Medical Radiation – More About Hadron Therapy


Routine treatment began in 1989 on the p(66)/Be neutron therapy unit, while proton therapy was first undertaken on iThemba’s 200 MeV beam in September 1993. The iThemba facilities were planned specifically to provide research opportunities in natural sciences for users from all over the country and other parts of the world, to supply high-energy particles for radiation therapy and to produce radioisotopes, primarily for medical applications.

All the major facilities, with the exception of the neutron therapy unit, were locally designed. The main accelerator is a variable-energy separated-sector cyclotron, capable of accelerating protons to a maximum energy of 200 MeV. The medical complex includes three radiotherapy treatment vaults, laboratories, offices, full medical physics and radiobiology facilities as well as a 30-bed on-site hospital. One of the treatment vaults contains the isocentric neutron therapy unit in which neutrons are produced by the reaction of 66 MeV protons on a thick beryllium target (designated p(66)/Be). The 200 MeV horizontal beam proton therapy facility occupies a second vault. Additional proton therapy beam lines are currently being designed for the third vault. Neutron therapy patients are usually treated during the day on Tuesdays, Wednesdays and Thursdays , while proton therapy takes place during the day on Mondays and Fridays.

Most patients, including those from other parts of the country and from neighbouring territories, are referred to iThemba through one of the local university teaching hospitals, viz, Groote Schuur Hospital (University of Cape Town) or Tygerberg Hospital (University of Stellenbosch). Both hospitals are about 25 minutes by road from the iThemba. Some private patients are also treated. At present initial patient assessment and preparation is undertaken at the referring hospitals while proton treatment planning as well as neutron treatment planning is undertaken at iThemba. Although many patients are housed in the on-site hospital for the duration of their treatments, others attend as out-patients.


The p(66)/Be neutron therapy facility incorporates an isocentric gantry capable of ± 185 deg. rotation. A rotating collimator (360 deg.) with a continuously variable rectangular aperture provides field sizes between 5.5 cm * 5.5 cm and 29 cm * 29 cm at a source-to-axis distance of 150 cm. A manually-controlled moving floor permits full rotation of the gantry. Downstream of the target are, in order, a pair of steel flattening filters (for small and large fields respectively), three tungsten wedge filters and a 2.5 cm thick polyethylene hardening filter, which removes unwanted low energy neutrons from the beam. A multiblade trimmer (blocking system) has recently been installed on the collimator to provide more flexible shielding. Neutron dose rates are typically about 0.50-0.60 Gy/min. A portal x-ray tube in the treatment head upstream of the collimator can be inserted on the beam axis and is used in conjunction with a neutron beam exposure for verification of the treatment field. The physical characteristics of the iThemba neutron beam are rather similar to those of an 8 MV x-ray beam.

In order to verify the dosimetry and treatment prescriptions, international radiobiological and national and international dosimetry intercomparisons have been undertaken. The results obtained were highly satisfactory, showing good agreement between participating centres. Several other radiobiological measurements have been made and the RBE (relative biological effectiveness) and OER (oxygen enhancement ratio) of iThemba’s neutron therapy beam have been found to be similar to those measured at other high-energy p/Be neutron therapy facilities. The energy spectra of the neutron beams for various irradiation conditions have been measured in air using the pulsed beam time-of-flight technique and in phantom using recoil methods.

Several clinical trials are currently being undertaken at iThemba, including treatments of tumours of the head and neck, salivary gland and breast and treatments of soft tissue and bone sarcomas, uterine sarcomas, paranasal sinuses and mesotheliomas. A protocol for prostate treatments is presently being implemented. A significant number of non-trial patients are also being treated.

Top of Page



The horizontal 200 MeV proton therapy facility is used mainly for irradiations of intracranial and head and neck lesions. The total distance between the beam line vacuum window and the isocentre is 7 m. A double scatterer plus occluding ring system is used to flatten the proton beam. The beam delivery system is designed for a maximum field diameter of 10 cm. The Bragg peak is spread out longitudinally using propellers made up of different thicknesses of acrylic and which are rotated in the proton beam. Since the beam energy cannot be reproduced exactly every day plastic range trimmer plates, 0.6 mm thick are placed just upstream of the modulator propeller in order to routinely produce a residual range of 24.00 ± 0.03 cm in water (distal 50% level) for patient treatment. This range corresponds to a proton energy of 191 MeV. For spread-out Bragg peak (SOBP) therapy the required is acheived by inserting the appropriate thickness of graphite (in a double-wedge system) into the beam.

The beam is controlled by two computerised feedback systems acting on two sets of XY steering magnets. The first system uses information from the multiwire ionization chamber, while the second one uses the information from the quadrant ionization chamber immediately upstream of the final collimator. The final (patient) collimator is located 27.5 cm upstream of the isocentre. Fixed inserts, which are usually custom-made of cerrobend, fit into the final collimator assembly which can rotate around the beam axis for alignment with the required treatment field. Clinical dose rates of about 3 Gy/min are routinely used. Individual calibrations are performed for each treatment field.

Several radiobiological and microdosimetric measurements have been made in both monoenergetic and modulated 200 MeV proton beams. A RBE of 1.00 was measured in the plateaux of both monoenergetic and modulated beams while RBEs of between 1.05 and 1.2 were obtained in spread-out Bragg peaks. Proton dosimetry and radiobiological intercomparisons have also been undertaken with various overseas centres and the results obtained were highly satisfactory. Proton spectra have been measured in the clinical beam under a variety of conditions using proton elastic scattering techniques.

The unique patient support and positioning system used at iThemba was designed by the Departments of Mechanical Engineering and of Geomatics of the University of Cape Town in conjunction with the Medical Radiation Group of iThemba. The system makes use of real-time digital stereophotogrammetry (SPG) techniques, which are commonly used in land surveying, and is linked to the patient support system which is a computerised adjustable chair with 5 degrees of freedom. When a patient undergoes a CT scan to locate the treatment volume, small radiopaque targets (1 mm diameter) are affixed to a custom-made plastic mask which fits the patient’s head precisely. From the scan information three-dimensional co-ordinates for all the targets are determined relative to a reference point in the treatment volume. This reference point is normally taken as the treatment isocentre. Retroreflective markers 8 mm in diameter are then fixed accurately on the mask exactly over the radiopaque markers. A close-fitting back is made for each patient mask and is fitted with a device for fixing the patient to the fully-adjustable chair headrest. In order to compensate for the chair’s lack of a roll motion the patient’s head often has to be tilted slightly from the vertical to allow positioning to be accomplished more efficiently.

During the patient positioning stage, a set of three charge-coupled device (CCD) TV cameras (out of eight which are positioned around the isocentre) captures video images through a frame-grabber of the retroreflective markers on the patient mask. These images are then analysed by a personal computer using SPG techniques. Since the positions of the video cameras and the direction of the proton beam are accurately known in space, it is possible to calculate the position of the centre of each of these reflective markers and hence the position of the reference point in the treatment volume, relative to the beam axis. The effects of camera distortions, aspect ratio and perspective are taken into account in the calculations. The co-ordinates of the beam entry point are also required and are obtained from the treatment planning programme.

Spatial corrections to align the vector between the beam entry point and the reference point (isocentre), which is in the treatment volume, with the beam axis are then sent from the SPG computer to a second personal computer which controls the patient support system (chair).

Computer-controlled stepper motors move the chair by the required amounts (X,Y and Z translation accuracy is within 0.1 mm and within 0.1 deg. for seat and vertical rotations) to bring the treatment vector directly into the proton beam. Additional information regarding the rotation angle of the final patient collimator, to align the collimator with the outline of the treatment volume, is also calculated by the first computer. It takes typically 2-3 iterations from an arbitrary position to align the patient in the required orientation.

Once the patient is properly positioned, according to preset tolerances of the beam entry, tumour and reference points (normally ±0.5 mm), the system is then set to monitor the positions of all the reflective targets and hence the position of the treatment volume. The beam can then be switched on and will be switched off automatically if any of these targets move by more than a preset amount. Analysis of the portal radiographs shows that the positioning accuracy of the SPG system is about 1 mm (1 standard deviation).

A sophisticated 3-dimensional non-coplanar treatment planning system (PROXELPLAN) is used for planning proton treatments at iThemba. The system is based on VOXELPLAN, obtained from the German Cancer Research Center, Heidelberg. For most treatments spread-out Bragg peaks are used but for smaller lesions (less than 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 , most commonly ateriovenous malformations, acoustic neuromas, meningiomas, pituitary adenomas, brain metastases and gliomas. Treatment sessions are currently on Mondays and Fridays.

Top of Page