FDA Reviews Approval Issues With Cancer Advocates

WASHINGTON—Data from patients granted "compassionate use" of an investigational cancer drug are examined but not usually pooled with controlled clinical trial findings when the Food and Drug Administration (FDA) considers whether to approve a new drug application (NDA), according to Richard Pazdur, MD, director of the FDA’s Division of Oncologic Drug Products.

Dr. Pazdur discussed the evaluation of "compassionate use" data and a number of other topics during a meeting between several FDA officials and representatives of cancer patient advocacy groups.

Although commonly used, "compassionate use" is not a preferred regulatory term, he said. It refers to the treatment use, as opposed to the investigational use, of unapproved drugs.

Treatment use of unapproved drugs may be divided into two main groups: single-patient treatment use and expanded access use, Dr. Pazdur said. Expanded access refers to multiple patients treated under a single protocol. In the single-patient use, individual treatment plans are drawn up for each patients.

The requirements for single-patient treatment include the existence of a drug supplier and sponsor, a qualified investigator, institutional review board (IRB) approval and informed consent, and FDA concurrence, Dr. Pazdur said.

There are two mechanisms for handling single-patient use of investigational drugs, he said. In the first mechanism, the single-patient investigational new drug (IND), a new sponsor files a separate IND for each patient.

In the second mechanism, called the single-patient exemption, there is already an IND, an existing sponsor, and an investigational protocol. With this mechanism, a patient who is ineligible for an investigational protocol is treated under a plan that is a slight modification of the existing protocol. "This is a more efficient mechanism for single-patient treatment," Dr. Pazdur said.

Companies provide FDA any data collected during the investigational phase of a drug, including treatment use data, Dr. Pazdur said. "We think it is important to get a sense of the adverse events or safety data from all uses of the drug," he said. "Most clinical trials are fairly small and narrow in terms of the defined population. Sometimes, there can be useful information in less restricted, more varied use of the drug."

Efficacy data from the clinical trials are examined separately. "Patients getting the drug under treatment use programs usually don’t fit the protocols of the clinical trials, and data from these patients are relatively uncontrolled," he commented.

Other topics addressed at the meeting included:

Toxicity: "In general, for oncology drugs, the primary reason for nonapproval is their lack of efficacy rather than their toxicity," Dr. Pazdur said. By the time a drug reaches the approval stage, FDA and the sponsor have a good idea of the drug’s toxicity and the acceptability of this toxicity to cancer patients, he said.

Generally, in oncology, Dr. Pazdur said, "we have a different risk-to-benefit ratio, compared with other diseases such as diabetes or arthritis." Toxicity problems that become evident in the clinical trials can usually be dealt with by instituting supportive care measures, drug labeling, and physician/patient education, he said.

"We are required to look at all the data on a drug before approving it," Dr. Pazdur said. "It has to be a convincing body of evidence. The bottom line is, we must feel comfortable approving the drug."

Targeted drugs: Increasingly, researchers are looking for specific genes or other biological entities at which to target therapeutic or preventive drugs. Two examples of approved targeted drugs are imitanib mesylate, also known as STI-571 (Gleevec), which inhibits an abnormal protein involved in causing chronic myeloid leukemia, and trastuzumab (Herceptin), a monoclonal antibody that targets the HER-2 receptor in breast cancer. Such medications pose their own challenges to both the FDA and the drug companies.

"It requires a certain creativity to recognize the limitations of a study of a targeted drug," said Patricia Keegan, deputy director of clinical trials design and evaluation at FDA’s Center for Biologics Evaluation and Research. "If you are going to identify a certain group of people for targeting, you have to be able to identify them reliably. We are becoming a little more savvy about evaluating such trials. We have certainly been developing experience working with these kinds of models every day."

Dr. Keegan noted that targeted drugs have both an upside and downside for the pharmaceutical industry. "The more narrow the population, the more narrow the market," she observed.

Foreign trials: FDA generally accepts data from cancer trials outside the United States, with a few caveats regarding ethnic differences, Dr. Pazdur said.

Ethnic differences in clinical trials can be classified as either extrinsic or intrinsic, he explained. Extrinsic differences may include differences in the practice of medicine, supportive care measures, subsequent therapies, and health care delivery practices that may ultimately influence clinical trial results.

Intrinsic differences may be viewed as differences in genetic polymorphism and population pharmacogenetics and pharmacokinetics. "If there are perceived differences between the population in which the study was conducted and the US population, the agency may ask for a ‘bridging study.’ This trial would more closely examine population differences and the feasibility of extrapolating foreign data to the United States," he said.

Most foreign clinical trials submitted to the FDA as part of the approval process take place in Canada or Western Europe. However, a large proportion of the patients in trials that supported the recent approval of a new aromatase inhibitor came from Russia and China.

"We were comfortable because the study had a hard endpoint as far as survival, it was a large trial, and it possessed internal consistency," Dr. Pazdur said. "When we analyze the data, we look at population differences and perform subset analyses to verify the presence or absence of population differences. This may be helpful in extrapolating foreign data to the US population."