PROTON THERAPY
The Promise of Precision

Dan T L Jones

Head : Medical Radiation Group
National Accelerator Centre, Old Cape Road
P O Box 72, Faure 7131, South Africa

email : jones@nac.ac.za

1. Cancer Incidence and Treatment

Cancer can broadly be defined as the uncontrolled growth and proliferation of groups of cells, the triggering of which is not yet fully understood. In industrialised societies about 30% of people suffer from cancer and about half of these die from the disease. More than half of all cancer sufferers receive radiation therapy (possibly in conjunction with surgery and chemotherapy). The prognosis in individual cases varies greatly and depends on tumour type, stage of diagnosis, environment, lifestyle and general health of the patient, etc. A patient who survives for 5 years after commencement of treatment without further symptoms is regarded as having been cured. The overall 5-year surival rate of all treated cancer sufferers is about 45% [1].

Cells from the primary tumour can metastasize (spread to other parts of the body) and about 30% of all cancer patients have metastases at diagnosis. Radiotherapy and surgery are both localised forms of treatment. They are used, alone or in combination, to treat the primary tumour and are responsible for about 90% of cancer cures (50% surgery, 40% radiotherapy alone or combined with surgery). In addition radiotherapy, even at moderate doses, is particularly effective for palliative treatment of metastases (i.e. for pain relief.) Chemotherapy is used to treat metastases and the 5-year survival rate is about 5% (about 10% of all cures) [1].

From the above it is clear that even modest improvements in cancer treatment will benefit a large number of people. A very important factor to also consider when assessing the cost-benefit of cancer treatment is the cost of not curing a patient. This can be very high and may involve risky salvage surgery, chronic health care, etc. The cost may be as much as 4-5 times the cost of curing a patient.

The objective of radiation therapy is to maximise the effect of the radiation on the target lesion and to minimise the effect on surrounding normal tissue. This is done by increasing either the physical dose differential or the biological effect differential between the target and normal tissue.

2. Proton Therapy : Technical Aspects

Robert Wilson first proposed the use of protons and heavier ions for therapy in 1946 [2]. The pioneering experimental work of Tobias and his associates at Berkeley CA a few years later confirmed Wilson's predictions [3]. Between 1954 and 1957 30 patients were treated on the 184 inch synchrocyclotron at Berkeley [4]. The machine was upgraded and the energy became too high for proton therapy and from 1957 alpha particles were used for therapy. Proton therapy continued in the USA (Harvard University, Cambridge MA) in 1961; began in Europe (Uppsala, Sweden) in 1957; Russia (Dubna) in 1967; Japan (Chiba) in 1979 and in South Africa (Faure) in 1993.

The main rationale for using proton beams lies in their physical selectivity (i.e. the ability to conform the dose to the target volume). Beams of these particles have unique dose distributions (Figs. 1, 2) which exhibit a relatively flat entrance dose region (plateau) followed by a sharp dose peak (the Bragg peak) in which the particles lose most of their energy. The dose distributions have sharp distal and lateral dose fall-offs, which are illustrated in the case of a 200 MeV proton beam in Fig 2. There is no radiation beyond range end. Ideally to adequately treat all lesions a range of greater than 26 cm in tissue (corresponding to a beam energy of 200 MeV) is required. The physical properties of proton beams are best utilized for eradication of well-delineated lesions close to critical structures, which in principle can be relatively easily avoided.

 

Fig. 1.   Depth dose curve for a 200 MeV proton therapy beam [5, 6]
compared with other typical radiotherapy beams.

 

Fig. 2.   Isodose (10%-90%) curves for a 10 cm diameter
monoenergetic 200 MeV proton therapy beam [6,7]

The biological effects of protons are not very different from those of conventional radiations (photons, electrons) and treatment protocols can be based directly on more than 100 years' experience with these latter radiations. To take full advantage of the potential of proton therapy requires, inter alia, an accurate beam delivery system, precise tumour and critical structure localization, accurate and reproducible patient set-up, accurate three dimensional treatment planning, including compensation for tissue heterodensities and allowance for organ movement during treatment sessions. If all these requirements cannot be met, the quality of the treatment could be compromised.

   Beam Delivery

Because the maximum dose occurs at the end of the range isocentric beam delivery (irradiation from any direction) is not quite as important as for x-ray and for neutron therapy, but is nevertheless desirable for proton therapy and three types of gantries, all with very different design criteria, have been built to date: the corkscrew gantry (Loma Linda University, USA [8]) the compact eccentric gantry (Paul Scherrer Institure (PSI), Villigen, Switzerland [9]) and the conventional 90 gooseneck gantry (Northeast Proton Therapy Center, Boston, USA and National Cancer Center, Kashiwa, Japan [10]). Non-orthogonal fixed beam arrangements (Hyogo Ion Beam Medical Center, Japan [11]) and at the National Accelerator Centre, South Africa [12] are also being designed. Together with a versatile patient support system and sophisticated beam delivery, such facilities provide a viable cost-effective alternative to isocentric facilities, albeit with limited applications.

The dimensions of high-energy proton beams emanating from an accelerator are quite small (10 mm in diameter) and the Bragg peak of monoenergetic beams is narrow (~20 mm full width at half maximum) and so the beam has to be modified to deposit the required dose over the whole 3-dimensional target volume. Either passive scattering (broad beams) or dynamic beam scanning (pencil beams) can be used. Modulating both broad and pencil beams in depth involves the superposition of suitably weighted proton beams of different energies (ranges) resulting in a uniform dose over the target region (Fig. 3).

 

Fig. 3.  Depth dose curve for a monoenergetic 190 MeV proton beam (thick line)
showing the Bragg peak at the end of the range.
The superposition of suitably weighted proton beams of different energies (ranges)
results in a spread-out-Bragg-peak (SOBP) which provides
a uniform dose over the target region

   Passive Scattering

The Bragg peak is modulated in depth over the longitudinal extent of the target volume by varying the energy (range) of the incident protons using a variable thickness rotating "propeller" [2, 14] (Fig. 3) or a ridge filter [15, 16]. In addition a field-specific absorber (compensator), shaped in 3-dimensions, can be used to tailor the range of the protons to conform the high-dose contour to the distal surface of the target volume. The problem with this passive technique is that there is constant modulation (neglecting inhomogenities) of the Bragg peak over the lateral extent of the target volume and no proximal surface dose conformation can be achieved (Fig. 4). Therefore, while the situation is still greatly superior to what can be realized in conventional therapy, some of the high dose region still falls in healthy tissue proximal to the target.

   Beam Scanning

Instead of using passive techniques to "paint" the target volume with a uniform high-dose, it is possible (since protons are positively charged particles) to magnetically deflect the elemental pencil proton beam. The dose is deposited by scanning the dose "spot" (Bragg peak) of the pencil beam in all three dimensions inside the target volume [17]. Through the superposition of a very large number of such individual elemental dose distributions conformation of the dose to the target volume can be achieved. Scanning can be done either in a continuous (raster scan) or discrete (spot scan) fashion. The beam can be scanned in the two dimensions perpendicular to the beam axis by two orthogonal magnets.

Alternatively, one magnet can be used to scan the beam in a strip in one dimension and either the patient or another magnet can be translated to advance the strip to the next position in the patient. Depth variation is done by interposing degraders in the beam (cyclotrons) or by changing the beam energy (synchrotons). Scanning is used for both beam "painting" and for intensity modulated beam delivery. The latter technique allows treatment planning to be optimized by delivering non-uniform dose distributions for each field to create a uniform dose in the target volume. Scanned beams reduce the integral dose to normal tissue because dose conformation to the proximal surface of the tumour can be achieved (Fig. 4). Because all the protons are used there is less activation of components and therefore less exposure of patients to background radiation.

 

Fig. 4.  Diagrammatic representation of the dose distribution
which can be achieved with scanned proton beams compared with
passively scattered beams [19]

Although beam scanning is likely to be the standard method of beam delivery of charged particle therapy beams in the future, at present only two scanning systems for ion beams are in routine clinical use, viz those at PSI, Villigen, Switzerland (proton beams) [9] and at the Gesellschaft für Schwerionenforschung (GSI), Darmstadt, Germany (12C beams) [18].   

Figure 5 shows comparative proton and x-ray treatment plans for a typical tumour (Ewing's sarcoma) using the best available techniques for proton therapy (scanned beams) and x-ray therapy (Intensity modulated beams). It is clear (editor notes: particularly in color originals) that the proton dose distribution is superior to the x-ray distribution in terms of conforming the maximum dose to the target volume and minimizing the dose to the surrounding normal tissue. Because of the physical characteristics of proton beams proton therapy will almost always be better than x-ray therapy if both are delivered under optimum conditions. Under such circumstances, as illustrated here, proton therapy will also be more efficient in terms of number of fields required to treat a lesion with optimum efficacy.

 

Fig. 5.   Comparative treatment plans (at 2 different levels) for scanned-beam proton therapy
(3 fields) [left side] and intensity-modulated x-ray therapy (9 fields) [right side]
for an Ewing's sarcoma [19].

3. Proton Therapy Facilities

Both cyclotrons and synchrotrons are used to produce proton beams for therapy. Tables 1 and 2  [11] show existing low-and high-energy proton therapy facilities respectively. The former are used almost exclusively for the treatment of eye lesions [mainly uveal melanoma and age-related macular degeneration (ARMD)]. Proton beams are most suitable for treating lesions (not necessarily malignant) close to critical structures such as uveal melanoma, ARMD, pituitary adenoma, meningiomas, arteriovenous malformation, acoustic neuroma, chondrosarcoma and chordoma; and prostate, cervix and paranasal sinus tumours. The latest available statistics (Tables 1 and 2) show that nearly 30 000 patients had received proton therapy up to July 2001 [11].

4. The NAC Experience

Proton therapy was first undertaken at the National Accelerator Centre (NAC) in 1993. The horizontal 200 MeV proton therapy facility (Fig. 6) is used mainly for irradiations of intracranial and head and neck lesions. Standard passive scattering techniques are used from beam modification.

 

Fig. 6.  Horizontal proton therapy treatment station.

A unique patient support and positioning system is based on real-time digital stereophotogrammetry (SPG) techniques (which are commonly used in land surveying) and CT scan information [20]. Patients are fitted with rigid custom made plastic masks, which carry radiopaque and retro-reflective markers. Position information calculated from the reflective marker images obtained by CCD cameras is used to automatically move the treatment chair to align the lesion in the beam with an accuracy of about 1 mm (1 standard deviation).

The treatment planning system is based on VOXELPLAN, obtained from the German Cancer Research Center, Heidelberg. Treatment plans based on Monte Carlo calculations are also done for special cases. For most treatments spread-out Bragg peaks are used but for smaller lesions (<20 mm diameter) crossfire plateau irradiations are given. Most treatments have been stereotactic radiosurgical procedures given in 3-4 fractions. Such fractionated treatments are possible because of the non-invasive nature of the patient immobilization and positioning system. Patients are treated for a variety of conditions (Table 3), most commonly ateriovenous malformations, brain tumours, pituitary adenomas, acoustic neuromas, meningiomas and brain metastases. Treatment sessions are currently on Mondays and Fridays.

A new dual fixed-beamline treatment station [12] is under development. Two fixed beam lines with a common isocentre will be installed : one horizontal line and one line inclined at an angle of 60° to the horizontal (Fig. 7).

 

Fig. 7.   Diagram of 2nd proton therapy treatment station.

Magnets from a dismantled physics experiment are being used. These two non-orthogonal beam lines and a scanning beam delivery system together with a robotic patient support system (with 6 degrees of freedom) will provide an extremely versatile treatment facility and permit the treatment of a wider range of lesions and increase patient throughput. A movable treatment nozzle, which can be used on both beam lines, is under consideration.

Beam utilization will be optimized by switching between the new treatment room and the existing one. In addition a dedicated proton therapy facility, based on a 230 MeV cyclotron, is currently being planned. This facility will include fixed beam arrangements, an isocentric gantry and both scattering and scanning beam delivery systems (Fig. 8). The existing treatment vault and the one currently under development will be incorporated in the new facility, but the beam delivery systems will be changed: the existing vault will contain an isocentric gantry and the other one will have a fixed horizontal beam. Both these stations will have scanning beam delivery. The two new vaults will contain non-orthogonal fixed beam arrangements with scattered beam delivery.

 

Fig. 8.  Layout of the NAC's proposed new dedicated proton therapy facility.
The three existing vaults are on the right (neutron therapy is in the middle).
The small vault between the cyclotron and the two new vaults
is for a possible future dedicated eye treatment facility, using the
66 MeV proton beam from the existing cyclotron.

 

References