By definition, rare diseases affect only a small number of patients. However, recruiting cohorts of sufficient size is not the only challenge in rare disease clinical trials. The lack of medicinal knowledge and ethical issues add to the complexity. Implementing novel statistical tools such as complex innovative designs (CID) can help clinicians speed up trials and obtain meaningful results.
About 7,000 rare diseases are known today, affecting more than 400 million people globally. For most of these diseases, no treatment is available yet. At the same time, precision therapy approaches divide cancers into subtypes, creating further conditions with small patient populations.
Both in oncology and inherited rare diseases, the tremendous therapeutic potential of novel gene, cell, and RNA therapies can be a game changer. Consequently, the number of clinical trials with small patient cohorts is expected to increase, and better tools are needed to overcome the associated challenges in generating evidence.
Table of contents
Challenges in rare disease clinical trials: what’s holding us back?
Recruiting a sufficient number of trial participants is challenging not only because of the small patient pools available but also because eligibility criteria are often difficult to define. Many rare disease patients are diagnosed only after years of wrong, unsuccessful therapies, creating heterogeneity in disease and treatment history within the patient population.
What is worse, rare diseases are often severe, reducing the willingness to enroll in a clinical study if there is a high chance of receiving only a placebo – an ethical dilemma only heightened by the fact that about 50% of rare disease patients are children.
Another issue is the lack of knowledge on disease symptoms and progression, resulting from the small number of reported cases. Missing information makes it hard to define meaningful endpoints and suggest appropriate therapeutic doses for the clinical trial. Overall, rare disease clinical trials require more administrative effort, time, and financial investment.
With small patient cohorts, result interpretation is also problematic because traditional statistical methods rely on a large number of observations. Novel, non-standard statistical methods may need to be applied to obtain meaningful results from rare disease clinical trial data.
Complex innovative designs: the key to overcoming these challenges?
Complex innovative designs (CID), a collective term for adaptive and Bayesian statistical tools and other features such as master protocols, have recently gained traction because of their potential to overcome the challenges in rare disease clinical trials. Featuring a combination of strategies, CID tools enable modifying study design elements, such as sample size and dosage, based on emerging data throughout the clinical trial.
“Sample size, the number of patients enrolled, is one of the major bottlenecks in rare disease clinical trials because patient identification and recruitment is so difficult. The smaller the number of patients needed, the better. Adaptive trials allow starting with a relatively small sample size that can be enlarged if needed, e.g., if the observed effect is smaller than anticipated, yet remains clinically meaningful,” explained Yannis Jemiai, Chief Scientific Officer at Cytel, a US company specializing in complex innovative designs.
Combining clinical phases with a gradually evolving population size in a so-called seamless design can answer multiple questions about the effect of a drug while using fewer resources. Seamless design is, therefore, another valuable tool that can speed up the trial considerably, as demonstrated in the evaluation of the Pfizer/BioNTech COVID-19 vaccine.
Complex innovative designs also support dose finding. Modeling and simulating pharmacokinetics and pharmacodynamics facilitates identifying the right doses to be tested. Small sample sizes are efficiently used by assigning just the number of patients needed to determine whether a dose is safe and effective. As data accumulate, doses can be fine-tuned, increasing the chances of a successful trial and ensuring that patients receive optimal treatment.
Focus on Bayesian adaptive clinical trial design
Sparse and limited datasets are a characteristic of rare disease clinical trials, and exact statistical methodologies should be designed to derive meaningful information even in these complex cases. Bayesian clinical trial design, in particular, facilitates the “borrowing” – or use – of external information, such as data from previous studies or real-world data. For example, data collected for an adult patient cohort can be used as prior information when starting a pediatric trial.
“There is a lot of unknown, especially in a rare disease setting. You are working with small numbers of patients or even subpopulations with different biomarkers. So, you need to get as much information as possible and make the right decisions to identify the right dose and possibly adapt clinical endpoints quickly. Using external data can greatly facilitate this task,” stressed Jemiai.
Borrowing from historical or external data also allows for the development of external control arms that reduce the number of patients in the control arm or even bypass it altogether. However, regulatory and statistical strategies are required to analyze the data landscape and identify the right sources to build the external control arm before designing it into the protocol.
In addition to enlarging the dataset obtained from the treatment arm, external control arms may also increase the willingness among patients to enroll in the trial. In particular, the ethical dilemma associated with pediatric clinical trials can be solved if all affected children can receive the therapeutic drug instead of a placebo.
Cytel’s role in rare disease clinical trials: your partner for success
Implementing complex innovative designs in clinical trials is not as straightforward as it should be, particularly for clinicians unfamiliar with statistical methods. Similarly, regulatory authorities are still hesitant and critically view the risk of bias, especially in trials with a reduced or no control arm. Hence, CID tools today are more frequently used in early development steps, which are not scrutinized as much as the later ones.
“While adjusting to the adaptive and Bayesian way of thinking may appear burdensome, the potential advantage of a principled way to reduce sample size, increase power, reduce costs, and reduce ethical dilemmas can outweigh the initial learning curve,” said Jemiai.
Cytel has decades of expertise in complex innovative designs – the company was involved in some of the first adaptive clinical trial designs ever conducted. Already in 2007, Cytel applied adaptive dose-finding to treat chronic diarrhea in patients with HIV/AIDS under the FDA Special Protocol Assessment (SPA) process. Today, Cytel has a broad offering that includes standalone software, comprehensive service packages, and consulting individually tailored to the client’s needs.
Case study: a rare disease trial success story
A recent Cytel client engagement case study showcased the diversity of applications for adaptive trial designs. A small biotech company in the rare disease space engaged Cytel to consult on a complex study design to address multiple concerns specific to the patient population and therapy.
Although the treatment showed promising results in earlier phases of the study, much uncertainty remained about the true underlying treatment effect observed in a comparative phase III setting. Four key adaptive designs and strategies were used, considering many variables. As a result, Cytel offered a variety of trial design candidates from which the clients could choose their preferred option.
“Our team at Cytel supports our clients every step of the way – from adaptive and Bayesian clinical trial design, through modifications during the trial and data analysis to discussions with regulatory authorities,” concluded Jemiai.
Learn more about adaptive and Bayesian design tools in rare disease clinical trial.