Chemotherapy

On average, carcinoid tumors seem to be less sensitive to chemotherapy than pancNETs.⁵ Single-agent capecitabine showed modest activity in a recent phase II study, with 13 of 19 patients experiencing stable disease (lasting > 1 year in 4 patients).¹⁶ Streptozocin-based combinations have been studied extensively, with response rates ranging from 16% to 33%. In a rare phase III trial, streptozocin/5-FU was compared with doxorubicin/5-FU. The response rate was 16% in both arms, but a statistically significant improvement in overall survival (24.3 vs. 15.7 months) was noted in the streptozocin-containing arm. Importantly, 35% of patients on the streptozocin-containing arm developed mild to moderate renal toxicity, suggesting that the risk/benefit ratio of this regimen is not particularly favorable.¹⁷

Temozolomide-based regimens have also been studied in carcinoid. Limited data suggest a possible benefit in bronchial carcinoids (4 of 13 patients with partial responses)¹⁸; however, other retrospective data suggest limited activity in carcinoid overall (1 of 44 patients with partial response).¹⁹ Oxaliplatin-based chemotherapy regimens have also been studied, but randomized data are lacking and the radiographic response rate seems to be relatively low in carcinoid.²⁰

Collectively, the data suggest that traditional chemotherapy is of limited value in carcinoid. Radiographic responses (e.g., according to Response Evaluation Criteria in Solid Tumors [RECIST]) are rare, and the value of stable disease is questionable in a disease that often progresses slowly even without treatment. No standard chemotherapy exists, and generally its use is restricted to patients with progressive, octreotide-refractory disease and no other treatment options.

Somatostatin Analogs

Somatostatin (SST) inhibits multiple glandular and exocrine secretions, particularly gastrointestinal secretions, such as serotonin, gastrin, and insulin. In addition to this effect on hormonal release, preclinical studies of its proapoptotic and antiproliferative functions suggest potential as a targeted therapeutic for NETs.²¹ Five G-coupled SST receptors have been identified (sstr_1–5), each with varying tissue- and receptor-specific activity. They are highly expressed on NETs regardless of functionality, although the subtype expression patterns vary widely.²¹

Somatostatin analogs (SSTa) have a proven role in the treatment of carcinoid syndrome because of their ability to bind these receptors and block the release of bioactive peptides and amines.^22,23 In the United States, octreotide is the only SSTa approved for this indication, with 2 preparations available: immediate and long-acting release (LAR). Octreotide binds sstr₂ and sstr₅ with high affinity (showing significantly less affinity for the remaining 3 receptors). The depot (LAR) form of octreotide allows for monthly, rather than daily, injections. It is currently FDA-approved for the control of symptoms related to carcinoid syndrome. Notably, lanreotide autogel, an analog with a similar receptor binding profile, is used worldwide for a similar indication but is not approved for use in the United States.

In addition to its role in symptom management, octreotide has been shown to have antitumor activity. The PROMID study was a double-blind, prospective, randomized trial of octreotide LAR versus placebo in previously untreated NETs of midgut origin.¹⁰ Interim results showed that octreotide LAR extended time to progression compared with placebo (14.3 vs. 6 months, respectively; P = .000072). Notably, 95% of patients had tumors with Ki-67 values of 2% or less; 75% had a hepatic tumor burden of 10% or less. No differences in overall survival were identified at analysis (33 of 43 placebo patients received octreotide LAR in the post-study setting), and the radiographic response rate (according to WHO criteria) was less than 3% in both arms.

The results indicate that octreotide LAR shows antitumor activity in advanced, well-differentiated midgut tumors (especially in the absence of bulky hepatic disease). The PROMID study is important because it provides proof-of-principle, showing that octreotide does, in fact, have cytostatic activity in midgut NETs. In the absence of an overall survival benefit, however, the results do not provide insight into the optimal timing of therapy (at diagnosis or after evidence for progression). Furthermore, the efficacy of octreotide in patients with bulky hepatic disease or NETs originating outside the midgut remains unclear.

Recognizing these limitations, the NCCN Clinical Practice Guidelines in Oncology (NCCN Guidelines) for Neuroendocrine Tumors include octreotide LAR as a treatment option for patients with progressive metastatic carcinoid of any primary site, well- to moderately differentiated NETs of unknown primary, and pancNETs (available in this issue; to view the most recent version of these guidelines, visit NCCN.org).¹⁴ Importantly, lanreotide autogel is currently under investigation in a phase III study (CLARINET) assessing antitumor activity in patients with nonfunctioning NETs of the pancreas, hindgut, and midgut (ClinicalTrials.gov identifier: NCT00353496). The results of this study should complement the PROMID data, allowing a deeper understanding of the antitumor activity of SSTa in NETs as a whole. For now, considering the available data and the cost and the relative inconvenience of monthly octreotide LAR injections, deferring initiation of therapy until disease progression is a reasonable option in an asymptomatic patient (as per NCCN Guidelines for Neuroendocrine Tumors; in this issue and at NCCN.org).

Beyond octreotide and lanreotide, SSTa with varying receptor binding profiles are currently in development. Hybrid somatostatin/dopamine compounds and nonpeptide analogs are also being investigated.²² Pasireotide (SOM230) is an SSTa with high affinity for sstr_1–3 and sstr₅. Relative to octreotide, it has slightly less affinity for sstr₂ but greater affinity for the remaining receptors. Phase II data suggest that 27% of patients for whom standard therapy fails experience symptom control on pasireotide.²⁴ A phase III trial to assess symptom relief in patients with octreotide-refractory carcinoid is ongoing, as are studies to assess antitumor activity (Table 1).

Peptide Receptor Radionuclide Therapy

Radiolabeled octreotide has been used for years as a diagnostic tool (I131-octreotide scintigraphy). More recently, peptide receptor radionuclide therapy with radiolabeled SSTa has been developed for therapeutic use. In patients with SST receptor–expressing tumors, this strategy can target radiation directly to advanced NETs. Clinical trials with radionuclide therapy for carcinoid have primarily focused on 2 radiolabeled SSTa: [¹⁷⁷Lu-DOTA⁰,Tyr³]octreotate (¹⁷⁷Lu-DOT-ATATE) and [⁹⁰Y-DOTA]-D-Phe¹-Tyr³-octreotid e (⁹⁰Y-DOTATOC).^11,25,26 Radiographic response rates as high as 30% have been observed in phase II trials. Notable adverse events include cytopenias and renal toxicity. Although encouraging, the results must be interpreted with caution, because prospective, randomized, controlled studies are lacking. Importantly, the development of this therapeutic strategy has been limited by issues related to drug availability.²⁷ Therefore, the use of this modality is currently limited to selected centers with the expertise to generate and administer the radiolabeled SSTa.^5,27 Additional studies are needed to definitely establish the safety and efficacy of radiolabeled SSTa. Accessibility also must be improved for this treatment to become routine in the United States.

Interferon

Interferon is another targeted therapy that has activity in carcinoid tumors. Interferon alpha-2a and -2b act directly on NET cells through binding to interferon receptors on their surfaces, resulting in altered gene expression and secretory function.²⁸ Several studies suggest that interferon can provide significant relief of symptoms, with biochemical response rates as high as 30% to 35%.^11,28 The data also suggest the potential for synergy when used in combination with an SSTa.²⁹ However, use in the first-line setting is limited by its side effects (including myelosuppression, depression, and fatigue) relative to octreotide alone.³⁰ Interferon is approved for the treatment of carcinoid syndrome in Europe but not in the United States.¹²

Table 1

Selected Ongoing Trials of Somatostatin Analogs in Carcinoid Tumors

View Table

In addition to its effects on hormone-mediated symptoms, interferon is believed to carry some degree of antitumor activity, through direct action on tumor cells and immunomodulation.²⁸ Small studies have shown trends towards improved time to progression.²⁹ However, a randomized trial of octreotide with or without interferon alpha in patients with progressive, metastatic NETs showed no significant differences in time to treatment failure or survival. Although several patients experienced stable disease, radiographic responses were infrequent (2%–5%).³¹ Therefore, use of interferon as an antitumor strategy remains controversial, and is typically reserved for patients with SSTa-resistant disease.^11,14,29 Notably, treatment with interferon is the control arm for an ongoing randomized phase III study exploring the use of bevacizumab in patients on octreotide (SWOG 0518 [ClinicalTrials.gov identifier: NCT00569127]; Table 2).

VEGF Pathway

In well-differentiated NETs, increased VEGF, a key regulator of angiogenesis, and VEGF receptor expression have been linked to metastases, angiogenesis, and decreased progression-free survival.³² Preclinical models provide support for the notion that VEGF is a valid target for therapy in NETs.³³ Numerous inhibitors have been approved for other indications, including the anti-VEGF antibody, bevacizumab, and the oral VEGF receptor tyrosine kinase inhibitors (TKIs), sunitinib, sorafenib, and pazopanib. Sunitinib, an oral inhibitor of c-KIT, VEGFR-, and PDGFR-signaling, was recently approved for use in progressive pancNETs based on improved progression-free survival.⁶ The results in pancNETs are encouraging, and underscore the potential value of the VEGF pathway as a therapeutic target in carcinoid.

In a randomized phase II study by Yao et al.,³⁴ 44 patients with advanced carcinoid received 18 weeks of bevacizumab or pegylated interferon alpha-2b. All patients were on a stable dose of octreotide. Up-front bevacizumab was associated with a higher response rate (18% vs. 0%), reduction of tumor blood flow by functional CT, and improved progression-free survival at 18 weeks compared with the interferon arm. These results prompted an ongoing phase III trial in patients with high-risk/poor-prognosis carcinoid (SWOG 0518; Table 2). Pilot studies of bevacizumab combined with chemotherapy, RAD001, or other agents have been performed, but no combination regimen has proven efficacy in NETs.^35,36

The activity of several different TKIs in NETs has also been evaluated in phase II trials (Table 3), with results consistently showing low radiographic response rates (≤ 10%) in patients with carcinoid tumors.^7,37 Despite the lack of significant tumor shrinkage, the fact that disease stabilization is observed (6-month progression-free survival ranging from 40%–73% across studies) suggests a potential cytostatic effect. Enrollment to the carcinoid stratum of a phase II trial of sunitinib was stopped because of insufficient evidence of radiographic response after the first stage.⁷ In retrospect, 82.9% of patients with carcinoid tumors experienced stable disease and 43.9% of patients showed at least some degree of tumor shrinkage (median time to progression, 10.2 months). The results hint at underlying antitumor activity, despite the fact that the prespecified radiographic response rate was not met.

Table 2

Selected Current Trials of Novel Agents/Combinations in Carcinoid Tumors

View Table

Collectively, the data suggest that disease stabilization (e.g., progression-free survival) may be a better end point than radiographic response when evaluating VEGF inhibitors in carcinoid. However, because the disease has the potential to progress slowly (e.g., median progression-free survival of 6–9 months in untreated patients),¹⁰ progression-free survival data should be interpreted with caution in the absence of adequate controls. The use of a control arm and the requirement for radiographic progression at baseline may improve one’s ability to assess antitumor activity when progression-free survival is the primary end point.⁵ At a minimum, stable disease (< 20% growth according to RECIST criteria) will be a more convincing indicator of drug activity in carcinoid if the study population is experiencing progression at baseline.

Table 3

Summary of Results of Clinical Trials With VEGF/mTOR Pathway Inhibitors in Carcinoid Tumors

View Table

mTOR Pathway Inhibitors

The results of several recent studies suggest that mTOR is also an attractive target in NETs and may be important in carcinoid (Table 3).³⁸ The mTOR pathway lies at the center of a complex network of growth factor pathways active in NETs, including the VEGF and insulin-like growth factor 1 (IGF-1) signaling pathways.³⁹ An analysis of primary pancNETs by Missiaglia et al.⁴⁰ revealed downregulation of phosphatase and tensin homolog (PTEN) and tuberous sclerosis-2 (TSC2) products, both of which are key inhibitors of the mTOR pathway. The precise mechanisms underlying the alterations of PTEN and TSC2 activity are unclear, but inactivating mutations have been identified in a subset of patients.⁴⁰ In well-differentiated carcinoid tumors, mTOR pathway components are frequently activated via hyperphosphorylation; similar activation profiles were not identified in poorly differentiated tumors.^41,42 Interestingly, activation of the mTOR pathway was strongly correlated with SST receptor expression in one such study.⁴¹

Similar to sunitinib, treatment with everolimus (an oral inhibitor of mTOR/TORC1, also known as RAD001) improves progression-free survival in patients with progressive pancNET (4.6 months with placebo vs. 11.0 months with RAD001; hazard ratio [HR], 0.35; log-rank P < .0001).⁸ However, the benefits of everolimus in carcinoid are less clear. In the RADIANT-2 trial, 429 patients with a history of symptoms attributed to carcinoid syndrome and evidence for radiographic progression within 12 months were randomized to receive either everolimus plus octreotide LAR or placebo plus octreotide LAR.⁷ The primary end point of the study was progression-free survival through adjudicated central review (which was not performed in real-time). Although a progression-free survival benefit in favor of everolimus was seen (HR, 0.77; P = .026), the predefined statistical boundary (P = .0246) was not met.^5,9 Importantly, the discordance between the central and investigator (local) review resulted in several patients being censored for the final analysis, leading to a loss of events and loss of statistical power for the primary end point. No impact on overall survival was noted (crossover from placebo to everolimus was allowed).

Everolimus is not approved for use in carcinoid, although additional trials with everolimus and related agents are ongoing and/or planned (including trials with everolimus combinations and TORC1/2 inhibitors). RADIANT-4 is slated to assess the activity of single-agent everolimus in nonfunctional tumors of gastrointestinal or lung origin in the context of a randomized phase III placebo-controlled study (ClinicalTrials.gov identifier: NCT01524783). Importantly, the current NCCN Guidelines for Neuroendocrine Tumors include everolimus as a category 3 recommendation, recognizing the lack of available therapies in this disease (available in this issue; to view the most recent version, visit NCCN.org).

Future Targets/Trials

Numerous new targets have been identified in carcinoid, which could lead to novel single-agent or combination regimens (e.g., with SSTa, VEGF- and/or mTOR-pathway inhibitors).^43,44 Although still in the pilot stages, several agents targeting components of growth factor signaling pathways are under investigation, including AKT, TORC1/2, and IGF-1R (Table 2).⁴⁵ In many cases, a rationale for combining agents can be readily identified (e.g., use of an AKT inhibitor to abrogate the feedback activation of the pathway that occurs with an isolated TORC1 inhibitor). Importantly, novel targets are also being explored in the context of symptom management. For example, blocking tryptophan 5-hydroxylase has emerged as a potential means for decreasing serotonin production in patients with carcinoid syndrome.⁴⁶ Preliminary results are encouraging, suggesting that future trials are warranted.

Challenges Related to Therapy Development in Carcinoid

Several barriers exist to identifying novel therapies for carcinoid tumors. Their rarity alone makes studying large numbers of agents and/or combinations difficult. The inherent biologic heterogeneity (e.g., presence or absence of carcinoid syndrome, site of the primary tumor, proliferative index, grade) augments this problem. Furthermore, although the prevalence of carcinoid tumors is relatively high, not all patients with the disease are appropriate for therapy. With a disease course measured in years, if not decades, attention must be paid to the appropriate timing of intervention. Recent trials have tried to address this problem through restricting eligible patients to those with evidence of radiographic progression.^6,8,9 Additional considerations should include the use of appropriate end points for trials (e.g., considering progression-free survival over response rate) and the incorporation of proper controls (e.g., randomized phase II trials). Studies are also needed to clarify the relationship between site of origin and underlying genetic changes, because this information could lead to the identification of novel targets and further refinements to the selection of patients for systemic therapy.^47,48 Moving forward, advances in molecular classification of carcinoid tumors may enhance the ability to identify patients most likely to benefit from therapy.

Concurrent or prior octreotide use is another potential confounder in carcinoid trials. Having established the fact that octreotide has antitumor activity, investigators must account for SSTa treatment in clinical trials. Common practices include stratification for concurrent octreotide, stratification for prior octreotide, or the requirement of an SSTa in all treatment arms. Octreotide LAR was required in both arms of the RADIANT-2 trial, and 80% of patients had received prior treatment with an SSTa.⁹ A subsequent exploratory analysis indicated that octreotide-naïve patients experienced better outcomes than those who had been previously treated, underscoring the need to account for SSTa use in the design of future trials.

A final barrier to identifying agents with antitumor activity stems from the fact that carcinoid tumors can be inherently difficult to image with standard cross-sectional imaging.⁴⁹ They can also grow extremely slowly, so that changes in tumor size may only be evident over many months (not weeks).^5,39 Even in the absence of dedicated antitumor therapy, many patients remain without evidence of overt radiographic progression for months. The discordance between the central and investigator reviewer interpretations in the RADIANT-2 study highlighted these challenges and called attention to the problem of informative censoring. In the future, real-time central review or identification of more reliable criteria for assessing response (compared with RECIST 1.1) may help eliminate this issue when evaluating cytostatic agents in carcinoid.⁵

Summary

Carcinoid tumors are generally slow-growing, and prolonged survival can be expected in most patients with optimal surgical management and symptom control. Octreotide is indicated for the treatment of hormone-related symptoms and delays progression in advanced small bowel carcinoid. Therefore, octreotide LAR is an obvious first choice when considering systemic treatment options for patients with advanced disease.¹⁴ Although often used in other well-differentiated NETs, its therapeutic value outside the setting of small bowel carcinoid has not been established. Cytotoxic chemotherapy and interferon are typically reserved for selected patients with progressive tumors given their limited efficacy and relatively unfavorable side effect profiles. The mTOR inhibitor everolimus is approved for use in pancNETs and showed a trend toward improved progression-free survival in a recent phase III trial in patients with a history of carcinoid syndrome (but is not approved for this indication). As such, there is a serious unmet need for approved systemic treatment options. Additional studies with everolimus (alone and in combination with other agents) are planned or ongoing, as are trials with VEGF inhibitors and other novel agents. Moving forward, trials should be undertaken with careful consideration of efficacy end points, criteria for response assessment, SSTa use, and patient selection (recognizing that carcinoid has the potential to be indolent even in the absence of treatment).