LOS ANGELES – The US Food and Drug Administration's accelerated approval pathway will be critical to getting gene therapies to patients with rare diseases, Peter Marks, director of the agency's Center for Biologics Evaluation and Research, said at the American Society of Gene & Cell Therapy's annual meeting here on Wednesday.
It's challenging for drugmakers to do randomized, placebo-controlled studies to demonstrate that investigational drugs that treat rare diseases improve patients' outcomes, since by definition these conditions affect a small population. In such circumstances, the accelerated approval pathway in which companies can gain approval for therapies based on how they impact a surrogate endpoint that is likely to correlate with clinical outcome improvements — for example, how a biomarker level changes from before and after treatment — can be helpful.
"When you're dealing with very small populations, we're probably willing to use some regulatory flexibility there," Marks said. "We'll try to do our best here, looking at each case individually."
His remarks come as the US Food and Drug Administration is mulling whether to grant accelerated approval to Sarepta Therapeutics' investigational gene therapy for Duchenne muscular dystrophy, SRP-9001, based on a surrogate endpoint — a highly anticipated decision the agency is expected to issue on May 29. An FDA advisory committee narrowly voted 8-6 in support of accelerated approval for the drug last week, after weighing concerns from FDA reviewers.
Marks in recent months has said multiple times that the agency wants to use the accelerated approval pathway to expedite getting investigational gene therapies to patients. Under this pathway, sponsors must demonstrate that a surrogate endpoint is "reasonably likely" to predict clinical benefit in patients. A sponsor, however, must commit to submitting a post-market trial that confirms the drug improves patient outcomes before the FDA can convert the therapy's accelerated approval into a full approval.
If such confirmation doesn't come, the FDA can withdraw the accelerated approval and remove the drug from the market. And in recent years, as part of an industry-wide review of the accelerated approval program, the agency has been following up on past approvals of cancer drugs to ensure post-market clinical data support continued commercial access, which has led several companies to pull their drugs from the market.
"We will have to have the confirmatory data at some point," Marks said. "Although we want to be flexible like a rubber band, … we can't have it to the point that we actually snap the rubber band."
When deciding whether to approve a new therapy, the FDA must weigh the nature of the disease, whether there are alternative therapies, and any uncertainties around the safety and efficacy of the drug in question. "All of those things have to be put together," Marks said. "It creates some very challenging situations."
If a confirmatory trial turns up negative results after a therapy has already received accelerated approval, the FDA will try to understand why it failed and review the study for any potential problems with trial design or methodology, he added. If the agency determines, however, that the therapy doesn't actually treat a disease as expected based on preliminary data seen during the accelerated approval process, it will remove the drug from the market.
"If we do our job well, it won't happen often, but it probably is going to happen occasionally," he said.
Marks assured, however, that for patients suffering from fatal, rare genetic diseases who have few or no other treatment options, the agency wants to get treatments to them quickly. But the agency also has the best interests of patients in mind, he said, in making sure that post-market data shows these treatments are beneficial, and not harmful, to them. "If, occasionally, we have something that we give an accelerated approval to, but it doesn't confirm out … and that happens occasionally — I'm not sure that's the worst thing at the end of the day," Marks said.