While the Oncologic Drugs Advisory Committee generally endorses the idea that progression-free survival can support full approval of first-line therapies in metastatic breast cancer, the panel left FDA with the challenge of qualifying how PFS represents clinical benefit and determining the measure's meaningfulness for regulatory purposes.

"Clinically there is no question among my colleagues and myself that PFS is clinically meaningful. Having said that, the challenge is how much," committee member Gary Lyman, Duke University, said at the panel's Dec. 5 review of Genentech's Avastin (bevacizumab) for first-line metastatic breast cancer. The firm is seeking approval based on the strength of PFS results, despite having failed to meet the survival endpoint in the same trial; the committee voted 5-4 against approval (¹ Also see "ODAC Split On Avastin Use In First-Line Metastatic Breast Cancer" - Pink Sheet, 10 December, 2007., p. 8).

Support of PFS as a direct rather than surrogate measure of clinical benefit in the first-line setting marks a reversal for ODAC, which determined at a 1999 meeting that time-to-progression was not an acceptable endpoint for regular approval of first-line cytotoxic therapy for breast cancer and that overall survival should remain the primary efficacy endpoint for trials supporting approval (² Also see "Breast Cancer Drugs Need Survival Benefit for Full Approval, Cmtte. Says" - Pink Sheet, 14 June, 1999., p. 4).

While the committee generally endorsed use of PFS in first-line MBC, panelists struggled to articulate specifically what is the clinical benefit achieved with an improvement in PFS.

"I think you are grappling with a concept here. You are trying to enunciate what is this clinical benefit. Perhaps you know in your heart that it may be, but you're having difficulty in enunciating or really clarifying what it is. I get this feeling from hearing several people here," FDA Office of Oncology Drug Products Director Richard Pazdur said.

Generally, the perceived benefits of PFS relate to the potential delay in symptom emergence and worsening, the delay in starting other potentially toxic agents and associated psychological benefits. "There's no question if I'm in the clinic, my psychology is much improved when my patient is not progressing, not to mention their psychology is, and there is no way to capture that as a clinical benefit," committee chair Maha Hussain, University of Michigan, said.

The inability to specifically characterize treatment effect as measured by PFS did concern some committee members.

"What we're hearing is that we don't have any way to say how it translates into quality of life, how it improves symptoms," temporary member Ralph D'Agostino, Boston University, said. "Our approval would rest completely on buying into PFS as an appropriate measure of efficacy, and I don't think we have that ability at the moment given the data that's before us."

Progression-free survival has typically been considered a surrogate for overall survival. "Conceptually, PFS has the desirable qualities of a surrogate endpoint because it reflects tumor growth, a phenomena likely to be on the causal pathway for cancer-associated morbidity and death," FDA's briefing material notes.

In October 2004, Genentech asked FDA if PFS would be an adequate endpoint for full approval. FDA's response was that it depends on the overall robustness and magnitude of the PFS finding and the overall survival results. At a pre-sBLA meeting in September 2005, FDA stated PFS would support accelerated approval but final overall survival would be necessary for full approval. However, the agency did not consider Genentech's application eligible for accelerated approval because the pivotal trial failed to show significant improvement in survival, FDA told "The Pink Sheet."

"Until recently, following ODAC's advice, the FDA's efficacy requirements for regular marketing approval for oncology drugs requires demonstration of clinical benefit, specifically, prolongation of life or better quality of life," the agency's briefing document states.

In its briefing material, FDA accepts that "PFS if properly measured might be an acceptable endpoint which may confer patient benefit." However, the agency seems skeptical that the observed benefit demonstrated by PFS will exceed the toxicity of cancer therapies in first-line MBC.

PFS has only been used once as the basis for approval in MBC: Bristol-Myers Squibb's chemotherapy Ixempra (ixabepilone) in second- and third-line therapy. TTP has served as the basis for approval for paclitaxel and for GlaxoSmithKline's Tykerb (lapatinib) for second- and third-line MBC.

Some panelists and FDA disagreed about the inherent meaningfulness of an effect on PFS in the first-line setting. Committee members noted that because many women are treated in the adjuvant setting, first-line MBC patients are rarely treatment-naïve and therefore similar to the second- and third-line populations, in which PFS has been accepted as an independent endpoint survival.

"The differences of managing first-line metastatic disease and subsequent second- and third-line ... in the past few years has become very cloudy," Lyman said. "I'm not really sure that population overall is any different than the patient who comes back after a recurrence after their first-line of metastatic failure."

FDA's position is that PFS is more clinically meaningful in second- and third-line patients due to differences in symptomatic burden. "Obviously in a more refractory disease setting there is a different risk-benefit than in a first-line setting," Pazdur said. Refractory patients are "a more symptomatic patient population. Hence, a delay in progression in a symptomatic population probably has a little more clinical meaning than simply a radiographic delay in an asymptomatic patient population."

"That's why we kind of question in the first-line setting where we do have obviously a different risk-benefit situation than in a more refractory disease."

Use of PFS as a primary endpoint of pivotal trials provides advantages over use of overall survival by allowing faster time to event and removing the potential for confounding of study results by subsequent therapies.

Although potential confounding by treatment regimens following a clinical trial is common in first-line MBC, FDA does not seem to be particularly receptive to such an argument. During the agency's presentation at the meeting, FDA cited Genentech's Herceptin (trastuzumab) and Pfizer's Camptosar (irinotecan) as examples where survival was shown despite post-protocol treatment.

The panel also acknowledged that use of the softer outcome endpoint poses a number of challenges. Committee member Gail Eckhardt, University of Colorado - who agreed PFS " is an adequate endpoint that does include clinical benefit" - noted that "what ... many of us are struggling with is the measurement in these kinds of studies, and I think that is going to be something that will have to be decided going forward."

"The use of PFS requires careful planning and attention to detail in assessing progression. Unlike the endpoint of overall survival, bias can be easily introduced in assessing PFS," Pazdur said.

"[PFS] is an endpoint that is very difficult to measure properly," Division of Drug Oncology Products Medical Officer Patricia Cortazar added. Common problems are patients without measurable disease, missed or incomplete assessments at baseline, infrequent or uneven assessments, and subjectivity in determination of progression.

Cortazar noted that trials with PFS as the primary endpoint should be blinded, and results should be adjudicated by an independent radiology review. "Because of the subjectivity of the endpoint, it might be necessary to have two trials to support each other," she said.

"We would require not only a look at survival, but we've also recommended basically powering for overall survival," Pazdur said. FDA has traditionally been strongly opposed to wholly substituting PFS for overall survival. At a workshop on oncology endpoints for ovarian cancer, the agency stressed that trials assessing PFS should be powered to identify an effect on survival as a secondary endpoint (³ Also see "Ovarian Progression-Free Survival Acceptable If Trial Is Powered For Survival" - Pink Sheet, 15 May, 2006., p. 19).

"The disadvantage of that ... is when you do overpower a trial for progression-free because you're powering it for survival, you may get a statistically significant finding for PFS, and it could be relatively marginal. And then we really have to have very careful discussion about what is the clinical meaningfulness of a very small impact on progression-free survival that may be highly statistically significant," Pazdur noted.

"In our guidance [on] clinical trial endpoints for the approval of cancer drugs and biologics, we have attempted to address problems associated with the use of PFS as a primary endpoint for registration trials," he said. The guidance, which was finalized earlier this year, reflects the growing shift towards non-survival endpoints (⁴ Also see "FDA Cancer Endpoint Guidance Abandons Placebos, Moves Beyond Survival" - Pink Sheet, 4 June, 2007., p. 5).

- Brian Marson (b.marson@elsevier.com)

Progression-Free Survival Represents Clinical Benefit, But How Much? – ODAC

More from Archive

More from Pink Sheet