E4206: AMG 706 and Octreotide in Patients with Low-Grade Neuroendocrine Tumors.

LESSONS LEARNED: Rate of progression-free survival at a particular point in time, i.e., a landmark analysis, is a difficult endpoint for a heterogenous malignancy such as neuroendocrine cancer.Landmark analyses can also be complicated by evolution in the standard of care during the conduct of a clinical trial.Improvements in biomarker development would be useful in developing future clinical trials in NET to better tailor individualized therapies and assess for possible efficacy endpoints.
BACKGROUND: Neuroendocrine tumors (NETs) are rare malignancies of the gastrointestinal (GI) tract that are highly vascularized and overexpress vascular-endothelial growth factor (VEGF). Sunitinib has demonstrated efficacy in the pancreatic subset of NET. This study explored the activity of another oral VEGF inhibitor, AMG 706 or motesanib, a multikinase inhibitor that targets receptor tyrosine kinases, including VEGFR1, VEGFR2, VEGFR3, KIT, RET, and PDGFR (IC50s = 2, 3, 6, 8, 59, and 84 nM, respectively).
METHODS: This was a single-arm, first-line, phase II study run through the Eastern Cooperative Oncology Group. Patients with low-grade NET (as defined by central confirmation of Ki-67 of 0%-2%) were administered a flat dose of 125 mg per day orally combined with octreotide long acting-repeatable (LAR) for patients who had been on a stable dose. The primary objective was to determine the 4-month progression-free survival (PFS).
RESULTS: Forty-four patients were evaluated per protocol. The 4-month PFS was 78.5%. The partial response rate was 13.6% (6/44), stable disease was 54.5% (24/44), 9.1% (4/44) had progressive disease, and 10/44 were not evaluable for response. Common toxicities included fatigue, hypertension, nausea, and headache, and most were grade 1-2. Median PFS was 8.7 months, and overall survival was 27.5 months.
CONCLUSION: Motesanib (AMG 706) demonstrated a 4-month PFS that met the per-protocol definition of efficacy. Fatigue and hypertension were the most common toxicities, and few grade 3-4 toxicities were encountered. The progression-free survival of 8.7 months in all NETs merits further study. © AlphaMed Press; the data published online to support this summary are the property of the authors.

Discussion

In this study, AMG 706 demonstrated the ability to prolong progression‐free survival for patients with advanced NET. The study met its primary endpoint with a 4‐month PFS of 78.5% and a median PFS of 8.7 months. The toxicity profile was compatible with earlier‐phase data and comparable to data from the studies investigating sunitinib and everolimus for NET. Although the response rate was low, it was greater than the 2% response seen in phase II studies with sunitinib in unselected NET populations. These data, as well as data from sunitinib and bevacizumab trials, provide proof of concept to support targeting the VEGF pathway as a treatment strategy in NET. The idiosyncratic toxicity of cholecystitis merits close consideration in future trials to sort out whether the toxicity (a known effect of somatostatin analogues) is attributable to AMG 706 or somatostatin analogue of choice.

When this trial was conceived, the phase III studies of sunitinib and everolimus in pancreatic NET had not yet been completed. The selection of 4‐month PFS as the primary endpoint in the study was based on historical controls prior to 2008. In retrospect, the study would have been more impactful if it had been powered to truly assess the PFS rather than PFS at a predefined time point. The PFS of 8.7 months in this study is reasonable in comparison with historical controls but would not be enough for consideration as a first‐line agent replacing sunitinib or everolimus. Additionally, the PFS estimate is hindered by the fact that many of the patients were receiving octreotide, which altered the rate of progression. This improvement in PFS was not well described prior to the activation of this trial. As a standard of care, based on RADIANT‐2 and RADIANT‐4, everolimus would be the preferred first‐line agent in this population, and any future trial design with a head‐to‐head comparison would have to include everolimus as a control arm.

The treatment paradigm for NET is in flux. Many patient‐specific (symptom burden) and tumor‐specific factors (primary tumor location, overexpression of somatostatin receptors) will play into the selection of somatostatin analogue versus oral targeted agents versus peptide receptor radiotherapy. Novel oral agents like AMG 706 will need to find their therapeutic niche with these factors in mind. By meeting its primary endpoint in this trial, AMG 706 demonstrated potential as a systemic targeted therapy for NET, but where it fits in a treatment algorithm remains unclear. Opportunities for this compound include a second‐line treatment against best supportive care alone or combination therapy with mTOR inhibitors. The series of real but modest gains in NET underscores the need for ongoing clinical trials in neuroendocrine tumors of any origin with patient‐ and tumor‐specific factors in mind.

Trial Information

Neuroendocrine – other

Metastatic/advanced

None

Phase II

Single arm

Progression‐free survival

Overall response rate

Overall survival

Toxicity

Active but results overtaken by other developments

Drug Information

Motesanib, AMG 706

Amgen

Small molecule

Angiogenesis – VEGF

125 milligrams (mg) per flat dose

Oral (po)

Daily

Patient Characteristics

24

20

Metastatic

Median (range): 65 (38–81) years

Median (range): 0

0 — 29

1 — 14

2 — 1

3 —

Unknown —

GI neuroendocrine tumors

Pancreatic neuroendocrine tumors

Primary Assessment Method

Total Patient Population

46

46

45

44

RECIST 1.0

n = 0 (0%)

n = 6 (14%)

n = 24 (55%)

n = 4 (9%)

n = 10 (22%)

9 months, CI: 6–13

28 months, CI: 14–45

4‐month PFS (defined as the primary endpoint) was 78.3% (95% CI: 65.8%–90.9%)

Assessment, Analysis, and Discussion

Study completed

Active but results overtaken by other developments

Low‐grade neuroendocrine tumors (NETs) are increasingly common, and the treatment paradigm has evolved over the last few years. Imaging and pathologic data suggest that NETs are highly vascular and overexpress vascular‐endothelial growth factor (VEGF) [1]. Phase II data suggested that bevacizumab and octreotide have an antitumor effect, but an improvement in survival was not seen in the phase III study [2], [3]. Somatostatin analogues have demonstrated improvements in progression‐free survival (PFS), but more potent antitumor agents were needed [4]. Sunitinib showed a response rate and survival benefit in pancreatic NET, leading to its approval, but a similar response was not seen in nonpancreatic NET [5], [6]. Everolimus has demonstrated improvements in overall survival in all gastrointestinal (GI) NET and has been approved for use in both settings [7], [8], [9]. We designed this trial to assay for a preliminary efficacy signal in GI NET using motesanib (AMG 706).

Our study met its primary endpoint of improving 4‐month progression‐free survival (78.5%), with a partial response rate of 13.6%. The treatment was well tolerated, with manageable side effects. However, it has not been moved into phase III studies, as it has been surpassed by other treatments. This study reflects some of the challenges of performing research in NET.

Our study was designed in 2007, before the differential tumor biology in pancreatic NET and nonpancreatic NET relative to its response to targeted agents was well described. Thus, the population was heterogeneous enough to confound the survival results. This study was designed with a short, predefined 4‐month PFS interval and a two‐stage design in the hopes of discarding an inefficacious treatment quickly but assaying for a potential signal in the combined population. That 4‐month time interval was based on rates of progression from the pre‐somatostatin analogue era, but perhaps did not provide a robust enough signal of efficacy. Additionally, some of the enrolled patients were lost to follow‐up for a lack of a confirmed response on imaging or coming off treatment too soon. If we were to redesign the trial, we would reassess our response evaluation to include these patients in an intention‐to‐treat analysis.

The hope is that with the advent of peptide‐receptor radiotherapy, the use of targeted agents can fit in the treatment algorithm to improve survival. Future trial design will look into the optimal sequencing of treatment for patients with NET between long acting somatostatin analogues, peptide receptor radiotherapy, and other chemotherapeutic agents. Additionally, for patients with tumors that do not overexpress somatostatin receptors, targeted agents like motesanib may have a role.

Adverse events observed at least once in 20% or more patients across all cycles of therapy.

Abbreviation: NC/NA, no change from baseline/no adverse event.