Since protons were first used to treat hospital cancer patients in the early 1990s, around 100,000 people have benefited from this alternative form of radiation therapy. While the X-rays used in conventional radiotherapy fully penetrate a patient and dump their maximum dose just after entering the body, protons deposit their biggest dose at a specific, energy-dependent depth and travel no further. This characteristic “Bragg peak” allows protons’ energy to be concentrated at the location of a tumour, so inflicting maximum damage to cancerous tissue while sparing surrounding cells.

These virtues, however, come at a price. Conventional radiotherapy generally involves accelerating electrons in a linear accelerator (linac) and colliding them with a tungsten target to generate X-rays. The apparatus, which is roughly 2 m long, is incorporated into a rotating gantry that allows X-rays to enter a patient’s body from a range of angles. But protons are much heavier than electrons, and require larger accelerators that generally serve multiple treatment rooms with 10 m-diameter gantries.

The greater size means that proton treatments typically cost about twice as much as those using X-rays. So while the therapy is becoming more popular – new treatment centres being built across Asia, Europe and the US, many of which are due to start up in the next year or two, will double the roughly 70 operating today – they are still only used in about 1% of all radiotherapy treatments. As such, says Bill Hansen, a marketing director at Varian Medical Systems in Palo Alto, California, many manufacturers of proton-therapy facilities are “struggling” to make a profit.

Developing more compact systems could change the economics of proton therapy for the better. Rather than having a large accelerator supply multiple rooms, the idea is to use a relatively small accelerator to serve a single room – so reducing civil engineering, component and maintenance costs. “Existing systems are large, expensive and very complex,” says Hansen. “Either the reimbursement for treating each patient has to be higher or the cost of treatment has to come down.”

Less is more

The workhorse of most proton-therapy centres is the cyclotron. This device consists of a pair of very powerful circular magnets placed above and below two semicircular electrodes with a gap between them. Protons emerging from an ion source in the centre of the device are forced to follow a spiral trajectory by the magnetic fields and gain a boost in speed at each turn as they cross the gap, thanks to an oscillating electric field. When the particles reach the edge of the magnetic field they leave the device with a very high energy – typically up to 250 MeV.

One way of reducing the cost of a cyclotron is to make the magnets with superconducting copper coils rather than ordinary ones. This means that cyclotrons can generate much greater magnetic fields and so bend the paths of the accelerating protons more tightly. Although the magnets themselves may be more expensive, boosting the field reduces the size of the cyclotron needed to generate protons of a given energy, thus lowering the total cost considerably.

The Belgian firm Ion Beam Applications (IBA), the world’s largest producer of proton-therapy systems, has adopted this approach, using superconducting technology to reduce the diameter of its accelerators from 4.3 to 2.5 m and to slash their weight from 200 to 45 tonnes. IBA chief research officer Yves Jongen says that, in principle, superconducting magnets could also be used to guide protons along the beamline, which runs from the accelerator up and over the gantry to the point of treatment. However, because the proton beam is scanned back and forth across the tumour during therapy, the magnetic field in the beamline needs to vary rapidly. This, he observes, removes some of the advantage of superconducting magnets: while superconductors incur no energy loss for a DC current, they do for an AC one.

As such, he says, IBA decided to use ordinary resistive magnets for the beamline, and to instead shrink the gantry (from 10 m to 7 m) by changing how the beam is scanned. To obtain the desired beam shape, the proton beam needs to be scanned at some distance from the patient, and in previous systems this scanning took place after the last of the bending magnets in the gantry. The new system instead scans the beam before it passes through the last magnet, making it possible to bring the magnets closer. As Jongen explains, this modification required 3D computer modelling to establish exactly what shape of magnetic field would be needed to bend the moving beam.

IBA completed its new design in 2012, and Jongen says that after testing it, the company has since installed 10 such systems, including five that are now treating patients. While the company’s large-scale facility costs between €40m and €60m, depending on the number of treatment rooms, the new system instead comes in at slightly below €20m. This means that one-room systems cost “slightly more” per room, but one room can still treat more than 300 patients a year – which is, he says, enough for most hospitals.

Another approach to reducing system size is to mount the accelerator directly onto the gantry. This strategy has enabled a Massachusetts-based company, Mevion Medical Systems, to develop a synchrocyclotron measuring 1.8 m across and weighing just 20 tonnes. Mevion’s first compact device has been operating since 2010, and it has since been updated to include a pencil-beam scanning system that allows protons to be delivered more quickly and more precisely, according to a company press release. However, placing an accelerator directly on the gantry comes with a couple of downsides. For one thing, the gantry can’t turn 360 degrees. According to Jongen, it also means that patients are exposed to more neutrons than they would be otherwise. These neutrons, generated when fast protons are lost in the device, irradiate healthy organs outside the tumour target.

On the straight and narrow

In the quest to improve the economics of proton therapy, however, not all companies are solely focused on size. Hansen at Varian says that his company is working to reduce the size of its gantries by making the beamline magnets both lighter and more powerful, possibly using superconductors. But a more important factor for them, he explains, is to boost power – in other words, to increase the dose rate of high-energy protons impinging on the tumour.

Hansen points out that patients must hold their breath during therapy on certain organs, particularly their lungs, to avoid the tumour moving in and out of the proton beam. Delivering the required dose as quickly as possible therefore limits the amount of time the patient needs to hold their breath, particularly if the tumour is big. This is less of a problem in traditional radiotherapy, given that X-rays do not target specific organs so precisely and therefore cause less variation in the received dose if the organ moves. In proton therapy, though, Hansen believes that a higher dose rate will improve results and therefore reduce the number of treatment sessions that are needed – lowering costs in the process. “That is really going to be the solution to the industry problem,” he maintains.

The problem Hansen refers to has been caused in part by doubts from health authorities and insurance companies about whether proton therapy really is superior to conventional radiotherapy. Harald Paganetti, director of physics research at the Massachusetts General Hospital in the US, says protons are definitely better in some cases – such as children with brain cancer, who can suffer a drop in IQ if healthy tissue is irradiated. But otherwise, he explains, it is often unclear whether dumping less energy outside the tumour translates into a clinical gain. In his 2017 ebook Proton Beam Therapy (from IOP Publishing, which also publishes Physics World), Paganetti observed that “we often do not know the importance of low dose radiation with respect to serious toxicities”, and Jongen adds that medical evidence on such topics is often slow to accumulate. “If you treat someone for cancer you have to follow them for five years to prove that you have effectively got rid of the tumour,” he says. “And then to have good statistics with a decent number of patients takes time”.

Undeterred, one British company is developing a novel system designed both to cut the costs and to improve the performance of proton therapy. So far, all proton-therapy systems have relied on circular accelerators – also including synchrotrons, which have been developed by Japanese multinational Hitachi. In contrast, Jonathan Farr, director of medical physics at London-based Advanced Oncotherapy, explains that his company plans to ramp up proton speeds using a linac.

In order to scan a tumour by reducing the depth of the Bragg peak in steps, proton energy must be varied rapidly. In a cyclotron, where protons are emitted with a fixed, maximum energy, this is done by placing lightweight absorbers of varying thickness in the beam path. A linac, in contrast, consists of a series of accelerating modules that can be individually switched on or off – a purely electronic process that Farr says in future could be carried out at up to 200 times a second. At that speed, he says, even lung tumours could be irradiated without requiring patients to hold their breath. Farr also argues that the modular nature of the accelerator means it should be cheaper to manufacture, assemble and maintain. Plus, he claims, the device should in future generate narrower beams than those from circular accelerators – yielding a radius of less than a millimetre, as opposed to 3 or 4 mm – allowing tumours to be more accurately targeted.

The technology behind Advanced Oncology’s Linac for Image Guided Hadron Therapy (LIGHT) was developed at the CERN particle physics laboratory in Geneva, Switzerland, and it relies on the world’s highest frequency RF quadrupole to initially accelerate protons in a very short space and therefore at low cost. In 2019 a prototype device will be moved from CERN to a thick-walled bunker at the Daresbury Laboratory in the UK for full-energy testing. The plan then is to start treating patients in a specially designed facility under an existing town house in Harley Street, central London, in 2020.

Farr says testing the machine “is going as expected” but admits that he and his colleagues still face some fairly stiff challenges, particularly when it comes to putting the components together on site – which must be done with millimetre precision over a distance of 24 m. “It is engineering. It is not science,” he says. “We don’t have to discover anything new, we just have to do the work. But we have to do that at a tight level.”

Farr believes that technical success will lead to commercial interest. Although he won’t put hard numbers on it, he is confident that the linac technology can “significantly reduce the entire project cost” for proton therapy. “Everybody in proton therapy is chasing the goal to get it down close to the cost of traditional treatment,” he says. “We may not get to that, but we hope to get close.”

• Enjoy the rest of the 2018 Physics World Focus on Instruments and Vacuum in our digital magazine or via the Physics World app for any iOS or Android smartphone or tablet.