At last week’s meeting, committee members reviewed mature results from those trials, and the trials failed to confirm the data that led to the accelerated approval. Women in the initial trial, called E2100 , received either paclitaxel (Taxol) alone or paclitaxel plus bevacizumab for locally recurrent or metastatic breast cancer. The women in E2100 who received the bevacizumab combination experienced a median improvement in progression-free survival (PFS)—time without tumor growth—of 5.5 months compared with those who received paclitaxel alone.
The FDA’s decision to grant accelerated approval went against the recommendations of ODAC, which voted five to nine against approval in December 2007. The FDA is not required to follow the recommendations of its advisory committees, although it does so more often than not. The results of E2100 had not shown an increase in overall survival and PFS remains a contentious surrogate endpoint in oncology clinical trials.
“FDA believes that, in accepting PFS as a regulatory endpoint , a close examination of the magnitude of improvement in PFS must be closely evaluated in a risk-benefit analysis,” said Dr. Richard Pazdur, director of the FDA’s Office of Oncology Drug Products, in his opening remarks at last week’s meeting. “Because treatment with Avastin is associated with considerable toxicity, the magnitude of PFS improvement—especially if not supported by an improvement in overall survival—should be substantial, clinically meaningful, and be able to be replicated in additional trials.”
The additional trials performed by Genentech, the manufacturer of bevacizumab, called AVADO and RIBBON1 , together enrolled almost 2,500 women. The women in AVADO were randomly assigned to receive chemotherapy with docetaxel (Taxotere) plus placebo or docetaxel plus bevacizumab. In contrast with the E2100 trial, the women in the bevacizumab arm of AVADO showed a median improvement in PFS of less than a month.
In RIBBON1, patients were assigned at the discretion of their physicians to receive chemotherapy with a taxane , an anthracycline , or capecitabine (Xeloda). Within those three groups, the researchers randomly assigned women to receive additional bevacizumab or a placebo. The improvement in PFS with the addition of bevacizumab was larger than in AVADO but still substantially smaller than that seen in E2100: 1.2 months for women in the taxane and anthracycline groups and 2.9 months for women in the capecitabine group.
In both trials, women in the bevacizumab groups had an increased risk of death. In an analysis of the trial data, prepared by the FDA, 0.8 percent of the women in AVADO and 1.2 percent of the women in RIBBON1 who received bevacizumab died from side effects thought to be related to the drug.
Given bevacizumab’s side effects and the lack of an overall survival benefit, “we need to consider what value [PFS] has to a woman with metastatic breast cancer,” said Dr. Mikkael Sekeres, associate professor of medicine at the Cleveland Clinic Taussig Cancer Institute and an ODAC member.
“Since PFS is the endpoint being studied,” said Dr. Wyndham Wilson, the ODAC chair, “what we are here to judge is whether or not there is a clinically meaningful—from a patient’s point of view—[difference in] quality of life between the patients who received Avastin and those who didn’t.”
The committee found any such difference impossible to ascertain, since the trials had not collected patient-reported quality-of-life data that could show an improvement in symptoms or psychological state in tandem with PFS. The AVADO trial collected a set of quality-of-life data, but only to show that the addition of bevacizumab did not decrease patients’ quality of life.
Sponsors of future trials evaluating PFS should take this issue to heart, stated Dr. Pazdur. “Sponsors really need to pay very close attention to [measuring quality of life], build it carefully into their protocols, and approach it with the same degree of caution and resources…as one would with a primary endpoint,” he said.
The committee’s decision to recommend removing metastatic breast cancer as an indication for bevacizumab does not mean that the FDA regrets awarding the accelerated approval, said Dr. Pazdur. “We do not look at this as a mistake,” he concluded. “There is a risk…in approving these drugs, and this [process today] is management of that risk.”