MEDICAL MANAGEMENT OF GEPNETS

Introduction

Neuroendocrine tumors have increased in incidence over time. A recent SEER database analysis showed that the age- adjusted incidence of NETs has increased nearly 6.4-fold from 1.09 per 100,000 persons in 1973 to 6.98 per 100,000 in 2012.¹ The increasing incidence was noted across all sites, grades and stages of neuroendocrine tumors. The median overall survival for all patients was 9.3 years. Localized NETs had better median overall survival (>30 years) compared with regional NETs (10.2 years) and distant NETs (12 months). On evaluation of survival trends over 3 time periods (2000–2004, 2005–2008, and 2009–2012), the improvement in survival was more pronounced in the subgroup with distant GI NETs (HR, 0.76; 95% CI, 0.67–0.86 for 2005–2008 and HR, 0.71; 95% CI, 0.63–0.82 for 2009–2012 compared with 2000–2004). In addition to factors like increased detection rates and awareness, the improvement in survival is also likely due to the addition of treatment options for neuroendocrine tumors in this time period. This chapter will cover the systemic therapy options for the management of well-differentiated GEPNETS. The management of high-grade or poorly differentiated neuroendocrine carcinomas will also be discussed briefly.

Treatment of all patients with GEPNETS should be individualized and approached in a multidisciplinary manner. Some important tumor related factors to consider include the site of origin (small intestine vs pancreas), grade of the tumor, presence or absence of symptoms attributable to the disease, hormonal hypersecretion, sites and burden of metastases. Some of the patient related factors that are pertinent in decision making include age, medical comorbidities and patient preferences. In a patient with hepatic metastases only, especially small bowel grade 1 or 2 NETs, surgical debulking should be considered if optimal cytoreduction can be achieved. In asymptomatic patients with low volume disease, surveillance with periodic clinical and radiographic assessment is also an appropriate first step of management. Upfront systemic therapy should be considered in patients who are symptomatic from tumor bulk or from hormone hypersecretion, especially if significant extrahepatic disease is note and not appropriate for surgical resection. In such patients, a combination of therapies can also be considered for rapid relief of symptoms.

While pancreatic NETs are more responsive to cytotoxic chemotherapy, small bowel NETs are relatively refractory to chemotherapy. Everolimus, an mTORC inhibitor is approved for pancreatic, lung and small bowel NETs while sunitinib, a multi tyrosine kinase inhibitor is approved for use in pancreatic NETs only at this time.^2,3 PRRT may be an option for all low or intermediate GEPNETs that express somatostatin receptors. While many therapeutic options have been added to the treatment armamentarium for GEPNETs, very little is known about the optimal sequence of these therapies.

Somatostatin analogues

For patients with symptoms of hormone secretion from GEPNETs, treatment with a somatostatin analog (SSA) is often the first step. Somatostatin, a hypothalamic peptide that inhibits growth hormone secretion was first discovered in 1973^4. Thereafter, its role in the regulation of multiple other hormones, neuropeptides as well as secretion of pancreatic enzymes was elucidated. Most well-differentiated GEPNETs express somatostatin receptors (SSTR) on their cell surface. There are 5 subtypes of somatostatin receptors (SSTRs; numbered 1–5), of which SSTR2 subtype is most commonly overexpressed by GEPNETs. Almost 80 to 100% of cases of well-differentiated grade 1 and 2 GEPNETS express SSTRs.⁵ The presence of these receptors is usually confirmed by an octreotide scan or the more sensitive ⁶⁸Ga-DOTATATE PET. Somatostatin receptors are G protein coupled receptors whose antisecretory function is modulated by inhibition of adenyl cyclase and regulation of calcium and potassium channels. However, they also have potent antiproliferative effects through multiple direct and indirect mechanisms.⁶ Some of the direct mechanisms include blocking cell division by blockade of mitogenic growth factor signals through interaction with the MAPK/ERK pathway or inducing apoptosis. Somatostatin analogues also exert many indirect antitumor actions through inhibition of secretion of growth factors, antiangiogenic effects and immunomodulatory effects.

There are currently two Food and Drug Administration (FDA) approved long-acting somatostatin analogues for the management of metastatic or advanced GEPNETs, octreotide LAR (Long Acting Repeatable) depot and lanreotide. Octreotide, an 8-amino acid peptide analogue, with more specific and longer duration of activity compared to native somatostatin was first introduced into clinical practice for the management of carcinoid syndrome in 1986.⁷ In this landmark study of 25 patients with symptomatic carcinoid syndrome and elevated urinary 5-HIAA levels, administration of subcutaneous octreotide led to palliation of symptoms in nearly all the patients and reduction in urinary 5-HIAA levels. The long acting depot formulation octreotide LAR ,administered intramuscularly once every 28 day was then compared to short acting subcutaneous octreotide in a randomized trial of 93 patients and showed similar efficacy.⁸ Subsequently, multiple studies of octreotide either alone or in conjunction with other therapies like interferon established the role of octreotide LAR in the management of carcinoid syndrome, leading to its FDA approval in 1995. Table 1 lists some of the important studies with somatostatin analogs.

Table 1.

Somatostatin analog trials

Trial	Treatment arm	Number of patients	Results
Phase 3 randomized prospective Arnold et cl 20059	Octreotide vs Octreotide +IFN -α	109	Median OS 32 months vs 54 months. Not significant. Partial tumor and stable disease at 1 year- 5.7% and 15% in both arms.
Phase 3 randomized prospective Faiss et al 200310	Lanreotide(L) vs IFNα (I) vs Lanreotide+IFNα(L+I)	83	Partial tumor regression in 1(L), 1(I) and 2(L+I)
Phase 3 randomized double-blind placebo-controlled ELECT Vinik et al 201611	Lanreotide depot vs Placebo	115	Percentage of days with rescue octreotide use 33.7% vs 48.5%
Phase 2 single –arm Kvols et al 19867	Octreotide subcutaneous Different doses	25	72% biochemical response
Phase 2 single arm in pancreatic islet cell Kvols et al 198712	Octreotide subcutaneous	22	63% biochemical response
Phase 2 single arm Saltz et al 199313	Octreotide subcutaneous	34	No objective response 50% stable disease 71% biochemical or symptom response
Phase 2 single arm prospective Arnold et al 199614	Octreotide subcutaneous	103	No objective tumor regression. 36.5% stable disease. 64% improvement in carcinoid syndrome
Phase 2 single-arm prospective Di Bartolomeo et al 199615	Octreotide subcutaneous in 2 dosing schedules	58	3% partial response 73% symptomatic control 77% biochemical response
Phase 2 single arm pilot Eriksson et al 199716	Lanreotide daily subcutaneous	19	5% tumor regression 70% stable disease 58% biochemical response
Phase 2 single arm prospective Wymenga et al 199917	Lanreotide depot	56	47% biochemical response 6% tumor regression 81% stable disease
Phase 2 single arm Ducreux M et al 200018	Lanreotide every 14 days or Lanreotide 30mg every 10 days	46	5% Objective response rate 70% stable disease
Phase 2/3 dose titration study Ruszniewski et al 200419	Lanreotide depot titrated from 90 mg to 120mg	71	38% overall response symptom 81% flushing improvement 75% diarrhea improvement

While octreotide’s role in the control of hormone hypersecretion is well established, its role as an anti-proliferative agent remained in question. PROMID was the first randomized, double-blind, placebo- controlled prospective study that examined the effect of octreotide LAR on tumor control.²⁰ Treatment naïve patients with functional or non -functional, locally inoperable or metastatic well-differentiated small bowel NETs received either placebo or long acting octreotide LAR at a dose of 30mg every 28 days. The median time to tumor progression in the octreotide LAR and placebo groups were 14.3 and 6 months respectively (HR=0.34;95% CI, 0.20 to 0.59; p=0.000072). At 6 months, stable disease was noted in 66.7% in the octreotide LAR arm vs 37.2% patients in the placebo arm. Though the long term median overall survival (OS) was only slightly different between the two patients groups (84.7 and 83.7 months), crossover of the majority of the patients on the placebo arm (38 out of 43) to the octreotide LAR arm may have confounded this.²¹ Only one partial response was noted in either group and no complete responses were noted. The Health- Related Quality of Life deteriorated more often and earlier in the placebo group as opposed to the patients who received octreotide LAR.²²

The PROMID study was limited to small bowel NETs. The antiproliferative action of somatostatin analogs in pancreatic NET as well as NET of unknown origin was established by the subsequent CLARINET study.²³ CLARINET was a randomized, double-blind, placebo-controlled study of lanreotide 120mg administered deep subcutaneously every 4 weeks compared to placebo in patients with well or moderately differentiated non-functional NETs of pancreas, small bowel or unknown origin. Median progression-free survival (PFS) was not reached in the lanreotide arm vs 18 months for the placebo arm (HR for progression or death, 0.47; 95% CI, 0.30 to 0.73; p<0.001). This pivotal study led to the approval of lanreotide for the treatment of metastatic or unresectable well to moderately differentiated GEPNETS by the FDA in 2014. The safety and efficacy of lanreotide as well as its anti-tumor activity was further confirmed in the open label extension study.²⁴ Lanreotide’s role in the control of carcinoid syndrome is also well established through multiple studies.^25,26,11

Higher doses of octreotide LAR or lanreotide are often used in practice for refractory symptoms or disease progression. Although well tolerated, it is hard to draw conclusions regarding their impact on PFS, OS or symptomatic improvement.^27,28 Ongoing trials like CLARINET FORTE (NCT 02651987) may provide an answer to these questions. The role of lanreotide following disease progression on octreotide LAR or vice versa is also not known. In a single institution retrospective study of 16 patients, treatment with lanreotide following disease progression or poor tolerance was reported to be favorable however many of these patients also received additional disease directed therapies.²⁹

Pasireotide, a somatostatin analog with avid binding affinity to SSTR 1,2,3 and 5, has been evaluated in multiple trials.^30,31 However, higher rates of grade 2 or 4 hyperglycemia without a benefit in PFS limited its usage and approval in NETs. Novel oral octreotide and subcutaneous depot formulations of octreotide are in early clinical trials and may add more therapy options in the future. Table 2 lists some ongoing clinical trials of octreotide or lanreotide.

Table 2.

Somatostatin analog ongoing or recently completed clinical trials

Trial name Phase	Disease site	Therapy	Number of patients	NCT
REMINET Prospective multicenter double blind randomized	Unresectable duodenopancreatic grade 1, 2	Lanreotide vs placebo as maintenance after chemotherapy	118	NCT02288377
PLANET Phase 1b/ 2	GEPENT grade 1/2	Pembrolizumab 200mg iv and Lanreotide 90mg every 3 weeks	26	NCT03043664
Phase 2 study of Ramucirumab with Somatostatin analog in advanced carcinoid tumors	Carcinoids Excluding pancreatic	Octreotide or Lanreotide plus Ramucirumab	43	NCT02795858
METNET-2 Pilot one arm open label prospective study	Advanced GI or lung well differentiated NETs	Lanreotide and Metformin 2550mg daily	20	NCT02823691
Sandostatin LAR and Axitinib vs placebo Phase II/III randomized double blind	G1 G2 NET of non-pancreatic origin	Sandostatin +Axitinib vs Sandostatin +placebo	148	NCT01744249

In summary, multiple studies have unequivocally established the role of somatostatin analogs in the management of both functional and non-functional metastatic GEPENTS making it a category 1 recommendation in National Comprehensive Cancer Network(NCCN) guidelines as well as in European Neuroendocrine Tumor Society (ENETS) and North American Neuroendocrine Tumor Society (NANETS) guidelines^32,33,34. The most frequent side effects are related to steatorrhea, flatulence, hyperglycemia and gallstones. Drug discontinuation due to side effects is infrequent. Most neuroendocrine experts consider either of the drugs appropriate for the management of carcinoid syndrome and for tumor control.

Historical results with chemotherapy

Before the discovery of agents targeting specific molecular pathways in NETs, the majority of the clinical trials examined the role of chemotherapeutic agents in the management of GEPENTS. Many of these early trials utilized end points like decrease in the size of hepatomegaly, reduction in urinary 5 HIAA or other biomarkers to assess response as opposed to the more modern and accepted response criteria such as radiographic Response Evaluation Criteria in Solid Tumors (RECIST) criteria. Thus, the response rates reported by these trials are likely an overestimation of the actual benefit from chemotherapy. Nevertheless, any discussion regarding systemic therapy options in GEPNETS is incomplete without reviewing some of these pioneering trials. To date, streptozocin (STZ) which was approved in 1982 is the only FDA approved chemotherapeutic agent for the management of advanced pancreatic islet cell tumors.

Chemotherapy in non-pancreatic GI carcinoid tumors

Early clinical trials in metastatic carcinoid tumors using the combination of STZ with 5- Fluorouracil (5-FU) or cyclophosphamide noted response rates of 44 % and 37 % respectively in carcinoids of small bowel origin with no difference in survival between the two arms.³⁵ A subsequent ECOG study looked at the combination of a modified schedule of 5-FU +STZ vs doxorubicin in 172 patients, where roughly one third of the patients had a confirmed NET of small intestine origin. The response rates in the combination 5 FU+STZ arm was lower than the prior ECOG study at 22%.³⁶ Dacarbazine as a single agent in metastatic carcinoid showed a response rate of 16 % (9/56 patients).³⁷ Based on these studies, a 3-arm study comparing 5-FU+ Doxorubicin, 5 FU+STZ and Dacarbazine was conducted using more objective radiographic criteria to assess response.³⁸ Response rates to all three regimens were modest (15.9% vs 16% vs 8.2% respectively). A randomized phase 3 clinical trial in 64 patients with progressive metastatic carcinoids comparing Interferon α to 5-FU/STZ combination showed no statistically significant difference in PFS and OS between the two arms and very few objective responses(9 and 3% respectively).³⁹ In an analysis of two phase II trials looking at the safety and efficacy of bevacizumab + FOLFOX (5-FU and Oxaliplatin) or CAPOX(capecitabine and oxaliplatin), response rates in the carcinoid arm was 13.6%(3/22 patients)with a median PFS 19.3 month while the response rates were higher in pancreatic NET at 41.7% (4/12 patients).⁴⁰ Phase 2 non-randomized studies of single agent capecitabine or in combination with bevacizumab have also shown much lesser response rates than historic data in SBNETs.^41,42,43 In a retrospective study of 18 patients with GI NETs, 4 of them of small bowel origin, combination capecitabine and temozolomide showed 1 complete response, 1 partial response and 1 stable disease.⁴⁴ Similarly temozolomide-based therapy in small bowel carcinoid was associated with no radiographic response and 1 biochemical response as opposed to response rates of 34% in pancreatic NET.⁴⁵ In a systematic meta-analysis of chemotherapy in non-pancreatic NETs by Lamarca et al, the authors concluded that the majority of the studies were of level C evidence with heterogeneous populations and treatments, limiting any conclusions.⁴⁶ There is a lack of evidence demonstrating objective response rates, overall survival or progression free survival benefit with chemotherapy in well-differentiated gastrointestinal NETs except of pancreatic origin. Thus, their use in small bowel origin NETs is generally considered only if patients have exhausted other standard therapy options. A summary of some of these trials is presented in table 3.

Table 3.

Chemotherapy in non-pancreatic carcinoids

Type of study	Regimen	Number of patients	Response rates	Median OS
RCT Moertel C et al 197935	STZ+ Cyclophosphamide Vs STZ+5-FU	89 total	26% vs 33%	12.5months vs 11.2 months.
RCT Phase 2/3 Engstorm et al 198436	5-FU+STZ vs Doxorubicin	172 total	22% Vs 21%	64 weeks vs 48 weeks
RCT Phase 2/3 study Sun et al 200538	5-FU+Doxorubicin Vs 5-FU +STZ Vs Dacarbazine	249 total Including pancreatic	15.9% 16% 8.2%	15.7months 24.3months 11.9months
RCT phase 3 Dahan et al 200939	5-FU+STZ Vs IFN-α−2a	64 total 36 midgut	1 partial response in midgut Vs 2 partial response in midgut	PFS 8.5 months vs 14.1months (midgut) PFS and OS difference not significant
Phase 2 single arm Burkowski et al 199437	Dacarbazine	56 total 28 mid gut	16% partial response total	20 months
Phase 2 combined analysis Kunz et al 201640 Included pNETs	FOLFOX+ Bevacizumab, CAPOX + bevacizumab	22 20	13.6% partial response, 5% partial response	33.1months 42.2 months
BETTER-1 phase 2 non-randomized trial Mitry et l 201441	Capecitabine + bevacizumab	49	No complete response 18% partial response	Median OS not reached Median PFS 23.4months
Phase 2 open label single arm Medley et al 201142	Capecitabine	20	11% biochemical PR No radiographic PR/CR	36.5months
Phase 2 single arm Berruti et al 201443	Capecitabine+Bevacizumab+Octreotide LAR	13	11.5% radiographic PR (including lung and unknown primary)	Median PFS 14.3months
Phase 2 trial Bukowski et al 198747	5FU+Adrimacyin+Cyclophosphamide+ STZ	63	31% response rate	10.8 months
2-arm non-randomized study Van Hazel et al 198348	Dactinomycin Dacarbazine	17 15	1 partial response 2 partial responses	28 weeks 47 weeks
Phase 2 randomized Oberg et al 198949	5 FU+ STZ Vs Interferon α	10 10	0 biochemical or radiographic response 50% biochemical response, 20% radiographic response	Not reported
Phase 2 randomized Janson et al 199250	IFN α Vs IFNα+ STZ+ Doxorubicin	12 11	1 biochemical response 2 radiographic response 0 biochemical or radiographic response	Not reported
Phase 2 single arm Kulke et al 200451	Docetaxel	21 11 mid gut	0% response rate in mid gut	Median PFS10 month Median OS 24 months
Phase 2 single arm Di Bartolomeo et al 199552	5 –FU+Epirubicin+Dacarbazine	38 total	2/ 20 carcinoid	Not reported
Retrospective study Kaltsas et cl 200253	5-FU+ Lomustine	31	21%	48 months
Phase 2 single arm Kulke et al 200658	Temozolamide + Thalidomide	29 total 14 carcinoid	25% radiographic response in all 7% radiographic response in carcinoid	Median OS not reached for entire cohort

Chemotherapy in pancreatic NETs

Broder et al were one of the first groups to report response rates of 37% with STZ in advanced pancreatic islet cell tumors in a single arm 52 patient study.⁵⁴ This was followed by a randomized ECOG study comparing STZ +5 Fluorouracil (5-FU) with STZ alone in advanced islet cell tumors.⁵⁵ Patients who received the combination chemotherapy were noted to have double the response rates compared to single agent STZ(63%vs36%) but no statistical difference in overall survival (24 months vs 17 months). Complete response rate of 33% was reported with the combination arm defined as the disappearance of all clinical or lab evidence of malignant disease and not by radiographic assessment. STZ/doxorubicin when compared to 5-FU/STZ in a prospective randomized trial of 102 patients with PNETs was noted to be superior with response rates of 69% vs 45% respectively.⁵⁶ Dacarbazine(DTIC),another alkylating agent like STZ, showed a response rate of 34% in a 50-patient study.⁵⁷ Despite the robust response rates noted in pNETs, the wide-spread use of dacarbazine and STZ is limited due to their toxicity. The advent of temozolomide (TMZ), an oral alkylating with a better toxicity prolife and tolerance has rendered the widespread use of STZ and dacarbazine in pancreatic NETs obsolete in modern times.

Capecitabine and temozolomide for PNETs

Alkylating agents like STZ work by inhibiting nucleoside incorporation during cell cycle division. The discovery of temozolomide, an oral alkylating agent that is converted to its active agent 5-(3-methy-triazeno) similar to DTIC, gave another treatment option with a more favorable toxicity profile. In a pilot phase 1 study of 30 patients with advanced neuroendocrine tumors excluding small cell carcinomas, temozolomide in combination with an antiangiogenic agent thalidomide showed an overall response rates of 25% (45% in pancreatic NET and 7% in small bowel NET).⁵⁸ Temozolomide, administered in a dose intense regimen of 150mg/m2 days 1–7 and day 15–21,was then studied in combination with bevacizumab, a monoclonal antibody targeting vascular endothelial growth factor in a 34 patient study.⁵⁹ 5 out of 15 patients with pancreatic NET had radiographic response while none of the patients with carcinoid tumor had an objective response. Another phase 1/ 2 study of everolimus in combination with temozolomide in patients with advanced pancreatic NETs showed a partial response rate of 40% and a median PFS of 15.4months.⁶⁰

The combination of capecitabine and temozolomide was first reported by Fine et al in 2005.⁶¹ They hypothesized that as most pNETs are wild type p53, the resistance to cell cycle specific cytotoxic agents is more likely a function of the tumor’s low ki-67 and indolent nature rather than a p53 mutation, a phenomenon called cytokinetic resistance. Thus, the synergistic combination of an alkylating agent that induces apoptosis in a static G0 cell phase and an antimetabolite like 5-FU has sound scientific rationale. They also noted a synergistic cell kill in BON1 cell line if these agents were delivered in a schedule dependent manner with 5-FU exposure preceding TMZ. Capecitabine is an oral pro-drug for 5-FU that gets activated in the liver to its active metabolites. Temozolomide mediates the majority of its cytotoxicity by methylation at the O6 guanine position. Tumors with high levels of the O6-methylguanyl methyltransferase (MGMT) repair enzyme are thus known to be more resistant to the action of temozolomide such as in glioblastoma and vice versa.^62,63 Pancreatic neuroendocrine tumors in general have been shown to have low or absent levels of O6- MGMT enzyme.⁴⁵ Even in pNETS with sufficient O6 MGMT, it is postulated that the initial administration of capecitabine may deplete the thymidine pools by inhibition of thymidylate synthase leading to decreased O6-MGMT enzyme repair activity and potentiation of temozolomide effect on the tumor. In their initial 10 patient series treated on this combination, they noted one complete response and 2 partial responses.⁶⁴

Following this, the combination of capecitabine and temozolomide, referred at times as CAPTEM, was reported by Strosberg and colleagues in grade 1 and 2 advanced pancreatic NETs.⁶⁵ In a single institution retrospective series of 30 patients who received capecitabine at 750mg/m2 twice daily days 1–14 and oral temozolomide at 200mg/m2 once daily days 10–14, the overall radiographic response rate was 70% (21 patients with partial response, 8 patients with stable disease).The combination as well as the dose of temozolomide used was well tolerated with lesser nausea and lymphopenia than the dose intense schedules used before. In a retrospective review of 18 patients including both pNETs and SBNETs, the majority of whom had also received prior chemotherapy (61%), the combination of CAPTEM still showed a response rate of 61% and median PFS of 14 months.⁴⁴ A meta-analysis of 15 published studies with pooled data from 384 patients, including some with grade 3 and non-pancreatic NETs, with this combination of capecitabine and temozolomide noted disease control rates of 72.89% (95% CI, 64.04% to 81.73%; p<0.01).⁶⁶ ECOG (Eastern Co-Operative Oncology Group) 2211, a two-arm, randomized, phase II clinical trial comparing single-agent temozolomide vs the combination of capecitabine and temozolomide in low-intermediate grade advanced pNETS was presented at the Annual American Society of Clinical Oncology (ASCO)GI meeting in 2018.⁶⁷At a median follow up of 20 months, the median PFS for the combination arm of CAPTEM was 22.7 months vs 14.4 months for single agent temozolomide (HR 0.58, 95%CI 0.36–0.93). Median overall survival was 38 months in the temozolomide arm and had not been reached in the combination arm. An objective response rate of 33% was noted with the combination, similar to what has been noted in prior retrospective studies. This trial is the first and only prospective trial that clearly validates the role of CAPTEM chemotherapy in patients with pancreatic NETs and also reported the longest PFS on any therapy in patients with pNETs to date.

Based on the early trials from 1980–1990s as well as the collective evidence through recent retrospective case series and the randomized ECOG 2211 trial, cytotoxic chemotherapy is an important treatment option in pancreatic NETs. Chemotherapy should be considered as an initial treatment option in patients with advanced pancreatic neuroendocrine tumors with a significant tumor burden, especially when symptomatic and if a rapid progressive clinical course is noted. Based on the robust response rates noted with this combination, their use in a neoadjuvant setting for borderline resectable or locally advanced pancreatic neuroendocrine tumor may also be considered.⁶⁸

Role of Interferons

Interferons were first introduced in the treatment of carcinoid tumors in 1982 by Oberg et al.⁶⁹ Nine patients with mid gut carcinoid tumors, six of them with symptomatic carcinoid syndrome were treated with daily intramuscular doses of leucocyte interferon- 3×10⁶ U per day for one month and 6 × 10⁶ U per day for another two months. Treatment with interferon resulted in prompt and decreased levels of urinary 5- hydroxyl indole acetic acid in six patients with liver metastases with resolution of the carcinoid symptom. Since then, multiple studies of interferon either as a single agent or in combination with somatostatin analogs showed biochemical response rates of about 40–50% and tumor response of 10–15%.⁷⁰ Studies have also looked at the role of interferon beta (IFN-β) as well as polyethylene glycol-modified (PEGylated) interferon, a once weekly long acting formulations of interferon.^71,72 The anti-tumor effects of interferons include a direct effect on cell cycle inducing arrest in G1 and G0 phase of synthesis, inhibition of growth factor production, antiangiogenic effects and immunomodulatory effect by increasing expression of class 1 antigens on tumor cells. Interferons have also been studied in combination with somatostatin analogs and bevacizumab. A list of some important randomized clinical trials utilizing interferons is summarized in table 4. A high incidence of flu-like symptoms, severe fatigue, depression and myelosuppression limit their wide-spread use in the management of neuroendocrine tumors. While NCCN guidelines list the use of interferons in neuroendocrine tumors as a category 3 recommendation, NANETs does not recommend the use of IFNα unless no other options are available.^32,33 ENETS include interferons as a second–line therapy for refractory carcinoid syndrome and as an anti-proliferative agent in mid gut carcinoid tumors with limited options.³⁴

Table 4 –

Randomized clinical trials with interferons

Trial	Treatment arms	Number of patients	Results
RCT Kolby L et al 200273	IFNα+ Octreotide vs Octreotide	68 metastatic SBNETs	No difference in 5-year OS. 56.8% vs 36.6%. IFNα treated patients with reduced risk of tumor progression (p=0.008)
RCT Arnold R et al 200574	IFN α+ Octreotide vs Octreotide	105 metastatic GEPNETs	Long term OS similar at 35 months for monotherapy and 51 months for combination. (HR1.19,95%CI [0.67–2.13; p=0.55). PFS similar
RCT Faiss S et al75	Lanreotide+ IFNα vs Lanreotide vs IFNα	80 therapy naïve metastatic GEPNETs	No statistically significant difference in rates of partial remission, stable disease or tumor progression in all 3 arms.
RCT open label Yao et al 201776	Octreotide+ Bevacizumab vs Octreotide +IFN-α−2b	427 advanced grade 1 and 2 NETs	No difference in PFS 16.6months vs 15.4months (HR,0.93;95%CI,0.73–1.18, p=0.55)

Molecular targeted therapies

Tyrosine kinase inhibitors (TKI)

GEPNETs are highly vascular tumors that have been shown to express high levels of Vascular Endothelial Growth Factor (VEGF), Vascular Endothelial Growth Factor Receptor (VEGFR) 2, VEGFR 3, Platelet Derived Growth Factor Receptor (PDGFR) alpha and PDGFR beta and the stem cell receptor c-kit.^{77,78, 79} Further, inhibition of VEGF pathway in a mouse pancreatic neuroendocrine model provided proof of concept for multiple studies utilizing angiogenesis inhibitors in GEPNETs. The VEGF pathway can be targeted either through tyrosine kinase inhibitors which inhibit multiple other growth factor receptors in addition to VEGF or through monoclonal antibody against VEGF receptors like bevacizumab. While sunitinib is the only drug under this category that has gained US FDA approval at this time, several TKIs have been tested or undergoing clinical trials.

Sunitinib for PNETs

Sunitinib is a broad-spectrum TKI that exhibits its action by inhibition of VEGFR1–2, PDGFRα, β, KIT, RET, FMS-like tyrosine kinase −3 (FLT3) and colony-stimulating factor receptor (CSF)-1R. Studies of sunitinib in Rip-Tag2 mouse models of pancreatic islet tumors suggested clinical activity.^80,81 Additionally, in a phase 1 clinical trial of sunitinib in multiple advanced solid tumors, antitumor activity was noted in one out of four patients with neuroendocrine tumors.⁸² This observation led to an open- label, multicenter, phase 2 study with 109 patients(n=41 carcinoids, n=66 pNETs) that showed an overall objective response(ORR) rate of 16.7% in pNETs (11 of 66 patients) and stable disease in 68% (45 of 66 patients).⁸³ The ORR was lesser at 2.4% in carcinoid tumors. Subsequently, a phase 3, randomized, double-blind, placebo-controlled trial was conducted in 171 patients with advanced and metastatic pancreatic NET with evidence of disease progression prior to trial enrolment.⁸⁴ Patients were randomized to receive 37.5mg daily dose of sunitinib or placebo. Due to clear differences in efficacy and the occurrence of more adverse events in the placebo arm, the trial was amended to allow crossover to the treatment arm. Median PFS was 11.4 months in patients treated with sunitinib compared to 5.5 months in the placebo arm (hazard ratio for disease progression or death, 0.42;95% CI, 0.26–0.66; p<0.001). Including 2 patients with complete response and 6 with partial response, the overall response rate on the sunitinib arm was 9.3%.The most common adverse events noted were diarrhea, nausea, asthenia, vomiting, hypertension, hand-foot syndrome and neutropenia. This landmark study led to the FDA approval of Sunitinib for the management of metastatic pancreatic NETs in 2011.

Pazopanib

Pazopanib, another oral TKI that targets VEGF receptors, PDGFR alpha and beta, FGFR 1 and 3 and cKIT has been evaluated in both pancreatic and small bowel NETs. In a parallel cohort study of patients with grade 1–2 carcinoid tumors (n=20) or pancreatic NET (n=32),7 of 32 patients with pancreatic NETs achieved an objective response.⁸⁵ In another nonrandomized, open-label, single-center phase II study in 37 patients with metastatic GEPNETs, an objective response rate of 18.9% was noted.⁸⁶ A more recent study presented at the ASCO 2019 demonstrated that VEGF targeting may still be a relevant therapeutic option in advanced non-pancreatic NETs.⁸⁷ 171 patients with low or intermediate grade non-pancreatic GI NETs were randomized to pazopanib vs placebo. Median PFS was 11.6 months in the pazopanib arm vs 8.5months in the placebo arm (HR 0.53; p= 0.0005). Although pazopanib was associated with more symptoms such as diarrhea, fatigue, grade 4 hypertension, the overall quality of life was similar between the treatment arms.

Cabozantinib

In preclinical models of pancreatic neuroendocrine tumors in RIP-Tag2 transgenic mice, VEGF inhibition by sunitinib was accompanied by strong immunoreactivity for c-met in tumor lymphatics. c-met blockade was shown to significantly reduce metastases to local lymph nodes.⁸⁸ Concurrent c-met and VEGF inhibition showed reduced invasion and metastases.⁸⁹ These findings led to a phase II two-cohort clinical trial to evaluate the efficacy of cabozantinib, a potent kinase inhibitor with activity against AXL, FLT-3,c-met, VEGFR 1,2,3 and RET in patients with progressive, well-differentiated grade 1–2 carcinoid or⁹⁵months(95% CI 8.5mnth-NR). 3 of 20 patients with pNET achieved PR (ORR 15%) with a median PFS of 21.8months (95% CI, 8.5–32 months). An ongoing phase 3, randomized clinical trial is evaluating the role of cabozantinib in advanced or metastatic NET with ki-67<20% of any origin (NCT03375320).

Other TKIS

Sorafenib showed partial response rate of 10% in carcinoids and pNETs in a phase 2 consortium study of 93 patients with metastatic GEPNETS.⁹¹ Axitinib, approved in the treatment of metastatic renal cell carcinoma, showed a median PFS of 26.7 months(95% CI,11.4–35.1)in an open-label, single-arm, phase 2 study of 30 patients with metastatic low to intermediate grade NET.⁹² Lenvatinib, another oral TKI, currently approved for use in advanced hepatocellular carcinoma and medullar thyroid cancers also showed significant antitumor activity in GEPNETs in a single arm phase 2 study with an overall response rates of 29%.⁹³

Bevacizumab

Bevacizumab is a recombinant humanized monoclonal antibody that blocks angiogenesis by binding and neutralizing VEGF. Bevacizumab has been studied in neuroendocrine tumors in many clinical trials in combination with somatostatin analogs, TKIs as well as chemotherapy.^{40.41,42,59,72,76,94} In a systematic review of bevacizumab-based combination therapies in neuroendocrine tumors, median PFS was reported in 8 out of 9 studies ranging from 8.2 months to 16.5 months.⁹⁵ However, it is hard to draw any conclusions due to the heterogeneity of the patients included and the different drug combinations used.

Everolimus in pNETs and SBNETs

Mammalian target of Rapamycin or mTOR is a serine threonine kinase that plays a key role in the control of cell growth, proliferation and cell death. It signals downstream of many receptor tyrosine kinases like VEGF, IGF, Akt /PKB, ERK1/2.⁹⁶ Patients with genetic cancer syndromes affecting the mTOR pathway genes like Neurofibromatosis (NF1), Tuberous sclerosis(TSC1 and TSC2) and Von Hippel Lindau disease (VHL) have an increased incidence of neuroendocrine tumors.^97,98,99,100 Further, multiple gene expression profiling studies indicate a role for PI3K/AKT/mTOR pathway in neuroendocrine tumorigenesis.^101,102,103 Everolimus, a selective mTORC1 oral inhibitor previously known as RAD-001, showed antitumor activity by in itself or in combination with octreotide in multiple preclinical studies in neuroendocrine tumor cell lines.^104,105 Temsirolimus, an intravenous agent that inhibits mTOR pathway was first studied in a phase 2 clinical trial in 37 patients with progressive neuroendocrine tumors of GI origin. Response rates of 5.6% was noted and the median time to progression was 6 months.¹⁰⁶ In a phase 2 study in advanced low to intermediate grade NETs, treatment with everolimus at 5mg or 10 mg daily in combination with Octreotide LAR showed promising antitumor activity with partial response rate of 22%.¹⁰⁷ RADIANT-1 (RAD001 in Advanced Neuroendocrine Tumors)was a multinational, open-label, phase II non-randomized trial to assess the antitumor activity of everolimus 10mg in advanced well to moderately differentiated pancreatic NET with progressive disease during or after cytotoxic chemotherapy.¹⁰⁸ Patients were stratified to receive either everolimus or everolimus and octreotide LAR. ORR was 9.6% by central radiology review with a median PFS of 9.7months in the everolimus arm (95% CI,8.3 to 13.3 months). RADIANT-2 was a randomized, double-blind placebo-controlled, phase 3 study in 429 patients with low or intermediate grade advanced neuroendocrine tumor associated with carcinoid syndrome.¹⁰⁹ Patients with disease progression within the past 12 months prior to enrollment were assigned to receive everolimus 10mg daily or placebo in conjunction with octreotide LAR. Treatment with everolimus was associated with a 5.1-month improvement in median PFS however the difference did not reach the prespecified threshold for statistical significance (HR 0.77, p=0.026). This was partly attributed to imbalances in the baseline characteristics of the arms and the cross over design, all favoring the placebo arm.

RADIANT-3 looked at everolimus 10mg daily vs placebo in advanced low-intermediate grade pNETs in a prospective, randomized, double-blind, placebo-controlled multicenter study of 410 patients.¹¹⁰ The median PFS was 11 months in the everolimus group as compared to 4.6 months in the placebo group representing a 65% reduction in the risk of progression by central independent assessment (HR for disease progression or death ,0.35;95% CI,0.26 to 0.44;p<0.001). The objective response rate was 5% as assessed by local investigators as compared to 2% in placebo arm. This study led to the US FDA approval of everolimus for the treatment of advanced pancreatic NET in 2011, the same year as the approval of sunitinib, nearly 30 years after the last drug approval for NETs. RADIANT-4 trial led to FDA approval of everolimus in well-moderately differentiated non-pancreatic NETs as well in 2015.¹¹¹ In this study, everolimus was associated with a 52% reduction in the estimated risk of progression or death. (HR 0.48;95% CI,0.35–0.67; p<0.00001) A trend towards improved overall survival was noted in the interim analysis. Though somatic mutations in the mTOR pathway are less frequently described in non-pancreatic NETs, the antitumor activity was postulated to be due to other factors like inhibition of growth factor signaling, epigenetic modulation and other undiscovered mechanism of action of everolimus. The grade 3 or 4 adverse events noted with everolimus across the RADIANT 3 and 4 trials include fatigue (2–4%), stomatitis (7–9%), diarrhea (7–10%), infections (2–7%), anemia (4–6%) and hyperglycemia (4–5%).Other pertinent side effects of all grades include rash, asthenia, peripheral edema, pneumonitis, nausea and weight loss. However, the drug discontinuation rate due to side effects in RADIANT −3 and RADIANT 4 trial was 13% and 12% respectively.

The RADIANT trials have firmly established everolimus as a therapeutic option for well-differentiated NETs of any GI, lung or unknown origin. Though RADIANT-3 and RADIANT-4 trials did not allow concurrent somatostatin analog therapy and enrolled only non-functional NETs, the combination of everolimus and somatostatin analogs is known to be safe and well tolerated. Several studies have also eluded to the role of everolimus in the control of a functional NET.^112,113

Combined mTORC and VEGF data or sequencing VEGF/mTORC

While targeting VEGF pathway and mTORC pathway individually has shown, the combination of these agents has been limited due to increased toxicity noted in multiple clinical trials. In a phase 1 standard dose escalation study of 12 patients combining everolimus with sorafenib, dose limiting toxicity was noted within the first cycle of therapy.¹¹⁴ Another phase 2 study that that combined temsirolimus and bevacizumab in well-moderately differentiated pNETs showed response rates of 41% and median PFS and OS were 13.2 months and 34 months (95% CI, 27.1 to not reached) respectively.¹¹⁵ Again, higher rates of grade 3–4 adverse events were noted. The combination of everolimus and bevacizumab in the CALGB 80701 study also noted higher rates of toxicity without additional benefits.⁹⁴ Combination of two angiogenesis inhibitors sorafenib and bevacizumab also showed unacceptable toxicity rates.¹¹⁶

Fewer studies have looked at sequencing therapies in the management of neuroendocrine tumors. Pazopanib was studied in a phase 2 clinical trial in advanced GEPNETS who had shown disease progression on at least one anti-angiogenic therapy or mTORC inhibitor.¹¹⁷Clinical benefit rate, as defined as complete response+ partial response+ stable disease was noted in 89% patients with prior mTOR inhibitor therapy, 83% in patients with prior antiangiogenic therapy and 60% in patients with both. An ongoing randomized open-label study in advanced pNETs to compare the efficacy and safety of everolimus followed by chemotherapy with STZ-5FU upon progression or the reverse sequence, SEQTOR (NCT02246127), may shed further light on the question of sequencing therapies. There are also multiple trials ongoing to help answer the question of where peptide receptor radionucleotide therapy (PRRT) fits in the sequence of treatment. The phase III COMPETE trial is a randomized, open-label, multicenter phase II study to evaluate 177Lu- Edotreotide PRRT in comparison compared to everolimus in patients with advanced somatostatin receptor positive GEPNETs (NCT03049189).Another ongoing randomized, open-label, phase II trial will assess the efficacy and safety of 177Lutetium-Octreotate vs sunitinib in well-differentiated pNETs (NCT02230176).

High grade neuroendocrine carcinoma

High grade neuroendocrine carcinomas are characterized by a very aggressive clinical course, high ki-67 and a propensity to metastasize early. The WHO classification of pancreatic neuroendocrine tumors was recently updated in 2017 to reflect the heterogeneity of grade 3 neuroendocrine neoplasms.¹¹⁸ Well differentiated grade 3 pancreatic NETs usually have a better prognosis than a poorly differentiated small cell or large cell neuroendocrine carcinoma. In large population-based studies of high-grade neuroendocrine tumors of GI origin, a ki-67 > 55% has been shown to be predictive of response to platinum-based therapy.¹¹⁹ There is a lack of prospective clinical trial data to guide the management of these high-grade extra pulmonary neuroendocrine carcinomas, both in the metastatic and locally advanced setting. Though surgery can be curative in a subset of patients with locally advanced tumors, the high risk of early metastatic dissemination raises the question of neoadjuvant therapy and careful patient selection in a multidisciplinary setting before such surgery is performed. In a SEER database study of 14,732 extra pulmonary NECs, the 5-year survival ranged from 58% to 60% in local stage NEC of female genital tract and small intestine to 25% for esophageal origin NEC.¹²⁰ Another large Nordic multicenter cohort study evaluating the role of surgery in 119 patients with high-grade pancreatic neuroendocrine tumors that included both well and poorly differentiated grade 3 tumors,14 patients underwent surgery in a non-metastatic setting.¹²¹ All these patients experienced recurrent disease with a median time to recurrence or metastases of 7 months (2–14 month). In another single-institution study of 44 poorly differentiated pancreatic NECs of all stages, 2-year survival rates were 22.5%.¹²²

Based on extrapolated data from small cell lung cancers, chemotherapy with a platinum combination, usually cisplatin/etoposide or carboplatin/etoposide is often the first line of therapy in metastatic poorly differentiated extrapulmonary NECs.IMpower 133 study led to the approval of the combination of atezolizumab, a checkpoint inhibitor and a platinum doublet chemotherapy in 2019 as first-line therapy for extensive stage lung cancer.¹²³ However, the upfront use of checkpoint inhibitors in extra pulmonary small cell cancers cannot yet be advocated for in the absence of well-designed clinical trials based on this. In a preliminary report of the neuroendocrine tumor cohort of a basket trial of dual immune checkpoint inhibitors using ipilimumab plus nivolumab in rare cancers, objective response rates were seen in 8 out of 19 patients with high grade neuroendocrine tumors in a second-or third-line setting. (ORR 42%).¹²⁴ However, two other clinical trials using single agent check point inhibitors in high grade extra pulmonary NENS only yielded 5% ORRs.^125,126 Thus, there is no clear evidence of overall survival benefit to recommend immunotherapy as a standard line of therapy yet and its role continues to be debated. The optimal second line chemotherapy in platinum-refractory setting is also not well established. Options include irinotecan-based regimens, topotecan, temozolomide or taxol. Due to paucity of data, participation in a clinical trial is encouraged.

Future directions

In addition to sequencing existing therapy options, there is an ongoing interest in combination therapies as well as novel therapeutic targets.¹²⁷ mTORC1/mTORC2 inhibitors have shown encouraging results in early phase clinical trials.¹²⁸ Early phase trails with CDK4/CDK6 inhibitors alone (NCT03891784) or in combination with everolimus (NCT03070301)are either underway or have been recently completed.^129,130 Ongoing trials are evaluating the combination of chemotherapy, immunotherapy or molecular targeted therapy in conjunction with PRRT (CONTROL NETS NCT02358356, NCT03457948, NCT03629847 respectively).Though reported trials have shown low response rates with single-agent immunotherapy, there are many ongoing studies of immune check point inhibitors alone or in combination with other therapies (NCT03457948, NCT03043664, NCT03095274).^125,131 Novel somatostatin analog-DM1 conjugate molecules like PEN-221, an antibody-drug conjugate containing the tubulin inhibitor mersantine, has shown preliminary evidence of antitumor activity¹³² and expansion cohorts in pancreatic and SBNETs are underway (NCT02936323). Phase 1 study of XmAb 18087, a bispecific antibody that engages the immune system against tumors by binding to somatostatin receptor 2 and CD3 is ongoing in neuroendocrine tumors and gastrointestinal stromal tumors (NCT03411915).

Conclusion

The systemic therapy options for the management of neuroendocrine tumors have come a long way from the discovery of somatostatin in 1973 to the most recent FDA approval of 177 Lutetium PRRT in 2017. A better understanding of the underlying molecular biology of neuroendocrine tumors has paved the way for the addition of many new therapies within the past 15 years. The significant improvement in overall survival even in advanced GEPNETs between 2000–2004 period to the 2009–2012 period as noted in the recent SEER database study may possibly be due to the addition of these new therapeutic agents in this timeframe.¹ With the addition of many therapeutic agents comes the challenge of appropriate treatment selection and sequencing. Patient and tumor centric factors remain paramount in the management of neuroendocrine tumors, making this the perfect example of how the science and art of cancer therapy come together to improve patient outcomes.

Key points.

• Octreotide LAR and Lanreotide both have demonstrated antiproliferative properties in somatostatin expression advanced well-differentiated GEPNETs in addition to their well-established role in the control of symptoms associated with functional NETs.

• Sunitinib is the only FDA approved TKI in advanced well-differentiated pancreatic NET.

• Everolimus, an mTORC inhibitor gained FDA approval for use in well-differentiated NET of lung, GI or pancreatic origin after progression on Octreotide analog therapy.

• Chemotherapy has limited therapeutic efficacy in small bowel NETs, however its role in pancreatic NETs is well established.

• Interferons may be considered for refractory carcinoid syndrome when there is a lack of other options.

• The question of appropriate sequencing of therapies, the safety and efficacy of combinations as well as novel drug targets remain to be explored.

Synopsis.

The increased incidence and prevalence of neuroendocrine tumors over the past few decades has been accompanied by an improvement in their overall survival as well. This may have been facilitated largely by the addition of multiple treatment options like somatostatin analogs, chemotherapy agents, targeted therapies like tyrosine kinase inhibitors and mTORC inhibitors and PRRT over this time frame. There are subtle differences in the management of small bowel neuroendocrine tumors (NETs) vs pancreatic NETs due to varying responses to some therapies, especially cytotoxic chemotherapies. The management of all patients with neuroendocrine tumors must be individualized based on patient characteristics as well tumor related factors like the site of origin, Ki-67 and the presence or absence of hormone hypersecretion. This chapter will review the role of somatostatin analogs, historical results with chemotherapy in gastroenteropancreatic NETs (GEPNETS) as well as more recent evidence for the use of cytotoxic chemotherapy in GEP-NETs. The chapter will also discuss molecular targeted therapies approved for use in GEPNETS and some ongoing clinical trials.