Cabozantinib Active in Neuroendocrine Tumors

In a single-arm phase II study, cabozantinib demonstrated clinical activity in patients with advanced carcinoid and pancreatic neuroendocrine tumors.

Jennifer A. Chan, MD

In a single-arm phase II study, cabozantinib demonstrated clinical activity in patients with advanced carcinoid and pancreatic neuroendocrine tumors (pNETs). Partial responses (PRs) were observed in 15% of each cohort treated with cabozantinib and stable disease was the best response in about two-thirds of patients.

Median progression-free survival (PFS) exceeded 20 months in the pNET cohort and was >30 months in the carcinoid tumor cohort, said Jennifer A. Chan, MD, at the 2017 Gastrointestinal (GI) Cancers Symposium.

“When you look at the other older phase II/phase III studies of other tyrosine kinase inhibitors, such as everolimus, the PFS durations are in the order of 8 to 16 months,” said Chan, clinical director, Neuroendocrine Tumor Program, Dana-Farber Cancer Institute. “We would obviously need to confirm what we see in a larger study.”

Preclinical works supported the study of cabozantinib, a dual inhibitor of VEGFR2 and c-MET, in neuroendocrine tumors. “There are data that suggest that MET may be a therapeutic target in neuroendocrine tumors,” Chan said. “There is some preclinical work showing that activation of MET is associated with growth in a xenograft model of pNETs. We’ve also seen expression of MET in pNETs in patients that we’ve treated at Dana-Farber. If you look at some of the preclinical work with cabozantinib in neuroendocrine cell lines, there’s inhibition of cell viability, and in mouse models of pNETs, cabozantinib can decrease metastases and decrease invasion.”

In the study she presented at the GI Symposium, 61 patients with progressive, well-differentiated, grade 1/2 carcinoid tumors (n = 41) or pNETs (n = 20) were enrolled. Patients had radiographic disease progression within 12 months of entry. A somatostatin analog was allowed if the dose was stable for 2 months. There was no limit to prior therapy, including anti-VEGF therapy. The median number of prior therapies was 3 in the pNET cohort and 1 in the carcinoid cohort. Prior therapies in the pNET group were sunitinib (60%), everolimus (65%), and temozolomide (55%), and in the carcinoid cohort, everolimus (29%), bevacizumab (15%), interferon (10%), and temozolomide (7%).

Patients received cabozantinib at 60 mg/day orally in 28-day cycles. Treatment continued until disease progression, unacceptable toxicity, or withdrawal of consent. Disease was restaged after every 2 cycles for the first 6 cycles, and then every 3 cycles.

Eighty-seven percent of patients completed at least 1 cycle of treatment. Of those who completed at least 1 cycle, 81% required dose reduction: the lowest cabozantinib dose was 60 mg in 10 patients, 40 mg in 24 patients, and 20 mg in 19 patients. The median number of cycles completed was 9. Eighty-two percent of patients reached the first reassessment point, after completion of at least 2 cycles of treatment.

The median number of treatment cycles completed was 10 in the pNETs cohort and 8 in the carcinoid tumors cohort. Median follow-up duration was 23.3 months. Twenty-five percent (5 patients) in the pNET cohort and 22% (9 patients) in the carcinoid cohort remained on treatment at the time of data analysis. The main reason that patients came off study was disease progression or death, experienced by 50% of pNET patients and 34% of carcinoid patients. Some 21% withdrew because of adverse events.

The primary endpoint was the objective response rate as measured by RECIST 1.1. “We stratified the analysis based on pancreatic tumors and carcinoid because we recognize that they’re different biologically,” said Chan. Treatment with cabozantinib was associated with objective tumor responses in both carcinoid tumors and pNETs.

Among the patients with pNETS, the best RECIST response was PR in 15% (3 patients) and stable disease in 75% (15 patients). In 2 patients (10%) in this cohort, response was unknown because treatment was stopped prior to restaging. Median PFS was 21.8 months (95% CI, 8.5-32.0). Among those with carcinoid tumors, best response was PR in 15% (6 patients), and 63% (26 patients) experienced stable disease. Two patients (5%) had disease progression. Response was unknown in 7 patients (17%) due to stopping treatment before restaging. Median PFS was 31.4 months (95% CI, 8.5 to not reached). Adverse events were consistent with those reported with the use of cabozantinib in other diseases. Grade 3/4 toxicity included hypertension in 13% of patients, hypophosphatemia in 11%, diarrhea in 10%, lymphopenia in 7%, thrombocytopenia in 5%, fatigue in 5%, and increased lipase or amylase in 8%.

Confirmation of cabozantinib activity in a randomized phase III trial in carcinoid tumors and pNETs is in development, Chan indicated. “We’re seeing signals in both, which is why we’re hoping to study it in both,” she said.

Invited discussant Vincent Picozzi, MD, director, Pancreatic Center of Excellence, Virginia Mason Medical Center, Seattle, said that the results are intriguing, but it’s uncertain how cabozantinib broadly impacts PFS or overall survival in neuroendocrine tumors.

He mentioned that given the wide confidence intervals around the PFS, and with respect to the overall response rate, “there seems to be a fairly substantial conclusion that rests on a small amount of change” in pNETs. In this group, 3 patients responded by RECIST, “but if we assume that each of the target lesions were all 5 cm in size, the total change in these 3 patients represents about a cm or less…and there’s a very large conclusion that’s made on a small amount of radiographic information.”

Chan J, Faris JE, Murphy JE, et al. Phase II trial of cabozantinib in patients with carcinoid and pancreatic neuroendocrine tumors (pNET). J Clin Oncol 35, 2017 (suppl 4S; abstract 228).


View more from the 2017 GI Cancers Symposium