INTRODUCTION — Neuroendocrine cells are distributed widely throughout the body, and neuroendocrine neoplasms of these dispersed cells can arise at many sites. The classification and nomenclature of neuroendocrine neoplasms arising within the digestive system has evolved over the past two decades (see "Pathology, classification, and grading of neuroendocrine neoplasms arising in the digestive system", section on 'Classification and terminology'):

Neuroendocrine tumors (NETs) are a heterogeneous group of malignancies characterized by variable but most often indolent biologic behavior. Clinical behavior and prognosis correlate closely with histologic differentiation and grade, as assessed by mitotic count and/or Ki-67 labeling index (table 1) [1]. (See "Pathology, classification, and grading of neuroendocrine neoplasms arising in the digestive system", section on 'Classification and terminology'.)

Systemic treatment approaches to control tumor growth and symptoms related to hormone hypersecretion for patients with advanced or metastatic well-differentiated NET arising in the pancreas will be reviewed here. Systemic therapy options for patients with advanced or metastatic well-differentiated gastrointestinal tract NET (carcinoid) are discussed elsewhere, as are the clinical presentation, imaging, biochemical monitoring, pathology, and classification of gastroenteropancreatic NET; localization and treatment of pancreatic NET; evaluation and management of NET of unknown primary site; management of symptoms of functioning pancreatic NET; local management options to control tumor growth and symptoms of hormone excess for well-differentiated metastatic gastroenteropancreatic NET; and the evaluation and management of patients with high-grade gastroenteropancreatic neuroendocrine carcinoma.

●(See "Metastatic well-differentiated gastrointestinal neuroendocrine (carcinoid) tumors: Systemic therapy options to control tumor growth".)

●(See "Metastatic well-differentiated gastroenteropancreatic neuroendocrine tumors: Presentation, prognosis, imaging, and biochemical monitoring".)

●(See "Pathology, classification, and grading of neuroendocrine neoplasms arising in the digestive system".)

●(See "Classification, epidemiology, clinical presentation, localization, and staging of pancreatic neuroendocrine neoplasms".)

●(See "Neuroendocrine neoplasms of unknown primary site".)

●(See "Insulinoma".)

●(See "Management and prognosis of the Zollinger-Ellison syndrome (gastrinoma)".)

●(See "Glucagonoma and the glucagonoma syndrome".)

●(See "Somatostatinoma: Clinical manifestations, diagnosis, and management".)

●(See "VIPoma: Clinical manifestations, diagnosis, and management".)

●(See "Metastatic gastroenteropancreatic neuroendocrine tumors: Local options to control tumor growth and symptoms of hormone hypersecretion".)

●(See "High-grade gastroenteropancreatic neuroendocrine neoplasms".)

CLASSIFICATION AND BIOLOGIC BEHAVIOR — The World Health Organization (WHO) classifies all gastroenteropancreatic NETs into low-grade (G1), intermediate-grade (G2), and high-grade (G3) categories based on mitotic count and proliferative index (Ki-67) (table 1) [1]. Poorly differentiated NETs are G3 neuroendocrine carcinomas. (See "Pathology, classification, and grading of neuroendocrine neoplasms arising in the digestive system", section on '2010 and 2019 World Health Organization classification'.)

As a group, well-differentiated gastroenteropancreatic NETs are generally indolent malignancies with a prolonged natural history. However, clinical behavior is heterogeneous, as evidenced by the following observations:

●NETs arising in the tubular gastrointestinal tract and pancreas may have similar characteristics on routine histologic evaluation, but they have a different pathogenesis and biology [2]. Pancreatic NETs in general pursue a somewhat more aggressive course than do other gastrointestinal tract NETs [3], although, conversely, most systemic agents have been associated with higher response rates among patients with pancreatic NETs than in those with gastrointestinal NET. (See "Metastatic well-differentiated gastrointestinal neuroendocrine (carcinoid) tumors: Systemic therapy options to control tumor growth".)

●G2 gastroenteropancreatic NETs have a slightly worse prognosis than do G1 NETs [4]. Although they are classified and treated similarly at present, as new treatment modalities become available, it is likely that the histologic grade of a well-differentiated NET will affect the selection of appropriate treatment.

●By contrast, poorly differentiated neuroendocrine carcinomas have a rapidly progressive clinical course and a poor prognosis. They are generally treated with platinum-based chemotherapy regimens according to guidelines established for small cell lung carcinoma. (See "High-grade gastroenteropancreatic neuroendocrine neoplasms".)

●There is a small subset of well-differentiated tumors with a proliferative rate that places them in the high-grade category (G3 NETs). These tumors have a clinical behavior that is midway between G2 NETs and neuroendocrine carcinomas. Management is discussed in detail elsewhere. (See "High-grade gastroenteropancreatic neuroendocrine neoplasms", section on 'High-grade, well-differentiated tumors (NET G3)'.)

GENERAL APPROACH TO THE PATIENT — The majority of patients with advanced pancreatic NETs have liver metastases [5]. Most tumors (between 50 and 75 percent) are nonfunctioning and are not associated with a hormonal syndrome. (See "Classification, epidemiology, clinical presentation, localization, and staging of pancreatic neuroendocrine neoplasms", section on 'Clinical presentation'.)

●Symptoms of hormone secretion – Patients with symptoms of hormone hypersecretion from a well-differentiated pancreatic NET should be managed with somatostatin analogs and other agents as appropriate to the specific syndrome. In general, initial therapy for insulinomas consists of carbohydrates and diazoxide, which directly inhibits the release of insulin from insulinoma cells. Careful monitoring is necessary in insulinoma patients when initiating therapy with somatostatin analogs, which may paradoxically worsen hypoglycemia due to decreased secretion of counterregulatory hormones. Everolimus can also be highly effective in improving glycemic control in patients with insulinoma. For patients with gastrinoma, high doses of oral proton pump inhibitors are the treatment of choice. Somatostatin analogs may be helpful for refractory cases. (See "Insulinoma" and "Management and prognosis of the Zollinger-Ellison syndrome (gastrinoma)".)

●Potentially resectable disease – For patients who have potentially resectable metastatic disease, resection may provide prolonged control of symptoms and tumor growth. Although the majority of patients recur, even if resection is complete, metastasectomy is generally preferred over medical therapy for patients with potentially resectable liver metastases. (See "Metastatic gastroenteropancreatic neuroendocrine tumors: Local options to control tumor growth and symptoms of hormone hypersecretion", section on 'Surgical resection'.)

●Unresectable disease – Somatostatin analogs and the molecularly targeted agents everolimus and sunitinib have been shown to improve progression-free survival (PFS) duration relative to supportive care alone in patients with metastases from a nonfunctioning pancreatic NET. For patients with low-volume disease who are asymptomatic, we suggest waiting until disease progression or development of symptoms related to tumor bulk before beginning systemic therapy. For most patients with a metastatic pancreatic NET who are felt to require therapy, somatostatin analogs may be preferred over a molecularly targeted agent as initial therapy because of the more favorable side effect profile. (See 'Benefits' below.)

●Therapy at progression – If an initial approach of observation is chosen, initiation of a somatostatin analog should be considered at the time of progression.

For patients with radiologic disease progression despite use of a somatostatin analog at standard doses, therapeutic options for those with hepatic-predominant disease include noncurative surgical debulking and nonsurgical liver-directed therapy (eg, transarterial bland embolization, chemoembolization, or radioembolization). (See "Metastatic gastroenteropancreatic neuroendocrine tumors: Local options to control tumor growth and symptoms of hormone hypersecretion", section on 'Hepatic-predominant metastatic disease' and "Metastatic gastroenteropancreatic neuroendocrine tumors: Local options to control tumor growth and symptoms of hormone hypersecretion", section on 'Surgical resection' and "Metastatic gastroenteropancreatic neuroendocrine tumors: Local options to control tumor growth and symptoms of hormone hypersecretion", section on 'Nonsurgical liver-directed therapy'.)

For patients with more widespread disease that is not eligible for liver-directed therapy, systemic therapy options include everolimus, sunitinib, cytotoxic chemotherapy, peptide receptor radioligand therapy (PRRT; eg, lutetium-177 [¹⁷⁷Lu] dotatate), or dose-escalated somatostatin analog therapy. Studies evaluating the optimal sequencing of therapy have not been conducted. Furthermore, interpretation of an attempted network meta-analysis to rank the relative efficacy and toxicity of a variety of systemic therapies for NETs is limited due to variability in patient populations, variability in eligibility and response criteria, a lack of trials directly comparing active treatments, as well as unclear efficacy of the comparator arm in some cases (such as interferon) [6,7]. (See 'Molecularly targeted therapy' below and 'Cytotoxic chemotherapy' below and 'Peptide receptor radioligand therapy' below.)

Randomized trials comparing different systemic therapy approaches are needed, and eligible patients should be encouraged to enroll in available trials, such as the following:

•A European phase III trial, SEQTOR (NCT02246127), is comparing the efficacy and safety of chemotherapy (fluorouracil and streptozotocin) followed by everolimus versus everolimus followed by fluorouracil and streptozotocin in patients with advanced and progressive pancreatic NETs.

•Other clinical trials are examining the sequencing of PRRT, including a randomized phase II trial of PRRT with ¹⁷⁷Lu dotatate versus sunitinib (NCT02230176) for progressive disease.

•In addition, the phase III COMPETE trial is evaluating PRRT with ¹⁷⁷Lu edotreotide versus everolimus (NCT03049189).

If trial participation is not available or desired, the following represents our approach to treatment:

•For patients who are highly symptomatic from tumor bulk or who have rapidly enlarging metastases, we suggest chemotherapy rather than molecularly targeted therapy or a dose-escalated somatostatin analog. For most patients, we suggest the combination of capecitabine and temozolomide (CAPTEM) rather than temozolomide alone. In the absence of comparative trials, the choice of CAPTEM over a streptozocin-containing regimen should be individualized, taking into account the convenience of oral rather than intravenous treatment, performance status, and the anticipated side effect profile of both combinations. (See 'Dacarbazine and temozolomide-based regimens' below.)

•The molecularly targeted agents everolimus and sunitinib improve PFS duration relative to supportive care alone in patients with metastases from a nonfunctioning pancreatic NET, and these drugs represent an appropriate option for patients whose disease has progressed while receiving a somatostatin analog. The choice of agent is often based on the expected side effect profile, which differs between these two agents. (See 'Sunitinib' below and 'Everolimus' below.)

•PRRT using radiolabeled somatostatin analogs is an option for patients with disease that expresses somatostatin receptors (as defined by nuclear imaging) and has progressed on therapy including at least a somatostatin analog. Although clinical trials are being planned, there are no data yet specifically comparing PRRT with other therapeutic agents, and the choice of therapy in this situation should be based on the availability of PRRT, expected side effect profiles, and patient comorbidities and preference. (See 'Lutetium-177 dotatate' below.)

•Another option for patients with indolent disease after an initial response or a prolonged duration of stability with standard-dose somatostatin analog therapy is dose escalation of the somatostatin analog. (See 'Control of tumor growth' below.)

If dose escalation is not chosen, the benefit of continuing therapy with a standard-dose long-acting somatostatin analog in patients whose disease has progressed while receiving such therapy is not well defined. For patients with nonfunctional tumors who experience unequivocal radiographic progression on a somatostatin analog and for whom a switch in therapy is planned, we would discontinue the somatostatin analog. For patients with functional tumors, we continue the somatostatin analog to minimize the effects of hormone secretion.

The following sections will discuss systemic treatment options to control symptoms related to tumor bulk, or hormone hypersecretion and tumor growth for well-differentiated pancreatic NETs. Local treatment options for gastroenteropancreatic NETs, including resection and liver-directed therapy (such as embolization), are discussed in detail elsewhere, as are systemic therapy for well-differentiated gastrointestinal tract NETs and treatment approaches for high-grade gastroenteropancreatic neuroendocrine carcinomas. (See "Metastatic gastroenteropancreatic neuroendocrine tumors: Local options to control tumor growth and symptoms of hormone hypersecretion" and "Metastatic well-differentiated gastrointestinal neuroendocrine (carcinoid) tumors: Systemic therapy options to control tumor growth" and "High-grade gastroenteropancreatic neuroendocrine neoplasms".)

SOMATOSTATIN ANALOGS — Somatostatin is a 14-amino acid peptide that inhibits the secretion of a broad range of hormones in vivo. Somatostatin and analogs of somatostatin (such as octreotide and lanreotide) act by binding to somatostatin receptors (SSTRs), which are expressed on the majority of NETs [8]. The ability of octreotide and lanreotide to inhibit the secretion of peptides from NET cells is mediated mainly through SSTR-2 and SSTR-5.

The presence of SSTRs can be determined by diagnostic imaging using a radiolabeled somatostatin analog (indium-111 [111-In] pentetreotide [OctreoScan] or PET using gallium Ga-68 DOTATATE [Ga-68 DOTATATE] or gallium Ga-68 DOTATOC [Ga-68 DOTATOC]). While either imaging with 111-In pentetreotide or Ga-68 DOTATATE/Ga-68 DOTATOC can be used to assess SSTR status in patients with NETs, the higher sensitivity of Ga-68 DOTATATE and Ga-68 DOTATOC PET suggests that they may be the preferred option in certain clinical scenarios, particularly in patients with smaller tumor volume. In general, uptake of radiotracer by the tumor is predictive of a response to therapy with somatostatin analogs. However, in some cases (eg, miliary disease), diagnostic imaging may be negative even if SSTRs are present on the tumor. In such cases, a trial of somatostatin analog therapy can be considered even in the presence of a negative scan. (See "Metastatic well-differentiated gastroenteropancreatic neuroendocrine tumors: Presentation, prognosis, imaging, and biochemical monitoring", section on 'Somatostatin receptor-based imaging techniques'.)

Benefits

Patients with symptoms from hormone secretion — Patients with metastases from functioning pancreatic NETs (which account for 10 to 30 percent of all pancreatic NETs [9]) often become symptomatic from hormone hypersecretion rather than from tumor bulk. For those who have octreotide-avid disease, symptoms of hormonal excess can often be well controlled with somatostatin analogs.

In series that include patients with only pancreatic NETs or combined series of patients with functioning gastroenteropancreatic NETs (gastrointestinal NETs and pancreatic NETs), somatostatin analogs provide symptom control in over 60 percent [10-18]. However, symptomatic benefit is dose-related, and also differs according to the type of functioning NET:

●Somatostatin analogs are highly effective in controlling the symptoms associated with VIPomas (watery diarrhea) and glucagonomas (especially improvement in the characteristic rash, necrolytic migratory erythema) [19,20]. (See "VIPoma: Clinical manifestations, diagnosis, and management" and "Glucagonoma and the glucagonoma syndrome".)

●Efficacy has also been shown for somatostatinomas [21,22], although the effects are may be less dramatic than with VIPomas and glucagonomas. (See "Somatostatinoma: Clinical manifestations, diagnosis, and management", section on 'Somatostatin analogue'.)

●Insulinomas and gastrinomas represent the most common types of functioning pancreatic NET, but the role of somatostatin analogs in controlling hormone-related symptoms is somewhat limited:

•Efficacy of somatostatin analogs for insulinomas with hypoglycemia is unpredictable [23-28], and somatostatin analogs should be used with caution in this setting. Only approximately one-half of these tumors express SSTRs and derive benefit from such therapy. Furthermore, somatostatin analogs may paradoxically result in transient worsening of hypoglycemia, presumably due to simultaneous inhibition of glucagon secretion. In general, initial therapy for insulinomas often consists of dietary modification. Treatment with everolimus may improve glycemic control in patients with insulinoma. Treatment with diazoxide, which directly inhibits the release of insulin from insulinoma cells, is another option. (See "Insulinoma", section on 'Medical therapy to control symptomatic hypoglycemia'.)

•The role of somatostatin analogs in patients with hormone-related symptoms from gastrinoma is also limited. High doses of oral proton pump inhibitors are the treatment of choice, as they are able to effectively control the hypergastrinemia-related gastric acid overproduction. Somatostatin analogs may be helpful in refractory cases. (See "Management and prognosis of the Zollinger-Ellison syndrome (gastrinoma)".)

Currently available somatostatin analogs include octreotide and lanreotide. Highly symptomatic patients may be initiated on short-acting octreotide with rapid transition to a long-acting formulation and subsequent titration of dose to optimize symptom control while the LAR formulation is starting to take effect. However, a depot preparation (Sandostatin LAR) has largely eliminated the need for daily octreotide injections and is now considered a standard approach for symptomatic treatment of advanced NETs [29,30]. Many clinicians initiate therapy with the depot preparation for most patients and do not give an initial dose of short-acting octreotide. Sandostatin LAR is typically initiated at a dose of 20 mg IM monthly with gradual dose escalation as needed for optimal symptom control [31]. Patients may use additional short-acting octreotide for breakthrough symptoms while doses are being titrated; therapeutic levels of octreotide are not reached until 10 to 14 days after the initiation of the LAR injection.

Lanreotide, another long-acting somatostatin analog, can be administered once monthly using a deep subcutaneous injection and appears to have similar efficacy to octreotide [14-16].

Control of tumor growth — In addition to controlling symptoms associated with hormone hypersecretion, somatostatin analogs have also been shown to control tumor growth. Given the variable and sometimes indolent disease course of some pancreatic NETs, the optimal time to initiate treatment with a somatostatin analog in asymptomatic patients remains uncertain. We suggest initiation of a somatostatin analog in patients with unresectable, asymptomatic, well-differentiated pancreatic NET and a high tumor burden, an approach that is consistent with guidelines from the European Neuroendocrine Tumor Society (ENETS) [30], North American Neuroendocrine Society (NANETS) [32], and the National Comprehensive Cancer Network (NCCN) [33]. For patients with asymptomatic, advanced, unresectable pancreatic NET and small-volume disease, we suggest initial observation alone rather than early administration of a somatostatin analog. In such patients, we initiate somatostatin analog therapy if there is evidence of clinically meaningful tumor progression.

Past studies have indicated that few patients with advanced gastroenteropancreatic NET (<10 percent) have objective tumor shrinkage with somatostatin analogs [30,34-37], even when limited to subgroups with progressive disease at the time of treatment initiation [11-13,34,38-40]. However, more reports have demonstrated that, in addition to an improvement in symptoms, treatment with somatostatin analogs may be associated with disease stabilization and significant prolongation of progression-free survival (PFS) [41-43]. Whether somatostatin analogs also increase overall survival is not yet known, although a correlation between PFS and overall survival in patients with advanced NET treated with single-agent somatostatin analog therapy has been shown [44].

Median duration of PFS and overall survival depends on several factors, including primary tumor location, Ki-67 percent, extent of liver metastases, presence of bone and/or peritoneal metastases, and the presence of symptoms when initiating treatment. A nomogram has been developed to estimate PFS in patients with well-differentiated gastroenteropancreatic NETs based upon these and other factors [45].

The PROMID study, which was the first randomized study demonstrating an antiproliferative effect of somatostatin analogs in NET, included patients with metastatic small intestinal NET who were treated with octreotide LAR versus placebo [41]. Because the study was limited to patients with small intestinal NET, applicability of these results to patients with pancreatic NET was uncertain. (See "Metastatic well-differentiated gastrointestinal neuroendocrine (carcinoid) tumors: Systemic therapy options to control tumor growth".)

However, support for the antiproliferative effect of somatostatin analogs in pancreatic NET was provided by the phase III CLARINET trial, which compared lanreotide versus placebo in 204 patients with advanced well- or moderately differentiated, nonfunctioning gastroenteropancreatic NET, including both gastrointestinal NET and pancreatic NET (45 percent of all enrollees) [42]. Patients were randomly assigned to receive either 120 mg lanreotide Autogel (n = 101) or placebo (n = 103) every four weeks for 96 weeks or until progressive disease (PD) or death. The trial’s primary end point was PFS as determined by Response Evaluation Criteria in Solid Tumors (RECIST) criteria (table 2). All patients had avid disease on somatostatin-receptor scintigraphy. Most patients (96 percent) had no tumor progression by RECIST criteria in the three to six months before randomization; approximately one-half had pancreatic primary sites. Compared with placebo, there was a highly significant advantage in PFS with the use of lanreotide. At a time-point of two years following initiation of treatment, median PFS was not reached with lanreotide compared with 18 months with placebo (hazard ratio [HR] for progression or death 0.47; 95% CI 0.30-0.73). Estimated rates of 24-month PFS were 65 versus 33 percent. There were no significant differences in quality of life or overall survival. The most common treatment-related adverse effect was diarrhea (26 versus 9 percent in the lanreotide and placebo groups, respectively). Based upon these data, lanreotide has been approved in the United States for the treatment of patients with unresectable, well- or moderately differentiated, locally advanced or metastatic gastroenteropancreatic NET.

The mechanisms by which somatostatin analogs control tumor growth may result from direct antiproliferative effects, including cell cycle inhibition and pro-apoptotic effects, and indirect antiproliferative effects, including inhibition of tumor angiogenesis and the release of trophic growth hormones [46]. Octreotide LAR and lanreotide are first-generation agents that primarily target SSTR-2 and SSTR-5; benefit for the addition of the second-generation somatostatin analog pasireotide, which targets multiple SSTR subtypes, with greater binding affinity for SSTR-1, SSTR-3, and SSTR-5, and lower affinity for SSTR-2, could not be shown in a randomized phase II trial of everolimus with or without pasireotide [47].

Dose-escalated therapy — Dose escalation is an option at the time of initial disease progression on a long-acting somatostatin analog, but it is not our preferred option. The benefits of escalating the dose and/or frequency of a somatostatin analog for disease control are not well established, and clinical practice is variable. At least some data suggest the potential for prolonged periods of stable disease when the dose intensity is increased (eg, by using higher than standard doses or reducing the dosing interval) after an initial period of long-acting somatostatin analog use [48,49]. However, whether this reflects indolent biology rather than an intrinsic dose-response relationship is not clear. Furthermore, how this strategy compares with other systemic treatments for disease control after progression on a long-acting somatostatin analog (eg, peptide receptor radioligand therapy) is not known. (See 'Peptide receptor radioligand therapy' below.)

Side effects — Somatostatin analogs are usually well tolerated, and side effects are generally mild [50,51]. Approximately one-third of patients may have mild nausea, abdominal discomfort, bloating, loose stools, and fat malabsorption during the first weeks of therapy, after which time, symptoms tend to subside. Use of pancreatic enzyme supplements can ameliorate symptoms associated with pancreatic insufficiency. Mild glucose intolerance may occur due to transient inhibition of insulin secretion. Somatostatin analogs reduce postprandial gallbladder contractility and delay gallbladder emptying; up to 25 percent of patients develop asymptomatic gallstones or sludge within the first 18 months of therapy [51].

CYTOTOXIC CHEMOTHERAPY — Well-differentiated pancreatic NETs are clearly responsive to cytotoxic chemotherapy (table 3). For patients who are highly symptomatic from tumor bulk or who have rapidly enlarging metastases, we suggest chemotherapy rather than molecularly targeted therapy or a somatostatin analog because of the higher objective response rate. For most patients, we suggest the combination of capecitabine and temozolomide (CAPTEM) rather than temozolomide alone. In the absence of comparative trials, the choice of CAPTEM over a streptozocin-containing regimen should be individualized, taking into account the convenience of oral rather than intravenous treatment, performance status, and the anticipated side effect profile of both combinations.

Dacarbazine and temozolomide-based regimens — Dacarbazine is an alkylating agent (like streptozocin), with activity against pancreatic NET. In an Eastern Cooperative Oncology Group (ECOG) phase II trial of dacarbazine in 42 patients with advanced pancreatic islet cell carcinomas, the objective response rate was 33 percent [52]. As with streptozocin, the toxicity of dacarbazine-based regimens has limited their widespread use.

Temozolomide is a less toxic orally active analog of dacarbazine with activity in pancreatic NET [53]:

●In prospective studies, temozolomide has been combined with thalidomide, bevacizumab, or everolimus, with overall response rates of 24 to 45 percent (table 3) [54-56]. In these studies, temozolomide has generally been administered using a dose-intense regimen of 150 mg/m² daily for seven days on an every-other-week schedule. Prophylaxis for Pneumocystis jirovecii pneumonia is recommended in light of the lymphopenia associated with long-term use of temozolomide according to this schedule. The relative contribution of temozolomide and the agent used in combination with it to the observed antitumor activity is unclear.

●Retrospective series and preliminary reports of a randomized phase II trial have also reported promising activity with CAPTEM in patients with pancreatic NET (table 3):

•In a retrospective series of 143 patients who were treated with CAPTEM, the response rate was 54 percent [57].

•Benefit of CAPTEM relative to temozolomide alone was shown in the phase II ECOG-American College of Radiology Imaging Network (ACRIN) study 2211, which randomly assigned 144 patients with unresectable or metastatic pancreatic NET and progressive disease within the past 12 months to temozolomide alone (200 mg/m² daily on days 1 through 5 every 28 days) or capecitabine (750 mg/m² twice daily on days 1 through 14) plus temozolomide (200 mg/m² orally daily on days 10 through 14), also given in 28-day cycles [58]. On-protocol treatment was prespecified as 13 cycles, after which time patients could receive any treatment at investigator discretion. In a preliminary report presented at the 2018 American Society of Clinical Oncology (ASCO) meeting, at a median follow-up of 29 months, median progression-free survival (PFS), the primary endpoint, was significantly better with combined therapy (22.7 versus 14.4 months, hazard ratio 0.58, 95% CI 0.36-0.93). Median overall survival was 38 months for temozolomide alone and had not been reached for the combined therapy group. The objective response rate was higher for combined therapy (33 versus 28 percent, p = 0.47), as was the disease control rate (objective response plus stable disease; 82 versus 68 percent). Treatment-related grade 3 or 4 adverse effects were twice as common with combined therapy (44 versus 22 percent); the most common were neutropenia (13 versus 4 percent), nausea and vomiting (8 versus 0 percent), diarrhea (8 versus 0 percent), and fatigue (8 versus 1 percent).

Role of MGMT expression — As has been noted in patients with glioblastoma, there appears to be a correlation between expression of methylguanine DNA methyltransferase (MGMT) and temozolomide responsiveness in advanced NET. MGMT is an enzyme that is responsible for DNA repair induced by alkylating agent chemotherapy. (See "Initial treatment and prognosis of newly diagnosed glioblastoma in adults", section on 'Temozolomide'.)

●In one study, of 21 patients treated with temozolomide-based regimens, none of the 16 with intact MGMT expression (including all 13 carcinoid tumors) responded to treatment, while four of the five patients whose tumors lacked MGMT expression (all with pancreatic NET) had a radiologic response [59].

●A second larger retrospective series also reported an association between MGMT status and response to alkylating agents (including temozolomide, dacarbazine, and streptozotocin-based regimens) in patients with NET [60]. In this study, the association between MGMT status and response to alkylating agents was observed in patients with either gastrointestinal or pancreatic NET.

MGMT status, however, has not yet been prospectively validated as a predictor of response in NET, and variability in the techniques and criteria used to assess MGMT status preclude its current use as a routine clinical test to select patients for temozolomide therapy. (See "Metastatic well-differentiated gastrointestinal neuroendocrine (carcinoid) tumors: Systemic therapy options to control tumor growth", section on 'Dacarbazine and temozolomide'.)

Streptozocin combinations — Streptozocin-based combination therapy has been a historical treatment standard for patients with advanced pancreatic NET (table 3). Antitumor efficacy can be illustrated by the following data:

●In an early randomized trial, streptozocin plus doxorubicin had a combined biochemical and radiologic response rate of 69 percent and a median survival of 2.2 years [61].

●A retrospective analysis of 84 patients with either locally advanced or metastatic pancreatic NET treated with streptozocin, fluorouracil (FU), and doxorubicin and using current standard response criteria reported a 39 percent objective radiographic response rate and a median survival duration of 37 months [62].

●Another retrospective analysis of 96 patients with pancreatic NETs treated with streptozocin plus FU reported an objective response rate of 43 percent, and an additional 41 percent had stable disease as the best response [63].

While streptozocin-based regimens are clearly active in patients with advanced pancreatic NETs, widespread use has been limited by a relatively cumbersome administration schedule and by concerns about toxicity, which can include myelosuppression, nausea, hair loss, and renal dysfunction [64].

Oxaliplatin-containing regimens — Reports suggest antitumor activity for some oxaliplatin-based regimens in pancreatic NET [65,66]:

●A combined analysis of two phase II trials examining oxaliplatin-fluoropyrimidine chemotherapy plus bevacizumab in advanced NET suggests antitumor activity for these regimens. The analysis included a study examining the efficacy of oxaliplatin plus short term infusional FU and leucovorin (FOLFOX) with bevacizumab (n = 36) and another examining capecitabine in combination with oxaliplatin and bevacizumab (n = 40) [65]. The best overall responses based on Response Evaluation Criteria in Solid Tumors (RECIST) criteria in 12 patients with pancreatic NET treated with FOLFOX-bevacizumab included 50 percent with a partial response and 50 percent with stable disease. The outcomes in 16 patients with pancreatic NET treated with capecitabine with oxaliplatin and bevacizumab were 18.8 percent with a partial response and 69 percent with stable disease.

●The contribution of bevacizumab to these results is unclear; similar benefit has been shown with capecitabine plus oxaliplatin alone [66].

MOLECULARLY TARGETED THERAPY — For patients with advanced pancreatic NETs whose disease has progressed on a somatostatin analog and who are not symptomatic from tumor bulk that would benefit from cytoreduction, we suggest treatment with a molecularly targeted agent (everolimus or sunitinib).

Progress in the understanding of the molecular biology of pancreatic NETs has revealed elevated expression of several cellular growth factors and their receptors (including vascular endothelial growth factor [VEGF] and the VEGF receptor [VEGFR]), and involvement of the mammalian Target of Rapamycin (mTOR) in pancreatic neuroendocrine tumorigenesis [67-69]. Many receptors like VEGFR function as tyrosine kinases (TKs). The finding in preclinical models that disruption of VEGFR signaling pathways inhibits neuroendocrine cell growth has prompted a number of clinical trials evaluating small molecule TK inhibitors (eg, sunitinib, sorafenib, pazopanib, and cabozantinib) and monoclonal antibodies (MoAbs) that target VEGF in patients with advanced NET.

Studies have demonstrated antitumor activity associated with bevacizumab (an anti-VEGF MoAb), and several TK inhibitors that inhibit VEGFR, as well as the mTOR inhibitor everolimus. As has been seen with cytotoxic chemotherapy, these agents appear to be more active in pancreatic NET than in advanced gastrointestinal NET (carcinoid). (See "Metastatic well-differentiated gastrointestinal neuroendocrine (carcinoid) tumors: Systemic therapy options to control tumor growth".)

Small molecule antiangiogenic TK inhibitors

Sunitinib — Sunitinib is a multi-targeted tyrosine kinase (TK) inhibitor that has shown activity against a range of signaling pathways and growth factors/receptors including VEGFR 1, 2, and 3 as well as PDGFR alpha and beta, KIT, glial cell-line derived neurotrophic factor, RET, FMS-like tyrosine kinase-3 (FLT3), and colony-stimulating factor receptor (CSF-1R).

In an initial phase II trial, sunitinib (50 mg daily for four of every six weeks) was administered to 109 patients with advanced NET [70]. Of 61 patients with pancreatic NET, 11 (18 percent) had a partial response, and 68 percent had prolonged periods of stable disease; median time to tumor progression (TTP) was 7.7 months. Rates of symptom control for patients with functioning tumors and refractory symptoms were not reported.

Continuous administration of sunitinib (37.5 mg daily) was compared with placebo in a phase III trial of 171 patients with progressing pancreatic NET [71]. Accrual was stopped prematurely prior to the first preplanned interim efficacy analysis. Median progression-free survival (PFS) was significantly longer with sunitinib (11.4 versus 5.5 months). There were eight objective responses with sunitinib (versus none in the placebo group), two of which were complete. Hand-foot skin reaction and hypertension of any grade occurred in 23 and 26 percent of patients receiving sunitinib, respectively, and the most common grade 3 or 4 adverse events in this group were neutropenia (12 percent) and hypertension (10 percent). Despite these side effects, there were no differences in the quality-of-life index with sunitinib. Rates of symptom control for patients with functioning tumors and refractory symptoms were not reported. In a later report, median overall survival favored sunitinib (38.6 versus 29.1 months), but the difference was not statistically significant, potentially due to crossover from placebo to sunitinib in 69 percent of the control group [72].

Largely based upon these data, sunitinib was approved in the United States for the treatment of progressive, well-differentiated pancreatic NET in patients with unresectable, locally advanced, or metastatic disease. Side effects may include hypertension, proteinuria, and other forms of renal toxicity, arterial thromboembolism, left ventricular dysfunction, and clinical heart failure, thyroid dysfunction, bleeding, myelosuppression, hand-foot skin reaction, delayed wound healing, hepatotoxicity, and muscle wasting. These side effects and their management are discussed in detail elsewhere. (See "Cutaneous adverse events of molecularly targeted therapy and other biologic agents used for cancer therapy", section on 'VEGFR/PDGFR inhibitors' and "Toxicity of molecularly targeted antiangiogenic agents: Non-cardiovascular effects" and "Toxicity of molecularly targeted antiangiogenic agents: Cardiovascular effects".)

Although few data are available, sunitinib provided control of refractory symptoms due to tumor hormone production in two patients with VIPomas [73]. By contrast, blood glucose concentrations were not increased in a single patient treated with sunitinib for a functioning metastatic insulinoma [74]. Furthermore, worsening hypoglycemia in patients with insulinoma and development of hypoglycemia in patients with a previously nonfunctioning pancreatic NET related to hyperinsulinemia have been reported [74-76].

Sorafenib and pazopanib — Two other orally active antiangiogenic TK inhibitors, sorafenib and pazopanib, have demonstrated modest activity in pancreatic NET in phase II studies:

●Sorafenib, which targets VEGFR-2 and PDGFR-beta, was evaluated in 43 patients with pancreatic NET [77]. In a preliminary analysis, responses were observed in 9 percent of the 41 evaluable patients.

●Pazopanib, which targets VEGFR-1, VEGFR-2, VEGFR-3, PDGFR-alpha and beta, as well as KIT, was evaluated in a prospective study of 51 patients with advanced NET (29 with pancreatic NET and 22 with carcinoid) on stable doses of octreotide-LAR [78]. Patients received pazopanib at a dose of 800 mg daily. The response rate among patients with well-differentiated pancreatic NET was 22 percent.

Cabozantinib — Cabozantinib, which targets VEGFRs, MET, AXL, and RET, was evaluated in a prospective study of 61 patients with advanced NETs (20 pancreatic and 41 gastrointestinal NETs). Patients started at a dose of 60 mg of cabozantinib daily. In a preliminary report, the response rate was 15 percent for patients with pancreatic NETs, with an encouraging median PFS of 21.8 months [79]. The CABINET trial, a randomized placebo-controlled phase III study, is currently enrolling patients and will evaluate the efficacy of cabozantinib in patients with advanced pancreatic and gastrointestinal NETs after prior therapy with everolimus (NCT03375320).

Lenvatinib — Lenvatinib, which targets VEGFRs as well as fibroblast growth factor (FGF) receptors, was studied in the phase II TALENT trial of 111 patients with advanced NETs (55 pancreatic and 56 gastrointestinal) [80]. All of the pancreatic NET patients had progressed after being treated with a targeted agent (everolimus in 69 percent, sunitinib in 29 percent). The objective response rate was 44 percent for the pancreatic NETs, and the median duration of response was 19.9 months (range 8.4 to 30.8).

mTOR inhibitors

Everolimus — Several nonrandomized studies have explored the activity of everolimus, with and without octreotide [81-83]. The activity of everolimus (10 mg daily) in pancreatic NET was initially explored in an international multicenter phase II trial of 160 patients, 45 of whom also received treatment with concurrent octreotide at the discretion of the investigators [81]. Median PFS was longer in patients who received octreotide plus everolimus compared with everolimus alone (17 versus 9.7 months), but whether the addition of octreotide contributed to the higher PFS is unclear since the study did not randomly assign patients to receive octreotide.

Everolimus monotherapy (10 mg daily) was compared with best supportive care alone in the placebo-controlled RADIANT-3 trial of 410 patients with advanced progressing pancreatic NET [84]. Everolimus was associated with a significant prolongation in median PFS (11 versus 4.6 months, hazard ratio [HR] for progression 0.35, 95% CI 0.27 to 0.45). There were confirmed objective partial responses (as defined by Response Evaluation Criteria in Solid Tumors [RECIST] v1.0 (table 2)) in 5 percent of patients receiving everolimus versus 2 percent of the placebo group. Drug-related adverse events were mostly grade 1 or 2, and included stomatitis (64 versus 17 percent of the placebo group), rash (49 versus 10 percent), diarrhea (34 versus 10 percent), fatigue (31 versus 14 percent), and infections (23 versus 6 percent), predominantly of the upper respiratory tract. The most common grade 3 or 4 drug-related adverse events were stomatitis (7 percent), anemia (6 percent), and hyperglycemia (5 percent). In a later analysis, median survival favored everolimus (44 versus 37.7 months), but the difference was not statistically significant [85]. The high rate of crossover of patients from placebo to everolimus (85 percent) may have confounded the ability to detect a difference in overall survival.

Largely based upon these data, everolimus was approved in the United States for the treatment of progressive NET of pancreatic origin in patients with unresectable, locally advanced, or metastatic disease.

Everolimus causes hyperglycemia, particularly in those with preexisting hyperglycemia. In analysis of data from the RADIANT-3 trial described above, the frequency of severe (grade 3 or 4) hyperglycemia was higher in those with preexisting diabetes mellitus or baseline hyperglycemia (15 versus 3 percent in those without diabetes or baseline hyperglycemia) [86]. Interestingly, rates of grade 3 or 4 hyperglycemia were similar in those with glucagonoma versus those without glucagonoma (9.1 versus 7.8 percent). (See "Glucagonoma and the glucagonoma syndrome", section on 'Laboratory abnormalities'.)

Largely because of this effect, everolimus may be of particular value in patients with functioning insulinomas and refractory hypoglycemia [87-90]. In one report, four patients with malignant insulinoma and refractory hypoglycemia normalized their glucose levels during everolimus therapy; two had an objective antitumor response [87]. Clinical improvement in the other two patients who had stable disease as the best response suggests a possible direct effect of the drug on insulin production and/or release. (See "Insulinoma".)

Though rare, everolimus has been associated with serious, adverse events, including pneumonitis [90]; tolerance should be carefully monitored during therapy. (See "Pulmonary toxicity associated with antineoplastic therapy: Molecularly targeted agents", section on 'Everolimus'.)

Everolimus plus bevacizumab — Both VEGF pathway and mTOR inhibitors are active in pancreatic NET. The benefit of adding bevacizumab to everolimus was addressed in a phase II trial in which 150 patients with locally advanced or metastatic pancreatic NET were randomly assigned to everolimus alone (10 mg by mouth daily) or with concurrent bevacizumab (10 mg/kg IV every two weeks) [91]. In a preliminary report presented at the 2015 American Society of Clinical Oncology (ASCO) meeting, combination therapy was associated with significantly higher response rate (31 versus 12 percent) and superior PFS (16.7 versus 14 months), but no overall survival benefit. Combined therapy was also associated with higher rates of serious grade 3 or higher toxicity, including diarrhea (14 versus 3 percent), hyponatremia (11 versus 3 percent), and hypertension (41 versus 12 percent). This combination warrants further investigation.

Temsirolimus — In a phase II study of 37 patients with progressive NET treated with the mTOR inhibitor temsirolimus, only 1 of 15 patients with pancreatic NET had an objective response, but 67 percent attained disease control (which included stable disease for at least two months) [92]. Higher baseline tumor levels of mTOR predicted for better outcomes.

Encouraging early results were noted in a phase II trial of temsirolimus (25 mg IV weekly) plus the anti-VEGF monoclonal antibody bevacizumab (10 mg/kg every other week) in 56 patients with RECIST criteria progression within seven months of study entry [93]. A confirmed partial response was documented in 23 patients (41 percent), and 44 (79 percent) remained progression free at six months. The most common grade 3 or 4 toxic effects related to therapy were hypertension (21 percent), hyperglycemia (14 percent), fatigue (16 percent), neutropenia, and headache (7 percent each).

PEPTIDE RECEPTOR RADIOLIGAND THERAPY

Radiolabeled somatostatin analogs — Targeted radiotherapy using radiolabeled somatostatin analogs is a reasonable option for patients with disease that expresses somatostatin receptors (SSTRs) and has progressed on therapy including at least a somatostatin analog.

Traditional external beam radiotherapy is beneficial for patients with painful bone metastases but has little utility for the more common visceral metastases. There has been substantial interest, however, in targeted radiotherapy using systemic administration of radiolabeled somatostatin analogs [94-104]. The most frequently used radionuclides for targeted radiotherapy include yttrium-90 (⁹⁰Y) and lutetium-177 (¹⁷⁷Lu), which differ from one another in terms of emitted particles, particle energy, and tissue penetration [105-107].

Radioembolization using ⁹⁰Y-radiolabeled glass or starch microspheres is a different procedure that is a local ablative treatment administered via the hepatic artery. Use of radioembolization to treat advanced NETs is discussed elsewhere. (See "Metastatic gastroenteropancreatic neuroendocrine tumors: Local options to control tumor growth and symptoms of hormone hypersecretion", section on 'Technique and outcomes'.)

Most of the series reporting efficacy and toxicity with radiolabeled somatostatin analogs have included both pancreatic NET and gastrointestinal NET [105,106,108,109].

Yttrium-90 dotatoc — The most extensive experience with ⁹⁰Y dotatoc comes from a large single-institution series of 1109 patients with metastatic gastroenteropancreatic NET and disease progression within 12 months of study entry, with visible tumor uptake on pretreatment SSTR scintigraphy [108]. After the initial dose, additional treatment cycles were withheld if there was tumor progression or permanent toxicity; otherwise, patients were offered retreatment. The specific interval was not specified. The median number of courses administered was two (range 1 to 10).

Overall, 378 patients (34 percent) had a "morphologic" response (defined as any measurable decrease in the sum of the longest diameters of all pre-therapeutically detected tumor lesions by computed tomography [CT], magnetic resonance imaging [MRI], or ultrasound), 172 (15 percent) had a biochemical response (defined as any post-treatment decrease in a tumor marker that had demonstrated progression prior to enrollment), and 329 (29.7 percent) improved symptomatically. The median survival from diagnosis was 94.6 months. Longer survival correlated with responses by any of the above criteria. Transient grade 3 or 4 hematologic toxicities developed in 142 (12 percent), and loss of renal function was the main dose-limiting toxicity. In all, 103 patients (9 percent) had permanent grade 4 or 5 (fatal, n = 35) renal toxicity. Older age, low baseline glomerular filtration rate, and high kidney uptake score were associated with severe nephrotoxicity.

Lutetium-177 dotatate — At least some data suggest that ¹⁷⁷Lu dotatate outperforms ⁹⁰Y dotatoc [107]. In a registry-based series of 450 patients with pancreatic or gastrointestinal NET, 54 percent were treated with ¹⁷⁷Lu dotatate alone, 17 percent were treated with ⁹⁰Y dotatoc alone, and combined therapy was administered to 29 percent [107]. The median progression-free survival (PFS) for the entire group was 41 months (27 months with ⁹⁰Y dotatoc alone, 40 months for ¹⁷⁷Lu dotatate, and 50 months for combined therapy). The nonrandomized nature of this series precludes drawing conclusions regarding the relative efficacy of these approaches; however, long-term side effects of peptide receptor radioligand therapy (PRRT) may include loss of renal function, pancytopenia, and myelodysplastic syndrome, although rates of severe (grade 3 or worse) toxicity are reported to be low (1 percent or less) [107].

Efficacy results with ¹⁷⁷Lu dotatate alone are available from the following reports:

●In one series of 443 Dutch patients with gastroenteropancreatic or lung NET, the objective response rate was 39 percent of whom an additional 43 percent had stable disease [110]. At a median follow-up of 78 months, median PFS was 29 months, median time to tumor progression was 36 months, and median overall survival was 63 months. Response rates were particularly high in pancreatic NET, ranging from 52 percent for nonfunctioning tumors to 62 percent for functioning gastrinomas, insulinomas, and VIPomas. The safety analysis included 610 patients who had received a cumulative dose of at least 100 millicuries (3.7 gigabecquerels). Acute treatment-related toxicity included grade 3 or 4 thrombocytopenia in 5 percent, lymphopenia in 50 percent, and elevated aminotransferases in 3 percent. Long-term toxicity included acute leukemia in four patients (0.7 percent, three fatal) and myelodysplastic syndrome in nine (1.5 percent, five fatal). There was no therapy-related long-term renal or hepatic failure.

●In a meta-analysis of 15 studies in combined populations of gastrointestinal and pancreatic NETs, totaling 872 patients with metastatic disease, the pooled objective response rate following a course of ¹⁷⁷Lu dotatate was 28 percent (95% CI 21-35 percent), and the pooled disease control rate was 79 percent (95% CI 76-82 percent) [111]. Early adverse effects were minimal, including nausea, vomiting, fatigue, and hormonal disorders.

●In addition to these data, among patients with advanced midgut NET, the benefits of ¹⁷⁷Lu dotatate were shown in the NETTER-1 trial, which demonstrated significant improvement in objective response rate, PFS, and overall survival with ¹⁷⁷Lu dotatate compared with high-dose long-acting octreotide in patients whose disease had progressed on standard-dose somatostatin analog therapy (median PFS not reached versus 18 months) [109]. These data are discussed in more detail elsewhere. (See "Metastatic well-differentiated gastrointestinal neuroendocrine (carcinoid) tumors: Systemic therapy options to control tumor growth", section on 'Radiolabeled somatostatin analogs'.)

Largely based on data from the NETTER-1 trial, in January 2018, the US Food and Drug Administration (FDA) approved ¹⁷⁷Lu dotatate for the treatment of somatostatin-receptor-positive gastroenteropancreatic NET, including those arising in the pancreas [112]. The recommended dose is 7.4 gigabecquerels (200 millicuries) as an intravenous infusion over 30 minutes every eight weeks for a total of four doses [113].

The optimal selection of candidates for ¹⁷⁷Lu dotatate is not established. Guidelines from the European Neuroendocrine Tumor Society, which largely mirror the eligibility criteria for the NETTER-1 trial, are outlined in the table (table 4) [114]. We agree with these guidelines.

The limitations of ¹⁷⁷Lu dotatate at present include the complexity of administration, the lack of trials comparing this agent with other systemic therapies, and the lack of widespread availability.

Long-term toxicity — The most serious long-term toxicity associated with PRRT is irreversible myelotoxicity and therapy-related myeloid neoplasms, including myelodysplastic syndrome, acute leukemia, myeloproliferative neoplasms, or any type of myeloid neoplasm. The available data from studies with long-term follow-up suggest a rate of myelodysplastic syndrome of approximately 2 percent and a rate of acute leukemia of approximately 0.5 percent. Given the risk and the poor prognosis after a diagnosis of therapy-related myeloid neoplasms, clinicians should closely monitor patients with periodic complete blood counts (CBC) after PRRT [115]. We suggest obtaining a CBC with differential at least every six months, and prompt referral to a hematologist if abnormalities are detected. (See "Metastatic well-differentiated gastrointestinal neuroendocrine (carcinoid) tumors: Systemic therapy options to control tumor growth", section on 'Myelotoxicity and therapy-related myeloid neoplasms'.)

Should the somatostatin analog be continued? — The added benefit of combining a long-acting somatostatin analog with PRRT compared with PRRT alone is not established. All patients in the NETTER-1 trial continued receiving the long-acting somatostatin analog during and after PRRT. At least some data from a retrospective analysis of 168 patients treated at a single institution for unresectable gastroenteropancreatic NETs suggest that combined therapy is associated with significantly better median PFS (48 versus 27 months) and overall survival (91 versus 47 months) [109,116] compared with PRRT alone.

For patients with a functional NET, we continue therapy with a somatostatin analog during and after PRRT. For patients with a nonfunctional NET, we typically continue therapy with the somatostatin analog, although we consider stopping it for a patient with a nonfunctional NET whose disease is progressing unequivocally on somatostatin analog therapy. This approach is consistent with guidelines from the European Society for Medical Oncology [117].

Iobenguane I-131 — Benefit also has been suggested for systemic radionuclide therapy using iodine-131-labeled iobenguane (iobenguane I-131) in individuals with metastatic gastroenteropancreatic NET whose tumors express the norepinephrine transporter and take up iobenguane, as determined by scintigraphy using radiolabeled iobenguane [118]. There are different iobenguane nuclides used for therapeutic and diagnostic purposes. Iobenguane I-131 (therapeutic) is also known as metaiodobenzylguanidine (MIBG or ¹³¹I-MIBG); iobenguane expression is demonstrated by uptake on iobenguane I-123 (diagnostic) scintigraphy. While iobenguane I-131 has been licensed by regulatory authorities in some countries and is approved in the United States for treatment of pheochromocytoma/paraganglioma and neuroblastoma, in our view, its use for the treatment of gastroenteropancreatic NET remains investigational.

Iobenguane is a compound resembling norepinephrine that is accumulated by some NETs. In retrospective series, biochemical responses were observed in 37 percent of patients with gastroenteropancreatic NET treated with iobenguane I-131, objective radiographic responses were noted in 28 percent, and symptomatic improvement was reported by 27 of 48 patients (56 percent) [119].

In another report, iobenguane I-131 (therapeutic) was "beneficial" in 73 percent of 20 patients with metastatic gastroenteropancreatic NETs (including eight with pancreatic NETs) [120]. Treatment was judged beneficial if clinical status improved, laboratory tests for secreting tumors improved by >20 percent, tumor progression was halted, the size of the most significant localization had decreased by >25 percent, and the dose of analgesic and cold somatostatin analog therapy could be lowered. Treatment was well tolerated, and only one patient had severe pancytopenia. Notably, at 12 months, 7 of the 14 responders were again symptomatic.

IMMUNOTHERAPY — The role of immunotherapy with immune checkpoint inhibitors is just beginning to be studied in patients with well-differentiated NETs. Early data suggest that anti-programmed cell death 1 (PD-1) antibodies have minimal activity as single-agent therapy. As examples:

●The efficacy of the anti-PD-1 antibody spartalizumab (PDR001) was evaluated in a phase II multicenter trial that enrolled 116 patients, including 33 with pancreatic NETs, 32 with gastrointestinal NETs, 30 with thoracic NETs, and 21 with poorly differentiated gastroenteropancreatic neuroendocrine carcinomas [121]. In a preliminary report presented at The 2018 European Society for Medical Oncology (ESMO) meeting, the clinical activity of spartalizumab was notable in patients with thoracic NETs (partial response rate 20 percent) but appeared marginal in other cohorts, including well-differentiated pancreatic NETs (partial response rate 3 percent, stable disease rate 55 percent).

●Activity of pembrolizumab in patients with programmed cell death 1 ligand (PD-L1)-positive advanced NET was evaluated in the KEYNOTE-028 study, which enrolled 16 patients with pancreatic NETs and 25 patients with gastrointestinal (n = 7), lung (n = 9), or other site (n = 9) NETs. Only one patient with a pancreatic NET had an objective response (6 percent); the stable disease rate in the 14 patients who were evaluable for response was 88 percent [122].

The activity of pembrolizumab was also evaluated in the phase II KEYNOTE-158 trial, which included a cohort of 107 patients with well- and moderately differentiated NET whose disease had progressed or who were intolerant of one or more lines of standard therapy [123]. PD-L1 expression was present in 16 percent of patients. Primary sites of disease included the pancreas (n = 40), small intestine (n = 25), other gastrointestinal sites (n = 18), lung (n = 14), and other organs (n = 10). The overall response rate was low and included four partial responses (3.7 percent), and the median progression-free survival (PFS) was 4.1 months. Partial responses were noted in three patients with pancreatic NET and in one with an unknown primary; all responding tumors were PD-L1 negative. Two of the responding patients had a sustained response approaching two years.

Trials are ongoing to evaluate checkpoint inhibitors in combination with other immunomodulatory agents, including vascular endothelial growth factor (VEGF) pathway inhibitors [124-126]. Eligible patients should be encouraged to enroll.

SPECIAL CONSIDERATIONS DURING THE COVID-19 PANDEMIC — The COVID-19 pandemic has increased the complexity of cancer care. Important issues include balancing the risk from cancer treatment versus harm from COVID-19, minimizing the number of clinic and hospital visits to reduce exposure whenever possible, mitigating the negative impacts of social distancing on delivery of care, and appropriately and fairly allocating limited health care resources. These and other recommendations for cancer care during active phases of the COVID-19 pandemic are discussed separately. (See "COVID-19: Considerations in patients with cancer".)

SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Well-differentiated gastroenteropancreatic neuroendocrine tumors".)

SUMMARY AND RECOMMENDATIONS — Neuroendocrine precursor cells are distributed widely throughout the body, and neoplasms of these dispersed cells, which are termed neuroendocrine tumors (NETs), can arise at many sites. The World Health Organization (WHO) classifies all gastroenteropancreatic NETs into low-grade, intermediate-grade, and high-grade categories based on mitotic count and proliferative index (Ki-67) (table 1). Poorly differentiated neuroendocrine carcinomas are all high-grade tumors. (See "Pathology, classification, and grading of neuroendocrine neoplasms arising in the digestive system", section on '2010 and 2019 World Health Organization classification'.)

Well-differentiated gastroenteropancreatic NETs have been traditionally referred to as carcinoids when they arise within the tubular gastrointestinal tract and pancreatic NETs (islet cell tumors) when they arise in the pancreas or, in the case of gastrinoma, the proximal duodenum. Although gastrointestinal and pancreatic NETs are morphologically similar on routine histologic evaluation, they differ in terms of pathogenesis, biology, and response to therapy. Systemic treatment options for advanced gastrointestinal NETs are discussed in detail elsewhere. (See "Metastatic well-differentiated gastrointestinal neuroendocrine (carcinoid) tumors: Systemic therapy options to control tumor growth".)

●Initial therapy

•Most patients with symptoms of hormone hypersecretion from a pancreatic NET other than insulinoma or gastrinoma should be managed with somatostatin analogs and other agents, as appropriate to the specific syndrome. (See 'Benefits' above and "VIPoma: Clinical manifestations, diagnosis, and management", section on 'Somatostatin analogs' and "Somatostatinoma: Clinical manifestations, diagnosis, and management" and "Glucagonoma and the glucagonoma syndrome", section on 'General measures'.)

However, somatostatin analogs may worsen glycemic control in insulinoma. Initial therapy for insulinomas consists of dietary modifications. Treatment with diazoxide, which directly inhibits the release of insulin from insulinoma cells, can be considered, as can treatment with everolimus, which has been reported to be highly effective in improving glycemic control in patients with insulinoma. (See "Insulinoma", section on 'Medical therapy to control symptomatic hypoglycemia'.)

For patients with gastrinoma, high doses of oral proton pump inhibitors are the initial treatment of choice. Somatostatin analogs are may be helpful for refractory cases. (See "Management and prognosis of the Zollinger-Ellison syndrome (gastrinoma)".)

For asymptomatic patients with low-volume disease, we suggest observation alone rather than early administration of a somatostatin analog. In such patients, we initiate somatostatin analog therapy at the time of clinically meaningful disease progression. (See 'Control of tumor growth' above.)

•For most patients with hepatic metastases that are resectable with curative intent, in the absence of extrahepatic metastases, diffuse bilobar involvement, or compromised liver function, we suggest resection rather than medical therapy (Grade 2C). (See 'General approach to the patient' above and "Metastatic gastroenteropancreatic neuroendocrine tumors: Local options to control tumor growth and symptoms of hormone hypersecretion", section on 'Surgical resection'.)

●Progressive disease and/or symptomatic from tumor bulk

•For patients who require therapy because of progressive disease or symptoms related to hormone production, and who are not already receiving treatment with a somatostatin analog, we suggest initiation of therapy with a somatostatin analog (Grade 2A); for those who are already receiving treatment with a somatostatin analog, we suggest a molecularly targeted agent, either everolimus or sunitinib (Grade 2A). Somatostatin analogs, everolimus, and sunitinib have all been shown to improve progression-free survival (PFS) compared with best supportive care alone, although none have been compared directly with each other. In the absence of comparative trials, the choice of initial agent may be influenced by the expected toxicity profile. Given the favorable toxicity profile, we suggest a somatostatin analog as an appropriate first choice for many patients and reserve use of a targeted agent if there is subsequent disease progression (Grade 2C). Everolimus may be of particular value in patients with functioning insulinomas and refractory hypoglycemia because of its association with hyperglycemia. (See 'Small molecule antiangiogenic TK inhibitors' above and 'Everolimus' above.)

•For patients who are highly symptomatic from tumor bulk and who have rapidly enlarging metastases, we suggest chemotherapy as an initial treatment because of the higher objective response rate compared with other approaches (Grade 2C). For most patients, we suggest the combination of capecitabine and temozolomide (CAPTEM) rather than temozolomide alone (Grade 2B). In the absence of comparative trials, the choice of CAPTEM over a streptozocin-containing regimen should be individualized, taking into account the convenience of oral rather than intravenous treatment, performance status, and anticipated side effect profiles. (See 'Streptozocin combinations' above and 'Dacarbazine and temozolomide-based regimens' above.)

•Hepatic arterial embolization is a reasonable alternative approach to systemic therapy for patients with hepatic predominant disease, particularly those who are symptomatic, who are not candidates for surgical resection. (See "Metastatic gastroenteropancreatic neuroendocrine tumors: Local options to control tumor growth and symptoms of hormone hypersecretion", section on 'Hepatic arterial embolization'.)

•For patients with a somatostatin-receptor-positive pancreatic NET, peptide receptor radioligand therapy (PRRT) using a radiolabeled somatostatin analog such as lutetium-177 (¹⁷⁷Lu) dotatate is a reasonable option, either at the time of progression on a somatostatin analog or following progression on other agents, including everolimus, sunitinib, or cytotoxic chemotherapy. Although clinical trials are underway, there are no data yet specifically comparing ¹⁷⁷Lu dotatate with other therapeutic agents, and the choice of therapy in this situation should be based on availability of ¹⁷⁷Lu dotatate and patient preference. (See 'Radiolabeled somatostatin analogs' above.)

•Another option for patients with indolent disease after an initial response or a prolonged duration of stability with standard-dose somatostatin analog therapy is dose escalation of the somatostatin analog. (See 'Control of tumor growth' above.)