Panel Favors Approval of Gemzar for Use In Patients With Late Stage Pancreatic Cancer

ROCKVILLE, Md--Because advanced pancreatic cancer is a devastatingdisease with no effective treatment, members of the FDA OncologicDrugs Advisory Committee (ODAC) voted to recommend approval ofthe nucleoside analog Gemzar (gemcitabine hydrochloride, Eli Lilly)as a first-line treatment for patients with nonresectable stageII, stage III, or metastatic (stage IV) adenocarcinoma of thepancreas.

The agent is currently available under a treatment investigationalnew drug program for selected patients with advanced or metastaticpancreatic cancer, which will continue while the company awaitsthe FDA's final decision on approval.

In the United States, 27,000 people die each year of pancreaticcancer, usually less than a year after diagnosis. "Previousclinical trials have shown that fluorouracil, the most commonlyused treatment for pancreatic cancer, in combination with otherchemotherapeutic agents is no better than fluorouracil alone,"said Daniel Von Hoff, MD, Cancer Therapy and Research Center,San Antonio. Therefore, fluorouracil was the drug against whichGemzar was tested.

Lilly presented two clinical trials: a phase III randomized, blinded(but not double blind) study comparing Gemzar with fluorouracil,and a phase II single-arm study of patients whose disease hadprogressed while they were receiving fluorouracil.

New Surrogate Endpoint

Dr. Von Hoff noted that it is difficult to evaluate advanced pancreaticcancer patients following surgery, radiation, and chemotherapyusing traditional endpoints such as the rate of tumor growth orshrinkage, since these often do not provide a meaningful analysisof a patient's overall health.

"Because clinical endpoints are so hard to measure in thisdisease, we wanted to set up a trial to determine the effect ofGemzar on what bothers patients most," he said.

This is easier said than done, according to John Andersen, PhD,Lilly Research Laboratories. The company attempted to solve theproblem by developing a complex matrix of components comprisingthe things that "bother patients most," including pain(as measured by patient reports and by analgesic consumption),Karnofsky performance status, and weight gain. Together, thesewere termed a clinical benefit response.

Based on a composite score of these established measures, patientswere classed as either clinical benefit responders or nonresponders."The response had to be marked and sustained, not transientor fleeting," Dr. Andersen said.

Clinical Trial Results

The phase III study of Gemzar vs fluorouracil included 126 patients,63 randomized to each arm, who had received no prior chemotherapy.Most of the patients had stage IV or metastatic disease.

At a dose of 1,000 mg/m² weekly for up to 7 out of 8 weeks,then weekly for 3 out of 4 weeks, patients receiving Gemzar showedsignificant improvement in the rate of clinical benefit response(24% vs 5% for fluorouracil) and time to disease progression.

The 6-month survival rates were 46% for Gemzar vs 31% for fluorouracil,and 1-year rates were 18% for Gemzar vs 2% for fluorouracil. Therewas an approximately 30% overall improvement in median survival(1½ months) in those patients who received Gemzar, the companysaid.

The phase II single-arm study of 63 patients who had not respondedto fluorouracil showed a clinical benefit response rate of 27%,with a median survival time of 3.85 months.

Adverse effects included anemia, leukopenia, neutropenia (withoutsystemic infection), thrombocytopenia, transient nausea and vomiting,low-grade fever, and skin rash. According to Mark Greene, MD,an oncologist at the University of California, San Diego, safetydata on a total of 979 patients treated with Gemzar in all studiesfound the agent to be well tolerated. Myelosuppres-sion is rapidlyreversible, and severe neutropenia is rare, he said.

The major issue in the discussion among ODAC members and FDA representativeswas whether Gemzar produces better palliation in pancreatic cancerthan can be achieved with other measures (including best supportivecare).

According to an FDA analysis of the phase III data, median survivalfor patients who stayed on study for 30 days was 173 days forGemzar and 129 days for fluorouracil. For patients who stayedon study for longer than 30 days, the difference in survival timebetween the two drugs narrowed: 175 days for Gemzar and 177 daysfor fluorouracil.

In terms of clinical benefit response, more than four times asmany patients on Gemzar as on fluorouracil showed a response,and these responses appeared more quickly and lasted longer.

Both ODAC and FDA complimented the Eli Lilly representatives onthe ingenuity displayed in creating the clinical benefit responsescale even as they acknowledged that it is difficult to analyzesubjective data like pain and quality of life performance scores.

They were also concerned that randomization might have been compromised,that is, that there might have been an imbalance between treatmentarms in the number of patients with liver metas-tases, baselineanalgesic scores, and numbers of dropouts within 30 days. Despitethese doubts, ODAC members expressed a belief that the randomizedtrial was adequate and well controlled.

They agreed that clinical benefit response, as defined in thecompany's protocols, is an appropriate primary efficacy endpoint,but they did not believe that certain other observations (objectivetumor response, time to progression, time to treatment failure,and survival in the study with only one arm) were supportive ofGemzar approval.

Despite the fact that there was only one randomized clinical trial(and that with a new surrogate endpoint), it was carefully conductedand demonstrated a definite increase in survival. For that reason,ODAC members voted 8 to 3 to recommend approval for Gemzar asa first-line treatment for patients with advanced pancreatic cancer.