Dejan Trbojevic, Stephen Brooks, François Méot and Nicholaos Tsoupas,
Brookhaven National Laboratory, Upton, NY, USA.
William Lou, Cornell University, CLASSE, Ithaca, NY, USA.
1 Work performed under Contract Number DE-AC02-98CH10886.
Worldwide, cancer is the second most common cause of death after cardiovascular diseases. There are many types of cancer and many ways of treating the disease, mostly through a combination of methods such as surgery, chemotherapy, immunotherapy, targeted therapy and/or radiation therapy. Conventional radiation therapy relies on X-rays, usually produced from accelerated electrons and focused onto the tumor to destroy the cancerous cells. Techniques such as Intensity Modulated Radiation Therapy (IMRT), 3D-CRT three dimensional conformal radiotherapy, or Image-Guided Radiation Therapy (IGRT) direct the beam from a variety of angles to maximise the effect on the tumor while avoiding surrounding healthy tissues. Success is seldom 100% and the possibility of side effects is a serious risk.
The advantages of hadron therapy - using protons or ions such as carbon instead of X-rays - has been recognised for some time. A beam of protons can penetrate tissue with very little diffusion. The particles deposit their maximum energy almost immediately before stopping. This property (the “Bragg peak”) allows precise location of the specific region that needs to be irradiated and damage to the surrounding healthy tissue can be minimised. This is especially critical in the case of children. Because of the success of the treatment, the number of hadron therapy centers has been growing strongly in recent years, and a total of about 200,000 patients have treated to date.
The importance of hadron therapy has been recognized by Governments worldwide. A recent workshop on Ion Beam Therapy  concluded by emphasizing the need for new technological developments particularly with a view to reducing the costs involved in further hadron therapy expansion. A major concern is the gantries used for the beam delivery systems which are huge and exceedingly heavy. They need to be far less massive and more compact. Ideally a technology is needed that can provide for rapid scanning (one or two seconds) of the beam over the tumor volume in three dimensions.
This is where the FFA linear gradient magnet comes into its own. The momentum acceptance of such a magnet can be very large, of the order , which gives a momentum range corresponding to the proton kinetic energy range, 65-250 MeV, required for patient proton radiation therapy. The large momentum acceptance allows fast energy change without any magnetic field variation. The patient treatment time can be made shorter, and the operation is simplified because there is no need to change the magnetic fields. There are corresponding advantages in lower power consumption (operating costs), and the simplified design has a considerable effect in reducing the gantry cost. The reduction in the weight and size of the magnets allows lighter, rotating structures.
For maximum effectiveness in focusing on the tumor while avoiding sensitive organs like the spine or the cardiovascular system, the proton gantry needs to be isocentric and able to rotate around the patient, allowing all possible angles of incidence. The design presented here assumes that at the entrance to the gantry protons within the required energy range (65-250 MeV) enter with zero offsets. Since the patient will usually be lying flat, the structure is such that the beam is bent to one side (left or right) or above the patient before being focused in onto the treatment area. Transverse spot scanning requires control over transverse distances of at least 10 cm. Previous studies  showed how to merge multiple energy orbits from an FFA structure into a single orbit. This single orbit can be obtained by introducing a triplet of combined function magnets in front of the merging point, as shown in Fig. 1.
The gantry itself is shown in Fig. 2 and comprises several modules. It rotates about the horizontal () axis. The beam enters from the left and exits at the vertical section at m. The first module, M1, the section to the left as far as coordinates , accepts all proton energies and is achromatic. Each of the different energy orbits reaches its maximum offset at the middle of the module, where the slopes of the optical betratron and dispersion functions are zero. This module is mirror symmetric about its mid-point, so all orbits are merged into a single orbit on exit with the same betatron and dispersion values as at the beginning.
The magnets bend in the opposite direction in the second module, M2, with the first half being the reverse of the second half of M1. This is followed by a matched FFA gradient arc, M3, which bends the beam into a perpendicular direction onto the patient, with the same repetitive structure of the FFA cells. The final section of the gantry, M4, merges the different energy orbits onto the same final path to the patient. The distance of the last magnet from the patient is 1.435 m, where the horizontal and vertical betatron functions have values as close as possible to 1 m. A patent  covers the design of the transverse scanning system, which uses two magnets, the first following the triplet magnets and the second above the patient, separated by a distance of about 1 m.
Figure 3 shows the layout of the magnets, the orbits (upper part) and the betatron and dispersion optical functions (middle and lower parts). The curves are for protons with energies from 65 MeV to 250 MeV in steps corresponding to momentum variations with respect to the central energy MeV of p/p= -0.38, -0.2, -0.1, 0.0, 0.1, 0.2 and 0.28. Similar details of the optical properties for the whole gantry are shown in Fig. 4.
The gantry magnets are Halbach type  built by the KYMA Italian-Slovenian company . Combined function magnets are chosen with most of the bending coming from the defocusing magnet (shown in Fig. 5a). The focusing magnet is effectively a regular focusing quadrupole as it has very small bending angle. The technological development benefits from experience of building permanent magnets for the FFA gradient test at BNL’s Accelerator Test Facility (ATF). Here electrons at four different energies were succesfully transported through an FFA gradient structure . BNL subsequently built 220 combined function magnets for the Cornell University and Brookhaven National Laboratory Electron Test Accelerator, CBETA .
Special assembly devices, shown in Fig. 5b, were built by KYMA to construct the gantry magnets. The magnets are very similar to the BNL Halbach permanent magnets but with the aluminium wedges replaced by 16 separate NdFeBr wedges.
In summary, there are significant advantages in using protons and ions in cancer therapy because of the confinement and consequent safety aspects afforded by the “Bragg peak”. However, for hadron therapy centers to be more widespread requires cheaper facilities, achieved principally through a reduction in the size, weight and therefore cost of the gantries. In this respect, the use of FFA-type linear gradient magnets offers several advantages:
(i) The magnetic field is fixed throughout the whole gantry, and the only variable aspect comes from the scanning dipoles at the end of the gantry. Transverse scanning is significantly slower as longitudinal energy scanning occurs automatically for each radial position.
(ii) The magnets are made of permanent magnet material (NdFeBr) and are small and light.
(iii) The large momentum acceptance of FFA’s means that the energy range needed for proton therapy of 65-250 MeV, can readily be met.
(iv) A rapid energy change can be carried out without need for magnetic field variation.
(v) The patient treatment time is shorter as there is no need to change settings of the gantry magnets.
(vi) There is a significant reduction in the electricity power bill as the magnets are made of permanent magnet material.
(vii) The overall cost is much reduced and the overall weight of the gantry is notably smaller.
(viii) As the rotating structure is so much lighter, it is easier to construct with consequent cost reduction.
(ix) The design allows very easy operation and no tuning is necessary.
We acknowlege the contribution of the KYMA company in producing combined function magnets of superb quality.
 Summary report on: Workshop on Ion Beam Therapy, Bethesda, Maryland, USA, January 9-11,
 D. Trbojevic, “Proton Therapy Gantry,” presented at the International Fixed Field Alternating gradient Workhop (FFA’18) in Kyoto, Kyoto University, 10-14 September 2018, https://indico.rcnp.osaka-u.ac.jp/event/1143/contributions/1219/
 D. Trbojevic, “Scanning System for particle cancer therapy,” United States Patent US 9,095,705
B2, August 4, 2016.
 K. Halbach, “Application of permanent magnets in accelerators and electron storage rings”, J. Appl. Phys., 1985, 57,(8), pp. 3605–3608
 “KYMA - High Quality insertion devices for light sources”, http://kyma.elettra.eu/
 Stephen Brooks, “Successful Test of Small-Scale Accelerator with Big Potential Impacts for Science and Medicine,” https://phys.org/news/2017-08-fixed-field-multiple-particle-wide-range.html
 CBETA Design Report, Cornell-BNL Test Accelerator: https://arxiv.org/abs/1706.04245