Exploring uncharted territories

Esther Thole | maandag 2 september 2019

The search for new or better therapies is opening whole new avenues in research, but also in regulatory assessment. EMA director Guido Rasi explores how the agency should prepare itself for the future.

No one knows what the future will bring. At the same time, there is no doubt that we will see an increasing supply of highly innovative medical products that no longer fit our conventional definition of drugs. These cutting-edge products are not only pushing the boundaries of our scientific, technological and clinical knowledge, but also of the ethical, legislative and regulatory frameworks that we build on to assess safety and efficacy.

The European Medicines Agency (EMA) has drafted a strategic reflection, 'Regulatory Science to 2025', to explore the upcoming changes in the drug development landscape and how these translate to EMA's regulatory responsibilities, operational approach and required expertise. In his keynote lecture at the Figon Dutch Medicines Days, EMA executive director Guido Rasi will elaborate on his vision for the future. Medicines got the chance to talk to him beforehand and get a glimpse of all the unknowns that are awaiting us.

What are your sources to spot future trends?

‘One of our initiatives to keep up with science and technology is our Regulatory Science Observatory, which is collecting information from many different sources, such as developers, investor meetings, academia, industry meetings and the various communities in our network. And we have our Innovation Taskforce, which acts as an access point for all developers, whether unexperienced or very senior, who want to discuss an innovative product or idea to see whether there is a possibility for approval within the current legislative and regulatory framework. For us, this is an important source of the ideas that are circulating. We come across highly interesting ideas that are still really forward-looking but have the potential to become very relevant in the future.’

How do these ideas affect the way EMA views its own tasks and responsibilities?

‘An important challenge for us is to ensure that we have all the necessary expertise in place, as such ideas are increasingly becoming reality. The expertise we will need in the future goes beyond the clinical pharmacology we are used to, it includes areas like nanotechnology, omics-applications, gene and cell therapy, modelling, big data analysis and so on.’

Innovative and personalised therapies like gene or cell therapy are conceptually very different from classical medicines. What does that imply for the regulatory process?

‘For one thing, it will mean that we are moving towards assessment of the procedure rather than the actual 'drug'. If you think about cell therapy where a patient's own cells are extracted, cultured and modified and then re-administered, you cannot put that unique treatment only through the classical assessment track. In addition to the formal benefit-risk assessment carried out for every medicine at the time of the authorisation, you have to focus on the procedure, on how the medicine will be delivered, and that is a huge change to what we have been doing so far.’

How can you assess safety in this kind of one-off therapies?

‘We need the same level of robustness in regulatory approval as we do with conventional medicines and these innovative medicines require a more proactive approach in post-market surveillance and pharmacovigilance. We need more planned, long-term monitoring of patients who have received such therapies. And in my view, we should also foster the assessment of performance, next to the traditional risk-benefit ratio. We cannot limit ourselves to safety alone.’

Why is performance important and how does it relate to long-term monitoring?

‘Take for example gene therapy, which relies on one intervention that changes the course of the disease for a lifetime. It is irreversible. That is a major change compared to the therapies we are used to. So, should we wait a lifetime before deciding on approval? Or should we just release it, cross our fingers and hope for the best? Of course not. We should prepare for and envisage lifetime monitoring, which means to focus on performance. This encompasses many aspects, from the patient's quality of life to the emergence of unforeseen diseases. Were these already present in the patient's genome or did they arise because of the intervention? Is the patient taking other medication that might lead to completely different responses in the short, middle or long term? There are so many questions, so many unknowns, but at the same time the severity of the disease and convincing efficacy in the short term might justify approval. But to learn more about all these unknowns, our only option is lifetime observation.’

Another complicating factor seems to be that you cannot simply stop the treatment anymore.

‘Yes, that also completely changes the framework on which you build your assessment. You cannot stop the treatment or undo its effects, you cannot compare patients with treatment to those without, because the treatment is so personalised. We therefore need new ways of monitoring the effects. It cannot just be based on serendipity, as is still largely the case. I mean for example that you take a pill and three hours later you get a headache; is that a signal of causality or is it coincidence? We need to get more proactive in monitoring and hopefully technology will help us out. Imagine an app that detects and reports certain signals that are registered by sensors or devices placed in the body.’

How can you include these other technologies in your assessment?

‘The increasing overlap between diagnostics, genetic profiling, apps, datasets, medical devices and medication creates yet another level of complexity. These converging technologies cannot all be assessed separately. We must come up with a new way of evaluating such integrated combinations. We as well as the developers are exploring uncharted territories.’

Should EMA's mandate be expanded to cover that?

‘Where these processes will happen is less relevant than having a clear platform to integrate them. One step could be to bring medicines and devices under the same roof, meaning that they should fall under the same EU Directorate. Then we can start to explore the room for integration, alignment of procedures and compatibility of data. The conversation has started on this, the awareness is there, so that is encouraging. We hope that under the new European Commission this integration will be fostered. There is no need to change everything, but at least we should avoid contradictory regulations and work towards compatible processes. That would be very helpful already.’

To what extent is the level of uncertainty truly different from that in the past? We are all prone to thinking that our future challenges are bigger than ever before, but 10 or 50 years ago, people were probably thinking the same.

‘When it comes to individual assessments, I think the level of uncertainty is about the same as it was in past. Years ago, we faced big uncertainties as well, but based on studies in 30,000 people we were comfortable enough to approve a painkiller that turned out to cause cardiovascular side effect some 20 years later. Now, with a new approach on pharmacovigilance, planning the needed evidence to be generated in the real world, we can discover adverse effects much faster. With the innovative therapies that are emerging now, such as targeted therapies in oncology, the picture is different. We are dealing with much smaller populations and thus have less data, but the process is probably more robust because the risk-benefit ratio is less controversial. These therapies are addressing very severe diseases and even though we don't know the long-term outcomes, we know exactly what will happen when we do nothing. In short, I think the overall uncertainty remains more or less the same.’

But still, smaller populations mean that you lose statistical power.

‘What we may lose in statistical power is countered by the level of detailed knowledge we have on the disease, the patients and the drugs. What we really need to do is to massively invest in profiling patients, because the more you can enrich the population, the higher the chances of success. When you start any clinical trial, you immediately have four populations: responders with no side effect, responders with side effects, non-responders with side effects and non-responders without side effects. With genetic profiling, you potentially have the tool to qualify the differences on a genetic basis, meaning that you can enrich the population to the first two groups as much as you can.’

Is profiling the way to go?

‘Well, it's one way to go. I strongly believe that if we massively invest in understanding the differences within a population, then we can target and use our resources most efficiently.’

Which other options do we need to pursue?

‘More planned monitoring and planned evidence generation post-marketing, rather than serendipity-based data coming in. We have started with that and with the new innovations we are expecting and the smaller patient populations, it will become more important. In essence, we have become much more aware of our own ignorance of all the unknowns. Better monitoring may help us to clear these unknowns much faster. When we are more aware, we become more committed to filling the gaps.’

Partners Medicines