Lluis Guasch is a lecturer at Imperial College London, where he researches novel ultrasound imaging algorithms for clinical applications and also teaches deep learning to MSc students. He began his research career working on geophysics, developing imaging algorithms to study the sub-surface of the Earth. A few years ago, he started working on the translation of seismic imaging algorithms to medical ultrasound applications, in particular for breast and brain imaging.
“QUSTom is a beautiful initiative that brings together outstanding scientists across Europe. For me, it is an opportunity to closely collaborate with people I have known and admired for years in a project that combines our skills to produce scientific outputs we could not have achieved individually”.
Why did you choose this profession and what motivated you to do it?
When I finished my degree in Physics, I knew that I wanted to do a PhD. I was very attracted to two seemingly unrelated fields, medical imaging and seismology, and I ended up pursuing a PhD in seismology. What attracted me to both topics was the magic of converting observable data from the surface of an object into a physical representation of its volume. This discipline is called imaging, and it combines my two favourite disciplines in maths and physics: inverse problems and wave propagation. The advances in computational power over the last decade made me realise that the large-scale algorithmically-complex solutions I was using to study the Earth’s interior were becoming feasible in smaller-scale ultrasound problems. This realisation allowed me to reconnect with my scientific interest in medical imaging, where I currently devote most of my research time. I have discovered that the two disciplines are in fact much more closely related than I thought, and that the mathematical principles used in both are almost identical, thanks to the proportional scaling of the sizes and wave frequencies used in both problems.
What will be your contribution to the QUSTom project?
My role in the QUSTom project is focused on demonstrating the first in-vivo application of the new imaging modality we are developing: uncertainty quantification of reconstructed images. In layman words, what we do is to provide additional information to clinicians so that they not only have an image of the breast, but also a map of how reliable this image is, which informs them on what features are real anomalies that can be associated to tumours and what features are likely to be artifacts of the reconstruction. Such a tool may prove invaluable because it will significantly increase accuracy in diagnosis in two ways. First, by reducing false negatives and improve survival rates of patients suffering breast cancer. And second, by reducing false positives, which will result in significant economic savings to healthcare systems, as well as avoid subjecting women to unnecessary stress during screening.
What do you hope QUSTom can achieve beyond the life of the project?
I strongly believe that the outcomes of QUSTom have the potential to radically transform how breast imaging is used to prevent and treat breast cancer.
Our vision and strategy are based on shifting the cost of screening and diagnosis from the acquisition devices to the image reconstruction algorithms. Current methods to image the breast require expensive (MRI equipment) or potential harmful devices (mammography exposes patients to harmful ionising radiation) which limit our capability to screen population at the rate breast cancer prevention and treatment demands. Ultrasound does not suffer from any of these limitations but requires more computationally-demanding algorithms to generate high-resolution images. The continued evolution of computers is, and will continue, decreasing the price of generating these images to offer a cost-effective high-quality solution that we expect can replace existing modalities. In addition, our method is much better suited to image women with high-density breasts, as ultrasound images are not obscured by dense tissues as is the case in mammograms.
The specific outcomes of the project will also have an impact in a shorter timescale. The technology outcomes of the project will generate new tools to complement existing breast imaging modalities. Besides the uncertainty maps I already mentioned, we will also produce multi-parameter quantitative images of the breast. This is an important breakthrough because these images provide three different tissue properties that can be directly used as cancer biomarkers.
What does QUSTom mean to you?
QUSTom is a beautiful initiative that brings together outstanding scientists across Europe. For me, it is an opportunity to closely collaborate with people I have known and admired for years in a project that combines our skills to produce scientific outputs we could not have achieved individually.
How do you like it so far?
If I had to describe it in one word, I would say buzzing! I am thoroughly enjoying the interaction with all the partners. I particularly like the end-to-end nature of the project, the scope covers everything required for this technology to be deployed in hospitals: the design and construction of the device done by the team in KIT to acquire data, the algorithmic development done by BSC, Imperial and ARCTUR, the image analysis from the radiologists at the Vall d’Hebron, all the way downstream to the commercialisation efforts led by Frontwave Imaging. It is a rare opportunity to be part of a project that has such visionary scope, and that has the potential to improve the life of millions of women by providing a better breast imaging solution.
Do you have any advice for young researchers who would like to follow in your footsteps?
I try not to take myself too seriously. I find that having a bit of fun and enjoying what I do helps me not only stay more focused on my work but also frees my mind to be more scientifically creative. I also try to always reserve time to continue developing my skills as well as learning new ones. In my case this is a bit hopeless because I am trying to keep up with the latest advances in deep learning, the fastest growing field in science currently. But nonetheless, being constantly exposed to new ideas and concepts completely unrelated to my day-to-day research helps me discover solutions where I least expect them. As a good friend of mine use to say: we don’t know what we are doing, but we are doing it!
On a more serious note, I find that what works best for me is to believe in what I am doing. I don’t mean the kind of naïve faith that relies on the idea that just wanting something and working at it hard enough will take you where you want to go, but rather to work on projects where your scientific intuition is telling you that this new crazy idea you had may actually work. For this is absolutely essential that your work colleagues provide you with the right environment for your creative mind to flourish; in the early stages of your career that means to choose a supervisor that is supportive and does not micromanage your daily activities, and as you progress it means that you need to be smart selecting the people you have in your team and choose as collaborators.