figure 1 presents the optimal role of sunitinib in PanNET therapy.

Targeted therapy should continue until the occurrence of disease progression or unacceptable adverse events. Side effects may require dose reduction to 5 mg of everolimus/day or 25 mg of sunitinib/day [149, 150].

The lack of direct comparative (head-to-head) studies of these drugs requires the selection of targeted drugs depending on the individual patient profile, taking into account comorbidities, adverse effects of drugs, and the availability of treatment.

The choice of everolimus may be limited by comorbidities such as uncontrolled diabetes or pulmonary disease, while sunitinib use may be limited by serious cardiovascular complications. Molecularly targeted drugs can be used as treatment options for first-line or second-line treatment following chemotherapy or administration of SSA [“cold” or “hot” RLT (PRRT)].

The best sequence of treatment, taking into account the impact on survival, symptom control, and quality of life, remains a subject of debate and controversy [161, 162].

At present, in Poland, it is not possible to apply treatment with other targeted or immunological drugs due to the lack of approval or insufficient evidence for the effectiveness and safety of these therapies, including the following: surufatinib, bevacizumab, sorafenib, pazopanib, axitinib, pembrolizumab, nivolumab with ipilimumab, and durvalumab with tremelimumab. The results of ongoing clinical trials are pending [163–168].

Minimal consensus statement on molecularly targeted treatment of PanNEN

4.3.3. Chemotherapy

Chemotherapy (ChT) is another antiproliferative treatment option for locally advanced and metastatic PanNENs.

In PanNENs, the results of ChT are much better than in other GEP NET locations (ORR 43–70% vs. 5–15%) [3, 13].

Cytostatics with particular activity in PanNENs include the following: STZ, 5-fluorouracil-based chemotherapy (5-FU, CAP — capecitabine), anthracyclines (ADM — adriamycin, EPI — epirubicin), dacarbazine (DTIC), temozolomide (TEM), irinotecan (I), and platinum-based chemotherapy (OXA — oxaliplatin, CDDP — cisplatin, CBDCA — carboplatin) administered as monotherapy or in combination with other cytostatic agents [169]. The widest clinical experience is related to PanNET treatment with streptozocin and CAPTEM (capecitabine + temozolomide) [170–179].

STZ has been approved by the Food and Drug Administration (FDA) for the treatment of patients with advanced PanNEN. The effectiveness of STZ was demonstrated both in monotherapy and in combination with other cytostatics: 5-fluorouracil (STZ/5-FU) and adriamycin (STZ/ADM) as well as the monoclonal antibody bevacizumab [180–184].

The oral, two-drug regimen of capecitabine + temozolomide (CAPTEM) is dedicated to the same group of patients with advanced PanNENs and is used in Poland as an alternative to STZ [14, 175].

Both drugs come in the form of tablets and are administered in 28-day cycles, where, from day 1 to day 14 the patient receives capecitabine 2 × 750–1000 mg/m2 (in practice 2 × 1000 mg), and on days 10–14, temozolomide is added in a dose of 150–200 mg/m2, followed by a two-week interval. Typically, therapy is continued either until disease progression, unacceptable toxicity, or the completion of 6–10 cycles.

The main adverse effects of CAPTEM are mucositis, neutropaenia, and thrombocytopaenia, but the treatment is well tolerated by most patients [170–176, 179].

Platinum-based chemotherapy regimens such as cisplatin + etoposide/carboplatin + etoposide (PE/KE) have demonstrated efficacy in the treatment of poorly differentiated large and small cell NECs and some NENs G3 characterised with high dynamics of disease progression or when other treatment options are exhausted [185, 186].

Despite the relatively high percentage of objective responses to treatment with the ORR in the range 30–67%, the median overall survival (mOS) is not satisfactory (11–13 months) [36, 128].

In the case of the intention to treat, we can distinguish adjuvant (postoperative) therapy, neoadjuvant or induction (preoperative) therapy, and palliative therapy (when radical treatment is no longer possible, the disease is locally advanced or generalised, i.e. with metastases). Indications for individual treatment methods are presented in Table 9.

Complementary (adjuvant) chemotherapy

In highly differentiated PanNETs G1– PanNETs G3 following radical surgical treatment, there are no indications for adjuvant therapy.

In some cases of pancreatic NETs G3, the decision about adjuvant treatment can be made on an individual basis (there is no evidence from randomised trials).

Pancreatic NEC — due to the high recurrence rate after radical surgery, adjuvant chemotherapy may be considered. The type of regimen and the number of treatment cycles are still under discussion. Due to the biological similarity of NEC to small cell lung cancer, therapeutic experience can be extrapolated, because there is still no evidence from prospective clinical trials supporting the disease management [13].

Induction or neoadjuvant chemotherapy

Despite little evidence, in selected patients with locally advanced, unresectable primary pancreatic NET G2 with a Ki-67 of 5–20% and in those with more aggressive course of disease (NET G3 and NEC) systemic induction or neoadjuvant treatment may be considered in order to reduce the extent of the disease and enable radical or cytoreductive treatment at a later stage [13].

Several retrospective studies have demonstrated the efficacy of the concurrent combination of cytostatics and radiotherapy (RCT) in locally advanced pancreatic NET as an alternative to surgery [187].

Palliative chemotherapy

Palliative chemotherapy is dedicated primarily to neoplastic disease in the stage of generalisation or local advancement, when it is not possible to apply surgical treatment of pancreatic tumours with a high cell Ki-67.

The use of chemotherapy can be considered in PanNET G2 with a Ki-67 value of over 10% and in NET G3 with a Ki-67 value of over 20%, while in the case of NEC it is often the only treatment option (treatment of choice).

The results of clinical trials conducted in patients with PanNET G1 and PanNET G2 indicate the possibility of obtaining a 43–70% objective response rate thanks to palliative chemotherapy with STZ or 5-FU-based therapy (CAPTEM), while in the NET G3 and NEC group the objective response rate was more favourable in the group of NEC patients following the use of platinum-based therapy (ORR 15% vs. 42%; p < 0.01), and in the NET G3 group after CAPTEM treatment (statistical data above) [128, 175].

The cell proliferation index value usually determines the choice of the type of chemotherapy.

5-FU-based chemotherapy regimens (CAPTEM, FOLFOX, CAPOX, FOLFIRI, FOLFIRINOX), i.e. drugs with moderate potency, are preferred in the treatment of NEN with Ki-67 < 55% [3, 13, 128].

More “aggressive” platinum-based regimens (PE, KE) are recommended in the treatment of NETs G3 and NECs with a high cell proliferation index Ki-67 > 55% [3, 13, 128].

The treatment algorithm for pancreatic NET is presented in Figure 1.

First-line chemotherapy

Slowly progressing PanNET G1 and PanNET G2

Chemotherapy in PanNET G1/PanNET G2 is not recommended as the first-line treatment.

Rapidly progressing PanNET G2

Chemotherapy with temozolomide (CAPTEM) or STZ+5-FU is considered in the first-line treatment for symptomatic and rapidly progressing tumours or in the second line of treatment following progression while taking SSA or PRRT.

PanNET G3

Chemotherapy with 5-FU (oral CAPTEM regimen or intravenous FOLFOX or FOLFIRI regimens) or alternatively STZ + 5-FU are the preferred first-line or second-line treatments following PRRT for NET G3 SRI (+), [18F]FDG PET (–).

PanNEC

Chemotherapy with PE regimen is the basic first-line treatment of NEC.

In the case of renal failure or other contraindications to the administration of cisplatin, the KE regimen can be used [3, 13, 14, 187].

Second-line chemotherapy

After the failure of first-line treatment, there is no strictly defined standard in selecting a second-line chemotherapy regimen. Clinical practice proves that the optimal treatment effects are achieved on the basis of similar stratification criteria as those used when selecting the first-line chemotherapy.

The qualification for subsequent lines of chemotherapy depends on the good clinical condition of the patient, i.e. 0–2 points in the World Health Organization/Eastern Cooperative Oncology Group (WHO/ECOG) performance status. Other parameters are also taken into account, such as the duration of progression-free time, persistence of complications induced by first-line treatment, comorbidities, location of metastases, SSTR expression, and patient’s preferences.

Each case requires an individual decision, preferably made in consultation with the multidisciplinary team (MDT).

PanNET G1 and PanNET G2

If ChT with STZ or CAPTEM is used in the first line, the following schemes may be considered in subsequent lines of treatment: FOLFIRI, FOLFOX, and CAPOX.

PanNET G3

If ChT schemes with STZ or CAPTEM are used in the first line, a platinum-based scheme is preferable in the subsequent lines: PE or KE.

PanNEC

After the failure of ChT with a platinum-based scheme (PE/KE), it is recommended that regimens containing STZ, CAPTEM, FOLFOX, FOLFIRI, CAPOX, and FOLFIRINOX are used in subsequent lines of treatment [3, 13, 14, 187].

If a good response to first-line chemotherapy is achieved and 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-based therapy may be considered using a PE/KE regimen.

However, in the case of long-term CAPTEM treatment in patients who have achieved objective response to treatment or long-term stabilisation of the disease, a so-called “therapeutic vacation” can be considered. The return to the treatment regimen can take place at the time of disease progression, provided that the progression occurs several months after the treatment was withdrawn.

Minimal consensus statement on PanNEN chemotherapy

Adjuvant chemotherapy (postoperative)

Palliative chemotherapy (advanced disease or distant metastases)

Minimal consensus statement on systemic treatment of PanNEN

4.4. Radioisotope therapy in PanNEN

The optimal algorithm for managing various forms of systemic therapies available in unresectable PanNEN is a constant challenge for clinicians. The primary goal of treatment is to prolong survival and ensure the maintenance of patients’ quality of life. PRRT with the possibility of using radionuclide therapy as an independent line of treatment or in combination with SSA, tyrosine kinase inhibitor (TKI), or chemotherapy is of particular importance.

Long-acting somatostatin analogues are usually recommended as first-line treatment in PanNETs, but in the case of tumours with Ki-67 ≥ 10%, chemotherapy with CAPTEM or biological therapy (everolimus, sunitinib) may be considered, with or without a long-acting somatostatin analogue. In the second-line treatment of PanNET with Ki-67 below 10%, PRRT may be considered as an alternative to the use of TKI and chemotherapy [European Society for Medical Oncology (ESMO) prefers PRRT] as the third-line treatment [13]. There are reports in the literature of patients in whom PRRT was used as a neoadjuvant treatment, enabling partial or complete resection of primary PanNET, and clinical improvement was obtained in functional pancreatic tumours, mainly insulinomas. Most polish centres treating patients with NEN have had positive experience in this area. Recently, there was a report on a case of a 26-year-old woman with disseminated glucagonoma after the administration of two cycles of PRRT, who achieved a partial response to treatment with almost complete regression of skin lesions in the form of necrotic migratory erythema [188]. According to our previous guidelines, the neoadjuvant treatment with PRRT in selected primary unresectable, secreting and non-secreting PanNET cases should be maintained as well as the possible use of PRRT as first-line treatment in malignant insulinomas to control symptoms of the disease.

PRRT is considered an effective and relatively safe therapeutic option in well-differentiated PanNET. In 2020, a combined analysis of two prospective and six retrospective studies using PRRT in PanNET was published, which determined the median PFS ranging from 20 to 39 months and the median OS from 37 to 79 months (PFS and OS were comparable to GEP-NETs in other locations) [189].

Based on the research published so far, we can say that there are several PanNET treatments, primarily second line, with proven efficacy, but treatment in resistant disease and the optimal sequence of therapy lines have not yet been established. Among the currently ongoing European studies, the results of which may help in this regard, is a phase 2 randomised trial under the acronym OCCULANDOM (antitumor efficacy of peptide receptor radionuclide therapy with 177lutetium — octreotate randomised vs. sunitinib in unresectable progressive well differentiated neuroendocrine pancreatic tumour: first randomised phase II; clinicaltrials.gov NCT02230176). The study is designed to evaluate the efficacy and safety of PRRT (7.4 GBq; max. 4 treatment cycles) compared with sunitinib (37.5 mg/day) in advanced PanNET G1/PanNET G2 with positive expression of SSTR, following progression on first-line treatment, regardless of the type of systemic therapy used: chemotherapy (only one line), everolimus, or long-acting somatostatin analogue. The results of this study are expected in 2023.

The main objective of the Italian randomised comparative phase II trial LUTHREE (optimising the interval between cycles of PRRT with 177Lu-DOTATATE in SSRT2-positive tumours; clinicaltrials.gov NCT03454763) is to assess the PFS and safety of PRRT (Lu-177 with SSA) including two arms of treatment: conventional PRRT therapy administered in cycles every 8–10 weeks compared to more frequent five-week cycles. The results of this study are expected soon (end of study: May 2021).

Most studies, the results of which are pending, aimed at establishing the role of PRRT, among others, in PanNETs, are studies comparing this treatment with other groups of drugs, mainly with everolimus (e.g. COMPLETE; clinicaltrials.gov NCT03049189), as well as in less differentiated GEP-NETs such as in the SSA arm of the NETTER-2 study (NETs G2/NETs G3 with SSTR expression; 177Lu-DOTATATE 4 cycles plus octreotide LAR 30 mg vs. octreotide LAR 60 mg; clinicaltrials.gov NCT03972488).

Only limited data on the efficacy of PRRT in PanNENs G3 with Ki-67 greater than 20% are available so far. In 2019, Professor Baum’s team published data on the results of PRRT with Lu-177/Y-90 (DOTATATE or DOTATOC) in 69 patients, including 46 patients with PanNET. Promising results have been obtained, especially in patients with Ki-67 index ≤ 55%, even in patients in whom chemotherapy was unsuccessful [89]. Similar results were obtained by Thang et al. in a group of 17 patients (17/28) with PanNET G3 without or after radio-sensitising chemotherapy [90] Nevertheless, it is necessary to perform well-designed prospective studies to fully define the role of PRRT in PanNEN, especially in well and poorly differentiated NET G3. The high uptake of [18F]FDG seems to correlate with a poorly differentiated tumour that is likely not to respond well to both PRRT and chemotherapy; in these cases, it may be beneficial to use a combination of PRRT and radio-sensitising chemotherapy [91].

The efficacy and safety of the locoregional treatment of liver metastases, using selective internal radioisotope therapy (SIRT), is similar to that of NET in different locations and was discussed in the first part of the recommendation [24] (including ArTisaN study, clinicaltrials.gov NCT 104 362 436).

The role of PRRT in the treatment of PanNENs

PRRT should be considered as second-line treatment in tumours with Ki-67 < 10%. PRRT may be an alternative to tyrosine kinase inhibitors/chemotherapy. PRRT can be used as first-line treatment in secreting malignant PanNENs to control disease symptoms.

In selected cases of primary unresectable secreting and non-secreting PanNETs, PRRT can be used as a neoadjuvant treatment. PRRT can be used in the treatment of PanNETs G3, with SSTR expression.

Minimal consensus statement on radioligand treatment in pancreatic NEN

4.5. Treatment monitoring

Follow-up examinations should include clinical symptoms, and biochemical and imaging examinations — anatomical (CT or MRI) and functional (SSTR) [V, 2A]#:

Evidence quality and the strength of recommendations

#Evidence quality and the strength of recommendations has been established on the basis of the following tables according to ESMO and NCCN [190, 191] (Supplementary File — Tab. S1 and S2).

Conflict of interest

For the Conflict of Interest Statement, please see the Supplementary File.

Authors’ contributions

All authors contributed to the idea, gathered the information, interpreted the data, and wrote and accepted the final version of the manuscript.