Note: This article was originally published on June 28, after the first day of the public hearing on Avastin. PGx Reporter's coverage of the second day of the hearing is available here.
By Turna Ray
While the US Food and Drug Administration would be willing to conduct research with Genentech to identify breast cancer patients who respond especially well to the oncologic Avastin, the agency appears unconvinced by evidence submitted by the Roche subsidiary to date that certain breast cancer patients are so-called "super-responders" to the drug.
"In the end, CDER's decision must be based on the available scientific data from adequate and well-controlled trials," John Jenkins, director of the Office of New Drugs at FDA's Center for Drug Evaluation and Research, said during the first of a two-day hearing to discuss publicly its proposal to withdraw accelerated approval for Avastin as a treatment for metastatic breast cancer.
Based on the totality of evidence submitted by Genentech, the agency maintained its previous position that Avastin does not provide substantial benefit to metastatic breast cancer patients compared to other available treatments, and therefore the drug's indication in breast cancer should be revoked. However, Jenkins pointed out that if the company is able to show in new studies that Avastin does benefit a subset of breast cancer patients, then the company can submit a new application for that specific population.
"We stand ready to work with Genentech and others to design trials to divine what, if any, subpopulations of patients with breast cancer might derive benefit from this drug that outweigh its risk," Jenkins said. "If such data are generated, a new science-based indication could be approved. Until that time, it is not appropriate for the drug to continue to be approved for the treatment of breast cancer when the totality of the available data does not support such an approval."
Genentech markets Avastin in combination with paclitaxel in patients with HER2-negative metastatic breast cancer who have not yet received chemotherapy for metastatic disease. Beyond its breast cancer indication, Avastin, which netted around $7 billion in sales last year, is approved as a treatment for metastatic colorectal cancer, non-small cell lung cancer, and glioblastoma.
In order to receive accelerated approval for Avastin in breast cancer in 2008, Genentech conducted two trials, the E2100 study and AVF2119g. The AVF2119g trial, which investigated Avastin plus capecitabine versus capecitabine alone in previously treated metastatic breast cancer patients, showed no progression-free survival benefit when Avastin was added. The FDA granted accelerated approval based on the results of the E2100 trial, which enrolled previously untreated metastatic breast cancer patients and treated them with either Avastin plus paclitaxel or just paclitaxel. In this study, patients in the combination arm experienced PFS for 5.5 months longer than those in the chemo-only arm.
However, sponsors that receive approval for drugs under FDA's accelerated approval scheme must conduct confirmatory trials to prove clinical benefit. After Genentech submitted data from two confirmatory studies – AVADO and RIBBON1 – the FDA notified Genentech in December of its intent to revoke Avastin's approval in metastatic breast cancer because it had determined that the drug/chemo combination did not improve overall survival over standard chemotherapy regimens, marginally improved PFS, and caused serious adverse reactions in breast cancer patients.
FDA's decision to remove the drug from the market followed that of the agency's independent expert panel, the Oncologic Drugs Advisory Committee, which in July voted 12 to 1 in favor of removing Avastin from the market in the breast cancer population (PGx Reporter 12/22/2010). Members of ODAC were present at the hearing and will vote on the second day as to whether FDA should remove Avastin's breast cancer indication.
In appealing FDA's decision, Genentech has requested FDA maintain Avastin's approval while the company conducts another confirmatory trial. Similar to E2100, this new trial would investigate Avastin in combination with paclitaxel compared to paclitaxel alone(PGx Reporter 01/26/2011).
However, unlike the open label E2100 trial, this new study would be double-blinded, and randomized patients would be stratified by high and low serum VEGF-A levels. As such, the confirmatory trial would have two co-primary endpoints, looking at Avastin's efficacy by improvements in PFS in the overall breast cancer population and in patients with high VEGF-A serum levels.
This proposal doesn't seem to have swayed FDA's decision to revoke Avastin's metastatic breast cancer indication. The agency believes that this trial will take years to complete since the protocol is still not settled and submission of trial results are projected to take three years or longer.
Specifically, "the proposed study, which by design requires the co-development of a validated biomarker assay for accurate and reliable measurement of serum VEGF-A levels that will be used to identify a subpopulation of women with metastatic breast cancer who might benefit from the addition of Avastin to standard chemotherapy, is [still] under development," Patricia Keegan, director of FDA's Division of Biologic Oncology Products at CDER, said during this week's hearing.
If FDA decides to revoke approval for Avastin in the metastatic breast cancer setting, Genentech stands to lose around $1 billion in revenue per year. Doctors would still be able to prescribe Avastin off-label for breast cancer patients, but insurers may not cover the drug, which can cost upwards of $80,000 per year.
At the end of the two-day public hearing, advisory committee members will vote on whether the available data support FDA's decision to revoke Avastin's breast cancer indication. Ultimately, the decision as to whether the drug will remain on the market for the broader breast cancer population, for only those patients who are responding, or not at all, will be up to FDA Commissioner Margaret Hamburg.
Meanwhile, breast cancer patients and oncologists spoke passionately during the hearing about keeping Avastin as an option for breast cancer patients while the company conducts additional trials to identify "super-responders" to the drug. Several women who identified themselves as Avastin super-responders described having positive outcomes while on the drug.
For example, Priscilla Howard, a triple-negative breast cancer patient, said that while on an Avastin and Xeloda combination regimen, her disease had not progressed for 32 months. Triple-negative breast cancer patients do not express the genes for estrogen receptor, progesterone receptor, or HER2 in their tumors, and have a more aggressive form of the cancer. Often women with BRCA1 mutations have this form of the disease.
"Research shows that certain women with metastatic breast cancer benefit from Avastin. Although we do not yet know who benefits most, we know that BRCA-associated cancers respond to certain treatments differently than sporadic cancers, and there is anecdotal evidence to suggest that BRCA mutation patients may be among those who respond well," said Lisa Schlager, director of community affairs for the patient advocacy organization Facing Our Risk of Cancer Empowered, at the hearing.
[ pagebreak ]
FORCE is working with the Moffitt Cancer Center to analyze data from a pilot survey among its hereditary breast cancer constituents with metastatic disease. Preliminary analysis of the data suggest that among women who have received Avastin, approximately half have said their cancer hasn't progressed. "While our survey is preliminary, it suggests that the hereditary cancer community responds well to the drug," Schlager said.
However, FDA officials maintained during the hearing that the available clinical data does not provide evidence of a super-responder population. One ODAC member noted that perhaps this subset of patients actually experiences a benefit from the chemotherapy that is combined with Avastin and that the efficacy contribution from Avastin is not so pronounced. Members of the advisory committee and the FDA agreed that it will be impossible to discern whether some patients actually see a substantial benefit from Avastin without a trial looking specifically at the drug as a single agent in the metastatic breast cancer population.
"We've heard today … from patients and doctors that they have seen what is termed 'super-responders' to these drug combinations [with Avastin]. It has been pointed out here that it's impossible to separate out the contribution of Avastin from the chemotherapy simply because they're being given together," reflected ODAC member Wyndham Wilson, who is also chief of the lymphoma therapeutics section at the National Cancer Institute.
Wilson was involved in the development of Taxol, the chemotherapeutic combined with Avastin in E2100. "I saw numerous patients go on for years on single-agent drug" with Taxol, Wilson said, raising the question of whether "super-responders" are really responding to the chemo agent.
"The best we can say about the super-responders is that if you look at the … clinical trials, we're not seeing a group like this," FDA's Keegan responded. "In addition to separating out the underlying treatment, there is also the patient's natural history of the disease. There is enormous variety of how patients diagnosed with first-line metastatic breast cancer are going to do. It's not like other diseases with very short and very predictable courses."
"In the absence of a clinical trial and control, we're having hard time identifying those [super-responder] patients from a survival curve, [or] a progression-free survival curve," she continued. "I think the more compelling thing is … in 2,400 patients there doesn't seem to be a group emerging that is behaving differently." During the hearing, Keegan took a moment to recognize that although there were many self-identified super-responders who spoke at the hearing, there were plenty of women who died while on Avastin and succumbed to sever toxicities.
George Sledge, president of the American Society of Clinical Oncology and one of the investigators in the E2100 trial, expressed doubt that a molecularly defined "super-responder" population can be readily identified form the trials done on Avastin thus far.
"In E2100 we saw PFS benefit it both triple-negative and ER-positive patients, so I cannot say that there is preferential benefit," Sledge told PGx Reporter. "The biologic argument is that triple-negative breast cancer and HER2-positive cancers tend to have higher tumor VEGF levels, though this of course does not prove preferential benefit."
"This is the critical question to which we have no answer: for whom will Avastin work?" posited Bob Erwin, president of the Marti Nelson Cancer Foundation, at the hearing. "Additional questions include: Why have Genentech and the FDA discussed a biomarker-guided clinical trial for Avastin for months but enrollment has not begun? Why has Congress not appropriated enough money for the FDA to expand its scientific staff and infrastructure to efficiently analyze and regulate drug/biomarker combinations?" Erwin continued.
Although pharmacogenomic analysis was done as part of a substudy within E2100 resulting in biomarker leads, questions remain as to how aggressively Genentech has pursued validation of these early findings. Erwin recommended at the hearing that FDA maintain Avastin's metastatic breast cancer indication under an accelerated approval mechanism "subject to serious carrot and stick" incentives that facilitate biomarker research into best responders.
[ pagebreak ]
Indiana University's Bryan Schneider, who led the genomic analysis in the E2100 substudy for Avastin, identified markers that conferred longer survival and protection against drug-induced hypertension. Specifically, in E2100, researchers uncovered preliminary PGx data linking VEGF-1154 AA and -2578 AA genotypes to an improvement in median overall survival.
Median overall survival in the undifferentiated E2100 population was 25.2 months in the paclitaxel arm versus 26.7 months in the Avastin/paclitaxel arm. In the genomic substudy, patients with the -2578 AA genotype had a median overall survival of 37 months, and patients with the -1154 AA genotype saw a median overall survival of 48.5 months.
Additionally, patients in the E2100 substudy with VEGF-634 CC and -1498 TT genotypes had protection from grade 3-4 hypertension, a common side effect of Avastin. This data were published in October 2008 in the Journal of Clinical Oncology.
Paul Schmidt, a Covington & Burling attorney representing Genentech, vigorously questioned FDA reviewers during the hearing as to whether the agency had considered any way in which it could keep Avastin on the market for those patients who appear to be responding to the drug. To this line of inquiry, Richard Pazdur, director of CDER's Office of Oncology Drug Products, answered simply, "No." Surprisingly, during this exchange, none of the company representatives brought up actual biomarker data from E2100 that could potentially identify super-responders if validated in further studies. The company may discuss its biomarker strategy in more detail during the second day of the hearing.
According to Genentech's appeal to FDA's revocation notice for Avastin, it appears that the company has continued to explore biomarker leads via retrospective molecular analysis in at least one study. Data presented on the AVADO trial at the December 2010 San Antonio Breast Cancer Symposium showed that patients with high levels of plasma VEGF-A who received standard doses of Avastin had a progression-free survival hazard ratio of 0.49. Meanwhile, patients with low levels of VEGF-A had a PFS hazard ratio of 0.86.
"This finding suggests that patients with high levels of VEGF-A may be more likely to derive a more substantial benefit from Avastin," the company said in its appeal. "The relevance of VEGF-A is scientifically plausible given Avastin’s inhibitory activity on the biologic actions of VEGF." The results from this biomarker analysis have not yet been presented.
Both Sledge and Schneider have called for prospective validation of the biomarkers identified in the E2100 substudy, as well as markers identified by other research groups, in order to account for discordant findings and marker heterogeneity. "We clearly need validation of Schneider's biomarkers, as no one would accept it based on a single study," Sledge said. "In addition, the SNP technology is changing at a rapid pace. The initial work was done with a handful of target SNPs, whereas current technology allows you to look at over a million."
Genentech has previously said that first- and second-generation VEGF assays used in different clinical trials yielded varying results as to whether identified markers were predictive of drug response or prognostic of patient outcome. The company previously told PGx Reporter that Phase III trials of Avastin that used a "first-generation" VEGF assay indicated that VEGF "was a strong prognostic — but not predictive — marker for Avastin’s efficacy." Studies using a "second-generation" VEGF test, however, showed that "VEGF at baseline demonstrated a potential predictive effect in [metastatic breast cancer] and pancreatic cancer for patients with samples available."
Schneider is validating response markers seen in E2100 in another trial, E5103. In this study, funded with a $5.8 million grant from Susan G. Komen for the Cure, Schneider will also try to find new response markers for Avastin by conducting genome-wide association analysis in 2,204 patients using the Illumina Infinium Human Omni1 array.
Speaking during a presentation at ASCO's annual meeting this year, Schneider said that during the E2100 study, "our group recognized that although this was touted as a targeted agent we really didn't know which patients would gain the most benefit with this drug, and clearly we didn't know which patients may gain an overall survival benefit."
Schneider believes that although E2100 didn't show an improvement in overall survival, the PFS benefit seen in patients treated with Avastin was "clinically significant."
Have topics you'd like to see covered in Pharmacogenomics Reporter? Contact the editor at tray [at] genomeweb [.] com.