What's the Future for Precision Radiotherapy?

Prof Karol Sikora


May 25, 2021

This transcript has been edited for clarity.

Hello, it's Professor Karol Sikora here. And I'm talking today about precision radiotherapy. Just like our medical oncology colleagues talk about precision chemotherapy: using molecular predictions of response, using biomarkers to determine whether a response is more or less likely in a particular case, can we do the same for radiotherapy? Where's the precision going in radiotherapy?

Well, radiotherapy is different from chemotherapy in that it's localised. And so precision means greater accuracy, greater delivery to the tumour, and sparing the organs at risk that are often critical.

If we look at the three things that have changed over the last 40 years that I've been a consultant in the NHS, the first thing is imaging. The precision of imaging is just fantastic compared with even 20 years ago.

And then there's IT. Treatment planning systems, which are the core of delivering precision in radiotherapy, have improved beyond any recognition, again, just over the last 10 to 15 years.

Finally, we're at the verge of something new: artificial intelligence coming in to the optimisation of radiotherapy. There's no doubt that AI is going to have a huge role to play. We talk about auto planning now, but it's going to be more than that. It's going to take over the way in which radiotherapy is actually delivered, because after all, a computer can do a million plans, if you like in an hour, even less, whereas we can't, we can't possibly take all that information.

Computers can learn the rules to deliver optimal planning, the most precise planning possible, with all the sort of nuances of organs at risk with different values in terms of sensitivity to short and long-term effects of radiation.

I began in an era doing radiotherapy when there were no computers. So I've seen it come and go right the way through. And I think at the moment, the problems we've got, that are still clumsy [include] the first one, the fusion of images - we're just not good at fusing the images, they still come in packets of different sizes, a lot better, but we've still got a way to go.

And once we get image fusion made perfectly, we should have better ways of using AI to identify the cancer.

You know, when you look at a glioma, for example, it is very woolly. If you get 100 oncologists to outline a glioma, they'll come back with some differences. There's always a few outliers, but most people are roughly in the middle. But none of us really know where the cancer ends, and the normal tissue begins. And that's a pretty critical decision in terms of outlining and planning.

The marking of organs at risk is a bit easier and already AI can do that. But in certain parts of the body, there's almost too many organs at risk, the base of the brain being classic. Too many organs at risk to actually set the rules in a convenient way, algorithms for automation.

But auto-planning systems are really the beginning, and we're just at the hinterland of increasing the precision of our plans by doing that.

If we look at the three technologies that are driving forward the precision of radiotherapy, there's SABR, stereotactic ablative body radiotherapy, proton beam therapy - and today I'm giving this talk from the control room of the proton beam unit at Newport, South Wales, in the Rutherford centre - and then the third one is the MR-LINAC, and whether the combination of real time imaging with therapy can actually improve the precision.


So let's look at these three modalities. They all have different uses. SABR is attractive, especially for oligo metastases. Patients with oligo metastases have a relatively poor prognosis, a matter of months in some cases, certainly unlikely to be more than a couple of years, and therefore having something very quick that can keep them out of hospital, and keep their symptoms at bay, and keep them responding to chemotherapy in other areas is quite attractive.

SABR has had a major push within Britain's NHS. The equipment to deliver SABR has been present for some years now, but it's not being used, for a variety of reasons, mainly training, governance, and so on. Now, it's going to be. By March of next year, there will be a release of a whole series of SABR units around the country, and that will improve the quality for patients, some of whom are nearing the end of their lives, others will have remarkable responses to oligo-metastatic disease. There are some circumstances where near SABR, if you like, in other words, reducing the fractionation to nearly SABR doses, is going to happen. I think we're going to see a blending of conventional fractionation to SABR-type fractionation, where the doses are high, and therefore the precision required with each dose delivery is also high.

Proton Beam Therapy

Then there's proton beam therapy, and that's controversial. What's the controversy? Well, it depends. The private centres such as the Rutherford here, obviously have to market their wares. The NHS centres have to restrict their wares. That's part of the rationing process of public health systems. The real answer is somewhere in the middle and getting the two groups together is clearly the way forward.

This machine here at Rutherford costs £15 million, the one at Christie and the one that's being installed at UCH, that one costs £125 million, so totally different costs.

And do you get value for your money? Probably not in reality, because although this machine here is a single gantry, and the Christie one is a four gantry machine, you could only treat one patient at once, and so there's a limitation on numbers going through and throughput.

Proton therapy is attractive, because the Bragg peak allows you to shape the contour of the high dose delivery system to that area more flexibly. Estimates around Europe suggest that 10% of radical radiotherapy would be optimally delivered by protons, and that's the sort of figure we're working for. What does that mean, in Britain? It means something like 20 centres around the country. At the moment, we have essentially four: three with Rutherford, one with the NHS. And then two more possibly: University College, which I've mentioned, with the NHS, and Advanced Oncotherapy (AVO) in Harley Street, which is a different technology to produce protons of high energy using a series of linear accelerators. Whether it's capable of producing adequate energy, the sort of 230 million volts that we have here, with a cyclotron production, remains to be seen.



The last of the technologies to deliver precision is MR-LINAC. And intrinsically that seems attractive. The question is, how often can you improve things by having an MR image on the same machine as you're going to deliver the high energy beam. And that really remains to be seen.

The similarity of all three technologies is they rely heavily on computer power to ensure precision. Without the computer power, they just wouldn't work.

The thing that's going to be more and more impressive as we go through the next decades is the way in which computers drive everything. And there's no doubt, predicting the dose, personalising the dose, what I've talked about so far, is only personalising in a geometric sense, for which you have to know where the cancer is, where the normal tissue [and] critical structures are. And then you can develop the best plan using some sort of artificial intelligence.

But the step after that is dosimetric optimisation. Mrs Jones and Mrs Smith both have breast cancer, they're going to have post-operative radiotherapy, do they get the same dose? At the moment they do. Almost certainly they shouldn't be. There should be differences, differences based on genomics, not just the cancer, but also their genetic structure, and their family structure.

So, as we move into the next decade, I can predict we'll come out of almost perfect geometric precision into an era where we'll get dosimetric precision, and when you look at the fractionation schemes, and the dose ranges given, they'll differ from patient to patient based on logical science, and not the heralded time of the literature, and how many fractions that sound convenient, that oncologists love to talk about.

So, the future is great, precision is the word for cancer, both with chemotherapy and with radiotherapy. Any points you'd like to make, please contact us. Thank you very much.

You can follow Prof Karol Sikora on Twitter


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: