FDA’s dose-optimization requirements for cancer therapies are sure to shake up clinical trial designs and clinical development strategies, and companies are calling for more clarity on how exactly the processes will change.
In public comments submitted in response to FDA’s draft guidance on dose optimization for oncology therapies, which was released in January, representatives from biotech companies and the NIH raised questions about where the dose optimization steps should be built into the clinical development process, how optimized doses should be established for combination regimens, and whether the guidance applies to rare and pediatric cancers.
How FDA addresses these questions in the next version of the guidance, and in its practice, will help biotechs understand how much dose optimization is going to change the costs and timelines of clinical development.
The draft guidance redefines the way dose-finding experiments will be carried out in cancer clinical development. It moves the field away from the maximum-tolerated dose (MTD) model that has become standard and instead focuses on finding an optimal dose for both safety and efficacy.
The benefits to patients are clear. Newer targeted therapies and immunotherapies often reach an efficacy plateau in dose escalation long before dose-limiting toxicities are reached, which means the new paradigm could lead companies to advance lower doses that minimize toxicities.
Though some biotechs see the dose-optimization framework as an opportunity for product differentiation, others have raised concerns about how it will impact the ability of small biotechs to run Phase I trials.
Those concerns were echoed in a public comment on the draft guidance from Jeffrey Moscow, chief of the Investigational Drug Branch for the Cancer Therapy Evaluation Program at NIH’s Division of Cancer Treatment and Diagnosis.
“Our concern is that the proposed implementation of the FDA’s guidance will significantly slow oncology drug development and will substantially increase the cost of oncology drug development,” he wrote, adding that it “runs counter to FDA initiatives to streamline clinical trials by requiring additional correlative studies on additional patient cohorts.”
The comment argued that dose-optimization studies shouldn’t be carried out in Phase I trials, but in the later stages of development. It reflects a broader confusion that runs throughout the public comments about exactly when dose optimization should take place.
It isn’t clear in the draft guidance just how much Phase I trial designs will need to be expanded and modified, which is especially relevant to small biotechs bringing therapies to the clinic with limited resources.
The draft guidance recommends a randomized, parallel dose-response trial design to compare dosages. However, it leaves open whether the randomized trial should happen prior to a registrational study, or as part of a registrational study. Given the heavy use of accelerated approval in oncology, running a randomized dose-response study before a registrational trial may mean incorporating it into Phase I.
It’s also unclear from the guidance whether the dose-finding part of the process, which will identify the doses for randomization, needs to be expanded, and by how much. If dose finding must be expanded, companies wonder whether that can take the place of a randomized study. The guidance states that “it may be reasonable to add more patients to dose-level cohorts in a dose-finding trial which are being considered for further development.”
The NIH comment argued that the added burden of dose optimization in Phase I goes beyond the effects on small biotechs; it also exposes a larger number of patients to unnecessary toxicities and ineffective therapies. Most therapies tested in Phase I studies do not prove to be beneficial for patients, Moscow wrote.
Multiple comments also raised questions about the design of the randomized trials. The guidance states that the studies don’t need to be powered to demonstrate superiority or non-inferiority of a dose, but stops short of recommending criteria for dose selection based on that data.
A response from Friends of Cancer Research states that “without statistical power, methods for evaluating data collected to adequately justify the optimal dosage(s) should be discussed.”
Other responses call for language that allows flexibility in dose-optimization study design. For example, a comment from Amgen Inc. (NASDAQ:AMGN) recommended that non-randomized dose-optimization designs, such as backfilling of cohorts, be considered on a case-by-case basis.
A comment from Certara Inc. (NASDAQ:CERT) raised the issue of patient populations. Most Phase I solid tumor studies evaluate a therapy across a range of tumor types, and the optimal dose may be different across cancer settings. “Can a randomized dose optimization study be conducted in a heterogeneic tumor population to identify a dosage for a registrational study in a specific tumor subtype?” the company asked.
One of the most frequent questions was around dose optimization for combination therapies, with many companies wondering whether dose optimization needs to be conducted for each agent separately and for the combination.
“Should monotherapy dose finding/dose-response/exposure-response be established prior to investigating combinations? Does the dose of the combination agent need to be optimized,” Certara asked.
Several commenters were also interested in learning about the exemptions from dose optimization. Given that recruitment for pediatric and rare cancer trials is already difficult, they questioned if and how dose optimization would apply in those settings.
The 60-day comment period closed Monday.