'Beyond Bricks and Mortar - Building Quality Clinical Cancer Services' Symposium 2011

Advanced Technology in Radiotherapy: How can we ensure Maximum Benefits for Patients? - Associate Professor Tomas Kron

Up to Radiation Oncology

prev pageprev page| TOC |next page

Peter MacCallum Cancer Centre

Download powerpoint presentation by Associate Professor Tomas Kron (PDF 4457 KB)

Norman Swan:

Our first presentation today is from Tomas Kron, Principal Research Physicist at Peter Mac, has academic posts in several other institutions. He has particular interest in dosimetry of ionising radiation, and he’s going to talk today about new technologies that may assist in safety and quality. Tomas.

Tomas Kron:

Thank you very much Norman, and I’m delighted that you’ve started off talking about measurement, that’s really right close to my heart being a physicist. I would also like to thank the organisers and particularly Abel and his team for allowing me to speak, as a physicist here the first session in the morning. I would also like to thank you all for actually being here, for a physicist first thing in the morning.

What I’d like to do first is really declare a possible conflict of interest, and ACDS will be talked about later today, but TROG is the clinical trials organisation running clinical trials with radiation oncology here in Australia and New Zealand. So if I say something about clinical trials and it sounds too positive and too enthusiastic, please take it with a grain of salt.

What I’d like to talk about in the next 20, 25 minutes is really fairly basic. I first wanted to talk about and introduce two types of new technology, IMRT and IGRT, which are sort of covering this spectrum of new technology. So for many of you this is really bread and butter already and you know everything about it, so please have a cup of coffee and enjoy yourself. But I think it is very important that we all try to speak the same language. One of the big problems in radiation oncology we have is communication, communication with all the stakeholders and the acronyms sort of IMRT, IGRT and our language is not particularly helpful to actually foster that communication. So I will try to come up with some illustration and potentially a definition of these terms, so we can build the rest of the day on that, probably a bit better than this tower was built.

Then I want to come to the theme of this whole workshop which is really why should patients be interested in that, and then sort of give a couple of ideas how we should introduce these technologies.
To start with I just want to sort of recap from a simple physicist's point of view what I think radiation oncology is all about; it is about achieving local regional control, and I was delighted to hear a surgeon say yesterday that we are all in the same boat. Indeed. It’s local regional control we are after, and that really requires two things. It requires the identification of the target, which is shown on the left and huge achievements have been made there in the last couple of years through imaging, PET scanning, MRI, there’s a whole variety of new technologies, which allow us to identify the target much better, not only image wise but also genetic and with a variety of other tools: molecular oncology, which give us more information about the target which could help us to deliver radiation more effectively.

But that still leaves us with the problem of actually delivering radiation, and that breaks down again into two different aspects. One aspect is to actually generate dose distributions which are fancy, and which can cover just the target and try to spare the normal structures around it as much as possible. And that’s really what IMRT is all about. And then we have to take this dose distribution, this very conformal dose distribution and actually put it in the right spot. There’s no point in having the right dose distribution in the wrong spot. So that is really what image guidance stands all about, and these are really sort of the themes of this presentation.

If I look at traditional radiotherapy then this is sort of illustrating that what was done 10 years, 20 years ago. If you look at the target it’s a complex target here on the left side, and we needed to try and cover that with radiation but trying not to give as much radiation, as much as possible, outside. The way we used to do that was a four field box, and that’s a four field box, as you can see here we have radiation fields coming from four directions and covering the target in a box shape. Very nice. It's not too difficult to hit the target. But we treat a lot of normal structures with this, and conformal radiotherapy then about 20 years ago started to emerge, and that’s now standard of practice, I think all over Australia. Conformal radiotherapy basically takes each of these radiation fields and shapes it to the target as it is seen from that particular direction. So that basically will cut out a lot of normal structures already.

What we would like to do and this is something we can do nowadays much more automatically than we used to do using multileaf collimators and I’ve sort of shown a multileaf collimator there; they’re just little leaves which allow us more or less to arbitrarily shape our radiation fields.

Why we would like to do anything more? And that really is informed by the fact that radiotherapy’s not just about treating the targeted, it’s also sparing the normal structures. And that's a sort of a typical scenario, it's actually more common than we think, paraspinal tumours, head and neck tumours wrapping around the spinal cord, the prostate and seminal vesicles going around the rectum. All these structures have similar type of shapes.

If I now take six radiation fields it becomes very soon clear that we can only achieve good sparing of the organ at risk if we modulate the fields, at least the one coming from the top and the one coming from the bottom, where we have a dose distribution, a fluence distribution which varies and gives more dose to areas where no critical structures are in the way and less dose in the area where the critical structure is in the way. And that’s basically intensive modulated radiation therapy, and you can see that this field at the bottom and also the one on the top would be in terms of its fluence, modulated. So that basically allows us now not just to treat convex surfaces, but also concave surfaces. And that’s really the trick what IMRT is all about. It allows us now to also spare normal structures which are very close or potentially even encapsulated with normal tissues.

Now what characterises IMRT treatment, a lot of different multileaf positions, lots of complexity throughout, it is something which can only be generated in with computer help, and you can see here a head and neck treatment, you can see the yellow is the treatment area, the pink in the middle is the spinal cord, and the treatment fields with the different fluences used to treat that are shown around there, there are seven fields, and each field is sort of sub segmented in all these little pencils, which are turned on and off depending on the intensity we would like to use. More than a thousand leaf positions. This is not something which can be generated by a human. This is not something we can put on a piece of paper and take from treatment planning into the treatment room and then someone entering that data on treatment. This all needs computers, needs computer networks, needs databases.

These intensity modulated fields are again generated by multileaf collimators just by taking this up, several of these fields at least three fields here, are generated by multi lift collimators putting them together and basically generating any fluence map you would like to achieve. But in real life this is obviously much more complicated. And we need to ask ourselves now what is in this for patients? And I hope I have illustrated that the idea of also being able to shape concave surfaces and follow the dose around a concave area means that we have better tissue sparing, and this is really what IMRT is all about, at least to start with, normal tissue sparing.

But that basically allows us to reduce potentially the toxicity, and if we say "Well we want to keep the toxicity the same", that will allow us then on the opposite sides to increase the dose to their target. If I can lower the dose toxicity and the dose to the critical structure due to IMRT, I could also do the reverse and actually increase the dose to the target, hopefully increasing the tumour cure. And I have also the opportunity potentially to give higher doses per fraction and make the treatment more intense, therefore give fewer fractions, potentially reduce the hospital stay and through this what is called hyper fractionation, potentially reduce the resource requirements, reduce the number of times the patient has to go into the hospital, and potentially even the cost.

Now the question obviously we have is that clinical evidence for the use of IMRT and there was a better analysis in Lancet Oncology in 2008. In summary, at the time they were saying there was emerging evidence for a number of areas, mostly related to organ sparing. This is basically the summary of that paper, and all I want to say in this complex structure where they look at different treatment arrangements and different objectives, in a matrix is that green means that IMRT is better. Red means IMRT is worse, and there are some potential areas where IMRT could be worse, so case selection is important, and yellow means there is really no difference.

But overall I think the evidence is emerging that at least for normal tissue sparing, like porotic sparing in head and neck treatment, for rectal sparing in prostate cancer and a variety of other areas. IMRT just adds this complexity and therefore flexibility which allows us better sparing of normal tissues.

These are sort of typical dose distributions you can achieve, these are dose distributions we had in Canada when I was working with Tomotherapy, and you can see that you can basically just treat the skin off a foot wall, you can have a virtual body and spare the actual spinal canal, and then the treatment on the left spares both the larynx and the spinal cord at the same time. Really fantastic dose distributions which can be achieved with sophisticated means of doing IMRT.

But what are the drawbacks? Well if you look at this dose distribution here with spinal cord sparing, then obviously you don’t want to get that a centimetre to the right, because then the spinal cord sparing is all gone. You want to get that to the right spot, and obviously there is more complexity in this, and there is more cost, because we have to invest in that, and obviously more QA requirements.

So particularly the idea of generating these highly conformal dose distributions implicitly also requires us to get them better positioned and more reproducibly positioned in the patient, and that’s really what IGRT is all about, where we basically use imaging at the time of treatment delivery to put the radiation dose in the right spot. You wouldn’t build an operating theatre these days without some sort of imaging attached to it, because you just don’t want to go blind, and I think it just makes imminent sense also not to go blind in the treatment room if we have the imaging tools.

And the imaging tools become available, they actually have been in Australia already since more than 100 years, this is the first x-ray ever taken in Australia, in Bathurst. And what happened there is less than a year after Rentkin’s discovery, Father Slattery had the same equipment there and a young chap shot himself into the hand with a spring gun, and the Father, being a good man said, “Oh well we’ll take an x-ray”. That’s the resulting x-ray you can see. The quality’s not quite great by our standards, but you can see there the shot gun there really quite nicely. It is not reported that this changed the clinical management. But you can see straight away that the potential is really there.

And if one looks at modern image guidance in radiotherapy, the images look actually very similar. If you compare that previous image with this, this is a prostate, and again in order to make it better visible we’ve implanted fiducial markers, three gold seeds as you can see here, and we can image these very easily at the time of delivery, and therefore adjust our radiation beams to the right spot.

There are now, since this is sort of a concept which really caught on, there are now a huge number of different imaging tools available in the clinic, and mostly our modern linear accelerators come not only with electronic portal imaging where the treatment beam is used for imaging, but also with an attached x-ray unit to the side, which can be used to acquire cone beam CT images, three dimensional high quality images at the time of treatment delivery.

These are other image guidance solutions, and I think many of these are more driven by the fact that one manufacturer needs to avoid conflicting with a patent of another manufacturer; that’s why we see this huge variety of different imaging solutions. But in principle they all do the same thing at the time of treatment or just before treatment imaging the patient, getting the dose in the right spot.

But the question is really does that make a difference for patients? And this is sort of illustrating fantastic tools on the left. If that doesn’t result in any improvement in any change in management, we are really not making a real impact. So the question is really what can we do with these images, and I think the easiest way is really just to detect the systematic errors, growth arrows from planning to treatment. The CT scanner where the patient is planned, is not identical to the treatment unit - we can see that now. Systematic errors can be avoided, we can also image the patient every day and just put the target in the right spot, not put the patient in the right spot, but put the target in the right spot, makes perfect sense.

We might even think of modifying the treatment plan or choosing appropriate plan and potentially eventually you’ll see changes which may occur over these seven weeks of radiotherapy treatment. An example is given here, this is a lung cancer treatment (inaudible) at Peter Mac, we do weekly cone beam CT’s of patients, and you can see here, that’s at the start of treatment - you can see clearly the tumour there. Now seven days later this looks still very much the same, two weeks later, three weeks later and one month later, under treatment the tumour shrinks like that. Well obviously this is good news, and I thought I’d show that radiation actually does work, and it does do something. I think most people don’t doubt that here, but you can see that it also changes the anatomy.

Now I’m not saying that we can now change our treatment fields and reduce our field sizes, we don’t know that as yet. I also can’t say "Is that a predictor for good outcome?" Is that something where we now need to be more radical, because lung cancer has only has a cure rate or five year survival of 20%, is that a patient by dose escalation we actually can push up that hill and get it to 60% probability? I don’t think we know, and as a physicist I shouldn’t comment anyway. But I think we can generate images which are thought provoking which provide us with a lot of more information for the patient’s benefit.

And if we’re to modify our treatment this would take this picture all the way around and basically say that this image guidance where we acquire high quality images at the time of treatment can actually go back into shaping our treatment approach and to change the planning, which is something we talk about as adaptive radiotherapy, and that’s sort of just knocking on our door.

This is really at the heart of things, why would this be interesting to a patient? To a physicist it’s not - of course this is interesting. This is great stuff. And it comes down to something like that. Is that something which is obvious? Is this new technology obvious to you? Does it make sense? Is it clearly obvious or is it so complex that we would say "Oh mmm, not quite sure"? There is a good example in the literature where a matter analysis was made, and I don’t know if you’ve seen that paper, I’m sure it has been used a couple of times. This matter analysis unfortunately didn’t find any papers prospectively which actually analysed the problem about the use of parachutes for trauma prevention, and it becomes very clear that this parachute or not, in this case is really right on the side of the obvious. I don’t think there is a clinical trial needed.
Top of page
Blunt scalpels you can do a bit more damage with a sharp scalpel, so that I’ve put it a bit more to the right, but again that’s not something we would attest. However many drugs which are injected systemically they are really complex, and the complexity there obviously means we need to test them in phase three trials as we do. And most of our trial culture really has developed on the back of these complexities. Our question is obviously IMRT, IGRT where does that sit on this? And I’m not suggesting it’s sitting there, I sort of actually there was an animation in the Macintosh environment, and it sort of moves forwards and back, it’s because I don’t think we are really perfectly clear about that.

So the question is, is it ethical to run a clinical trial between a four field box as we’ve seen, and an IMRT treatment as you see down there, and obviously if you just show the pictures to someone (unclear word) or the ability of the investigator to say to a patient, "Well I honestly don’t know if you’re better off there or there, so we needed to test it", it’s something which is obviously in doubt. However we need to also ask ourselves why should it not work? Isn’t it obvious it should work and I should as a physicist say "that’s obvious". But there are hidden problems, and I think we need before we introduce this technology, be very careful that cancer is not one disease, it’s not one cancer patient.

There are many, many different ones and what is good for one patient is not necessarily good for another. Can it be implemented widely? Is it something which can be done in a small country clinic, and that’s really sort of what we needed to come to about. The technology certainly is not the hindrance in this case. The technology in many of the modern regional centres is at least as good and advanced as in other centres. Is it too complex, too expensive? And I think just that simple example can we actually trust what we see, because so far I’ve just shown you pictures, and the pictures show you a nice dose distribution, which obviously looks good. But if we actually test that and that’s - I think Ivan is going to talk a bit more about that this afternoon in terms of the ACDS, can we verify that what we think we give, what thing we think it looks so great, is actually, like that’s a quality assurance is important, and that’s an Anthropomorphic phantom which the RPC which is a group in Houston in Texas, centres around all over the US and North America, and to some Australian centres. And we are asked to plan that phantom and this is a plan, you can see the dose distribution here in the phantom sort of curving around things, giving two different dose levels, and then they measure that and you can this is then the measurement compared to the plan, and they compare then, is what you see really what you get? Independently audited.

These are the results and they were published five years after the event, so these are really old results, but I think they are very, very sobering. Oh you can also say, "That's it's good - half full, half empty". You can say "Oh three quarters of the radiation therapy centres did a great job". Well that also means one quarter didn’t, and that’s really not very reassuring indeed. These figures were now published because they actually have been remedied and through this audit process, through application of auditing and standards, these numbers look now much better. But I think certainly with the introduction of this new technology, this is a sobering thought.

There are also other more hidden problems with IMRT, which you might not appreciate to start with and which are very difficult to communicate to patients, since we use all these small radiation beams, we actually have the beam on for much longer, which means we have more leakage radiation, and we sort of bath the whole patient a bit in radiation doing that. We give the dose to the target much better, we spare the critical structure’s much better, but there is an increased dose.

Now that is something which is worrying if we get good, if radiotherapy cures patients, secondary cancers become a problem because 20 years down the track patients may still have side effects and develop new cancers due to the fact that they had radiotherapy, and you can see from these numbers in a theoretical calculation of a rather influential paper five years ago, that whatever you do, the rate of secondary cancers using IMRT sort of doubles in these scenarios.

Again this is five years old. I don’t think this is true anymore. Papers like that have the effect of shaking us up and we are doing now IMRT differently, taking this into account. We give less monitor units, optimise not just in terms of the number of, in terms of the dose distribution, but also reduce the monitor units, make it more robust, spare organs which are at risk of secondary cancer. But these are problems.

So are we actually on the other side? Is it now ethical not to run a clinical trial? And I think that is really something where the jury is out, and I think we have to do and think about this really on a case by case basis.

Now just two more minutes, in terms of trying - what do we need to do in terms of introducing new technology? And if you look at that list here, then a lot of these things are actually part of clinical trials as well, and obviously clinical trials can be a tool, but not the sole tool to introduce new technology, and help introduce new technology. And I think it always starts with "Is it actually needed?" The manufacturer will tell you "Yes it is needed", but it may not be something which is actually then reflected locally. I think, being a physicist here obviously, you could expect that multidisciplinary teams are important, and I think this meeting as such reflects that really very nicely. This bicycle by the way is the Tripartite bicycle as you can see. I’m not saying that this is reflecting the speed with which the Tripartite is sort of moving ahead, but it certainly is a good sign. Learning is important, teaching, training, and we are not working in isolation and training things like this workshop, it’s also essential.

I put this slide up here just as an example of training in this case about cone beam CT for an adaptive bladder protocol, also to say that I have to give my apologies for the rest of the day, I have to go up to Darwin. We are running a summer school on functional imaging from our college. So this is a bit not a hit and run talk, but it’s a talk and run. But I’ll be staying here for the tea break, so if you have any questions you’re welcome.

And I end up with this proceeding of caution, this is another very famous paper from a Belgian group where they looked at modern technology, and I’m not sure if you can actually read that. I can’t even read it myself. It is about conformal arc radiotherapy, which they introduced in a small in-house clinical trial, and they found yes indeed, dose escalation was good and the higher the stage the worse the outcome. But they also found surprisingly that once they introduced image guidance and fiducial marker, the outcome got worse. And that's sort of the highlighted areas there, which again are probably difficult to see, so this is really quite unexpected.

Why would things get worse with image guidance? And the reason is really that due to the image guidance they actually reduced the margins by half, thinking "Oh we used the old formulas to calculate the margins, did that and found that oh they had gone too far. So proceeding with caution is important, and in summary I think it’s fair to say that radiotherapy does rely on technology. There are good reasons to believe, hopefully after this presentation, not just for physicists, to believe that it’s actually making a difference, and there are a wide variety of approaches available.
I think this picture sort of shows exactly what we have ahead. It's a road which sort of leads into a good future, it’s a hilly road, it goes up and down as you can see, but there’s a double line, there’s no turning around. So I’ll leave you with that, and ...

(applause)
Top of page
Norman Swan:

I mean that just that trial you quoted at the end was actually the point I was going to make first. Given you don’t know how far the tumour has actually spread in local tissues, how do you know what is the geometry to, you know, how far do you go, and surely that’s a huge risk in IMRT that you actually don’t spray too widely, or you should be spraying more widely than you are.

Tomas Kron:

I think that that’s certainly a very valid point, I’ll leave that to the clinical colleagues, thank goodness they draw the lines. But the modern imaging gives us certainly a lot of ideas. What I want to point out about the paper ...

Norman Swan:

But it can tell you whether there’s cancer cells a centimetre away from your target.

Tomas Kron:

No. There are increasing number of studies for example, there’s a very influential paper on breast cancer where they actually took mastectomy specimens and pathologically analysed the actual visible tumour and then looked at the tumour cells away from that. There are more papers, more studies like that coming out, and I think they will influence it. We also have chemotherapy and other tools available to address these two different issues. I agree though I think at this point in time we need to tread carefully, the evidence - so far at least the reported evidence is positive, and I put that paper in particularly because I think it will be absolutely essential, if there are negative experiences to report that as well. I think this is a fantastic example of good clinical practice despite the fact that that is that ...

Norman Swan:

Do you think so, are you sure?

Tomas Kron:

... that they published it.

Norman Swan:

No, no, no. So them publishing it, I’m now going back to IMRT, IGRT. The question is, it sounds great, but there, you know, there is a history in radiation oncology of new technologies, you know, the proton beam therapy, you know, and big companies trying to flog very expensive technologies. When you actually balance the equation with increased complexity, you’re going to get more errors, but and you’re not quite sure, there haven’t been many randomised trials here, are we absolutely sure this is becoming state of the art, this is best practice?

Tomas Kron:

I think I - I don’t think we can ever say we are absolutely sure, because as I illustrated with a case of late, of secondary cancer induction, that is something which happens 20 years down the track. So I don’t think we will have any clinical evidence about that in the foreseeable future. If you ask me though, if I have a head and neck cancer if I want to have IMRT or not, I would run out screaming if I don’t get IMRT. And the rationale is also there is now sufficient clinical evidence for some areas to justify that.

Norman Swan:

Any comments or questions please come up to the microphone, John?

Question (John):

Thanks very much Tomas. You asked the question is it ethical to run a clinical trial. I think it is ethical to run a clinical trial. Would I go on it? Probably not. Radiotherapy to me, through all my treatment it was the most scariest. It is an intangible, you can’t see it, you can’t smell it, you can’t see it. With a drug, you know, there’s - you have the chemo nurse, they’re protected, and it doesn’t feel too good, but it is there. And so I think that’s going to be one of the real issues for RTs and also for patients, because there is all this new technology which is giving better outcomes, but as we get further into the technology area, how is it going to be proven to be actually the best?

Tomas Kron:

I think that that’s a very, very good question. Clinical trials though have another limitation in this context. As we get better and better in terms of diagnosing cancers and identifying cancers and their molecular profile, running a randomised trial and say 100 patients there, 100 patients there, they are identical, and having very few cases just let’s say five patients in one arm makes a 5% difference in clinical outcome. That is a huge number, 5% in survival curves going apart, everyone goes "Yes, that is fantastic". There are very few patients, and they are so different and we know about that difference. So it will become increasingly difficult to actually run phase three randomised clinical trials because of the additional knowledge we have, and I think databases, registries, are sort of ideas which are currently thrown around to inform patients still about outcomes.

Norman Swan:

But you’re going to have the classic problem of a device industry which is enormous pressure of industry for their latest gizmo and, you know, everybody wants one, they’ve been to, you know.

Tomas Kron:

I’m the first one, if my credit card unfortunately as my wife tells me is ...

Norman Swan:

If you could. (laughs) Maxed out long ago.

Tomas Kron:

Yeah, maxed out long ago, but okay.

Norman Swan:

Tomas thank you very much.

(applause)

Top of page

prev pageprev page| TOC |next page