SAN FRANCISCO—Recent Phase III trials show that upfront treatment with bevacizumab prolongs progression-free survival in patients with ovarian cancer. Whether the anti-angiogenic agent should be considered a new standard of care, however, was up for debate here at the most recent Oncology Congress.
The current standard of care for ovarian cancer patients is surgery followed by systemic therapy with paclitaxel and carboplatin. Additional options are available, said Bradley Monk, MD, but are less commonly used, including the addition of a targeted agent. “I am going to try to persuade you that a targeted agent is not only acceptable but is the standard,” he said.
The pivotal data come from an 1800-patient randomized Phase III cooperative group trial, GOG0218. Women who received chemotherapy plus bevacizumab with bevacizumab maintenance therapy had a median progression-free survival of 14.1 months, compared with 11.2 months for women who received chemotherapy plus a short course of bevacizumab or 10.3 months for women who received chemotherapy alone.
“This was the number one abstract at ASCO this year ,” Dr. Monk said, emphasizing the community's excitement over the data.
Moreover, there were no unexpected adverse events in the trial. Thirteen patients (2.1%) in the bevacizumab maintenance arm had a gastrointestinal perforation compared with eight patients (1.3%) in the short-course bevacizumab arm and five patients (0.8%) in the chemotherapy-alone arm.
“Bevacizumab does cause bowel obstruction in ovarian cancer, but the rate is low if the number of prior regimens is low, and this was a chemotherapy-naïve group of patients,” Dr. Monk said.
“Dr. McGuire is going to tell you that progression-free survival is not important. Progression-free survival is important. This hazard ratio, 0.72, is clinically meaningful.” Overall survival data are not yet mature, with approximately 75% of patients still alive. “So you can't say that bevacizumab doesn't improve overall survival. It is too early to tell.”
Previous analyses have shown that progression-free survival correlates tightly with overall survival in ovarian cancer and, therefore, is a good surrogate marker in this disease setting. “In fact, the FDA, in our consensus conference with ASCO and AACR, confirmed that progression-free survival is a valid independent measure of benefit when measured appropriately, such as in a placebo-controlled trial like GOG0218,” Dr. Monk said.
The results of a second large randomized Phase III trial, ICON7, also showed improvement in progression-free survival with the addition of bevacizumab. In this British-run trial, women who received chemotherapy plus bevacizumab with bevacizumab maintenance therapy had a median progression-free survival of 19.0 months compared with 17.3 months for the women who received chemotherapy alone.
The difference was statistically significant, with a hazard ratio of 0.81. However, the progression-free survival curves cross between months 21 and 24.
Dr. Monk cautioned that a lot is going to be made of these crossed progression-free survival curves, but that they are not as important as some people will make them out to be. The trial used a lower dose of bevacizumab than was used in GOG0218—7.5 mg/kg every three weeks instead of 15 mg/kg—and maintenance therapy was for 12 months, compared with 15 months in the US trial.
“Maybe it should be given for a longer time,” he said. Also, he noted, the overall survival curves are trending in the right direction, though more than 80% of patients remain alive.
The next research priority is to evaluate bevacizumab's impact on quality of life, Dr. Monk said. As the Quality-of-Life Chair for GOG0218, Dr. Monk declined to provide specific information, but said the data should be available very soon.
The optimal duration and dose of bevacizumab also needs to be determined. “Less is always less money, but more is not always more effective. This is the urgent scientific priority,” he said.
Finally, he said that there are four other trials testing the drug in various subgroups of ovarian cancer patients, including those with platinum-sensitive and platinum-resistant disease. “So if you don't believe two positive trials, ICON7 and GOG0218, that is fine. But we have four other trials that will also be positive and we will get there. But hopefully your patients can begin to receive the benefits of bevacizumab today.”
To launch his argument, William P. McGuire, MD, noted there are basic requirements that must be met in order for a therapy to be practice changing: similar or improved toxicity and quality of life; significant improvements in outcomes, including overall survival; a consideration of pharmacoeconomics; and replication.
He noted that the toxicity associated with bevacizumab is acceptable in previously untreated ovarian cancer patients, but that no data are yet available regarding quality of life. “My understanding from having my ear to the ground is that there is no improvement in quality of life,” Dr. McGuire said.
And, he suggested, patients' financial quality of life should not be ignored. If patients have a 20% copay, which is common, and bevacizumab is given until progression, “then I think none of the people who have ovarian cancer are going to send their kids to college.”
As for efficacy, the data show a clear increase in progression-free survival with maintenance bevacizumab therapy, but not with a short course of the drug. However, half of the progression events occurred while patients were on therapy, suggesting that the overall survival data will mature quickly. “I, therefore, am looking forward to some mature survival data,” he said.
As expected, Dr. McGuire raised the issue of progression-free survival as a marker of benefit. He concurred that there is a strong correlation between progression-free survival and overall survival in first-line therapy, but pointed out that the correlation is based on trials that tested cytotoxic agents.
“In trials of biological therapy, in both primary and recurrent disease, this correlation appears less strong,” he said.
Therefore, despite the FDA's agreement that progression-free survival is an adequate surrogate for overall survival, Dr. McGuire said caution should prevail where biologic therapies are concerned. “I believe we must wait for overall survival data here.”
In support of that view, he pointed to the controversy regarding bevacizumab in breast cancer, where progression-free survival fails to translate into overall survival benefit. And if one looks at bevacizumab trials across different types of solid tumors, the correlation between progression-free survival and overall survival in other solid tumors breaks down, with an R2 value of just 0.61.
“I think we have to be wary of assuming that this difference in progression-free survival is going to translate into an overall survival advantage,” he said.
If the progression-free survival hazard ratio seen in GOG0218 does hold for overall survival, and one assumes that the median overall survival in the control arm is 39 months, based on previous trials, then the median overall survival in bevacizumab maintenance arm would be 54 months. “Looking at early survival curves it is unlikely we are going to see a median survival of 54 months on this trial,” he said.
Dr. McGuire was also skeptical about the cost-effectiveness of bevacizumab therapy. A “back of the envelope” calculation indicates that bevacizumab would cost $72,576 for the median patient treated in the maintenance therapy arm of GOG0218. As of today, there appears to be a 3.8-month gain in progression-free survival. That translates into a cost of $229,187 per year of progression-free survival gained.
In general, treatments that cost $50,000 per year of life saved are considered a “good buy,” he noted. Therefore, even if the progression-free survival gain holds for overall survival, bevacizumab “is unlikely to be considered cost effective, at least at its current cost.”
The good news is that the ICON7 trial does essentially replicate the GOG0218 results, though the difference is less robust than the earlier trial. And, again, the overall survival data are not mature.
“Bevacizumab does have a clear clinical effect, and the adverse events are acceptable, but on the basis of today's study, however, with a progression-free survival gain of this magnitude alone, it is of unclear benefit to patients,” Dr. McGuire concluded. “We clearly need mature overall survival data. We need quality-of-life data. And we clearly need a formal pharmacoeconomic analysis to see, if, in fact, even if there is a survival advantage, this is a good buy.”
A Final Word
In a talk at the Congress on targeted therapies later in the day, Richard T. Penson, MD, Clinical Director Medical of Gynecologic Oncology at Massachusetts General Hospital, revisited the GOG0218 and ICON7 data. He said that while bevacizumab increased progression-free survival in the two trials, the magnitude of benefit is not very large and will not translate into improved overall survival. “The worry is that the [early] overall survival curves are overlapping” in GOG0218, he said. “And the word on the street is that ICON7 has worse survival curves.
“Bevacizumab is unequivocally a great drug,” he said, “but it is not for everyone.” Therefore, the community should be doing more to identify the subsets of patients who benefit from the agent.
As for the importance of cost in medical decision making, he said, “I think bevacizumab is going to be the test of cost-effectiveness in the United States.”
Post-Debate Discussion & Audience Participation
Post-debate audience participation questions showed that neither debater carried the day. Nineteen percent of the audience said GOG0218 data would change their practice as of today, 33% said it would not, and 48% said they were not sure.
When asked how many trials practice changing regimens should be based on, 7% of the audience said a single trial was adequate, 42% said more than one trial, 40% said more than two trials, and 12% said they were unsure.
The vast majority of audience members (79%) thought that overall survival was the best endpoint for measuring efficacy in previously untreated ovarian cancer, compared with 11% who said response rate, 5% who said progression-free survival, and 5% who were not sure.
During the post-debate audience questions, Drs. Monk and McGuire continued their discussion. With Dr. Monk emphasizing that the drug is active and that the key thing that needed to be determined was not whether it should be used but for how long and in what patients.
Speaking up from the front table, Session Chair Deborah K. Armstrong, MD, Associate Professor of Oncology at the Sidney Kimmel Comprehensive Cancer Center at Johns Hopkins and Associate Professor of Gynecology and Obstetrics at Johns Hopkins University School of Medicine, noted that just having an active drug might not be all the community needs.
“I treat breast cancer, and whenever I look at this issue I just replace ‘bevacizumab’ with ‘trastuzumab,’” she said. “It is true that trastuzumab is an active agent, but what it is, is that in a fraction of patients it is very active and in the majority of patients it is inactive. To me, what we need to find is the HER2 expression equivalent for bevacizumab,” referring to the HER2 protein expressed in trastuzumab-sensitive cancers.
Dr. McGuire agreed that finding a biomarker for responsive patients would be helpful but noted that even 18 years after paclitaxel became the standard of care, researchers have yet to find a good biological correlate to identify patients who should or should not receive the drug. “It is not a hopeless idea, but we have a long way to go,” he said.
In an interview at the close of the session, Dr. Monk sounded more tentative than he did during the debate when asked if this should be practice changing. “Not every patient is treated the same way,” he said. “So it is about timing, it is about dose, and it is about duration. And integrated into that are the economics of a very expensive drug, which admittedly has a relatively small benefit, but fortunately is not very toxic.”
Also afterwards, Dr. McGuire said that when he first reported the pivotal cisplatin-paclitaxel data, “there was a lot of pushback from physicians from European countries with national health plans, because paclitaxel was so expensive—even though we had overall survival data. And it was nowhere near as expensive as bevacizumab.”