New US Food and Drug Administration draft guidance will help sponsors use external data like real-world evidence instead of factorial clinical trials to demonstrate the individual effects of each drug in a combination therapy, provided they can overcome the limitations inherent in the data.
A factorial clinical trial remains the most effective way to measure the effects of combination drug treatments, the FDA said, and an adaptive design “decreases the overall number of participants needed for the trial, while also limiting the number of participants exposed to potentially less effective therapy,” the draft guidance states.
Companies and patient advocacy groups have argued that even factorial trial designs, which are multi-arm, randomized studies with a combination arm, the monotherapy arms, and potentially a standard-of-care arm, could expose patients to a less efficacious drug in a single-agent arm. They also are complex and costly for drugmakers, requiring a longer trial with more patients.
The draft guidance, released July 17, allows sponsors to potentially use “external data” to satisfy the evidence requirement for combination products, provided:
- “There is strong biological plausibility for the combination regimen
- The natural history of the disease is highly predictable
- The drug as a single agent has been demonstrated to not be as effective as compared with its use in combination with other classes of drugs, and/or
- The magnitude of the treatment effect of the combination is expected to be large.”
The alternate approaches “may accelerate development of novel combination regimens and decrease participant exposure to potentially less effective therapies,” the FDA said in the guidance.
Guidance Sets High Bar, Has Limited Scope…
“I think with this new guidance document, the FDA has made clear that factorial trial designs continue to be the gold standard, but they did acknowledge flexibilities,” Mark Stewart, VP of science policy at the Friends of Cancer Research told the Pink Sheet.
“Even then, the FDA has obviously set quite a high bar when it comes to thinking about how to use real-world data and other external data sources outside of the clinical trial,” Stewart said. “I’m curious to see what the reception might be from the FDA when companies come to them with trial designs leveraging external data.”
“There are a couple of things that the guidance didn’t touch on that also are frequently encountered by the companies such as expectations for dosing when it comes to these types of therapies and the tolerability and safety piece,” he added. “Those are important factors when you’re talking about combining two agents that have their own toxicities.”
“Being able to tease out the adverse events is important too, from a clinician standpoint, if they need to make modifications,” Stewart said. “Which of the therapies in those combinations is driving the safety events, and how can we mitigate that through dose modifications, or some other strategy?”
In addition to not addressing safety or dosing, the FDA added that the guidance “does not address contribution of effect in settings where an investigational drug is being developed in combination with a drug approved for the same indication for the purposes of comparing the approved drug to the combination, i.e., ‘add-on’ trials to standard of care (SOC), or to fixed combinations of previously approved drugs for the approved indication(s).”
…But Could Have Broad Impact On Growing Trial Category
A growing understanding of the science behind the causes of cancer has increased the popularity of a multidrug approach targeting multiple mechanisms underlying the disease, the FDA said.
Tens of thousands of combination drug trials are underway, according to clinicaltrials.gov. The guidance would impact several companies, including those testing immunotherapies or antibody drug conjugates along with experimental treatments.
Johnson & Johnson in 2017 gained approval by using external data along with findings from its Phase 1b EQUULEUS trial of Darzalex (daratumumab) in combination with pomalidomide and dexamethasone for relapsed or refractory multiple myeloma patients with two or more prior lines of therapy who were refractory to their last treatment.
The new FDA guidance broadened the scope of prior guidelines to include investigational drugs combined with drugs approved for different indications and combinations of two or more drugs already approved for different indications. Its 2013 predecessor focused solely on developing two or more new drugs not previously approved for any indication.
The draft guidance is open for comment until Sept. 15.
FDA Sets Quality Tiers For External Data
The FDA said external data can provide “varying levels of evidence” for the contribution of effect.
“External data from clinical trials (same setting, same indication) may offer a high degree of relevance, especially when clinical trials overlap in time and the data is contemporaneous, as compared with data from a previously conducted clinical trial which may introduce temporal biases,” the agency wrote.
“Prospectively collected patient-level data, e.g. registry data, that includes demographics, disease characteristics, and treatment and outcomes of interest,” is listed next, followed by “other patient-level real world data.”
The FDA also lists “summary-level evidence from previously published trials or from previously published observational (non-interventional) studies,” but states elsewhere in the guidance that it “should be considered only hypothesis generating for a prospective trial.”
In terms of the most appropriate endpoint, the FDA said “overall survival is a well-defined and objective endpoint, however, certain real-world data sources can be incomplete in collection of death data, and the outcome may be confounded by anti-cancer therapy that an individual receives subsequent to trial participation.”
Other time-to-event endpoints may be subject to bias “for not only the baseline measurements, but also for the measurement of event time,” the FDA wrote in the guidance. “The comparability of index date and follow-up time between arms should be cautiously considered.”
Patient-reported outcomes and biomarkers, “when validated,” also could be used, the draft guidance states.
The agency said sponsors interested in using external data “are encouraged to consult the responsible FDA review division as early as possible.”