Ninon Burgos wins 2019 ERCIM Cor Baayen Young Researcher Award for her work on computational imaging of neurodegenerative diseases

Date :
Changed on 25/03/2020
Currently working in the ARAMIS Research team of the Brain and Spine Institute (Institut du cerveau et de la moelle épinière – ICM), a multidisciplinary laboratory affiliated to several research institutions, including Inria Paris, CNRS, Inserm and Sorbonne Université, Ninon Burgos’ research involves developing computational imaging tools to improve our understanding and diagnosis of dementia. Inria Paris We interviewed Ninon on the occasion of her winning the prestigious ERCIM Cor Baayen Young Researcher Award, which she received at the end of October in Rome during the European Computer Science Summit (ECSS).

The Cor Baayen Young Researcher Award recognizes the outstanding scientific quality of your research and the impact it has already had on science and society. For what aspects of your research in particular were you awarded this prize?

What is the ERCIM Cor Baayen Young Researcher Award?

The Cor Baayen Young Researcher Award is awarded each year to a promising young researcher in computer science or applied mathematics. Worth € 5000, it is named for the first president of ERCIM (the European Research Consortium of Informatics and Mathematics), Cor Baayen. This year it was jointly awarded to two outstanding young researchers Ninon Burgos and András Gilyén of Caltech.

I have been developing image synthesis techniques to enable the analysis of medical images for use in diagnosis and treatment planning since the beginning of my PhD. I developed a method to generate computed tomographic (CT) images (or “slices” of the body that can then be digitally “stacked” together to form a 3D image of the patient) from magnetic resonance imaging (MRI), which provides high-resolution anatomical information. The method was initially developed to enable the quantitative analysis of positron emission tomography (PET) images acquired on a hybrid PET/MRI scanner. PET is a nuclear medicine imaging technique that uses small amounts of radioactive compounds called radiotracers to identify changes in organ and tissue function at the cellular level to detect disease onset early on. 

I then extended this approach so that healthcare professionals could plan radiotherapy treatments for patients based only on MRI alone.

At the moment, I am focusing on analysing images to improve our understanding of dementia and how we diagnose this neurodegenerative disease. To this end, I have developed an image synthesis technique that generates healthy-looking images specific to a patient. When we compare these images to a real image of the patient, we can use the pseudo-healthy model to detect the areas of the image that display abnormalities. These “abnormality maps” could help clinicians better diagnose disease by highlighting pathological areas in a data-driven fashion and so improve how they interpret subsequent analyses. The award rewards all this work.

Ninon Burgos

Ninon Burgos: five key dates

2012: Graduates with a MSc from Imperial College London

Sep 2012: Begins PhD at the Centre for Medical Image Computing, University College London

Feb 2016: Starts post-doc at the Centre for Medical Image Computing, University College London

Jan 2017: Joins the ARAMIS Research Team

Oct 2019: Receives the ERCIM Cor Baayen Young Researcher Award 

Some of your techniques have already met with success in the medical world. Could you describe the most important ones you have developed?

The approach I developed during my PhD was one of the first image synthesis methods applied to the medical domain. With the advent of deep learning, this field is advancing every day but the growing number of papers being published on the topic does not necessarily imply that these techniques are increasingly being translated to the clinic.

I am proud of the fact that my image synthesis method to generate CT images from MRI has been transferred to clinical research. Indeed, it is currently being employed at the Dementia Research Centre at UCL’s Institute of Neurology in a four-year project called Insight-46 – a neuroscience sub-study at the MRC National Survey for Health and Development (a longitudinal survey of people born in Britain in March 1946 still going strong today)that involves 1000 PET/MR image acquisitions. What is more, Oncovision, a company that develops, manufactures and distributes medical imaging devices, is interested by the method and has signed a commercial agreement.

Why did you choose to pursue this area of research?

I studied electronic and computer science at ENSAE (École nationale supérieure d’électronique et de ses applications), but have always been interested in medicine thanks to my parents who are both nurses. After discussing my options, two of my professors at the time encouraged me to apply for the MSc in Biomedical Engineering at Imperial College London. Going on to do a PhD was not part of my original plan, but I wanted to learn more about medical imaging (as well as stay in London). During this time, I discovered that methodological developments are driven by clinical applications – and this led me to pursue my career in this area. 

Could you tell us more about the laboratory in which you work?

The ARAMIS Lab at the ICM, which I joined in January 2017 after being awarded a PRESTIGE postdoctoral research fellowship (a Marie Skłodowska-Curie fellowship programme), is led by Olivier Colliot and Stanley Durrleman. We are a multidisciplinary methodological group with strong expertise in brain image computing, statistical learning and neurodegenerative disorders. The fact that the ICM is hosted within the Pitié-Salpêtrière Hospital, which is the biggest hospital for adults in Europe and a leading neurology medical centre, helps in this respect. ARAMIS is affiliated with several research institutions, including Inria, the CNRS, Inserm and Sorbonne Université. The team boasts roughly 40 members - PhD students, postdocs and engineers, four permanent researchers and two clinicians, a neurologist and a neuroradiologist. It is a very stimulating and pleasant environment in which to work.

What are your plans for the future?

I intend to continue working on the analysis of medical images to improve differential diagnoses for applications in personalized medicine.

To this end, I will be developing advanced computational representations of multimodal imaging data and building flexible decision support systems that are more sensitive and produce results that are easier to interpret. Thanks to my springboard chair at PRAIRIE (the PaRis Artificial Intelligence Research InstitutE), I will (soon) be able to offer a PhD fellowship to pursue this line of research.