Medical 3D Players: Fighting Cancer, When Physics Meets 3D Printing

Pieter Slagmolen  00:02

Welcome to the 3D Players Podcast where we explore personalization in healthcare through advancements in 3D technology. We'll talk to leaders championing more predictable and sustainable patient care. I'm Pieter Slagmolen, and I'm joined by my co-host, Sebastian De Boodt. Our guest today is Tom Depuydt. Tom is the head of Radiation Physics at University Hospital in Leuven, and we will take the opportunity to talk about mass personalization in radiotherapy. Now, regular listeners of our podcast may find this a surprising twist in topics. But radiotherapy is one of the foundations of cancer treatment. For us, this is an inspiring field that we are eager to dive into. In Europe, cancer is the number one cause of death, affecting approximately one in four people. So chances are that every one of our listeners knows somebody who is suffering, has suffered, or has died from cancer. The thought of learning how 3D technology, and specifically 3D printing, can help cancer treatment is rather exciting for us. Tom, it's an absolute pleasure to have you here with us today.

Tom Depuydt  01:09

Thank you, Pieter. The pleasure is mine.

Pieter Slagmolen  01:11

Maybe for the people in our audience who are less familiar with it, can you start by explaining what radiotherapy is actually all about?

Tom Depuydt  01:20

Well, you already mentioned that it plays an important role in the treatment of cancer, and indeed, a little bit more than half of the patients facing cancer treatment will have radiation therapy in their treatment at some point. So, it plays an important role in the cure for cancer. So, what we do, we use, as the name suggests, radiation; we use beams of ionizing radiation to treat cancer. This is our agent; this is what we work with. And what we try to do is to expose the tumor to a high level of radiation dose damage to eradicate it. But you know that ionizing radiation causes some damage. Our goal, and certainly medical physics (it's my specific profession), deals with the fact that we want to give the tumor a very high dose but the surrounding healthy tissue a very low dose. And that's all technique; technology that involved that from imaging to treatment delivery technology. That goal had a lot of research, a lot of developments from the last 40 – 50 years to reduce the exposure of surrounding healthy tissues from certain critical, sensitive structures in the body.

Sebastian De Boodt  02:51

All right, now Tom, you are leading a team of physicists, not physicians. It seems a significant outlier in the medical field that we have a technical team so directly involved in the patient treatment. What is your role exactly as a physicist in the day-to-day radiotherapy world?

Tom Depuydt  03:09

In radiation oncology, we have a shared responsibility, as we have the physician as the medical doctor who will follow up on the patient and is involved in the diagnosis, but also, in a way delineates where the disease is, the affected tissue and the surrounding tissues, determines the sensitive structures around that volume, and then prescribes a certain dose. And that depends on certain protocols, on knowledge share, on clinical studies, where they determine that, okay, certain cancer types need a certain dose. And then we come into play, the medical physicist. What do we do? We have a few aspects: we manage the equipment (the entire lifecycle of the equipment), so when it comes in, it's installed, we accept it, we commission it, and we keep the quality of the machine up to the standards that are required for treatment. And the other aspect is that definitely in radiation oncology, we play a role for every individual patient and we are involved in their treatment, because my team creates the treatment plans, designs the treatment plan, as they're very specific and customized to each individual patient. We create these treatment plans. And that happens in the digital world, we create kind of a digital twin of the patient based on CT images. Now the CT is the backbone of the digital twin. And in that environment, that digital environment, we can simulate the treatment beams. We can simulate all models of our treatment units, and the development of these models — that's also our responsibility. In this environment, we can basically create the treatment. Treatment is then sent to the linear accelerators proton therapy, or HDR after loads for brachytherapy, and is then delivered to the patient but the actual ladder, the computational process to develop these treatments — because it's very complicated, we have 1,000s of parameters, it’s not a manual process, it's fully automated nowadays — are automated up to the highest level. That's our responsibility.

Sebastian De Boodt  05:26

All right. Sounds very, very interesting and complicated. We're very familiar with personalization on surgical applications. And I think most of our episodes, so far, have revolved around it. To a less extent, we're familiar with personalization in radiotherapy. We already briefly touched upon some of the parts where you're involved in. Now, why is personalization important in radiotherapy? And what are some of the specifics that really need personalization in radiotherapy?

 

Tom Depuydt  05:54

So, as I explained, it is already highly personalized. The treatment is custom-built for the patient. And that relates to how the treatment systems deliver treatment. But okay, we have a digital model, and you can extend that customization. Also, certain devices require that we perform the treatment be customized as well, can be personalized. That already happens up to some level, with certain thermoplastic sheets, for example, that are molded around the patient, with certain cushioning and vacuum cushions, for example, that are personalized. But now we see the evolution towards 3D printing as a technique to make these personalized devices help in treatment. And that's the evolution that is ongoing right now, I would say.

Pieter Slagmolen  06:48

And that's the next step, in a way. But today, if I understand it correctly, for our audience, the personalization is, is already happening by basically tweaking how the treatment machines behave in relation to the patient. Right? That's what happens when we talk about treatment planning and how it will be brought to the patient.

Tom Depuydt  07:06

Yeah, exactly. So we are already highly personalized models — not only do we take into account the specific anatomy of the patient and the target volume, the surrounding structures — but we also take into account breathing motion for example, that is also very specific to a patient so that if we have moving tumors, we also take that into account. So, you see we use as much information about the patient as we can, and we build it into this digital twin, and then we create the treatment planning specifically for that patient.

Pieter Slagmolen  07:40

It does feel like a very resource-intensive approach. Can you illustrate how many people are involved in treating a patient? How much effort is being spent in preparing all of this for an individual patient?

Tom Depuydt  07:55

Well, for an individual patient while you need from the CT scanner — which I already mentioned is the backbone of this patient model, the digital twin — until we start treatment you have to count that's one to two weeks, depending on the complexity. And you have some anatomy segmentation and nowadays, okay, there's a lot going on with artificial intelligence to speed that up and to make that easier, more consistent even. And then the treatment planning process itself. Okay, it's not a manual process, and we have a lot of technology that helps us there. Already some machine learning and AI is also finding its way to that part of the process and to make that speedier and maybe increase quality even. But that process takes take some time, it can go from an hour to a few hours to maybe a few days to create a treatment plan that is patient specific. And so that depends on the complexity. And then there's a part where we do quality assurance, quality control for each individual treatment plan. And there we strive towards automation as well. Things should go more automatic rather than manual and trying to avoid more the experimental validation of things and more go to computational validation of these treatment plans. But it's labor-intensive, yes. But we do see that technology helps us there. And that will further increase in the coming years.

Pieter Slagmolen  09:23

I think it's, I mean, you've touched on a number of things that we're looking at as well, because indeed, what we're trying to do is make... well, at least the conversation that we're trying to have... is how do we make personalization more scalable? We're also seeing on the surgical side that, yeah, technologies like AI, for example, are definitely going to help there. The way you introduce it, that's also one of the aspects that you're using is automation on the software level, are there other things that are being done in the radiotherapy field, that maybe make radiotherapy as a whole, more scalable?

Tom Depuydt  10:00

Well, we have the patient model, the digital twin, and we are injecting as much information as we have. We started off using no CT imaging, and now we use CT imaging, we go beyond that. So more and more information is put in there and also more knowledge on biology. This AI will be trained with biology, as well, now I'm going a bit out of my expertise. But that's where I anticipate it will go. And that's, that it's logical that's the AI. Maybe that's not the discussion you will have today. But AI will use all the information that is available to make a good decision. So that patient model, that digital twin, will become richer and richer and richer. And in that, we will support automated treatment and decision making.

Sebastian De Boodt  10:50

I would like to take it back as well to the 3D printing, and go a little bit deeper into that, and maybe just taking a step back, because I'm kind of a novice in the field. And my colleague Pieter here, you know, it's kind of his background. For me, I'm stepping into this quite new, you talked about certain devices where you're using thermoplastics. Now looking into 3D printing, what's the role of those devices actually in the treatment? Why are they needed?

Tom Depuydt  11:15

Well, let me maybe first give you a small overview of what is done with 3D printing nowadays in radiation oncology, and then we'll get into more detail. So 3D printing is available in radiation oncology; a lot of centers buy 3D printers, have them on site. But what is done with it... that varies over the different sites. It's used for phantoms, and what we call phantoms are devices that we create to do, for example, experimental dosimetry. We can create our own phantoms now and we can, develop and design the shape ourselves, depending on the experiment that you do. In fact, a phantom for that specific experiment, you can print it, use it. And so that gives a lot of freedom. Certain people use it to print even the geometry of the patient and validate the treatment in the geometry of the patient or the printed phantom. So the shape of the phantom gets the shape of the patient, and the treatment is validated in these circumstances.

Pieter Slagmolen  12:14

So you're basically copy-pasting the patient, right?

Tom Depuydt  12:16

Exactly, yes. We can do dosimetry in that object. And in this we validate the treatment before we deliver it. Okay, that's one use. Then in brachytherapy, there's also a lot of applications. What we do there, for example...

Pieter Slagmolen  12:31

For clarification, for people not familiar, can you explain what brachytherapy is?

Tom Depuydt  12:36

Yeah. So brachytherapy is where we bring in radioactive sources close to or in the tumor itself. You can imagine if you want to avoid surrounding structures, and you are able to bring the radiation source inside the tumor, that's a big advantage because you treat the tumor from inside. It's also used for superficial lesions, skin lesions, etc. And that's where there are a lot of applications with 3D printing because you want to bring the sources close to the patient and certain ducts. You have to place these ducts exactly at the same position, at the right position on the skin. And for example, if you have treatment of the nose, it's very interesting to print a device that shapes around that local anatomy, where you have these ducts where you bring in these sources. So you can imagine that 3D printing is a real advantage. We use it also in UZ Leuven [teaching and research hospital in Belgium], and that's an application used In a lot of brachytherapy sites. Some people look into other sites for brachytherapy, as well, concept is the same but there the device is brought into the patient in intracavitary treatments, for example. That’s under development; it's a lot more complicated as you bring in these devices into the patient. Another application is bolus. An external beam, it's practically similar to what I explained with the nose. We have certain complicated anatomy, and we can place material on top of it. We do that because with megavolt beams, the highest dose is not delivered at the surface, but at a certain depth of a few centimeters. But for some indications, we want to pull the dose towards the surface. And that we do by putting some material on top of the skin. But in some parts of the body, it's very hard to have like a generic bolus. So, if we can print it based on the CT scan, where it's patient specific and it’s really shaped around the patient’s anatomy, and we get much better quality. So that's another application. So these three applications are the main applications nowadays. And yeah, we use these routinely already. And a lot of hospitals have invested in that.

Sebastian De Boodt  14:50

And are those, where are they now in terms of maturity of using 3D printing? Is everybody at all the centers, kind of figuring things out for themselves with more of a DIY approach, I guess? ...Yet, of course, a professional approach! But creating things yourself, or is there already more established products that’s a bit more boxed in terms of what you have that’s available.

Tom Depuydt  15:13

There's beginning to be some commercial interests at the moment, but it is fairly limited still. Also, how these are produced, I think hospitals and industry are still looking for ways to collaborate. We have today the famous MDR, the medical device regulation that plays a role there, as well, and will have an impact on how we organize ourselves in the future. But what I'm pretty sure about is there will be a collaboration between hospitals and industry. That's obvious. This approach will make it possible to roll out 3D printing technology in radiation oncology in a much broader way. And that's necessary because, okay, the way it goes, these things are used (if you look in literature), they are used but with small cohorts, not the large clinical studies. To really prove that they have an advantage, you need a huge number of patients and larger studies to prove [the advantage] that’s the way medicine goes. It should be evidence-based. And I think, at the moment, there’s just not enough available to be able to do that.

Pieter Slagmolen  16:28

One thing that I wanted to ask, as a follow up on your applications, maybe for the people in the audience who don't know that I have a history in radiotherapy myself. So, I started my PhD with Tom, already 17 years ago in Leuven [Belgium] and I remember, the first week there, the former head of oncology forced me to spend time with patients for a full week. Basically, I had to be part of patient interviews and also take part in the planning process and everything. When I was doing research, I could at least relate it to the reality — which I found very interesting. But also, I could see that the technology at that point in time was not very mature yet. I mean, in terms the planning, a lot was still being done in 2D, but also a lot of the devices that you described now, were being used in treatment, because these people were already treated with radiation therapy, right? So, these devices were already required. They were manufactured, maybe in a more archaic way at that point in time. Can you relate that, the reality of how these devices were made before to what 3D printing can do now in terms of efficiency in terms of way of working?

Tom Depuydt  17:38

Well, one other example is patient immobilization. So, we treat patients not in a single fraction, however, that is sometimes done as well. But usually you have multiple fractions, 20 to 30 fractions. And every time we need the patient to be in exactly the same position, as in the planned CT image basically. So that means you need certain devices that reproduce or help reproduce the position of the patient on a daily basis. And also, the time that you started. We did it with thermoplastic masks, and these are devices that are sheets of plastic, thermoplastic. But it's also used in 3D printing, but in another form. We heat it up, and then we mold it around the patient, it cools down and it becomes hard. You get a mold for the patients that completely encompasses the patient. That's a process that patients, well, don't find very comfortable. It's a bit claustrophobic, they have to wait for 20-25 minutes for it to cool down and to harden. I’ve been in one of these masks just to have the experience, to know what it is. And it's really claustrophobic! It shrinks a little bit so the process is not very pleasant, the patient is not looking forward to that. Also, they have to get into this mask every time for each individual fraction. What we're looking into today with 3D printing is, I guess, an option. The freedom that you have with 3D printing is that you can create a device that instead of really encompassing the entire anatomy, and squeeze it into a certain shape, you can really focus on the parts of the anatomy. For example, in the head and neck region that are really stable and the nose bridge. In other parts of that region, well, they're pretty stable. And they're not really depending on weight loss of the patient, for example, or other things, and they're very stable. You can build a device that immobilizes the patient based on the specific features, and then it becomes a device that might be more open, less claustrophobic, and for the patient that makes a big difference. We've done a small study in the past. One of my students, Steven Michiels, did a small study. And again, it's a small cohort, but it already showed that there were some advantages. The patients were more comfortable in these devices. And in terms of immobilization quality (because that's, of course, also important that they should be at least as good as the technology we used before) they gave similar results. And that's based on a very small study. I know from a scientific point of view that’s not very valuable, but it shows that there is some promise in there, and we should keep on working on that solution.

Sebastian De Boodt  20:10

That’s already a fourth application out of the three that you previously mentioned. And what's striking me is that they're spread over a lot of different ways, like phantom. And then as fixation devices. Looking at it from an industry point of view, is there any one of the four where you say, okay, “If we need to collaborate with industry...” Or if there is potentially a good fit for industry to really take this on and look at commercial opportunities, then that's the one of the four that I would see emerging first?

Tom Depuydt  20:40

Well, looking at what we're doing in Leuven, the example of brachytherapy that I mentioned, patient bolus, that's, in my opinion, a no-brainer, where we should have solutions for that as soon as possible. Because they are used, and it's certainly in brachytherapy, it is quite obvious that it's better than the solutions we have before. Sometimes with one material that, well, if you have a device that you cannot imagine that we did it the way we did in the past. So, it's pretty obvious that these kinds of solutions should become available. What is important in radiotherapy, specifically, that is combining the 3D printing with the clinical workflow and having it integrated. That's important because, okay, we have a certain workflow that where we have one to two weeks to prepare the treatment. And the production of these devices is patient-specific, and should perfectly fit into that workflow. I think, from a commercial standpoint, that's where the gain is. There we should create solutions because okay, it also is connected to rolling it out in radiation oncology in a broader way that you do not only have (like in academic sites with a lot of resources that can develop their own solutions...) but that all hospitals doing radiation oncology have this technology in-house and have the software and these solutions can be available to use for the patients. And that will play a big role.

Sebastian De Boodt  22:14

And when you say integration, how should I see it? Is it mostly on the software side that you're looking for a smooth way to go from your therapy planning and maybe to design, to manufacturing? Or is it also the 3D printing itself that you feel needs to be inside the hospital? Or can it also happen outside? What's most critical when you say it needs to be more integrated?

Tom Depuydt  22:35

Yeah, well, let's start with the design. I already explained a few times already the digital twin that we have because that's the core. And well we do the development of the treatment in silico. So it's obvious that you create devices that are needed for the treatment together with the treatment planning, and they should be combined — you should create a treatment plan. For example, immobilization, you should, maybe introduce certain bolus material in some parts. Maybe you should avoid that so that the device is created together with the treatment plan, bolus material brachytherapy. The same should be created with the treatment plan. And sometimes it becomes a treatment plan. For example, trajectories of the sources through these devices, well, they should be in a way created in the treatment planning. You create the device together with the treatment plan, it's part of that process. So it's not that... I hope that at some point, it's integrated in the treatment planning software that we have in radiation oncology, and radiotherapy. Where it's basically an end result of your treatment planning process, and you have the treatment plan that is sent to the delivery unit, and you have a device that is created. And that can be either printed in house, maybe outsourced, we will see about that. But as long as it's produced within the timing that we need to start the treatment for the patient, because in cancer, it's really important to start in time, obviously, then, then all solutions are possible.

Sebastian De Boodt  24:01

So one of the things that in the 3D printing in the hospital community is now a hot topic — and also driven by the MDR — is quality management of in-house produced devices. How do you look at that in this 3D printing scenario? And is this something that you feel you're confronted with already at the moment? Or is this more something we should take into account in the future? What's your experience with that?

Tom Depuydt  24:25

Well, we have to discuss with industry how we divide the work of this quality measurement system, and they will be okay. If I'm looking at it from a medical physics point of view, at some point, we have to look at the aspects that are related to medical physics and whether the device is functional. You guys might look at the geometry of the device and say, “Okay, we’ve got a model,” and we create the device within the specs of the geometry that was delivered to us. That's one thing. But at the point that we get it into the hospital, we should do some additional quality checks that are related to the treatment plan. Whether the device, in terms of interaction with radiation, or if the homogeneity of the device is really within specs, can be used for treatment. But that's yeah, under discussion, it's under development. I think the next coming years, we'll have to find solutions for that.

Pieter Slagmolen  25:23

You talked already about two of the barriers maybe. Integration into clinical workflow, and then probably quality and quality assurance as two aspects of expanding the adoption of 3D printing. Are there other aspects that we should think of that are important that are inhibiting you from doing more 3D printing?

Tom Depuydt  25:44

Well, we should definitely look into the cost as well. That's an important aspect. Now, 3D printing over the years, I've been into that field for a few years and I saw the prices go down tremendously. But okay, we should take that into account, too. We have to build new solutions; we have more possibilities now to improve treatments. What's important is that we go look for these opportunities to improve treatment for the patients. But the cost should be within the limits. We can find additional functionality in these devices, etc. Still, it's important that it remains more or less at the level of what we're doing today. That's an important aspect we should take into account if we develop new solutions. That's for sure.

Sebastian De Boodt  26:27

And previously, you also mentioned the importance of clinical evidence and that it should be evidence based. Do you see barriers there, as well? Or is it more a matter of, you know, time, and more institutes building cases?

Tom Depuydt  26:40

Well, there are some challenges and MDR has put a big challenge in front of us — because, okay, it's challenging — and discussions are ongoing, I think everywhere, you have to develop these devices, use them in a clinical study framework and improve their added value. And to use these devices, create these devices, before they become like a product. And that's a challenge. We have to do it with the new regulations that are in place, we have to find solutions. Otherwise, it's impossible to do this kind of research and prove that they have added value. We have to be able to do that. And we have to find ways to do that. But within the regulation framework, obviously,

Sebastian De Boodt  27:21

The challenges may be more on the quality and patient safety side, rather than on being able to do the experiments to get to what's the added value of using the 3D-printed device, for example.

Tom Depuydt  27:32

Yeah, it will be step up. I mean, we'll start patient safety, obviously and with limited studies, and then build up towards a device that’s really functional, that has the potential of, for example, immobilization, immobilize the patient properly, and then do clinical studies with these devices. And that should be possible in order to prove the added value.

Pieter Slagmolen  27:55

Maybe on another technology note, you already mentioned that AI is also entering into your space. How do you see this impacting the future of how treatment is being prepared for radiotherapy patients?

Tom Depuydt  28:12

Well, there's a lot of research going on. It's not my specialty, but there's a lot of ongoing research, I mean, AI, I'm a bit personally...while I'm not skeptic towards AI, I think it's, it's an interesting technology, and it's already proven its worth but I do hope we go beyond the hype and use it where it's really added value, maybe today, it's used in a lot of applications where the added value is limited. But in some applications, the added value is really, really obvious. I do believe that it will change also a lot of things we do in radiation oncology. I mean, that's the way to go forward. It will show up in a lot of software packages. And techniques that we use in the future. That is obvious; there's no way around it.

Sebastian De Boodt  28:58

Tom, for a couple of times, you mentioned that this digital twin is so important in doing proper planning. How is this digital twin?  I guess it's a model of how a patient would behave? Is it something that you are developing yourself? Or are there common models that the radiotherapy community is, you know, contributing to, how should I look at that?

Tom Depuydt  29:21

Well, I use digital twin because it is also used in industry and manufacturing. Actually, it's the same which means it's something digital that has the same features as the real patients. Okay, it's a model which means it only has a limited number of features. Not all of them, but more and more, you'll see that the digital twin becomes richer and richer. Motion is already in there, and biology will be in there, as well. Some other information that we have available, even genetic information, will come together into that digital model and can be used to design and develop treatment. It's a lot of the preparation. Almost all of it does not involve the patient anymore. It involves the digital model of the patient and even immobilization will be developed based on the digital twin, while in the past, we used the patient as a model. We use a thermoplastic sheet. Now when that's a study already done, that's possible, which means that the immobilization device and the patient see each other the first time during the first fraction of treatment. That's why I call a digital twin, because it's we do the preparation phase of the treatment.

Pieter Slagmolen  30:30

Awesome. I think digital twin analogy we've heard it a few times already in this in this account, it's always been used in different contexts. I think this is a very tangible one to understand a very interesting use case. Maybe looking beyond, we talked a lot about digitalization, future applications of digital twin integration, radiation therapy and 3D printing and how it fits in there. I want to take it back maybe to what a lot of our audience is familiar with, which is the surgical side of treatment probably also on the oncology side. Do you see the coexistence of the different treatments in cancer, radiation therapy and surgery maybe as two primary ones? Do you see that change in the future as radiation therapy becomes more digital, more advanced? For example, will it put additional demands on surgical treatment afterwards? Do you see that dynamic change between these two?

Tom Depuydt  31:24

That's an interesting question. Well, there's always an interaction between the different ways to treat cancer. Every issue oncology you have, even today, and now we're getting a bit out of the scope, I believe, but anyway, you have surgery, you have chemotherapy, you have immunotherapy, so you have a lot of ways to treat cancer, and they interact continuously. And okay, evolution or development in one of these always impacts the others and directl interaction with others — there's always some dynamics. And how 3D printing specifically might impact that, we’re getting better and better at radiation therapy, more precise, over the last 30-40 years have become really, really precise. So that has an impact, for example, on surgery. These studies have been ongoing; it's not always easy to make these comparisons. But that will have an impact. But as I said, it's continuous interaction and evolutions within these separate fields that will impact the dynamics.

Pieter Slagmolen  32:30

Everyone is driving the performance of their individual treatment forward.

Tom Depuydt  32:36

And there is some competition which is always good.

Sebastian De Boodt  32:39

So Tom, thanks so much for joining us and sharing all those useful insights. Three key takeaways from our conversation are, first of all, that a personalized approach is already ingrained in your day-to-day within radiotherapy. So it seems like it's only a logical next step to venture into 3D printing. And that's where secondly, I would say it's very insightful to see how there's already various applications of 3D printing that seem to be successful, such as brachytherapy, and then the bolus, are maybe the ones that will see reach broad adoption first. But then, lastly, my third takeaway is probably the importance of the collaboration between industry and the hospitals, to really, take it to the next step, and that one of the things you should work on is combining the 3D part of your clinical workflow with the design is part of the treatment plan. It's something we can't separate. Tom, anything to add?

 

Tom Depuydt  33:40

No, I really liked your summary. Can I use it? Thank you very much. It was a pleasure to be here to discuss these things with you guys.

Sebastian De Boodt  33:50

We appreciate you sharing your insights with us on 3D players, a podcast where we explore trends, insights, and innovations in personalized and sustainable health care. We are your hosts Pieter Slagmolen and Sebastian De Boodt. Thank you so much for listening, and join us for the next edition.