Background

Bevacizumab (Avastin®) is the first cancer therapy approved by the Food and Drug Administration (FDA) to inhibit angiogenesis, which refers to the development of new blood vessels needed to carry vital nutrients to a tumor.  Bevacizumab inhibits angiogenesis by blocking "Vascular Endothelial Growth Factor" or VEGF (pronounced Veg-F), a protein that is excessively produced by tumor cells and a key factor in the growth of many types of cancer. 

This therapeutic agent is a monoclonal antibody, a type of highly specific antibody created in the laboratory that selectively binds to substances in the body, including cancer cells.  Each monoclonal antibody is created to find and bind to one specific substance.  Bevacizumab selectively binds to VEGF, blocking the protein's ability to bind to its receptors on the surface of endothelial cells, which are the main cells in the inside lining of blood vessels.  By blocking the activity of VEGF, bevacizumab can reduce the development of blood vessels and the delivery of oxygen and nutrients needed by the tumor to grow and spread.

Bevacizumab was first approved by the FDA in 2004 and 2006 respectively for the first- and second-line treatment of metastatic cancer of the colon or rectum (colorectal cancer) in combination with 5-fluorouracil (5-FU)-based chemotherapy.  In 2006, it was also approved as first-line treatment in combination with chemotherapy (i.e., carboplatin and paclitaxel) for a form of inoperable (unresectable) metastatic lung cancer (i.e., non-squamous, non-small-cell).  For both colorectal and lung cancer, bevacizumab was approved by the FDA based on randomized clinical trials that demonstrated improvement in overall survival (OS).

In contrast, in 2008, bevacizumab was FDA approved for another indication, but this time, approval was not on the basis of overall survival as the trial endpoint--which brings us to the story of bevacizumab and metastatic breast cancer.

Bevacizumab and metastatic breast cancer

When the FDA approved bevacizumab for the first-line treatment of metastatic breast cancer in combination with paclitaxel, the agency based such approval on improvement in progression-free survival (PFS).  In a Phase III randomized clinical trial called the E2100 study, patients with locally recurrent or metastatic breast cancer who received bevacizumab and paclitaxel (Taxol®) lived approximately twice as long without their cancer worsening when compared to those who were treated with paclitaxel alone.   But what does "twice as long" specifically mean?   It does not mean that women lived longer, in fact, there was no increase in overall survival.  An improvement in PFS means that it took twice as long for the tumor to progress.  Patients treated with bevacizumab and paclitaxel demonstrated a median PFS of 11.3 months, whereas those who received paclitaxel alone had a median PFS of 5.8 months—that is, a 5.5 month improvement in the length of time before tumor growth and disease progression.  However, a significant improvement in overall survival was not shown.  Furthermore, those who received bevacizumab experienced a higher overall incidence of toxicities as well as more severe toxicities.   

So in the case of metastatic breast cancer, FDA approval was given to bevacizumab based solely on increased progression-free survival, suggesting the agency's acceptance of such a surrogate endpoint as an appropriate measure of efficacy. 

A surrogate endpoint is a physical sign or laboratory finding used in clinical trials that may not itself be a direct measure, but one that is considered reasonably likely to predict therapeutic benefit and a clinically meaningful outcome, such as survival.   However, there are few "true" surrogates, and it is crucial to note that a treatment that delays tumor growth and cancer progression may not actually lengthen survival. 

It is clear that treatment with bevacizumab is associated with an increased risk of serious adverse effects.  And though it is not known whether the treatment will lengthen life, it is known that serious toxicities may very well negatively impact quality of life. 

A controversial decision with profound implications

The FDA has previously accepted progression-free survival (PFS) as a surrogate endpoint to support regular drug approval for adjuvant therapy in breast cancer.  Adjuvant therapy refers to additional treatment following primary therapy (i.e., surgery)—such as chemotherapy, radiation, hormone therapy, and/or biologic treatment--to decrease the risk that the cancer will recur.   In 1998, for example, based on PFS, the FDA approved trastuzumab (Herceptin®) in combination with paclitaxel, and in 2004, approved gemcitabine (Gemzar®) combined with paclitaxel.  However, importantly, survival data were subsequently added to the labels in both cases.

But some in the oncologic community feel that the bar was lowered when approval for bevacizumab was sought in the first-line metastatic setting.  In December 2007, the FDA requested the advice of its Oncologic Drugs Advisory Committee (ODAC) concerning whether significant improvement in PFS alone without a demonstrated survival advantage should be used to measure direct clinical benefit in the first-line metastatic setting.  The panel voted 5-4, recommending against approval in this setting, indicating that the data were insufficient to show that the benefits outweighed the risks (FDA ODAC Meeting 2007).

The members considered two phase III randomized clinical trials in their decision:

• The first study, AVF2119g, a multicenter phase III trial, evaluated the efficacy and safety of bevacizumab and what the body does to the drug (pharmacokinetics) when administered in combination with the chemotherapy agent capecitabine (Xeloda®) as second- or third-line treatment of metastatic breast cancer (Miller, 2005).  The study evaluated 462 women with metastatic breast cancer.  The dosage for bevacizumab was 15 mg/kg once every 3 weeks.  Capecitabine was given at the FDA-approved dose of 2,500 mg/m2 in two divided doses for 2 out of every 3 weeks with appropriate dose reductions for toxicity. The study found no additional benefit on overall survival or PFS when adding bevacizumab to capecitabine therapy.  The NBCC partnered with Genentech on the design, oversight, and outreach for this trial. 
• A second study, E2100, described briefly above, was a randomized open-label (i.e., non-blinded) phase III trial in which women with HER2-negative metastatic or locally recurrent breast cancer were randomized to receive bevacizumab plus paclitaxel or paclitaxel alone (Miller, 2007).  The primary endpoint was PFS.  Other end points included objective response rate, toxic effects, overall survival, and quality of life.  These women did not have previous chemotherapy treatment.  Some women received taxane-based chemotherapy (e.g., paclitaxel, docetaxel) as adjuvant therapy after surgery but were required to have a disease-free interval of at least 12 months after receiving the taxane.  As noted earlier, those patients who received bevacizumab and paclitaxel had a median PFS of 11.3 months compared to a median PFS of 5.8 months in those treated with paclitaxel alone--i.e., a 5.5 month improvement in the time before disease progression.  However, again, there was no statistically significant increase in overall survival, showing that PFS was in fact not a surrogate endpoint for survival. 

Toxicities

A crucial consideration is the fact that those who received bevacizumab experienced a significantly higher overall frequency of toxicities that were also more severe than seen in those who received paclitaxel alone.  The addition of bevacizumab to paclitaxel was associated with a 20.2% increase in serious side effects classified as "Grades 3-5" Adverse Events.  In addition, death occurred in 1.7% of patients (6/363) in the bevacizumab arm compared to 0% (0/348) for those who received paclitaxel alone. 

Based on these data, whether PFS in this setting is an effective surrogate for clinical benefit such as survival remains questionable.

Yet, despite ODAC's recommendation and the concerns its members raised regarding lack of survival data, severe toxicities, and trial design, the FDA approved bevacizumab (Avastin®) in 2008 as a first-line treatment in combination with paclitaxel for women with metastatic HER2-negative breast cancer.  The approval was extremely controversial and surprised many, particularly since the FDA typically follows the advice of its advisory panel.  Instead, on this occasion, the FDA overruled ODAC's recommendation and granted "accelerated approval," showing acceptance of PFS in the absence of overall survival—and, therefore, based on what some consider a lower level of evidence.

The FDA requires confirmatory trials for agents that have received accelerated approval, and several additional trials of bevacizumab for women with metastatic breast cancer are currently under way, including the AVADO, RIBBON-1, and RIBBON-2 trials.  However, all study details and outcomes are not yet available.  And in each of these studies, the primary endpoint is, once again, progression-free survival. 

Disturbing possibility – anti-angiogenesis agents may promote more invasive tumor growth

In March 2009, the results from a study in mouse models suggested a disturbing possibility--that although anti-angiogenesis agents may initially shrink tumors, they may later promote more invasive tumor growth as an adaptive response of the cancer as it seeks the necessary oxygen and nutrients (Paez-Ribes, 2009).  The investigators studied the effects of the anti-angiogenic drug sunitinib in mouse models of glioblastoma and pancreatic neuroendocrine cancer.  They confirmed that the tumors stabilized or shrank during the first weeks of treatment with sunitinib, but later had an adaptive response, with increased invasion into adjacent tissue and metastasis. 

The results of this research are similar to those seen in earlier animal studies.  They are also consistent with some early results seen in a small number of clinical trials suggesting that anti-angiogenic therapy may alter the natural history of tumors.  This includes a clinical trial in which a subset of glioblastoma patients had recurrence at multiple sites confirmed by MRI imaging after initial tumor shrinkage during treatment with bevacizumab (Zuniga, 2009). 

It is important to note that the mouse models were not treated with a chemotherapy agent.   Additional clinical studies are crucial to determine whether bevacizumab affects tumors in patients as seen in these mouse models and to combine anti-angiogenic agents with chemotherapy to evaluate whether patients gain the early benefit of the former without promoting subsequent invasiveness and metastasis.

So where does this leave us?

Clinical trials are critical to improve breast cancer care; they produce the best evidence of how well new interventions will work.   Trial endpoints are the basis for new drug approval by the FDA, and endpoint selection determines what information is available to women and their doctors as they make these crucial decisions.  But it is important to understand that not all endpoints provide the same amount of evidence about the benefits of an intervention, and long-term safety, efficacy, and quality of life issues might not become apparent with shorter duration trials.  Study endpoints must be chosen in a manner that maximizes the value and meaning of the clinical trial in addressing the burden of breast cancer for women.