Diagnosis and treatment of bronchopulmonary neuroendocrine tumours: State of the art

Abstract

Bronchopulmonary neuroendocrine tumours (BP-NET) are a heterogeneous population of neoplasms with different pathology, clinical behaviour and prognosis compared to the more common lung cancers. The management of BP-NET patients is largely based on studies with a low level of evidence and extrapolation of data obtained from more common types of neuroendocrine tumours. This review reflects our view of the current state of the art of diagnosis and treatment of patients with BP-NET.

Nomenclature and classification

We exclusively focus this review on bronchopulmonary neuroendocrine tumours (BP-NET), i.e. typical carcinoids (TC), atypical carcinoids (AC) and large cell neuroendocrine carcinomas (LCNEC), excluding small cell lung cancer and other histological subtypes of lung cancer with features of neuroendocrine morphology and differentiation [1–3], i.e. non-small cell lung cancer with expression of neuroendocrine markers, and LCNEC combined with small cell, squamous cell or adenocarcinoma components, as diagnosis and management of these tumours should comply with the guidelines for small cell and non-small cell lung carcinomas, respectively [2–6]. Furthermore, very rare subtypes of lung tumours with neuroendocrine features including primitive ectodermal tumours, desmoplastic round cells tumours, paragangliomas and carcinomas with a rhabdoid phenotype and likewise [3] are not included.

Incidence

BP-NET (including TC, AC and LCNEC) comprise 1–3% of all lung cancer cases and 20–30% of all NET cases [3,7,8]. The incidence has increased considerably over the last 30 years. According to the North American Surveillance, Epidemiology, and End Results (SEER) database the annual age- adjusted incidence is 1.35 cases per 100 000 individuals, but the true incidence of BP-NET is higher as the inclusion of carcinoids in the SEER database is not complete [7,9,10].

Histopathology of BP-NET

BP-NET include a broad spectrum of neoplasms, ranging from indolent, well differentiated carcinoid tumours to highly aggressive, poorly differentiated neuroendocrine carcinomas [3,11]. The currently used IASLC 2007–2010 tumour, node, metastasis (TNM) classification for lung cancer, recently revised 2015 WHO classification of BP-NET and Travis‘s histopathological classification is based on the growth pattern of the tumour cells, morphological differentiation and anatomic extent of the disease [3,12–15] (Table I).

Table I. Summary of diagnostic criteria of BP-NET (the 2015 WHO classification) [3].

	Typical carcinoid (TC)	Atypical carcinoid (AC)	Large cell neuroendocrine carcinoma (LCNEC)
Mitotic count	0–1 per 2 mm^2 (10 HPFs)	2–10 per 2 mm^2 (10 HPFs)	> 10 per 2 mm^2 (10 HPFs) (median of 70)
Necrosis	No	Usually focal or punctiform	Large zones
Ki-67 proliferation index	≤ 5%	≤ 20%	40–80%
Differentiation and grade	Well differentiated Low grade	Can be both well- and moderately differentiated Intermediate grade	Poorly differentiated High grade

HPFs, high-power fields.

Due to their rare and complex nature, the pathological diagnosis of BP-NET can be difficult. Therefore, a pathologist with experience in neuroendocrine thoracic malignancies should be consulted [2]. Positive immunohistochemical staining for one or both of the neuroendocrine markers synaptophysin and chromogranin A (CgA) is essential for the diagnosis. Neuron-specific enolase (NSE) and neural cell adhesion molecule (NCAM)-CD56 suffer a poor specificity and are not recommended as diagnostic markers for the neuroendocrine differentiation [16]. However, CD56 may complement a standard diagnostic antibody panel for discrimination of the high-grade tumours from well differentiated BP-NET [3,17]. Unlike common subtypes of lung carcinomas, the thyroid transcription factor 1 (TTF-1), together with a low molecular weight cytokeratin 7 (CK-7) expression in BP-NET may be weak, thus complicating a differential diagnosis in the metastatic cases [1–3]. Difficulties in BP-NET classification commonly occur due to poor reproducibility of the mitotic counts and/or assessment of necrosis, especially in small biopsies with extensive crush artefacts. In cytological specimens mitotic counts are unreliable [1–3]. The Ki-67 proliferation index has been proposed in the 2015 WHO classification for the differential diagnosis of BP-NET [3] (Table I). The Ki-67 index is reproducible [18–20] and can be estimated with great accuracy even in small biopsies [21]. Therefore, we recommend that the Ki-67 index is determined as a supplement to aid in distinguishing the main subtypes of BP-NET [18]. However, prognostic significance of this marker within individual subtypes of well- and moderately differentiated BP-NET is not provided [3,21].

Diffuse idiopathic neuroendocrine cell hyperplasia and tumourlets

Diffuse idiopathic neuroendocrine cell hyperplasia (DIPNECH) is a primary disorder which is recognised by the 2015 WHO classification [3]. DIPNECH is considered as a potential but not necessary premalignant condition for tumourlets, TC and AC, but not for high-grade tumours [2], and comprises a proliferation of pulmonary neuroendocrine cells confined to the basement membrane of the respiratory epithelia. These lesions have a multifocal pattern, are located peripherally and may accompany a malignant tumour and be misinterpreted as metastatic disease [22]. Tumourlets are small, less than 5 mm groups of neuroendocrine cells that extend beyond the airway mucosa and are morphologically similar to TC. Tumourlets are frequently seen in patients with bronchiectasis, interstitial fibrosis and other chronic lung diseases [23]. The clinical significance of tumourlets is uncertain.

Biochemical markers

Serological biomarkers in BP-NET are poorly studied and their predictive value is considered to be lower than in patients with gastro- entero-pancreatic NET (GEP-NET) [24]. Plasma CgA (p-CgA) may, however, provide some information on tumour burden, recurrence after radical surgery and treatment response [24]. Measurement of p-CgA is therefore recommended in BP-NET patients before and during follow-up after surgical intervention as well as during medical treatment [25].

Other biomarkers, such as NSE, are not specific for BP-NET and are not routinely recommended for use as biochemical markers.

Screening for hypersecretion of tumour-specific hormones may be relevant in the diagnosis and follow-up of patients with the carcinoid syndrome or Cushing's syndrome.

Aetiology

The high-grade BP-NET demonstrate a strong correlation with heavy smoking history [3,15]. Risk of LCNEC in never-smokers is less understood. The role of active tobacco smoking in the aetiology of TC and AC is uncertain [26,27]. The importance of driver gene mutations of common types of lung cancer, such as the epidermal growth factor receptor (EGFR) mutations or the echinoderm microtubule-associated-protein-like 4 and the anaplastic lymphoma kinase genes (EML4-ALK) translocations, is still poorly explored in BP-NET [28].

BP-NET are usually sporadic but approximately 5% of the cases are associated with the familial hereditary syndrome multiple neuroendocrine neoplasia type I (MEN-1) [26,29]. Mutations in the MENIN tumour suppressor gene located on chromosome 11q3 can be identified in 80–90% of probands with familial MEN-1 [26,29,30].

Clinical presentation

The location of BP-NET lesions and tumour stage determine the clinical features. TC are often centrally located and patients usually have a history of years of coughing, dyspnoea, recurrent pneumonias and haemoptysis [15,26]. AC and LCNEC are most often localised peripherally presenting clinically with non-specific symptoms, such as flu-like symptoms, night sweats and dyspnoea [26,31]. Hormone-related symptoms occur in functioning variants of BP-NET [32,33]. In MEN-1 associated BP-NET signs and symptoms related to affected endocrine organs may occur [29,30,34,35].

Carcinoid syndrome and Cushing's syndrome

Less than 5% of BP-NET is functioning, i.e. release biologically active amines and peptides and exhibit hormone-related symptoms in the patients, such as the carcinoid syndrome and Cushing's syndrome. Patients with TC and AC, who have hormone- associated symptoms, should be screened for both conditions [32,33,35].

Two of the most common manifestations of the carcinoid syndrome are flushing and diarrhoea [26]. Measurement of p-CgA and 24-hour urinary 5-hydroxyindolacetic acid (5HIAA) excretion should be performed on suspicion [26,33]. At diagnosis and during follow-up pro-brain natriuretic peptide (pro-BNP) and an echocardiography is indicated to detect early signs of carcinoid heart disease [36].

Ectopic production of adrenocorticotropic hormone (ACTH) occurs in approximately 2% of BP-NET patients. The frequent clinical features of Cushing's syndrome secondary to ectopic ACTH production includes central obesity, proximal muscle weakness, hypokalaemia, glucose intolerance, hyperpigmentation and infections [32,37]. The diagnosis can be confirmed by the determination of an elevated level of ACTH and cortisol in blood and of urinary-free cortisol [37].

Multiple Endocrine Neoplasia 1 syndrome

MEN-1 syndrome is an autosomal dominant disorder, characterised by parathyroid adenomas, pancreas NET and pituitary adenomas. Less common, MEN-1-associated tumours are broncho-pulmonary, thymic or adrenal tumours as well as cutaneous lipomas. Symptoms arise from the organs involved or may be related to hypersecretion of hormones (i.e. parathyroid hormone, prolactin, follicle stimulating hormone, luteinising hormone, thyroid stimulating hormone, gastrin, insulin and other) or local or metastatic disease [34,35]. In BP-NET patients with a family history of MEN-1 or with presence of at least one of three cardinal diseases (parathyroid adenomas, pancreatico-duodenal tumours and pituitary adenomas), hormonal and biochemical screening and an MEN-1 mutation test should be performed [29,30,34,35].

Diagnosis, workup and specific imaging

Initial diagnosis and staging of BP-NET is based on the same principles as common types of lung cancer [4–6]. The primary standard workup after clinical suspicion of lung cancer includes a contrast enhanced chest and upper abdominal computed tomography (CT) scan, a CT- or ultrasound-guided transthoracic needle biopsy (TTNB) in peripheral lung lesions and flexible fiberoptic bronchoscopy with bronchial washing, bronchial brushing and forceps biopsy in central tumours [6,38]. An endobronchial ultrasound (EBUS) with the real-time transbronchial needle aspiration (TBNA) may significantly increase the diagnostic yield of traditional diagnostic techniques without adding morbidity [27,38–40]. EBUS-TBNA can provide sufficient materials for the immunohistochemical stains to establish a more accurate histopathological diagnosis in centrally located peribronchial and paraoesophageal BP-NET which are not visible by conventional bronchoscopy and are inaccessible to CT-guided TTNB [3,17,39,41]. The risks from bronchial brushing and the bronchial biopsy in the BP-NET are similar to bronchoscopy in general and can therefore be safely repeated [27,39]. In suspicious intrabronchial lesions the combination of the standard bronchoscopy with EBUS-TBNA can reduce the amount of endoscopic failures in order to obtain the adequate tissue samples [41]. Mediastinal lymph nodes can be evaluated with EBUS-guided TBNA, and an endoscopic ultrasound-guided (EUS) fine needle aspiration (FNA) should be performed in patients with probability of mediastinal spread to carefully determine the stage of disease [4,6,13,40] (Figure 1). The bronchoscopic appearance in low-grade BP-NET may be sufficient to make a presumptive diagnosis but its histological confirmation is preferred [26,39]. Even though severe haemorrhage after issued FNA or TBNA which requires operative interventions is rare, in some specific cases with macroscopically definitive findings of richly vascularised endobronchial tumours, the preoperative biopsy may be avoided [26,39,42–44]. EUS can be useful to exclude an airways invasion and provides information on the endobronchial tumours relation to surrounding structures [41]. Mediastinal lymph node involvement in TC does not impact the prognosis and therefore preoperative mediastinal staging in TC may be avoided [14,39,45].

Figure 1. Schematic drawing illustrating a suggested pathway for diagnosis and staging of suspected BP-NET in patients with no distant metastases.

In clinical practice, the diagnostic workup for further evaluation of pulmonary nodules is frequently completed with an ¹⁸F-fluorodeoxyglucose positron emission tomography (¹⁸F-FDG-PET) scan before the histological diagnosis is confirmed [4,6]. Due to varying ¹⁸F-FDG uptake patterns, the preoperative ¹⁸F-FDG-PET is considered optional in biopsy proven, well differentiated BP-NET [5,46–48]. However, several studies have shown a high sensitivity of ¹⁸F-FDG-PET in well differentiated BP-NET [48–50]. A more accurate assessment of ¹⁸F-FDG-PET uptake rates in well differentiated BP-NET, especially in TC, could be obtained by using lower cut-off values (< 2.5) of the maximal standardized uptake value (SUVmax) rather than the widely applied SUVmax in other lung carcinomas [48,51,52]. In contrast, ¹⁸F-FDG-PET is a valuable tool for the preoperative assessment of disease extension in patients with intermediate- and high-grade tumours [4,5] (Figure 1). Most TC, AC and some of LCNEC express somatostatin receptors (SSTR) that can be effectively visualised by isotope-labelled somatostatin analogues (SSA). ¹¹¹Indium octreotide scintigraphy using a single photon emission computed tomography (SPECT) has been replaced in many centres by PET scans most often using the ⁶⁸Ga-DOTA-peptide, as imaging with SSA-PET demonstrates a higher sensitivity with detection of smaller lesions than SSA SPECT [53,54]. New tracers including ⁶⁴Cu-DOTATATE, are underway that may further increase the spatial resolution, sensitivity and specificity of the scans [55]. Although only a few studies have explored the diagnostic value of SSTR-PET scans in BP-NET specifically, their results indicate a high rate of detection of primary tumours and small metastatic lesions in well differentiated tumours [46,53,55–57].

We recommend that the staging of patients with well- and moderately differentiated BP-NET should include a SSTR-PET scan combined with a diagnostic CT scan of the chest and abdomen (Figure 1). In higher grade BP-NET an ¹⁸F-FDG-PET scan with integrated diagnostic CT fusion is a very sensitive imaging method with high sensitivity for detection of small metastatic lesions [47]. Combined PET tracers using ¹⁸FDG and ⁶⁸Ga/⁶⁴Cu-DOTA-peptides may be complementary for disease staging in the certain cases [47,50,54–57]. The choice of appropriate diagnostic imaging modalities should be discussed on a multidisciplinary tumour board meeting. The physiological fitness of patients for surgery predicts postoperative complications and morbidity [58]. Therefore, all potentially operable patients should undergo a lung function test measuring the forced expiratory volume in 1 second (FEV1), the forced vital capacity (FVC) and the diffusing capacity of the lungs for carbon monoxide (DLCO), and a cardiac workup in case of suspected decreased cardiac functional capacity. A multidisciplinary tumour board should be consulted to approve diagnosis and staging as well as to discuss indications for additional examinations and management [4,58].

Treatment

As a result of the low incidence of BP-NET, it is difficult to conduct prospective studies with an adequate number of cases, and patients with BP-NET are often included only as a minority in NET clinical trials. Therefore, current treatment recommendations are based primarily on single institution cohort observations, retrospective studies, extrapolation of results from other tumour types and a few, sometimes underpowered prospective studies (evidence level IIB, IIC , IIIA and IIIB) [59,60].

Localised disease

Surgery is the treatment of first choice for BP-NET patients with resectable tumours. Standard lobectomy with systematic lymph node dissection is recommended. A limited resection (wedge or segmentectomy) may be considered appropriate for TC in stage I, provided the resection is complete (R0) [14,58]. Due to lower morbidity and better survival rates a bronchial sleeve resection is preferred over pneumonectomy in case of centrally located tumours, whenever possible [61]. The reported five-year survival rates after radical surgery are 89–100%, 58–88% and 27–75% for TC, AC and LCNEC, respectively [14,27,31,39,45,62–70] (Table II). The 10-year survival for TC is more than 90% in N0-stage, but 75% if hilar (N1) or mediastinal (N2) lymph nodes are involved. There is no significant difference in 10-year survival rates for TC with N1 or N2 involvement and therefore surgery for N2-stage can be recommended, whenever possible [13,14,39,58]. In contrast, the five-year survival rates for AC with N1 and N2 involvement are highly different, around 70% and 22–60%, respectively [39,45,63–66]. If radical surgery for locally advanced AC is possible this is recommended. Patients with high-grade tumours with loco-regional lymph node metastases have a high risk of recurrence and poor prognosis [27,31,67,70–72], and surgery in locally advanced LCNEC is generally not recommended.

Table II. The reported five-year survival rates in cohorts of BP-NET patients.

	Localised disease	Locally advanced disease	Metastatic disease
TC	88–100% [14,27,63,64,66]	60–78% [14,63,66]	42–57% [14,64]
AC	61–100% [27,63,66]	22–60% [63,66]	25–42% [26,64,81]
LCNEC	27–75% [62,67–70]	24–45% [68–70]	0–15% [26,69,81]

AC, atypical carcinoids; LCNEC, large cell neuroendocrine carcinomas; TC, typical carcinoids.

A minimally invasive approach, such as video-assisted thoracoscopic surgery (VATS), is recommended in experienced centres due to fewer complications and potentially increased survival rates [38,73,74]. Segmentectomy and wedge resections with systematic lymph node dissection should be considered especially in patients with compromised pulmonary function and/or severe comorbidities [38,75].

Occasionally carcinoids present as a polypoid structure without extension through the cartilaginous wall. Under these circumstances and with careful patient selection endobronchial resection may be a reasonable alternative to surgical resection. This requires careful follow-up including bronchoscopy for local recurrence [42,43].

Stereotactic body radiation therapy (SBRT) and/or radiofrequency ablation (RFA) are local modalities useful for treating medically inoperable stage I BP-NET patients [76].

There are sparse published data addressing the effect of adjuvant treatment after radical (R0) or microscopic non-radical (R1) surgery for TC and AC [77]. In case of an incomplete resection a re-operation should be considered in patients with TC and AC [39,72].

After R1-surgery, complementary radiotherapy (RT) alone or concomitant with chemotherapy should be considered in the certain AC and LCNEC patients with localised and locally advanced disease. These therapies can be administrated with similar dose and fractionation schedules used to treat stages IIIA–IIIB non-small cell lung cancer [4,13], albeit the evidence is sparse [67,78,79]. Patients with macroscopic residual disease (R2) and without possibilities for neither re-operation or curatively intended RT or chemo-RT should be offered palliative treatment when needed.

Several retrospective studies and one prospective study suggest that adjuvant platinum-based chemotherapy in radically resected stage IA–IIIA LCNEC patients may improve the outcome and significantly reduce the risk of recurrence. According to these studies, the five-year survival averaged 51–59% after adjuvant chemotherapy versus 33–37% after surgery alone [31,67,68,70], and fit patients are, therefore, usually offered adjuvant treatment with four cycles of cisplatin/carboplatin and etoposide.

Patients with neurological signs and symptoms should be examined for brain metastases, but prophylactic cranial irradiation is not advocated in BP-NET.

Locally advanced disease

Less than 10% of TC patients have locally advanced disease at the time of diagnosis and approximately 2–10% develop loco-regional or distant recurrence during follow-up [27,66,72]. There is no evidence that combined chemo-RT improves progression-free survival (PFS) or overall survival (OS) in these patients [78,79]. Therefore, TC patients with non-resectable, local or locally advanced disease should be assessed for palliative tumour reductive surgery and/or systemic treatment.

Approximately 30% of AC and up to 45% of LCNEC patients have involvement of the hilar and mediastinal lymph nodes at diagnosis, and about 25% of AC patients, and more than 35% of LCNEC patients develop loco-regional recurrence in the course of the disease [65,69,80,81]. A definitive dose RT (sequential or concomitant) combined with chemotherapy could be considered for such patients with locally advanced disease, but the optimal sequence of modalities is still unclear [78]. The reported five-year survival rates in cohorts of BP-NET patients with locally advanced disease are shown in Table II.

Metastatic disease

Biotherapy with somatostatin analogues

SSAs have symptomatic, biochemical and antiproliferative activity in NET [82–85]. In BP-NET patients SSA have mainly been used to provide relief of carcinoid symptoms and to reduce the risk of a carcinoid crisis [84].

The effects of SSA are due to the binding of the drug to SSTRs, which are widely expressed in NET [86–88]. Recently, in vitro studies showed that metastatic TC and AC tend to have diffuse overexpression of SSTR, whereas high-grade tumours only have focal SSTR overexpression [87,89].

The most commonly used SSA are Octreotide-LAR^® and Lanreotide-Autogel^®, which have high binding affinity for the SSTR2 subtype. Pasireotide^®, which is a new SSA, has a more universal binding profile (SSTR1-3 and 5) and is currently being tested in BP-NET in clinical trials.

Several phase II-studies and the RADIANT-2 phase III-study investigating the efficacy of SSA have included small number of BP-NET patients. An objective response (OR) was observed in 5–18%, while a stabilisation of tumour growth was seen in over 70% of the patients. In an exploratory subgroup analysis of the RADIANT-2 trial, median progression-free survival (mPFS) was 5.6 months for BP-NET patients whereas the mPFS in the whole patient group was more than 11 months [90–93].

In previous studies, SSA has been used alone or in combination with interferon alpha (IFNα) in patients with GEP-NET. This combination has shown both symptomatic and antiproliferative effects but no improvement of PFS [16,94–96]. There are no published data that support the efficacy of IFNα used as monotherapy or in combination with SSA in BP-NET patients.

In conclusion, SSA treatment can be considered in low-grade BP-NET patients with unresectable or advanced disease.

Chemotherapy

Current European guidelines recommend chemotherapy with streptozotocin (STZ) and 5-fluorouracil (5-FU) for advanced TC and AC, however the evidence is limited (IIIA, IIIB) [33].

STZ-based chemotherapy in BP-NET is associated with response rates (RR) of 20–25% and a median overall survival (mOS) of 18–24 months. The combination of STZ with 5-FU seems to provide a survival benefit compared to combinations with other agents [97–99]. The most common serious adverse effect of STZ is kidney toxicity.

Patients with LCNEC are usually offered platinum and etoposide as first line of treatment. RRs varies but are lower than in patients with SCLC, and the mOS is 6–20 months [100–102]. The treatment is frequently associated with transient bone marrow suppression and neutropenic infections.

Five non-randomised studies have demonstrated the efficacy of temozolomide (TMZ) as a single agent or in combination with other chemotherapeutic agents and/or angiogenesis inhibitors in patients with well- and intermediately differentiated BP-NET [103,104]. TMZ is administered orally and is usually well tolerated, although bone marrow toxicity and lymphopoenia with opportunistic infections may occur. Some investigators advocate the use of TMZ as the first choice of chemotherapy in TC and AC based on the promising results in GEP-NET [105,106]. However, the combination of TMZ and capecitabine used in GEP-NET has not been investigated in prospective studies of BP-NET. In BP-NET RRs of TMZ-based regimens was 14–33%, and more than 30% had stable disease (SD). PFS was approximately 6 months [104,107,108]. There was no significant difference in efficacy of combined therapy and TMZ alone [109–112].

No established second-line treatment for patients with high-grade tumours exists. In fit patients with tumour progression and previous long lasting response to chemotherapy, re-induction of first-line treatment with platinum and etoposide can be attempted.

Sometimes BP-NET of low- and intermediate-grade may exhibit a clinically aggressive course of disease [113]. In some cases, a re-biopsy may show progression to a higher grade of malignancy, and a biopsy from a progressing lesion should therefore be considered in recurrent or progressive disease before initiation of next line of treatment.

Peptide receptor radionuclide therapy

Peptide receptor radionuclide therapy (PRRT) exploits the ability of NET to bind radioactive β-emitter-labelled SSA to SSTR2 on tumour cells and thereby deliver cytotoxic radiation. This approach is primarily indicated as second- or third-line treatment in non-resectable and metastatic BP-NET with a tumour uptake higher than the physiological liver uptake at SSTR2 imaging.

¹⁷⁷Lu-DOTATATE and ⁹⁰Y-DOTATOC are the applied radiopharmaceuticals for PRRT. Although randomised clinical trials are lacking, it seems that there is no difference in anti-tumour efficacy between the two isotopes. Studies have shown that response is achieved in 20–25% of the patients and more than 60% achieve SD. The median TTP for these patients is approximately 20–30 months [54,114–116]. Morphological, biochemical and clinical response to PRRT as well as mOS are significantly higher in patients with well differentiated compared to poorly differentiated BP-NET, and response is correlated with high tumoural uptake of SSA assessed by pre-therapeutic SSTR imaging [117].

Serious adverse effects occur in 10% of patients and include bone marrow suppression, nephrotoxicity and, rarely, aplastic anemia. Nephrotoxicity after PRRT is dose and fractionation dependent [118] and evaluation of renal function before and accurate dosimetry during the therapy is needed. The rate of decline in the renal function is lower in patients treated with ¹⁷⁷Lu-DOTATATE than with ⁹⁰Y-DOTATOC. Therefore, treatment with ¹⁷⁷Lu-DOTATATE is preferred in patients with increased risk for renal impairment [119].

Targeted therapy

The mTOR inhibitor everolimus has shown antiproliferative effect in in vitro studies on BP-NET cell lines and in vivo on xenograft mouse models [120,121]. The response to everolimus is assumed to be dependent on the expression of proteins of the PI3K/AKT/mTOR pathway [122]. In a phase II study of everolimus, 55% of the patients achieved SD and the mPFS was around 6 months [122]. Both a phase II study [123] and the RADIANT-2 phase III study [92,93] [where 7% (N = 4) and 10% (N = 44) BP-NET patients were included, respectively] have suggested an additional effect of the combination of everolimus with octreotide LAR as compared to octreotide LAR alone. An explorative subgroup analysis indicated that patients with low- to intermediate-grade BP-NET had longer PFS in the combination arm compared to the control arm (13.4 vs. 5.6 months), but the difference was not statistically significant [92]. Thus, the place of everolimus in treatment of BP-NET is still not defined.

Several ongoing phase II and phase III studies (Identifiers: NCT 01524783; NCT 01563354; NCT 01317615) are designed to clarify the efficacy of everolimus in patients with BP-NET.

Surgical debulkning and treatment of liver metastases

Non-randomised studies suggest that aggressive treatment of liver metastases in patients with indolent NET, such as TC and AC, may prolong survival [124,125]. Five-year survival in selected patients is up to 61–94% [126,127]. However, surgery and local ablative therapy (e.g. RFA, microwave ablation and SBRT) is only possible in less than 10% of the patients due to bulky disease [124,126]. In patients with multiple liver metastases liver selective palliative treatments (e.g. transcatheter arterial bland- or chemoembolisation, and selective internal RT treatment of liver metastases) may be considered [126–128].

Even when radical resection cannot be achieved, patients with metastatic disease in the liver (of maximum 75% of the liver volume) should be evaluated for debulkning, particularly if hormone-related symptoms are difficult to control. In selected LCNEC patients with a relatively low- to intermediate Ki-67 index and clinically SD, local treatment of liver metastases in combination with systemic therapy may also be considered, whereas in patients with LCNEC and a high Ki-67 index, local treatment is generally not recommended [126,127,129].

Follow-up

Radically operated TC patients in stage I–II should generally be followed up yearly for at least 20 years due the indolent behaviour of the tumour. A bronchoscopy should be considered annually in patients with centrally located tumours, and in other cases with increased concern of local recurrence [39,44]. Patients with locally advanced disease as well as AC and LCNEC patients in good performance should be followed with a CT scan of the thorax and upper abdomen every 3–6 months. Treatment factors including type of surgery, systemic and local treatment, tumour factors such as tumour aggressiveness and burden, as well as patient factors such as age and comorbidity should be taken into account to provide the most appropriate, individualised follow-up programme. Measurement of p-CgA in BP-NET patients during follow-up can provide complementary information for early detection of the disease recurrence. In BP-NET patients with previously demonstrated high uptake on SSTR imaging, a combined CT and SSTR-PET may be considered if recurrence or progression is suspected. The prognostic impact of SSTR imaging in a routine BP-NET surveillance programme remains to be clarified.

In conclusion

The current management of BP-NET suffers from lack of studies of sufficient level of evidence. The multidisciplinary tumour board has a crucial role due to challenges in pathological classification, complexity of imaging modalities [3,8,12,47–57] and the rapid development of the new therapeutics [89–93,104,110]. The major potentially curative treatment is surgical resection. Controversies regarding the value of current and new prognostic or predictive factors should be solved in prospective studies [45]. A number of studies suggest an improvement in survival with postoperative adjuvant treatment in high-risk patients, but the optimal treatment has yet to be established [67–71]. The benefit of chemotherapy and/or RT with curative intent in inoperable patients with localised or locally advanced BP-NET is unclear. Results from ongoing multicentre studies, including RADIANT-4 (Identifier: NCT01524783), LUNA (Identifier: NCT01563354) studies and investigator-initiated Phase II trial RAD001 with paclitaxel and carboplatin in first-line treatment of patients with advanced LCNEC (Identifier: NCT01317615) are awaited to clarify the role of the mTOR-inhibitor everolimus alone or in combination with other agents in metastatic BP-NET. While SSA are useful for symptom control in functioning BP-NET [9,33,60,84], their antitumour activity in BP-NET still needs to be shown. Although no randomised trials of PRRT exist, PRRT can be valuable in BP-NET patients with progressive metastatic disease as second- or third-line treatment [54,114–119]. Patient selection for resection of lever metastases and local ablation therapies should be carried out by the multidisciplinary tumour board [126,127]. Further adequately dimensioned, clinical studies designed especially for BP-NET are warranted.

Acknowledgements

Morten Ladekarl: Novartis advisory Board; Henning Grønbæk: Novartis advisory Board, received research grants form Novartis and IPSEN pharmaceutical companies; Ulrich Knigge: Novartis Advisory Board, IPSEN Advisory Board; received research grants form Novartis and IPSEN pharmaceutical companies.

Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.