Remarks by Scott Gottlieb, M.D. RAPS 2017 Regulatory Convergence Conference September 11, 2017 Washington, DC
Excerpt:
Last week, I spoke about some of the steps we’re taking to make the process for the pre-clinical development of new technologies more scientifically advanced and efficient. We’ll have much more to say about other new steps we’re taking to achieve these pre-clinical goals. But today I want to talk about some of the steps we’re taking to address the clinical part of drug development -- the traditional three phases of trials.
To address these issues, our Center for Drug Evaluation and Research, under the leadership of Dr. Janet Woodcock, is taking steps to modernize its Office of New Drugs. The goal is to make sure that our workflows and policies are rooted in the best science and management principles, and that our staff has the support and tools they need to fully achieve their public health mission. I plan to talk more about some of these steps, later this month, at the National Press Club.
I’ll also advance a Strategic Policy Roadmap that will detail additional steps we’re pursuing. These, and similar, efforts are aimed, in part, at making sure that FDA is able to adopt the modern scientific tools we need to maintain the rigor of our programs. We need to make sure that our approach to regulation is efficient, and doesn’t become an obstacle to the translation of scientific discoveries into practical solutions for patients. We need to make sure that we’re using the best science so we maintain our gold standard for determining safety and benefit.
I’ll have more to say on these topics soon. Today I want to outline some of the new steps we’re taking when it comes to our policies related to clinical development. I want to focus on two efforts, in particular.
First, I want to address some of the steps we’re taking to modernize our approach to how we collect the clinical information that we use to make decisions about the safety and effectiveness of new drugs. And second, I’ll discuss other steps we’re taking to modernize how we evaluate that information. This includes improvements in the tools we use and parameters we adopt as a way to measure safety and benefit and properly evaluate new drugs.
I want to start with the first issue; the steps we’re taking to modernize the ways that clinical information is collected. FDA continues to advance the use of new tools and clinical trial designs. For example, we’re seeing wider use of adaptive approaches, which allow scientists to enrich trials for patient characteristics that correlate with benefits, or that help predict which patients are least likely to suffer a certain side effect.
This predictive information is valuable. It can be incorporated in a new drug’s label and help inform more careful prescribing.
As part of these approaches, we’re also seeing more use of combined-phase studies, what’s referred to as seamless trials. Instead of conducting the usual three phases of study, seamless trials encompass one adaptive study where the phases are separated by interim looks. By using one large, continuous trial, it saves time and reduces costs. It also reduces the number of patients that have to be enrolled in a trial.
These methods are increasingly prominent in oncology drug programs. Under the leadership of Rick Pazdur, our Oncology Center of Excellence is taking steps to better evaluate and cultivate these new approaches as one part of our ongoing efforts to modernize our approaches.
Owing in part to these leadership efforts, we’ve seen more sponsors develop oncology drugs that forgo the conventional three sequential phases of drug development. They opt instead for seamless approaches. Under these trial designs, they’ll typically add cohorts to a first-in-human trial to investigate doses and activity in a variety of cancers.
These methods have been used with some of the newer immunological therapies. It may also be used in other drugs targeted against specific molecular defects. Seamless designs are particularly advantageous for drugs that work in a variety of diseases, allowing rapid evaluation of the drug and potential approval under our accelerated approval pathway. These new approaches are also highly consistent with the goals of the 21st Century Cures Act and the recently passed FDA Reauthorization Act.
We’ve seen examples where this approach has allowed the rapid development of drugs in multiple different tumor types. If we had to stop and start formal Phase II trials in each different organ system where a cancer arose, it could have been a protracted process. This approach is well suited to the kinds of drugs that are being developed now, where drugs intervene on common elements found across multiple kinds of disease states. At FDA, we’ve identified more than 40 active commercial investigational new drug applications for large first-in-human oncology trials alone that use these seamless strategies.
Other concepts that we’re seeking to better adapt to clinical development, using new guidance and policy work, include common control studies and the wider use of large simple trials.
We’re also advancing the use of ‘Master Protocols’ to enable more coordinated ways to use the same trial structure to evaluate treatments in more than one subtype of a disease or type of patient. This approach is particularly relevant when it comes to targeted drugs. These are drugs that may intervene on markers that are relevant across many different disease subtypes. We may, for example, want to evaluate these different targets simultaneously, as part of one large study. This could give us a better way to understand the comparative benefits of a drug across different settings. To enable these master protocols, it’s often important to do molecular patient screening. This can lead to the development of a diagnostic that can also be used to guide patient care.
We’ll be talking more about all of these concepts and initiatives. We need to be mindful that these and similar new strategies also bring new uncertainties. These new uncertainties can create new risks. So, as in all of our regulatory efforts, we must also take additional steps to make sure that we’re protecting patients and ensuring drug safety.
For example, with adaptive and seamless clinical trial approaches, informed consent used in clinical trials must be updated as the trial progresses, not only to reflect new safety data, as is already standard practice, but also to incorporate data on the evolving view of efficacy.
First-in-human trials with expansion cohorts may also encompass an entire drug development program in a single trial. So potentially important regulatory interactions, such as the standard guidance meetings held at the end of Phase 2 and before the initiation of Phase 3 trials may not automatically occur. Comparable regulatory milestones need to be built into the new seamless clinical trial process. We need to ensure we provide comparable interactions and oversight.
The Agency also needs to engage in more communication between sponsors, investigators, IRBs, and other stakeholders involved in the development program. This is not a “business as usual” approach. It may require a much more iterative process, with greater communication between all of the stakeholders involved in the clinical trial processes. On the second point that I wanted to highlight today, we’re also taking new steps to modernize how sponsors can evaluate clinical information, and how FDA reviews this data as part of our regulatory process.
This starts with better use of more advanced computing tools, and more sophisticated statistical and computational methodologies, as part of the drug development and the drug review process. This includes more widespread use of modeling and simulation, and high performance computing clusters inside FDA.
FDA already has high performance computing clusters. These tools help us develop more sophisticated methods for evaluating the data that’s submitted to us from clinical trials. The computing tools also enable us to properly evaluate the more sophisticated components that are submitted to us as part of product review applications.
I’m directing an effort to try and increase our investment in these computing tools. They’re an increasingly important part of our work. But access to them at FDA is limited. We’re taking new steps to make sure our review staff has more access to these computing platforms.
These tools are especially important to our use of modeling and simulation as a part of drug review, not only in our Division of Pharmacometrics, in the Office of Clinical Pharmacology, but across our review program. Almost 100 percent of all new drug applications for new molecular entities have components of modeling and simulation.
Typically, these modules are focused on similar uses. They include modeling dose response as a way to better evaluate the safety and efficacy of different doses, and help select the optimal dose for the general population or subgroups. They also include methods to estimate a new drug’s effect size to develop the appropriate sample size for pivotal trials. Finally, modeling and simulation is also commonly used to evaluate the reliability of endpoints, such as helping to demonstrate a relationship between a biomarker and a clinical endpoint relationship.
We’re going to be making more advanced use of these and similar tools as one part of our overall efforts to make our review process more efficient and scientifically advanced. Among other things, we plan to convene a series of workshops, publish guidance documents, and develop policies and procedures for translating modeling approaches into regulatory review. We’ll also be conducting a pilot program on these approaches. As part of this pilot, FDA will grant meetings to participating sponsors who use these approaches, as a way to provide more collaboration on the model-informed drug development issues.
Our goal is to see how we can accelerate methods that improve our ability to use advanced tools to meet FDA’s gold standard for regulation.
These methods are being applied to both common and rare diseases. FDA is also collaborating with scientists to use similar computational tools to develop natural history models, based on placebo arms in Parkinson’s disease, Huntington’s disease, Alzheimer’s disease, and muscular dystrophy. If we’re able to make better use of rigorous, reliable natural history models, especially for rare diseases, it can help make clinical trial recruitment more efficient.
We’re also using similar methods to develop predictive algorithms, like those found in applications of artificial intelligence. To take one example, we’re developing classification algorithms in lung cancer, as a way to enable machine reading of CT scans based on widely accepted standards. These tools are still in development. But they can eventually be used to classify tumor dynamics like response to treatment, and increase the precision of our assessments. All of these advances help make the development process more efficient, and hopefully, less costly.
Once we design and publish these algorithms, depending on their stage of development, they can either be incorporated into clinical studies as an exploratory endpoint for validation or, if already validated, as means of helping to better inform the results of a primary endpoint.
Additionally, to better delineate how we’re going to approach the overall development and evaluation of drugs targeted to certain unmet medical needs, we plan to begin work on at least ten new disease-specific guidance documents over the next year. Some of these documents are already underway. Among the diseases we’re targeting are areas of significant unmet need like Amyotrophic Lateral Sclerosis (ALS).
This latter guidance is an outgrowth of a scientific document developed by the ALS Association that laid out some relevant principles. That document was developed with funds that the ALS Association generated from their Ice Bucket Challenge. FDA is grateful for their efforts, and the agency is responding with its own version of that guidance document.
These are just some of the steps we’re taking to make the drug development process more scientifically modern and efficient. We need to make sure that we’re giving our review staff access to the best scientific tools and opportunities to advance their work.
It’s also the case that drug development doesn’t stop with drug approval. We can and should learn a lot more about how treatments work in actual patients in real world settings, so that patients and payers are getting the highest possible value for their money.
In all these things, we need to make sure our own policies and approaches are keeping pace with the sophistication of the products that we’re being asked to review, and the methodologies being brought to these endeavors. Our work with high performance computing, and simulation, is a good example of an area where we need to make sure our methods match the sophistication and resources of the tools and approaches being adopted by sponsors. If we fall behind, it’ll be harder for innovators to bring forward novel products and approaches. It could also make it more challenging for FDA to adopt the most efficient and effective scientific methods, and continue to meet its high standards.
We need to do these things to make sure we’re providing an efficient path for the translation of cutting-edge science into practical treatments that are going to benefit patients. We need to do these things because the rising cost of drug development is unsustainable.
Unless we find ways to modernize how we approach our work, and make more efficient use of our resources, then we’re going to get fewer medicines, and higher costs. We’re not going to realize the benefits of the scientific advances we’re seeing as quickly, if we see them at all.
Most of all, we have to do these things to make sure that we’re efficient in bringing new scientific opportunities to patients who need them. And that we actively uphold FDA’s gold standard for safety and effectiveness.
