How do you approach clinical trial design for cell and gene therapies?
There are a lot of similarities, but also some notable differences that I have encountered the last 15 years working in multiple gene and cell therapies, compared to small molecules or other biologics like antibodies.
One main difference is that cell and gene therapies delivery is usually more invasive and inherently they have a higher potential risk because of long-term effects (for both safety and efficacy). Therefore, we usually have a combined First-in-Human phase 1/2 trial with patients, while other therapeutics usually have a phase 1 trial in healthy volunteers and then move to patients for phase 2. It can also be more challenging to implement a placebo arm that is standard in other therapeutics because cell and gene therapies frequently have a more invasive delivery route.
In addition, as cell and gene therapies are typically designed to provide extremely long to lifelong effect, clinical trial participants need to be followed for a minimum of 5 and in some cases up to 15 years, for durability of effect and safety, instead of a few weeks or months in other therapeutics.
Are there other challenges to be aware of?
Cell and gene therapies are frequently used for rare diseases that, by definition, pose challenges to recruitment. Also, rare diseases frequently do not have any previously approved drugs, validated endpoints or robust natural history studies. Because we often have a lot of gaps in rare diseases, we consider starting a natural history study before or in parallel with the phase 1/2 trial. This can provide confidence for both efficacy and safety as an “external comparator,” especially if there is no placebo.
Another difference is the level of evidence needed to go to registration for phase 3 trials. For rare diseases, health authorities appreciate that it is not feasible to recruit a large number of patients for phase 3 trials, compared to a common disease with typically two large phase 3 trials (and sometimes a good size phase 2b dose selection trial). That’s why we need a lot of confidence in the phase 1/2 results – to enable us to run a single relatively small phase 3 trial while generating enough evidence for registration and reimbursement. Pretty much the whole clinical development program can be based on only two clinical trials – a phase 1/2 and a relatively small phase 3 – so they need to be well designed and executed. And the program may need to be supported by a natural history study, in order to be adequate for regulatory approval and reimbursement, potentially worldwide.
"It is useful as a general strategy to go to the FDA and EMA as early as possible and identify disagreements, in order to find a solution that will be acceptable to both as a way forward."
How can you derisk development for such trials?
We need to thoroughly evaluate the patient population and endpoints as early as possible, even at the stage of indication selection. This is one of the key criteria we use when building the clinical pipeline. This is because often these diseases tend to progress slowly, which makes it even more difficult to show a therapeutic effect at a reasonable timeframe for drug development. It also depends on what is the expected therapeutic effect. For example, in ophthalmic slowly progressive neurodegenerative diseases, improvement of vision can be a relatively faster and “easier” endpoint to show, compared to slowing progression or prevention of visual impairment that is more challenging to prove with confidence in a reasonable timeframe.
We may also need to design a natural history study, not only for “external comparator” information, but also to identify the most sensitive endpoints (which may need to be designed de novo) before we integrate them into the phase 1/2 trial and validate them before the phase 3 trial. For example, we can check in the natural history study the signal-to-noise ratio, the annual disease progression and retest variability to figure out which of the potential endpoints has the highest sensitivity, in order to justify which endpoints to use in the phase 1/2 trial. We implemented this strategy successfully with a newly designed endpoint of delayed dark adaptation for slowly progressing retinal degeneration. We fine-tuned it in a natural history study and allowed the phase 1/2 trial to show efficacy in the first 3 patients, highlighting the potential to accelerate the clinical program.
Another way to derisk the program is to have discussions with regulators as early as possible, so that before we start the phase 1/2 trial, we have an agreement about the endpoints and the kind of results, such as magnitude of treatment effect, that would be eventually acceptable for registration. This is where natural history results become extremely important, because regulators can evaluate data from actual patients and hopefully agree with the proposed endpoints and magnitude of effect.
How are interactions with the FDA different than with other regulatory agencies like the EMA?
The main message is that they don’t always agree with each other. We tend to focus much more on the FDA responses because FDA controls such a huge market and is one of the most advanced and experienced agencies. I have also had very good experiences with EMA, and they are similarly extremely well-versed, knowledgeable and helpful. I have seen that other regulatory agencies – like UK, Canada, Australia etc – generally agree with either FDA or EMA, besides Japan or China, which sometimes have a rationale for differences or additional requests in specific topics.
It is challenging to make sure that the various regulatory agencies agree with each other, but it is useful as a general strategy to go to the FDA and EMA as early as possible and identify disagreements, in order to find a solution that will be acceptable to both as a way forward. If we don’t know that in advance and we find out that there’s a disagreement in phase 3 or even at the end of phase 2, it may be too late to change the trial design without significant delays and budget increases.
"CRO selection is one of the most critical things for the success of the whole program. You need to start selection several months or a year before you move into the clinic."
How do you approach interactions with payers?
I have had a lot of interesting discussions during the last 10 years with payer advisory boards and payers for gene therapies for ophthalmology, which can be pretty expensive. The key is to be able to demonstrate with data the value over time, because it is the long-term effect that matters here. It’s difficult to balance running clinical trials as quickly as possible to get to registration and commercialization, and at the same time also show long-term preservation of efficacy. By design, we try to treat as many patients as realistically possible in the phase 1/2 trial, so that by the time we register after a phase 3 trial, there are patients from the earlier phase 1/2 trial who can demonstrate durability of treatment effect. Without durability of effect, payers are not going to see the long-term value and pay the premium price of cell and gene therapies.
What is your strategy when it comes to building out the clinical team?
I have found out that it is very important, especially for biotech smaller companies, to hire people who have prior work experience in this field, but also to be “nice” collaborative flexible people. Personality matters, sometimes even more than specific work experience. I try to build teams of people with different backgrounds and different personalities. I don’t want for example all introverts or all extroverts. We need a variety of backgrounds and a variety of approaches, so that we can all bring different viewpoints and collaborate to bring therapies to patients.
In the beginning, especially in small biotechs, I usually try to cover the workload with consultants, because there is not enough work for several full time people, but we still need expertise in all different functions. It isn’t cost efficient for the company, and it isn’t fair to the people to hire them full time when there isn’t enough work to be done. When we get closer to the clinical trial, we start to need some full-time people, and it becomes a matter of how much work there is, what is the indication, the pipeline etc. For companies with only one compound, they will obviously need fewer FTEs than companies with a series of compounds at different stages of development.
Also, based on my experience, it’s usually not productive, and it may even become a disaster, to have only the external CRO without supervision running a trial and just expect the CRO to deliver the results at the end. You need to have a few internal people to manage the CRO, for example clinical development, clinical operations, medical affairs and similar critical line functions.
"Preparation, intense monitoring, direct communication, fast and accurate evaluation and decision making with quick implementation are keys to dealing with the unexpected."
How do you evaluate potential outsourcing partners?
Compared to when I started in industry gene therapy clinical trials, more than ten years ago, nowadays there are many more CROs with cell and gene therapy trial experience. CRO selection is one of the most critical things for the success of the whole program. You need to start selection several months or a year before you move into the clinic. What I find very useful is to make sure that as a sponsor I talk with the actual team that is going to support the project, and not just with the best team that they send during the RFP process, which is frequently then switched to a much less experienced team after contract signing.
The second thing is to try to get realistic cost estimates, because budget is very critical for every company, smaller or larger. Figuring out the details around what exactly are the services that the CRO offers (which sometimes are not obvious or leave room for different interpretations) is critical in order to compare CROs. In order to provide a compelling and competitive cost estimate, the CROs frequently may omit key tasks needed for the specific trial that a CMO can identify, that are easy to be missed or misunderstood at the time of the original proposal and may prove very costly or even impossible to add on later during an ongoing trial. Also, a cost estimate of potential adjustments that may be needed in the middle of the trial can help a better understanding of budget options, show the work experience and philosophy of each CRO and help to make a well-informed decision.
How can you prepare for something to go wrong?
Things almost never go exactly as expected and planned. It’s a fact of life, and in our industry and probably any industry, you need a plan B and maybe a plan C. It’s also critical to be able to implement those plans very quickly when needed.
On a previous program, for example, we had a relatively unexpected adverse event, still one that could happen, so we had a plan B. We were able to successfully address it quickly, because of close monitoring and direct communication. I was personally communicating with the key investigators and asking them to call my mobile phone immediately, while the patient was still in their office, if they were concerned. When that specific adverse event occurred, the investigator called me on the spot, which enabled me to identify the issue very quickly. Based on that, I did a review of all available relevant data practically overnight, discussed it with the team and we adjusted very quickly several key points to prevent it and to support investigators and patients. That pretty much saved the validity and robustness of the trial results. It allowed us to get positive results, even ahead of the original timeline, and finance the company for the next stage. If we hadn’t been monitoring so closely or hadn’t been able to implement changes quickly, we would have lost the rest of the patients from accurate evaluation due to adverse events, had slow enrolment from concerned investigators and could not get conclusive results on time. Preparation, intense monitoring, direct communication, fast and accurate evaluation and decision making with quick implementation are keys to dealing with the unexpected.
Anything else?
The best advice I can think of is to do what we have learned to do and have practiced as doctors. We are all doctors, and this is what we need to bring to our companies and to everybody, because nobody else has this unique perspective. This is especially what the investigators also appreciate.
Some of us have spent decades of our lives treating patients before coming to industry, and not only do we understand the patients and the physicians clearly, but we also have an instinct for what is going on. If you feel that something is wrong or does not make sense, or that it may be an incredible breakthrough or therapeutic benefit, follow that intuition and check it out. It may make all the difference to eventually bringing new treatments to patients who need them. Our patients are waiting!
