Equipment for Radiation Surgery Using Narrow 185 MeV Proton Beams: Dosimetry and design

Abstract

The purpose of the present work was to optimize and standardize irradiation conditions and dosimetry methods in order to investigate the prerequisites for the routine use of narrow high energy proton beams for cerebral radiation surgery. Particular importance was laid on the design of the arrangements for defining and controlling the path of the beam in the laboratory in view of the desirability of working with well-defined parallel beams containing a minimal contribution of scattered and secondary radiation. At the same time it was intended that this apparatus should be used to adjust the beam reproducibly onto an ideal beam path prior to each irradiation session. It was also considered desirable to be able to supervise the centering and the structure of the beam during the irradiations in view of the variations which can arise due to varying running conditions in the synchrocyclotron. The beam was collimated and led to the place where the irradiations were to be performed, 25 metres from the synchrocyclotron, with a system of focusing quadrupole magnets and bending magnets. Final collimation of the beam penetrating the object was arranged with a system of accurately aligned cylindrical and plane-parallel metal apertures. The energy of the protons in the beam was 185±0.2 MeV and the maximum total fluence was 5·10¹⁰ protons/s. A method of ¹¹C activation dosimetry was developed with which an absolute determination of the fluence and the dose in the proton beam could be made with good accuracy. These determinations were performed by irradiating small polystyrene cylinders placed along the beam axis at the isocentre. The activity induced in the cylinders by the protons was measured in a well-type crystal detector. The efficiency of the detector for the detection of annihilation photons was determined before each measurement by means of calibration with standardized activities of ²²Na and ⁶⁰Co. The overall uncertainty in the dose determinations using ¹¹C activation dosimetry was ±8 per cent. Monitoring of the dose during irradiation was performed with an ionization chamber. Since it was most important to minimize the scattering of the protons, this chamber was designed as a ‘free air chamber’ with the electrodes parallel to the beam. The chamber could not be used for absolute measurements, however, since there was a contribution to the ionization in the chamber from scattered protons and secondary radiation from the walls of the collimator defining the cross section of the ionization volume. Since the proton beam at the isocentre was also affected by scattered radiation from the beam-defining aperture and to some extent by the other apertures, correction factors for the ionization chamber readings were determined for various collimator geometries in the beam by means of ¹¹C activation dose measurements. Other corrections to the ionization chamber measurements were either found to be very close to 1 or were calculated. However, it was found that, in spite of the facilities which were provided for controlling the path of the beam in the system, there was a contribution of scattered and degraded radiation which varied somewhat from day to day. Therefore, for accurate dosimetry it was necessary to continue to calibrate the chamber prior to and after each irradiation session with ¹¹C activation dosimetry. The properties of the narrow proton beam as it passed through brain tissue were studied by means of measurements on the density distributions of photographic films exposed in phantom material. For these experiments, a beam size of 5.0×2.5 mm² was used (at the time when the experiments were performed this beam size was considered suitable for radiation surgery of patients with parkinsonism). Collimation of the beam centrally in the brain can be considered optimal. The contributions to the dose outside the ideal geometric beam path caused by imperfections in the apparatus such as geometric penumbra and scattering in the collimator walls were comparable to the unavoidable contributions from diffusing Coulomb electrons and protons undergoing multiple scattering in the brain tissue. The feasibility of making clinical application of the dose distributions is considered favourable in view of the gradient in the transverse dose distribution, the secondary radiation dose to the brain tissue outside the geometric beam, and the central depth dose distribution. Matching of the beam angles and beam cross sections in order to obtain optimum three-dimensional dose distributions using rotation therapy was achieved with the aid of a specially developed analogy method.