FDA's Oncologic Drugs Advisory Committee is being asked to weigh the impact early termination of a pivotal study had on the risk/benefit assessment of Pfizer's Sutent (sunitinib) in pancreatic cancer.

At an April 12 meeting, the agency seeks ODAC's views on whether sunitinib's risk/benefit profile in pancreatic neuroendocrine tumors is favorable given that early termination of the Phase III study created uncertainty about the magnitude of improvement in progression-free survival. FDA is also framing the question against a backdrop that includes the absence of a significant benefit on overall survival and multiple drug-related toxicities.

FDA meeting briefing documents reflect the agency's growing discomfort with reliance on PFS as an endpoint in cancer drug approvals absent a clear benefit on either overall survival or quality-of-life factors.

Some of FDA's other concerns about the Sutent application also mirror issues that have arisen with other recent oncology product reviews, including the impact of early pivotal study termination as well as whether a drug's toxicities are potentially worse than the disease itself, particularly for conditions with long periods of indolence.

Moving Beyond Renal And GI Cancers

Pfizer is seeking clearance for Sutent's use in unresectable pancreatic neuroendocrine tumors (PNET). Approval would mark the third cancer indication for the tyrosine kinase inhibitor, which is already labeled for advanced renal cell carcinoma and gastrointestinal tumors. Having failed in several other settings, PNET would appear to offer the last hope for expanding sunitinib beyond the renal and GI cancer settings, although the drug is being studied as adjuvant treatment for RCC (Also see "Pfizer Clears Out Pipeline: After Oncology Disappointments, Talks Up Diabetes" - Pink Sheet, 4 October, 2010.).

Sutent is one of two drugs ODAC will consider for pancreatic tumors at its meeting. Earlier in the day the committee will review Novartis' Afinitor (everolimus) for neuroendocrine tumors of gastrointestinal, lung or pancreatic origin (Also see "Discordant Trial Results Among Issues FDA Bringing To Advisory Panel Review Of Novartis' Afinitor" - Pink Sheet, 8 April, 2011.).

PNET, also known as well-differentiated pancreatic islet cell tumors, is a rare disease, accounting for approximately 1.3% of all pancreatic cancers by incidence, FDA noted. Though typically slow growing, most PNETs are malignant and metastasize commonly to lymph nodes and the liver, and less commonly to bone, lung, brain and other organs. The 10-year survival rates are estimated at 58% for localized disease, 46% for regional disease and 11% for distant disease, FDA said.

There are two FDA-approved systemic treatments for PNET. Teva Parenteral Medicines' Zanosar (streptozocin) is indicated for treatment of metastatic islet cell carcinoma of the pancreas. Labeling states that due to the drug's inherent renal toxicity, treatment should be limited to patients with symptomatic or progressive metastatic disease. Novartis' Sandostatin (octreotide acetate) is indicated for symptomatic treatment of diarrhea and flushing associated with carcinoid syndrome and vasoactive intestinal peptide tumors.

An Early End To The Phase III Study

The Phase III trial (A6181111) for sunitinib in PNET was a randomized, double-blind study of patients with locally advanced or metastatic well-differentiated pancreatic islet cell tumors that had progressed on a scan within two to 12 months prior to baseline.

Patients were treated until progression by RECIST (Response Evaluation Criteria in Solid Tumors) criteria, as determined by the investigator, or unacceptable toxicity. At the time of progression, placebo-treated patients were unblinded and allowed to cross over to sunitinib.

The study was originally designed to enroll 340 patients to observe 260 PFS events, with an interim analysis planned at 130 events.

Approximately eight months after accrual began, the protocol was amended to install an independent data monitoring committee to monitor patient safety and be involved in the conduct and interpretation of the pre-specified interim efficacy analysis at 130 PFS events.

The protocol amendment did not state that the DMC would review efficacy data prior to the pre-specified interim analysis, according to FDA's briefing documents.

The DMC met three times and reviewed both safety and PFS data, although the efficacy reviews were not pre-specified. In February 2009, the DMC recommended the study's termination based on an observed median advantage in PFS of 5.6 months with sunitinib compared to placebo after only 73 PFS events. At the time the study was closed, there were 171 patients enrolled and 81 PFS events (31.2% of the planned events) observed.

The investigator-reported study results, published in the Feb. 10, 2011, issue of the New England Journal of Medicine, showed sunitinib to be associated with a significant 5.9-month benefit in PFS, 11.4 months versus 5.5 months on placebo (Also see "FDA Sends Sutent sNDA For Pancreatic Cancer To ODAC" - Pink Sheet, 1 March, 2011.).

The original NDA submission in December 2009 resulted in a May 2010 "complete response" letter. FDA requested a post hoc, blinded, independent central radiologic review of PFS events for all 171 subjects enrolled and a recalculation of investigator-determined PFS. Pfizer resubmitted the efficacy supplement in December 2010.

Benefit Uncertain Despite Strong Study Results

FDA concluded that in all PFS analyses, including sensitivity analyses performed by the agency and the sponsor, sunitinib improved PFS in patients with well-differentiated PNET. FDA's analysis, which was more conservative than the investigator-reported or central radiology analyses, found a significant 4.8-month benefit in median PFS with sunitinib, with a 57% reduction in relative risk.

However, uncertainty remains about the magnitude of benefit due to the study's early termination and the relevance of PFS in this disease setting, the agency said.

"Interpretation of results based on early data looks is challenging, especially with a relatively subjective primary endpoint such as PFS," FDA's briefing document states. "General issues with early PFS analyses include inadequate follow-up, missing assessments, high rates of censoring, and inability to provide an accurate and reproducible estimate of the treatment effect size."

FDA said it is well-recognized that the true treatment effect of a drug can be over-estimated in studies with multiple data looks that are stopped early for efficacy. "In group sequential designs, when data are assessed repeatedly over time, a random high that overestimates the true effect is more likely with early analysis of the data. Thus, the PFS results from Study A6181111, which was stopped early based on the observation of less than 30% of the planned number of events, likely overestimated the true treatment effect size."

FDA expressed concern about overestimated treatment effect in proposing to withdraw accelerated approval of Genentech's Avastin (bevacizumab) for metastatic breast cancer. Office of Oncology Drug Products Director Richard Pazdur said the magnitude of PFS benefit seen in the pivotal E2100 trial of bevacizumab in combination with paclitaxel, which was stopped early based on an interim analysis, may have represented a "random high" (Also see "FDA's Avastin Decision Puts Pressure On Progression-Free Survival Endpoints" - Pink Sheet, 3 January, 2011.).

FDA has repeatedly insisted that oncologics seeking approval on the basis of a PFS endpoint also demonstrate a benefit on overall survival or quality-of-life factors, but neither appears to have been shown in the case of sunitinib and PNET.

Although there appears to be an advantage for sunitinib on overall survival, the most recent study update failed to show significant benefit, FDA said. Furthermore, low event rates and the high crossover from placebo to sunitinib arms make the OS data difficult to interpret.

There also was no clinically or statistically significant benefit on global health-related quality-of-life or functional scales in a patient-reported outcomes tool used in the study, and subjects in the sunitinib arm reported statistically significant worsening of one of the elements on the symptom scale, diarrhea.

Considering Safety In The Context Of Indolent Disease

Sunitinib's adverse event profile in the 1111 study was consistent with its known safety profile in RCC and GIST, FDA said. Common adverse events included diarrhea, nausea/vomiting, fatigue/asthenia, hypertension, hand-foot syndrome, stomatitis, neutropenia and thrombocytopenia. Rare but serious events included leukoencephalopathy, intracranial hemorrhage, renal failure, hepatic toxicity, ventricular arrhythmia and cardiac failure.

Despite the absence of any new safety signals, the agency appears concerned about the drug's toxicities because PNET can have a relatively indolent natural history. FDA raised similar concerns - that the drug's toxicities may be worse than the disease itself - in connection with its review of AstraZeneca's recently approved medullary thyroid cancer treatment vandetanib (Also see "Toxicity Of AstraZeneca's Vandetanib Expected Focus For FDA Advisory Committee" - Pink Sheet, 30 November, 2010.).

-Sue Sutter (s.sutter@elsevier.com)

Early Study Termination Is Key Issue In ODAC Review Of Pfizer's Sutent For Pancreatic Cancer

More from Archive

More from Pink Sheet