WASHINGTONThe Food and Drug Administration (FDA) plans
to review potential new endpoints for use in approving cancer drugs. The
decision-making process will include public discussions involving practicing
oncologists, academic physicians, the pharmaceutical industry, and advocacy
Richard Pazdur, MD, director of FDA’s Division of Oncology
Drug Products, described the agency’s plans during a meeting between several
FDA officials and cancer patient advocates. The session provided insights into
current FDA thinking and activities regarding the oncologic drugs approval
Traditionally, FDA has relied largely on survival data as the
primary endpoint for approving drugs. But drug companies and advocates have
urged the use of other endpoints, including time to progression, response rates,
and quality of life. "We know there is a significant debate, and we would
like to discuss this in a larger forum," Dr. Pazdur said, adding that there
is discussion within FDA itself on the issue.
If the FDA concludes that new endpoints may be appropriate, it
will then take them to a meeting of the agency’s Oncologic Drugs Advisory
Committee (ODAC) for further discussion and a recommendation vote on whether the
committee believes the FDA should adopt the endpoints.
The question-and-answer session touched on a number of other
Expanding Trial Eligibility
Dr. Pazdur said he would
like to see a broadening of eligibility of many trial protocols to include
patients representative of those who will eventually use the drug in the
"Once a drug is approved, it is not limited to patients
with 0 or 1 performance status," he said. "By having less restrictive
entry criteria, we will get a better picture of how a drug works in the
population that will actually use it."
By accruing only excellent-performance patients, trials cannot
assess symptom improvement because such patients have few or no symptoms, Dr.
Pazdur noted. "Our idea in approving a drug is to demonstrate clinical
benefit," he added. "People have focused on survival advantage, but
clinical benefit may be the improvement or delay of symptoms."
FDA has talked with the pharmaceutical industry about the
possibility of doing two trials of a drugone in less advanced stages of a
disease and the second in symptomatic patients to assess clinical benefit, Dr.
"A drug’s efficacy may not be optimally observed in
patients who have a poor-performance status or are highly symptomatic. This is a
concern of many sponsors," he commented.
Patricia Keegan, MD, deputy director of clinical trials design
and evaluation at FDA’s Center for Biologics Evaluation and Research, agreed
with Dr. Pazdur’s statement. She added that there is also a perception by
clinical researchers and drug companies "that it’s got to be this
way." Dr. Keegan emphasized that FDA does not write clinical protocols.
"We give comments about things that are not acceptable from a scientific
point of view, and I think we are willing to rethink these issues," she
Need for Two Phase III Trials
FDA wants pharmaceutical
companies to conduct phase III trials of a drug, but many now carry out only one
such trial and hope that it will yield "eye-popping" results, Dr.
Pazdur said. Often, the results prove marginal or there are serious
methodological problems with the single study, but the sponsor nonetheless
pursues submission of a new drug application (NDA).
"We are then in a quandary because we must have proof of
the drugs’ safety and efficacy to approve it. Most importantly, patients lose
because a potentially effective therapy is not available. The company loses
because of the revenue loss if the drug is not approved."
Problems With Chemoprevention Trials
Drugs to prevent
cancer or treat precancerous conditions pose difficult issues for FDA in terms
of deciphering how to judge their effectiveness and long-term safety. "What
are the trade-offs of giving a drug or product to people who, at this point,
have no disease? That is what we really need to know," Dr. Keegan said.
"Some issuesproblems in terms of sample size, follow-up, and other
thingsseem almost insurmountable."
Such studies would require trials of great length and patient
size. "Proof has to be very rigorous, because you are asking people to take
medications for a prolonged time," Dr. Pazdur said. "Surrogate markers
for chemoprevention trials, for the most part, are still under development and
ODAC’s Evolving Role
Dr. Pazdur acknowledged that his
division is bringing fewer cancer drugs before ODAC and that the committee’s
role may evolve from that of prior years.
"One has to remember that in the name ‘ODAC’ is the
word ‘advisory,’" he said. "We are using our consultants
throughout the review process and not simply at the ODAC meeting. We are
discussing thematic issues of how a drug should be approved and what are
appropriate endpoints. We will take drugs to them with which we have specific
problems and feel we need a public hearing."
FDA brings in outside advisors at several stages of its review
of a cancer drug, allowing the agency to make more rapid decisions, Dr. Pazdur
said. He cited as an example the recent rapid approval of imatinib mesylate,
also known as STI-571 (Gleevec), for chronic myeloid leukemia. "It is
obvious from the trial data that STI-571 is effective," Dr. Pazdur said.
"We do not want to unnecessarily delay approval of drugs when there is
compelling data to expedite our review process."
Advocates and pharmaceutical
companies both want quality-of-life (QOL) data considered by FDA during the drug
review process, but the agency has been reluctant to use such data in granting
"It certainly seems like a logical way to go," Dr.
Keegan said. "Nobody wants to approve a drug that doesn’t make people’s
lives better, and this looks like the obvious way to do it. We have just had a
lot of difficulty figuring out the best way to get reasonable information."
The agency’s concern in looking at QOL data, she said, is that
"it is often very hard to interpret exactly what it means. In a trial
directed at arthritis, you have a large population that stays on study a long,
long time, and you don’t lose anybody. That is not often the situation with
cancer trials. We end up with lots of missing data points, and we know, in our
hearts, they are not just missing at random."
Dr. Pazdur said that FDA is interested in QOL measures as a way
of assessing clinical benefit. "The problem is that the tools to measure
qualify of life have methodological problems, including missing data and the
difficulties of interpretation," he added.
He pointed out that the agency has been encouraging sponsors to
"pick a few symptoms to measure from the patient’s perspective. Let’s
get them involved in a prospective evaluation of symptoms. Is your cough
improved? Can you walk better? Can you sleep? Obviously, the patient is the only
person who can measure these things."