Monica Chirila 1
Monica Chirila 1
Monica Chirila 1
MC: Could you please tell us a few words about your current activity and about your
main accomplishments in the field of radiotherapy?
MvH: Well, I work on improving the precision of radiotherapy. I've been very active in image
guided radiotherapy, quantifying organ motion, 4D, that kind of stuff, and I like to look for things
that are unsolved. One thing that wasn't solved, was delineation. People are delineating tumors
but do not agree, so I spent some time observing variation and teaching on that. Now, I think the
big questions remaining are regarding tumor biology - what do we need to radiate? Where are
the targets? Where's the microscopic disease? Currently, the focus of my work is big data - trying
https://doi.org/10.53011/JMRO.2021.02.10 86 Published by Casa Cărții de Știință
2784 – 0131/Copyright 2021 The Author(s). This is an open access article under a CC-BY license
to learn from the data of thousands of patients that were treated in the past. We try to see for
instance, which parts of the patient are the most sensitive to radiation? What happens if you get
a little bit more dose here and there? Do you see a difference in outcome? I think it's kind of a
natural progression.
MC: Is it difficult to gather this amount of data that you need to analyze?
MvH: I started in Amsterdam, but I work in Manchester now, the biggest Cancer Center in
Europe. But treatment data, outcome data is harder to get, so we've linked up with external
databases, like the NHS (National Health System). So there is a lot of data. One thing that we
can always find out is if the patient is dead or alive. So actually, a lot of the activities aim to
predict the outcome. And sometimes you just find really surprising things. For example, we were
looking at data from older 3D plans and correlated with the “dead or alive '' status, and we found
a little spot in the heart which got more dose and was correlated with a worse outcome. That
was kind of an eye opener. And that gives us an opportunity, because it is not always possible
to spare all the heart, so we can try to avoid the structures which are the most sensitive. It might
be the electrical system, but we don't know. So we have to figure it out. We have to run
prospective studies, looking at the effects of radiation - we're just finishing one. We looked at
cardiac CT, before and after radiotherapy, and so we will hopefully report our findings in the near
future.
MC: So you are not only searching for factors to predict the patient outcome, but also
to identify toxicity-sensitive structures?
MvH: Yes, both. But it can be very complicated. If you want to compare data from different
patients you have to put them together. The organs at risk can be mapped together, but the
tumors are in different parts of the body, have different shapes and different sizes.
MC: How do you deal with the variability of the data? There are different devices,
different types of scanners, and they produce images that are slightly different.
MvH: Actually, we do a lot of things with the planning data. And that's fairly well standardized.
So that's not such a big deal. But obviously, statistics is actually quite complicated, making
corrections for confounding factors.
MC: How did you identify these problems that need to be solved, and which later
become your research focus? Do you find them by yourself? Or do you collaborate with
clinicians and ask their opinions on what may be more difficult in their clinical work?
MvH: Of course, I am collaborating with great clinicians and their input is really important.
But it's also just… listening. I’ve been around for so long now. I’m just picking up.
MC: How did you get to be involved in this field? Were you interested in that from the
beginning of your career or was it something that started to interest you in time?
MvH: It's kind of funny story with that, because I studied physics. In school I was good at
maths, but also physics. Computer science didn't exist yet. While I was in University I had a one
year project. And so I went to the Cancer Institute in Amsterdam and there was a guy building
his own computers and doing a lot of technology work. So that is how I got interested in this
topic.
87
MC: What about image guided radiotherapy? Which was the main progress in the last
years, in your opinion?
MvH: Well, obviously, we've got all sorts of new scanners, new reconstruction algorithms,
MRI guidance, but the fundamental problem is the patient. And I think people often forget that,
you know, you can get such beautiful equipments, but if the patient’s organs are moving during
the imaging, or during treatments, it's never going to be very good. So I think we need to find
ways to deal with that. And it’s funny, because the medical physicists want to measure stuff.
They put a phantom on a scanner and they make measurements, but this does not say anything
about the usability, because if you put a patient there, the things are totally different.
MC: Why do you think about image enhancement, in particular the use of the Cone
Beam CT(CBCT) as a predictive tool?
MvH: Well, there is something in there, but the CBCT is not very good. And the main reason
why it is not very good is because the patient is moving, the organs are moving. So, the best
images actually, strangely, you get in the lung where you can do 4D CT. The contrast is very
high. We’ve been looking at how these images change over time, but I don't think it's very
predictive. But we're looking at those images to understand what is happening, for instance, to
predict whether you need to adapt to treatment. So it's a workflow thing.
What we are now going to do with the cardiac substructures is doing what we call a rapid
learning study. So we're treating one way and then at a certain point to kind of change it and
we're treating another way and then what we hope to see is a jump in the output. We have to be
really careful because if something else changes at the same time, obviously then that's what
you measure is not true.
MC: In what fields of radiation oncology do you think artificial intelligence will be most
helpful? Is it going to be autocontouring or decision making or something else?
MvH: I think it is contouring, but also this one I mentioned before - the biomarkers - I think
that would be quite cool as well. The computer should say: I've measured densities in your body
and looked at the picture and you're very weak. I do not like radiomics because it's garbage in
and garbage out in a way, and then because the box is so big, you have so many variables. Yes,
you can always find something but if you don't correct for the other variables…Actually it might
be the same with genetics. Yeah, lots of zillions of people are working on them. But very few
studies have found something that actually works.
MC: Did you or do you work for the industry, for the companies which produce
radiotherapy devices?
MvH: No, not for, but with the companies. My first project was the portal imaging device, so
we created the Varian PortalVision. We worked for decades with Varian, and then we started
88
working with Elekta for the cone beam CT. I was out with a friend and he invited me to join him
to write a grant. The deadline was in just a week. It was on image guided radiotherapy in prostate
cancer. I did a little software things. And when Elekta started to build those machines, we were
already in the circle, and we wrote the software from scratch and then we provided bug fixes.
We were subcontractors. That was big.
I also work with Aquilab on training contouring system. I wrote some software there, but I will
probably collaborate with other companies too, in the future. I also have some ideas, but if they
are not translated into projects... Even if I give away the software for free, it’s not going to be
used clinically, it has to be in a company, and it has to be supported, FDA approved and stuff…
But companies are difficult in the sense that they often change very quickly. So different people,
new head of the group, different vision, priorities… I'm involved with MRI Linac, but only very
little.
MC: What you say is very nice, because we need this human part, also.
MvH: And I know you were asking about the companies and that's often difficult because it
really depends on a few people. And if they are not there, or they got a new boss, suddenly
people don't understand what they have.
MC: About understanding – this raises another question: how do you make people
understand what you're thinking about in case there's a gap between your training and
their possibility of understanding the concept which you're explaining?
MvH: I just try to explain it very, very simply. But yeah, I think the main problem is people
who don't want to listen, they don't understand. If people have a preset idea, then they really
don't understand the issue, and you should have started at another level. To explain this, you
go back to the basics.
MC: I was thinking that the studies on motion management are becoming even more
important, because we are giving a higher dose in less fractions. So, the movement has
more impact.
MvH: It is totally true. The amount of technology that you use depends on what you need it
for. But if there is a simple solution, that is much better.
MC: Do you think that the MRI Linac will be the best solution for accurately delivering
the radiotherapy ?
MvH: There are other tools for image guided radiotherapy, like fiducials and many more. If
the MRI Linac could provide imaging in 3D in a fraction of a second, then it could pretty much
solve all the problems, but that's not how it is. It's a complicated machine and workflows are
really, really difficult. And it's tight, it's claustrophobic. And MRI is slow, and slower than the CT
and slower than the CBCT. Because you can take a single slice very quickly, but if you want to
take the volume, it takes much longer. Not all the organs move the same, and the longer the
workflow is, the more is the chance that the target or the organs at risk would move.
MC: What are your plans for the future? Do you have anything in particular that you're
working on that you would like to share with us?
MvH: Yes, no doubt… As I mentioned, one project is focused on measuring the electrical
effects of radiotherapy on the heart. Another thing that we're starting up is to use routine imaging
for biomarkers. For example – a change in the performance status can influence the change of
the treatment, but evaluating the performance status can be subjective. But the images actually
89
have a lot of information because you can see how the muscles are developed. You can see the
liver, you can see the heart (whether it has calcification or not), so there is a lot of information in
routine data that we can exploit. It’s already there; we just have to find out how to use it. But
that's not the most important thing. The most important thing is that this amount of data exists in
the past. So we can try out all sorts of algorithms on the old data to see whether that is predictive
of what's happening to the patient. So, that is a cool thing. I like that a lot.
MC: You mentioned the lockdown times with all the challenges, and I am wondering,
how do you find ways to overcome challenges? What drives you? What motivates you to
solve your research or work related problems and challenges? How do you manage not
to give up?
MvH: Giving up isn't an option, because we can give up on a project, but that does not mean
that we can give up on a student. You can give up on an idea, but you can't give up on the
students. It is not an option, because you're responsible. And yes, I take that very seriously. One
of the postdocs in our university said: I have two supervisors: a good supervisor (my colleague)-
he tells me to go to bed early, read, etc; and I have a bad supervisor, who tells me to go to
parties. That’s me. And I was very, very proud to be the “bad” supervisor.
MC: I recently asked a young researcher in the field of artificial intelligence what he
would ask you if he would meet you. His answer surprised me. He said that he would not
ask anything, but would only thank you for all that you have done. So, I will do the same,
and end by thanking you for your valuable contribution for delivering quality radiation
treatments to cancer patients.
90