Nothing Random About It: The State of Cancer Drug Development at FDA

FDA is getting tougher on cancer drug approvals. But stakeholders on all sides-industry, regulators, and patient groups-say that's probably a good thing. Under the leadership of Richard Pazdur, MD, the controversial director of FDA's newly formed Office of Oncology Drug Products, the agency is raising the bar for getting cancer drugs on the market.

FDA cancer drug head Richard Pazdur says the agency wants to see more rigorous randomized trials. That is a good thing for oncology drug developers.

Ramsey Baghdadi

"Good drug, bad development." That is how Food & Drug Administration director of the Office of Oncology Drug Products Richard Pazdur, MD, described ImClone Systems Inc. ’s colorectal cancer drug cetuximab (Erbitux) during a 2002 Congressional hearing.

The insinuation was that Erbitux—a hyped-up novel cancer therapy—was probably effective, but ImClone fumbled the clinical trial design and failed to prove it. Many attendees of the hearing and those reading media reports of the Erbitux probe were flabbergasted. The head of FDA’s cancer division was admitting that he thought the drug was likely a therapeutic advance, but he was unwilling to put his support behind it because of what was essentially a technicality.

"If the head of the oncology division thinks a drug is good…you should approve the drug on a provisional basis so people who are dying everyday at least have a chance to get it while you get to do more studies," one senior official at a major cancer company says.

But that was not the thrust of Congress’ investigation at the time. Pazdur was a key figure in the probe because he allegedly contacted an outside consultant who would later notify Erbitux partner Bristol-Myers Squibb Co. that ImClone was going to get a "refusal to file" letter for the drug. The rest is history. A high-profile insider trading case involving one of the former top executives at ImClone—and his friend, Martha Stewart—resulted in serious jail time. Cetuximab was eventually approved in February 2004 for colorectal cancer, several years after the companies, and cancer patients, thought it would be.

Three years later, a lot has changed. FDA has consolidated all of its cancer review efforts under an office-level group within the Center for Drug Evaluation & Research, including biologics, chemo-preventative drugs and diagnostic imaging. After a national search for a head of the new oncology office, the agency went with who they knew best: Pazdur.

Pazdur is now the single most important person to oncology approvals in the country. He serves as the gatekeeper to a $24 billion cancer drug market in the US, which grew 17% in 2004 and is now the third largest therapeutic class, according to IMS Health Inc. The FDA’s cancer drug czar has headed up the newly formed oncology office for less than a year, but has already put his stamp on the most critical review division in the country. He is the rare government regulator that is larger than the office he supervises due to his hands-on approach and bold, straight-talk demeanor.

Pazdur also has almost exclusive jurisdiction over one of the most important drug development tools for large pharma and smaller biotech companies—accelerated approval (AA) (See "The State of AA"). The regulatory shortcut allows companies to get a conditional approval based on surrogate endpoints that "reasonably" suggest clinical benefit, with the promise of postmarket Phase IV confirmatory trials. (See "Hurry Up and Wait: The Slowdown in Accelerated Approvals," IN VIVO, July 2005 (Also see "Hurry Up and Wait: The Slowdown in Accelerated Approvals" - In Vivo, 1 Jul, 2005.).)

One thing hasn’t changed: "Good drug, bad development" continues to sum up the Pazdur’s mindset, at least in the view of some oncology drug developers. One thing is clear: Pazdur won’t sign off on a drug based on mediocre clinical studies and a gut feeling.

But three years after Erbitux, more people seem to be buying into Pazdur’s vision. Is Richard Pazdur good for oncology drug development? The answer, according to a number of current and former high-level cancer company execs, regulators, and patient advocates is a measured "yes." Most think Pazdur is establishing a more rigorous set of data expectations, and thus, creating more hurdles to winning FDA approval. But even drug sponsors acknowledge that Pazdur’s high standards may ultimately help the quality of oncologics on the market.

Cancer Drug Reviews

Cancer drug reviews are taking longer than they used to. In 2005, it took an average of 204 days from the time of submission to an FDA approval decision, according to data compiled by the Stanford Washington Research Group. That was approximately 21% longer than the 169-day average in 2004. (See Exhibit 1 and Exhibit 2) Add to that the fact that there were only three cancer drugs approved last year compared to eight drugs in 2004. Simply put, FDA is taking longer to approve fewer novel oncology agents.

Nevertheless, oncology R&D is on a roll right now. FDA has approved three new therapies over the last three months, a remarkable clip for any drug class.

At the end of December, FDA approved Celgene Corp. ’s lenalidomide (Revlimid) as a first-line therapy for transfusion-dependent anemia patients with myelodysplastic syndrome (MDS). Bayer AG /Onyx Pharmaceuticals Inc. ’s Dec. 20 approval of sorafenib (Nexavar) was one of the first for renal cell carcinoma in over a decade. In late January, FDA simultaneously approved the Nexavar competitor sunitinib (Sutent), manufactured by Pfizer Inc. , for two indications—a first in oncology. Sutent was cleared for both advanced renal cell carcinoma and as a second-line treatment for gastrointestinal stromal tumor (GIST).

And all three drugs come to market with compelling evidence that they will have an impact on the course of disease.

Revlimid helped 67% of patients in an open-label, single-arm Phase II trial of 148 MDS patients with red blood cell (RBC) transfusion-dependent anemia achieve independence from RBC transfusion. The median duration of transfusion independence was 44 weeks post-treatment (See "Building A Business in Drug Safety," The RPM Report, January 2006 (Also see "Building a Business in Drug Safety" - Pink Sheet, 1 Jan, 2006.)).

A randomized, placebo-controlled trial of Nexavar found a median time to tumor progression or death in the treatment arm of 167 days compared to 84 days for the control group.

Median time to progression for GIST patients treated with Sutent was 27 weeks compared to six weeks in the placebo group. In the second indication, metastatic kidney cancer patients treated with Sutent experienced a 26%-37% response rate.

Those are compelling results in oncology, where the slightest hint of progress can create a buzz in the clinical community and garner attention from Wall Street. For example, the last kidney cancer drug before Nexavar—Chiron Corp. ’s aldesleukin or interleukin-2 (Proleukin)—was approved in 1992 based on response rates in 13%-15% range and a toxic death rate of 4%. Interferon is also used off-label in kidney cancer but demonstrates sporadic efficacy and is also very toxic.

Pazdur highlighted Nexavar’s impact on progression-free survival (the time from randomization to progression or all-cause death) as clinically meaningful and highly statistically significant. Pazdur called the data "one of the most impressive results I’ve ever seen in oncology" during a media call announcing the approval.

So if the drugs have stronger data, why are they taking longer to reach to the market? The answer: better data takes more time to gather, and often involves more rigorous trial designs than once were common in oncology.

Randomized Trials, Not Random Decisions

When asked what is the number-one thing oncology drug developers can improve upon to secure approval, Pazdur is unhesitant in his answer: randomized clinical trials.

In a perfect world, clinical development experts say, randomized trials would be the rule, not the exception. However, in oncology, ethical constraints, the difficulty in enrolling patients and raising cash—particularly for smaller biotechs—make conducting such studies difficult.

From a regulator’s point of view, the key advantage of randomized trials is to avoid reliance on tumor response rates, a notoriously subjective endpoint that often turns approval decisions into guesswork. Instead of analyzing the percentage of patients with shrinking tumors, randomized trials can look at event endpoints such as survival, time to progression and progression-free survival. From a risk perspective—a relative term in cancer—randomized studies give FDA reviewers a more comprehensive look at drug toxicity since toxicity in the treatment arm can be compared to toxicity of a control arm.

Pazdur says emphatically that he is not only encouraging sponsors to use randomized studies for standard full approvals, but for accelerated approvals as well. Even if a company is compelled to use a non-randomized trial for its first indication, conducting multiple randomized studies in different diseases can give FDA more confidence in the evidence of the drug’s clinical benefit, Pazdur says.

In fact, Pazdur claims that Nexavar and Sutent would have been in danger of getting turned down had it not been for their rigorous clinical development plan. "The more recent drug approvals have shown with both Nexavar in the renal cell application and also Sutent in the GIST application the need for randomized studies," he says.

For example, if Bayer had conducted a single-arm trial, Nexavar would have demonstrated a 2% response rate, which certainly would have drawn a "not approvable" letter from FDA. And Pfizer conducted a randomized GIST study to support Sutent. The company also performed a 55-patient, single-arm study for the GIST indication looking at progression on or intolerance to Novartis AG ’s breakthrough leukemia drug imatinib mesylate (Gleevec). That study showed a 9% response rate. "If we were relying just on response rates in these two applications, it would be very difficult to ascertain the benefit of the drug," Pazdur says.

Still, some companies and patient groups argue that placebo-controlled studies are medically unethical because they prevent patients from taking a potentially beneficial new drug. Pazdur says placebo studies are somewhat of a misnomer in cancer because patients are receiving best supportive care along with placebo—they’re just not getting the experimental drug itself.

Until recently, placebo-controlled studies were not common in oncology mainly because the toxicity of conventional chemotherapy agents is so obvious that it is not realistic to have a truly blinded study. In addition, dosing schedules and intravenous dosing regimens made placebos impractical. All of that changed with the advent of oral chemotherapeutic agents, and especially with the development of more targeted therapies with the potential for lower toxicity.

In any event, Pazdur stresses that FDA does not mandate placebo-controlled studies and says there are a number of safeguards in place because of the agency’s sensitivity to the issue. For one, FDA looks at populations where there is no effective therapy—either approved or those recognized by the medical community. FDA also allows patients to cross over if the experimental agent is deemed clinically beneficial.

In the Nexavar study, FDA met early (April 2005) with Bayer after an initial interim analysis and urged the company to stop randomization and allow cross-over to active treatment. FDA then ensured Bayer that a meaningful impact on progression-free survival would warrant a full approval.

On the other hand, companies that come to FDA with a single trial, particularly a single-arm study, are in jeopardy before they even get through the door. Revlimid was an exception. "The magnitude of benefit was important in making a decision regarding that drug," Pazdur says. "If we were talking about a much smaller magnitude of difference, albeit that it might have been statistically significant, it might have been very difficult to understand the benefit of that drug on transfusion reduction."

Approval Standards are Moving on Up

Indeed, a quick look at several drugs that failed to win FDA approval indicates that the agency has raised its review standards.

For example, Johnson & Johnson ’s acute myeloid leukemia (AML) drug tipifarnib (Zarnestra) demonstrated a remission rate of 15% in 135 patients. Enzon Pharmaceuticals Inc. /Inex Pharmaceuticals Corp. ’s anticancer drug vincristine sulfate liposomes (Marqibo) exhibited a remission rate of 21% in 119 patients. Both drugs were rejected by FDA in 2005.

Earlier approvals had much lower response rates. Pfizer’s metastatic colorectal cancer drug irinotecan (Camptosar) and Roche ’s metastatic breast cancer drug capecitabine (Xeloda) demonstrated low double-digit response rates and were approved in 1996 and 1998, respectively.

Pazdur is unapologetic: "We’re not making any progress in cancer simply by lowering the standards by which we approve drugs."

More recently, Abbott Laboratories Inc. ’s metastatic prostate cancer agent atrasentan (Xinlay) was shot down by FDA’s Oncology Drugs Advisory Committee in September. That drug exhibited a 14% reduction in disease progression compared to placebo in approximately 1,000 patients as part of a meta-analysis of two studies, which failed to meet their primary endpoints. Pazdur and the ODAC members were clearly underwhelmed by the data—and the development plan.

Although randomization is helping FDA assess other critical clinical endpoints for approval, survival remains the gold standard in oncology. But even there, some sponsors suggest, FDA is nudging the bar higher.

What is now considered a significant advancement in survival? Pazdur says there is a lot of controversy, but no real answer. "We have generally looked at any advantage in a survival endpoint because of the difficulties in showing it," he says. But even that is open to interpretation when toxicity is taken into consideration. For example, if one drug showed a survival advantage of a few days but was very toxic, FDA would send that agent before an advisory committee.

The idea that FDA would not immediately approve a drug that shows a survival benefit over standard treatment—regardless of the toxicity—illustrates that approval standards are rising, some companies say. It "looks on the surface to be demanding and setting the bar higher," one former drug cancer executive says.

Not everyone sees it that way. There is an undercurrent of FDA oncology medical reviewers—former and current—who allege that Pazdur has lowered the bar for approval. Those critics point to Erbitux and Iressa, which exhibited an 11% response rate in trials, as proof that Pazdur hasn’t done anything to make the standards for approval more rigorous.

One former FDA reviewer points out that progression-free survival—not survival itself—is now the first-line clinical endpoint of choice in colorectal cancer. Achieving progression-free survival is considered an easier task than demonstrating survival. While oncology drug development is inherently political, the former reviewer claims that Pazdur has amplified the political tone in the oncology office. "It’s a much more political process than it was previously."

Ellen Sigal, PhD, chair of Friends of Cancer Research, a powerful organization that comprises most of the major cancer advocacy groups, is also adamant that approval standards in oncology haven’t been raised. She says expectations for FDA have never been clearer and the elevation of the oncology division to an office within the Center for Drug Evaluation & Research means cancer is getting the attention it has long deserved.

"I think the bar is being raised by technology," Friends of Cancer Research analyst Jeff Allen, PhD, says. Allen claims that the improved ability of companies to pick and choose the most effective molecules to develop has led to fewer novel agents being presented to FDA, and therefore, a perceived slowdown in approvals and loftier requirements for market clearance.

Sigal is more concerned about the effect of heightened awareness of drug safety—the specter that is casting a shadow across FDA review units—will have on cancer applications. She does, however, acknowledge that FDA is increasingly asking sponsors to conduct placebo-controlled studies, but that’s a good thing in her opinion. "What I’m really worried about, which is denied by oncology, is this impact of safety and more data being required because the public debate on safety is so misunderstood. That is a big concern."

And sponsors acknowledge that FDA has good reasons for pushing for better trials. "If [FDA] accepts a poor program but they are convinced it is a good drug, then obviously they are setting a precedent of being accepting of poor development," the former company official maintains. And that is precisely the dilemma Pazdur is trying to avoid. "Pazdur wants to get away from quick and dirty studies with very limited patient numbers and resorting everything to post-approval commitments—I think he wants a little more data before he is ready to approve it."

Some companies say their conversations with Pazdur and the oncology office clearly indicate the agency’s standards are now higher. "There’s a sense of urgency about cancer and a lot of drugs have been approved in the past based on very sketchy types of data," one current senior official at a cancer company says. "I think the way Pazdur looked at his mission was to put more proof into the way drugs get approved."

Sponsors are being asked to use more scientifically valid controlled trials instead of the usual open-label Phase II studies, representing a shocking change in culture in the oncology arena. "I think that cause a lot of people to get a little bent out of shape because it’s a change in the whole paradigm," the current company official suggests.

Cancer’s Public Enemy?

But no one can talk about FDA’s attitudes towards drug development without talking about Pazdur himself. Pazdur’s comments at the Erbitux hearing represent just one example of how he has served as a lightning rod within the agency and the pharmaceutical industry. He is quotable—and often quoted—and not always favorably.

Some who have dealt with him inside the agency view him as a grandstander. Others see him as refreshingly willing to tell it like it is. FDA Office of Medical Policy Director Robert Temple, MD, suggests that Pazdur is often misunderstood: people sometimes take his off-the-cuff remarks at an advisory committee meeting and assume that new official policy has been presented.

Whatever the reason, Pazdur has found himself a regular target of the editorial pages of The Wall Street Journal.

For example, the Journal ran a scathing editorial after AstraZeneca PLC ’s lung cancer drug gefitinib (Iressa) was effectively taken off the market in summer 2005. FDA’s decision to severely restrict labeling for Iressa to those patients already on the drug left Genentech Inc. /OSI Pharmaceuticals Inc. ’s competing erlotinib (Tarceva) as the only other indicated option on the market for lung cancer patients. "Dr. Pazdur’s decision to give future cancer patients—even those with the Iressa-response gene—only the option of Tarceva could well amount to a death sentence for many," the July 6 editorial asserted.

Just a few months earlier, the Journal again took the gloves off with an editorial on FDA’s "dawdling" on cancer approvals, using a female patient dying of kidney cancer to prove its point. Pazdur "seems to be more worried," the editorial charged, about preventing drug companies from getting away "with a so-called ‘race to the bottom’ of trial design than he is with getting good drugs to patients."

Pazdur’s selection to head the new Office of Oncology also ignited controversy. Some, including members of his own review staff, felt Pazdur should have been punished, not rewarded, for his role in the Erbitux scandal. "With any painful experience it is also an opportunity for growth and I think [Erbitux] has allowed us at FDA to really look…at ways of improving our communication with sponsors," Pazdur says, looking back on his experience and involvement with cetuximab (See "Post-Erbitux: How Martha Stewart Changed FDA" (Also see "California Court’s Inaction On TiO2 Prop 65 First Amendment Case Breeds New Lawsuits" - HBW Insight, 24 Apr, 2024.)).

And it appears that most cancer advocacy groups—and drug development sponsors—have rallied around Pazdur. Representatives from the Friends of Cancer Research and a number of companies, including Bristol-Myers Squibb and GlaxoSmithKline PLC recently met with the Journal editorial team to discuss the staff’s views on oncology regulation and drug development.

As a former clinician at the MD Anderson Cancer Institute, Pazdur is also part of a network of Texas oncologists who have influential roles in the Bush Administration. Pazdur himself was considered by the White House during its first round of interviews for the FDA commissioner position in 2001. (The position is currently held by another MD Anderson alum, Andrew Von Eschenbach.)

For his part, Pazdur says he’s not running in a popularity contest. "I understand that there are many people that have different viewpoints and I appreciate those viewpoints," he says. "I wish my critics treated me with the same courtesy I’ve afforded them."

In general, Pazdur believes his critics are confusing two different issues: access to unapproved drugs and approval standards for new drugs. He believes FDA is appropriately addressing those questions separately: giving patients with no other options the chance to try experimental drugs without sacrificing the gold standard of FDA approval. Sigal agrees: "You can’t just throw drugs at these people."

For example, expanded access programs allow unapproved therapies to be given to patients who have failed approved therapies or are unable to participate in clinical trials. Treatment investigational new drug (IND) programs and single-patient INDs are two routes for cancer patients to receive experimental treatment. Regulators, patient groups and drug companies readily acknowledge that expanded access is not the perfect solution. For one, there is no financial reward for manufacturers if drugs are available only through an expanded access program. Although sponsors are allowed to charge for drugs on a cost-recovery basis, they are hesitant to do so because they would set a de facto wholesale market price for the drug (See "Is the Oncology Bubble Ready to Pop?").

And it isn’t as easy to separate access to unapproved medicines from approval standards as Pazdur suggests. If standards for drug approval are indeed going up, then logic would have it that patients will have fewer therapy options because fewer drugs are hitting the market. In order to give cancer patients as many viable choices as possible, FDA needs to expand access to unapproved drugs.

Nevertheless, Pazdur is unflinching when it comes to defending approval standards. "How can you sell these drugs to the American public based on a few responses and a glimmer of activity that we see in clinical Phase I testing? I think that would be a mistake until we get more information on the drug."

The oncology office director says FDA is committed to bolstering the expanded access mechanism. With Nexavar, FDA encouraged Bayer to open up an expanded access program to include untreated or metastatic renal cell carcinomas. Bayer obliged in May, enrolling almost 2,000 patients—roughly half of those had no prior treatment. In Europe, a similar program opened in October had enrolled 90 patients as of the end of 2005. In January, FDA met with the American Society of Clinical Oncology and the National Coalition of Cancer Survivors to discuss how to improve expanded access.

The New Frontier in Oncology

Politics aside, Pazdur says the next major advance in cancer will be the ability to define patient subsets that will most likely benefit from targeted cancer therapies through the incorporation of validated biomarkers. So far, only one relatively new cancer biomarker has been predictive of a specific drug’s efficacy: the HER-2 protein test for Genentech’s breast cancer agent trastuzumab (Herceptin). Another once-promising biomarker, the epidermal growth factor receptor (EGFR) test, does not appear to work for Erbitux or Iressa.

The major issue standing in the way of progress in biomarker discovery and validation, according to Pazdur, is that drug companies are in the business of developing drugs—not diseases. The identification of biomarkers is essentially an attempt to re-define disease on a meaningful practical level, he says.

"That’s going to require earlier development of biomarkers and it being an inherent part of the drug development process at an earlier stage." Pazdur stresses the importance of incorporating biomarkers into the earliest stages of the clinical program—not after a drug has failed in the general population and the sponsor is trying to resurrect the experimental agent.

FDA and other government health agencies look like they’re ready to get serious about biomarkers. The agency, along with the National Cancer Institute and the Centers for Medicare & Medicaid Services announced Feb. 14 a federal health initiative called the Oncology Biomarker Qualification Initiative (OBQI) to collaborate on biomarker development.

The first OBQI initiative is intended to validate and standardize the use of fluorodeoxyglucose-positron emission tomography (FDG-PET). The imaging technology will be used in non-Hodgkin’s lymphoma drug trials to evaluate FDG-PET as a predictor of tumor response. The agencies expect to unveil a number of other cancer biomarker projects over the next several months. FDA is also expected to release a list of 75 "opportunities" in March as part of its Critical Path initiative; a significant percentage of those projects are sure to be in cancer.

While FDA is partnering with other governmental agencies in oncology, it has also reached across the Atlantic to its European counterpart, the European Agency for the Evaluation of Medicinal Products (EMEA), for input regarding cancer drug development and regulation. FDA conducts a monthly conference call with EMEA officials to discuss specific applications. Those open discussions with EMEA, and other agencies including Health Canada and the Japanese Ministry of Health, Pazdur says, are making valuable contributions to improving the US drug approval process.

One cancer company official, however, finds the increased interaction with European authorities troubling. The official claims the partnership is the equivalent of too many chefs in the kitchen and is leading to inconsistent advice from FDA to sponsors. One company claims that inconsistent advice led to a competitor getting approval more quickly due to differing guidelines on endpoints and trial design even though the competitor was further behind in development.

Reading the Tea Leaves

Though there are different opinions of Pazdur’s impact on drug development, one thing is certain: he takes the war on cancer personally. There will be an estimated 1.4 mil. new cancer cases in 2006, according to the American Cancer Society. Approximately 560,000—or 40%—of those patients will die. Those sobering statistics argue for access to cancer drugs—any drugs with a hint of efficacy—patient advocates, drug companies and other health care stakeholders argue.

Pazdur’s argument is clear: The FDA is not standing between the patient and the drug. There are a number of avenues available to patients to get access to unapproved drugs. That is separate from putting a drug on the open market. After all, FDA’s stamp of approval in essence says to the American public: this drug is safe and effective and the benefits outweigh the risks.

FDA says that it looks at drug reviews on a case-by-case basis, and that is no doubt true to a certain degree. But when you take the cases and put them together, it’s clear that the overall paradigm is shifting in oncology.

Pazdur wants to see multiple studies and randomized trials. That could cost cancer developers more on the front-end but it is the best way to reduce the risks inherent in FDA approval and could save money on post-market research. Companies that are willing to take their chances with a single-arm study should be sure that the magnitude of effect is strong enough to overcome skeptical reviewers, skeptical outside experts and an FDA that is adamant that it wants to see more data.

Revlimid was a clear-cut case where the effect of the drug was unmistakable. Panel member after panel member applauded the clinical effect of the drug during its ODAC review. In fact, an internal analysis by the company showed that Revlimid was the most effective cancer drug since Gleevec. Nevertheless, Revlimid still had to go before an advisory committee and FDA had to extend its review deadline by three months to evaluate the company’s post-approval risk management program.

There’s no doubt that the agency—and Pazdur himself—are serious about examining new surrogate biomarkers as the basis for approval. In the same vein, FDA is looking at ways it can contribute to making expanded access programs more effective.

But make no mistake: the bar for approval is being raised. Companies prepared to meet heightened standards will be rewarded by FDA with full approvals in the place of accelerated approvals tied to follow-up trials. They may also receive more favorable labeling, as evidenced by Nexavar and Sutent. However, oncology drug developers willing to risk going to FDA with one single-arm study better have compelling efficacy data. Otherwise they may be left wondering if they had a good drug, but a bad development plan.