It's time to revive personalised medicine.

What you can do to make medicine more personalised.

Scott Wagers (sign in restaurant in Leuven, Belgium)

We abandoned the concept of personalised medicine in favour of precision medicine 15 years ago; that was a mistake.

The problem with precision medicine is that it still lumps people together; it’s just smaller groups.

What most do not realise is that truly personalised medicine is becoming increasingly feasible.

It will be the future, unless we fail to do the work that needs to be done to change the current system.

Why we walked away from personalised medicine.

It was a grey autumn morning 15 years ago when I sat leaning forward towards the desktop speaker. T

We were debating how to frame a grant funding proposal. Specifically, we were having a dialogue about the term "personalised medicine".

"I just can't imagine we would ever be able to have truly personalised medicine. There just wouldn't be enough money to fund the clinical trials you would have to do, and how could you make a medicine that takes years to formulate and manufacture personal?"

I could hear, or at least imagine, the nodding of everyone on the call.

This was not just our small group; this shift happened across many different fields. The thinking that personalised medicine was not possible became codified when the National Research Council asserted that personalised could be misinterpreted to mean developing treatments for each individual.

Instead, the idea was we should aim for precision medicine. Instead of treating, for example, all of asthma the same, you work to define different sub-groups of asthma. However, it would no longer be a misinterpretation to think we can develop treatments for each individual.

The future of medicine should be treating the individual, not disease.

Engineered cell therapy could make truly personalised medicine possible. Engineered cell therapy such as CAR-T involves removing our own cells, modifying them to better target a cancer, and putting them back into our body to do the work. There is also growing evidence that CAR-T or a similar approach can be used for other diseases as well.

CAR-T represents a shift. A shift away from discovering new medicines to engineering them. As such it represents the fuzzy border between science and engineering. With CAR-Ts it should be feasible to make modifications rapidly - days and maybe even hours. The biggest implication of this is that truly personalised medicine is no longer impossible.

Will the need for randomised controlled trials hold us back?

As the number of targeted drugs and the number of biomarkers grow, it becomes more difficult to run randomised controlled trials (RCTs). It becomes increasingly challenging to find patients with a particular biomarker.

When we talk about complete personalisation, we will need to have biomarker profiles unique to each individual. RCTs not only become difficult; they become impossible.

Are we willing to accept new forms of therapy without RCTs?

Without extensive toxicity testing?

The rising concerns about vaccines, which are some of the most tested interventions we have, should give us pause that moving away from RCTs for most therapies will be straightforward.

A placebo-controlled RCT is our best way of knowing in an objective manner if an intervention is working or not. Even so, RCTs are inherently flawed because they do not represent the real world. The real world, however, is messy, and it is difficult to know if the effect you see is due to the intervention or some other variable.

Let's say you have a cold and you have heard that a raw egg in a milkshake cures the common cold. You try it, go to bed and the next morning your cold symptoms are gone. You might logically conclude that an egg in a shake works. However, the common cold tends to resolve rather suddenly. You may have awoken with no symptoms, egg in a shake or not. A randomised controlled trial would reveal this false positive result when the majority of the study participants saw no effect with an egg in a shake. Individual experience is too compromised by confounding variables.

True personalisation would require us to have faith in our understanding of biology, real world data and the ability to predict the effect of a particular customised intervention without having the luxury to study it for ten years.

Maybe AI can save the day.

There is a lot of hype that AI can cure everything. This might be true, but what any AI needs is data upon which to train. It even goes deeper; it needs knowledge upon which to train. Considering that we are still uncovering new dimensions to measure and consider in biology, such as spatial relationships, single cell expression, as well as the impact of the environment, etc., we don't have the knowledge to train the AI. More troubling is that even if we had the knowledge, we would need to have a way to organise and share that knowledge or data. Right now, the practice of medicine and the practice of collecting complete data varies widely, introducing bias into datasets.

Furthermore, we will need to have the ability to study those molecular processes on an ever-increasing scale. Personalisation requires more molecular real-world data. It requires that such data is available and usable to build knowledge.

Progressing towards truly personalised medicine is a wicked problem.

There are too many features of the current system and too many interdependent stakeholders to easily progress towards truly personalised medicine. It is what is known as a wicked problem.

I believe we need to lean into multi-stakeholder problems while not expecting them to get dramatically better anytime soon. The only way they will get better is if a critical mass of stakeholders adopts a new paradigm.

To make this shift, there has to be an evolution.

First, we must build awareness about how the current system is not working. Then we to demonstrate the feasibility and the value of the change. The third step, which is often missed, is diminishing bottlenecks or barriers to change. The last step is developing the concept into a broad community that is ready to take over from the old system.

Big projects lead the way.

To move towards truly personalised medicine, there are several concepts that big projects need to develop:

1. Organising, structuring and molecular data available.

2. Developing and maintaining cohorts that include molecular profiling.

3. Advancing the engineering of cells and other effectors to manage diseases

4. Achieving a sufficient level of evidence for regulatory approval without a RCT.

5. Identifying biomarkers to guide treatment in the pre-disease phase

6. Developing generative AI to support medical research and clinical decision-making

7. Adaptive studies where treatments are tailored to the individual.

8. Building an understanding of stakeholder perspectives on personalised medicine

There are many large projects already working on these concepts. Here are a few examples:

P4O2: Building cohorts of lung disease (at risk, COPD, Long COVID, pulmonary fibrosis) that includes molecular profiling

T2EVOLVE: Developing preclinical models, stakeholder perspectives, and influencing the regulatory environment to drive the development of engineered cell therapy forward.

GOE: Organising the genome data for 500,000 people in Europe.

Please add your ideas for other concepts or projects working towards truly personalised medicine

What you can do.

No matter who you are, there is a role for you in reviving truly personalised medicine.

If you are a well-established researcher and you perceive a barrier to the realisation of truly personalised medicine:

1. Convene a consortium around the barrier/problem that you perceive. Do this even if there is no immediate funding opportunity.

2. Start a simple project together, a prototype project.

3. Seek funding for large projects

If you are a researcher with a technique or asset that you would like to develop further,

1. Develop a description of that technique and asset.

2. Contact relevant big projects and join as community members.

3. Contribute and build relationships to become involved in the next project.

If you are an early-career researcher involved in a big project

1. Step forward and volunteer to lead efforts in the project that are directly part of the project plan and require leaders.

2. Deliver more than expect to demonstrate your ability to lead

If you are a member of the public,

1. Seek out and get involved in projects - many will have educational videos and stakeholder groups you can join.

2. Offer to contribute your expertise regardless of what it is; problems are solved more quickly with a multidisciplinary effort.

What you achieve in big projects is what will be remembered. It is also what you will enjoy remembering the most.

I am creating a roadmap for guiding different stakeholders on what they can do to help drive forward the kinds of big projects that will make a real difference.

To get a copy when it is done, subscribe to this newsletter.