Research in progress: fighting diseases by more easily detecting them

A doctor’s medical expertise is not always enough to detect diseases. That is why it is important to combine artificial intelligence with traditional image processing techniques to facilitate detection and take medical imaging to new levels. This is an endeavor that is logically at the very heart of the EPITA Research and Development Laboratory (LRDE), thanks to its specialized researchers.

Behind the scenes of the new episode of the “Recherche en cours, EPITA Laboratoire d’innovation” (Research in progress, EPITA Laboratory of Innovation) series, dedicated to the Neuromed project managed by Élodie Puybareau, teacher-researcher, member of the Image Processing team at the LRDE and co-director of the Image Major, the school invites you to learn more about this exciting field designed to facilitate the daily lives of doctors and hence patients.

What is the Neuromed project?

Élodie Puybareau: Today, thanks to technology, doctors have access to high-performance machines that allow them to see everything that is going on in the patient’s body. However, certain changes in the body at the beginning of a disease are sometimes minute and therefore difficult to detect. The goal behind our research is to help doctors more easily see these changes in order to better anticipate how the disease will evolve and hence, provide better treatments. Nowadays, we often hear about neural networks as a solution to various problems, except when detecting these small lesions – in other words, abnormal changes in the human body – these neural networks are not yet sufficiently effective. Our aim is to increase this performance through new software provided to doctors in order to simplify their work, both in terms of diagnosis and patient follow-up.

Élodie Puybareau is also co-director of the Image Major

What do you like about research?

Élodie Puybareau: The great thing about research is that it is multifaceted: depending on your field, your objectives or your methods, your work will never be the same as that of another researcher. In medical imaging, for example, research consists above all of finding new innovations that will allow doctors to facilitate their work, patient follow-up and comfort, and allow for better and earlier treatment of diseases. Without the research already conducted in the past in the field of medical imaging, we would surely have encountered numerous roadblocks in our understanding of many diseases and not have been able to develop certain treatments! Of course, some companies can develop their own solutions. However, our Image Processing team strives to help doctors free of charge by providing them with all the tools we can imagine.

Do you feel any particular pressure when working on health-related issues?

Élodie Puybareau: In the broadest sense, research must use an ethical and responsible approach, and this is even truer when it concerns the healthcare sector. We cannot provide doctors with medical diagnosis software if there is the risk that a mistake has been made in its design! When we publish a scientific article, we must be as precise, honest and transparent as possible because our goal is not to replace doctors but to assist them. They must be familiar with the results and the tool’s performance in order to use them in the best possible manner.

Does working on these types of tools require joint efforts?

Élodie Puybareau: Collaborating with other experts is essential. When I arrived at EPITA in 2017, I teamed up with Isabelle Bloch (Télécom Paris / Sorbonne University). Together, we started working on a thesis topic regarding this project, applied to imaging in pediatric radiology, for the treatment and diagnosis of diseases affecting children. As a researcher, it is vital to share what you find and exchange information with the research community. Even if you can work alone, I believe that research is always better done as a team! Putting our heads together ensures multiple ideas, especially when we do not all have the same areas of expertise. This joint effort is what makes research extremely rich.

Is it also important to work with students?

Élodie Puybareau: You can’t work as a teacher-researcher if you don’t like working with students! Our job is to train them, of course, but also to share our passion for our work. They are researchers in the making and I think it’s great to be able to help them find their path. This is fully in line with my ideas about the research profession, which, in my opinion, is to be useful.  If what I do can help even one doctor more effectively treat a patient, then that is enough for me. It sounds a bit “Miss France” when you say it like that, but it’s true!

Finally, in the case of a thesis, the role of researcher takes on another dimension. As a thesis supervisor, we have to provide students with ideas so that they can implement them and begin their research project. Then, as time goes by, the relationship changes: the student masters the subject, sometimes much better than the supervisor, and becomes the driving force in giving ideas. The supervisor then makes sure that the direction taken is correct and that the scientific articles are well written.