Characterization of fragmented heavy-ion beams using a three-stage telescope detector: detector configuration and instrumentation

Accelerated heavy-ion beams used in biological and medical research are often utilized in conjunction with absorbers which lead to the fragmentation of the beam. The BERKLET, initially a two-stage solid-state telescope detector, was designed to make rapid, on-line energy and linear energy transfer (LET) measurements of individual particles in a heavy-ion beam, thus allowing characterization of fragmented beams. From data collected with the BERKLET, one is able to determine a number of important parameters. These include: residual energy and LET histograms for the full beam and for the individual Z components, relative number of particles with a given Z, and dose and track average LET's for the full beam and for the individual Z's. Improvements to the BERKLET design and changes in data analysis are discussed and contrasted with the results of an earlier BERKLET configuration. The most notable improvements are the addition of a thin scintillation detector for improved LET measurement, a tenfold improvement in the dynamic range of the event discriminator, reported here as 1:2000, and dual high-and low-gain amplification of the LET signals, permitting the identification of particles with Z's ranging from 12 down to 1.     

Depth absorbed dose and LET distributions of therapeutic 1H, 4He, 7Li, and 12C beams

The depth absorbed dose and LET (linear energy transfer) distribution of different ions of clinical interest such as 1H, 4He, 7Li, and 12C ions have been investigated using the Monte Carlo code SHIELD-HIT. The energies of the projectiles correspond to ranges in water and soft tissue of approximately 260 mm. The depth dose distributions of the primary particles and their secondaries have been calculated and separated with regard to their low and high LET components. A LET value below 10 eV/nm can generally be regarded as low LET and sparsely ionizing like electrons and photons. The high LET region may be assumed to start at 20 eV/nm where on average two double-strand breaks can be formed when crossing the periphery of a nucleosome, even though strictly speaking the LET limits are not sharp and ought to vary with the charge and mass of the ion. At the Bragg peak of a monoenergetic high energy proton beam, less than 3% of the total absorbed dose is comprised of high LET components above 20 eV/nm. The high LET contribution to the total absorbed dose in the Bragg peak is significantly larger with increasing ion charge as a natural result of higher stopping power and lower range straggling. The fact that the range straggling and multiple scattering are reduced by half from hydrogen to helium increases the possibility to accurately deposit only the high LET component in the tumor with negligible dose to organs at risk. Therefore, the lateral penumbra is significantly improved and the higher dose gradients of 7Li and 12C ions both longitudinally and laterally will be of major advantage in biological optimized radiation therapy. With increasing charge of the ion, the high LET absorbed dose in the beam entrance and the plateau regions where healthy normal tissues are generally located is also increased. The dose distribution of the high LET components in the 7Li beam is only located around the Bragg peak, characterized by a Gaussian-type distribution. Furthermore, the secondary particles produced by high energy 7Li ions in tissuelike media have mainly low LET character both in front of and beyond the Bragg peak.     

Dose contributions from large-angle scattered particles in therapeutic carbon beams

In carbon therapy, doses at center of spread-out Bragg peaks depend on field size. For a small field of 5 x 5 cm2, the central dose reduces to 96% of the central dose for the open field in case of 400 MeV/n carbon beam. Assuming the broad beam injected to the water phantom is made up of many pencil beams, the transverse dose distribution can be reconstructed by summing the dose distribution of the pencil beams. We estimated dose profiles of this pencil beam through measurements of dose distributions of broad uniform beams blocked half of the irradiation fields. The dose at a distance of a few cm from the edge of the irradiation field reaches up to a few percent of the central dose. From radiation quality measurements of this penumbra, the large-angle scattered particles were found to be secondary fragments which have lower LET than primary carbon beams. Carbon ions break up in beam modifying devices or in water phantom through nuclear interaction with target nuclei. The angular distributions of these fragmented nuclei are much broader than those of primary carbon particles. The transverse dose distribution of the pencil beam can be approximated by a function of the three-Gaussian form. For a simplest case of mono-energetic beam, contributions of the Gaussian components which have large mean deviations become larger as the depth in the water phantom increases.     

The LET spectra at different penetration depths along secondary 9C and 11C beams

Owing to the potentially therapeutic enhancement of delayed particles in treating malignant diseases by radioactive 9C-ion beam, LET spectra at different penetration depths for a 9C beam with 5% momentum spread, produced in the secondary beam line (SBL) at HIMAC, were measured with a multi-wire parallel-plate proportional counter. To compare these LET spectra with those of a therapeutic 12C beam under similar conditions, the 12C beam was replaced with an 11C beam, yielded in the SBL as well and having almost the same range as that of the 9C beam. The LET spectra of the 9C beam and its counterpart, i.e. the 11C beam, at various depths were compared, especially around the Bragg peak regions. The results show that nearby the Bragg peak lower LET components decreased in the LET spectra of the 9C beam while extra components between the LET peak caused by the primary beam and the lower components due to the fragments could be observed. These additional contributions in the LET spectra could be attributed to parts of the emitted particles from the radioactive 9C ions with suitable conditions regarding the LET counter. Integrating these LET spectra in different manners, depth-dose and dose-averaged LET distributions were obtained for the 9C and 11C beams, forming the basic data sets for further studies. In general, the depth-dose distributions of the 9C and 11C beams are comparative, i.e. almost the same peak-to-plateau ratio. The ratio for the 9C beam, however, has room to increase due to the geometric structure limitation of the present detector. The dose-averaged LETs along the beam penetration are always lower for the 9C beam than for the 11C beam except at the falloff region beyond the Bragg peak. Applying the present depth-dose and dose-averaged LET data sets as well as the essential radiobiological parameters obtained with 12C beams previously for HSG cells, an estimate concerning the HSG cell surviving effects along the penetration of the 9C and 11C beams shows that lower survival fractions for the 9C beam at the distal part of the Bragg peak, corresponding to the stopping region of the incoming 9C ions, can be expected when the same entrance dose is given. It is still hard to appreciate the potential of 9C beams in cancer therapy based on the present LET spectrum measurement, but it provides a substantial basis for upcoming radiobiological experiments.     

Elevated LET components in clinical proton beams

This paper assesses the contribution of secondary particles to pencil and passively scattered proton beams, in particular when considering the dose-averaged linear energy transfer (LET(d)) in biological treatment planning. Proton Monte Carlo simulations are performed in water phantoms and for two patients, considering all primary and secondary particles, including recoils from inelastic nuclear interactions. Our results show that secondary protons exhibit LET(d) values up to a factor 10 higher than those of the primary protons at the same depth. Thus, secondary protons have a significant impact on the LET(d). Their contribution increases the LET(d) by ～50% along the central axis and even >200% in the penumbra. Furthermore, the LET maximum after the peak changes from 12 to 15 keV μm(-1) when adding secondary protons to the primary contribution. This is important when modeling LET(d) with analytical methods. The contribution of recoils (A > 3) is observed to be 1.2% in the entrance region considering a prostate case. The degree of biological damage inflicted by recoils remains hard to quantify, but is discussed on the basis of detailed energy spectra. The results highlight the role of secondary protons in LET-based radiobiological effectiveness calculations for proton therapy and when analyzing radiobiological experiments. Furthermore, the findings demonstrate the impact of inhomogeneities on the LET and the subtle changes between the LET distributions of passively scattered and actively scanned beams.     

Production of clinically useful positron emitter beams during carbon ion deceleration

In external beam radiation therapy, radioactive beams offer the best clinical solution to simultaneously treat and in vivo monitor the dose delivery and tumor response using PET or PET-CT imaging. However, difficulties mainly linked to the low production efficiency have so far limited their use. This study is devoted to the analysis of the production of high energy (11)C fragments, preferably by projectile fragmentation of a stable monodirectional and monoenergetic primary (12)C beam in different absorbing materials (decelerators) in order to identify the optimal elemental composition. The study was performed using the Monte Carlo code SHIELD-HIT07. The track length and fluence of generated secondary particles were scored in a uniform absorber of 300 cm length and 10 cm radius, divided into slices of 1 cm thickness. The (11)C fluence build-up and mean energy variation with increasing decelerator depth are presented. Furthermore, the fluence of the secondary (11)C beam was studied as a function of its mean energy and the corresponding remaining range in water. It is shown that the maximum (11)C fluence build-up is high in compounds where the fraction by weight of hydrogen is high, being the highest in liquid hydrogen. Furthermore, a cost effective alternative solution to the single medium initially envisaged is presented: a two-media decelerator that comprises a first liquid hydrogen section followed by a second decelerating section made of a hydrogen-rich material, such as polyethylene (C(2)H(4)). The purpose of the first section is to achieve a fast initial (11)C fluence build-up, while the second section is primarily designed to modulate the mean energy of the generated (11)C beam in order to reach the tumor depth. Finally, it was demonstrated that, if the intensity of the primary (12)C beam can be increased by an order of magnitude, a sufficient intensity of the secondary (11)C beam is achieved for therapy and subsequent therapeutic PET imaging sessions. Such an increase in the intensity might be easily achieved with a superconducting cyclotron.     

Backscatter towards the monitor ion chamber in high-energy photon and electron beams: charge integration versus Monte Carlo simulation

In some linear accelerators, the charge collected by the monitor ion chamber is partly caused by backscattered particles from accelerator components downstream from the chamber. This influences the output of the accelerator and also has to be taken into account when output factors are derived from Monte Carlo simulations. In this work, the contribution of backscattered particles to the monitor ion chamber response of a Varian 2100C linac was determined for photon beams (6, 10 MV) and for electron beams (6, 12, 20 MeV). The experimental procedure consisted of charge integration from the target in a photon beam or from the monitor ion chamber in electron beams. The Monte Carlo code EGS4/BEAM was used to study the contribution of backscattered particles to the dose deposited in the monitor ion chamber. Both measurements and simulations showed a linear increase in backscatter fraction with decreasing field size for photon and electron beams. For 6 MV and 10 MV photon beams, a 2-3% increase in backscatter was obtained for a 0.5 x 0.5 cm2 field compared to a 40 x 40 cm2 field. The results for the 6 MV beam were slightly higher than for the 10 MV beam. For electron beams (6, 12, 20 MeV), an increase of similar magnitude was obtained from measurements and simulations for 6 MeV electrons. For higher energy electron beams a smaller increase in backscatter fraction was found. The problem is of less importance for electron beams since large variations of field size for a single electron energy usually do not occur.     

Linear energy transfer dependence of a normoxic polymer gel dosimeter investigated using proton beam absorbed dose measurements

Three-dimensional dosimetry with good spatial resolution can be performed using polymer gel dosimetry, which has been investigated for dosimetry of different types of particles. However, there are only sparse data concerning the influence of the linear energy transfer (LET) properties of the radiation on the gel absorbed dose response. The purpose of this study was to investigate possible LET dependence for a polymer gel dosimeter using proton beam absorbed dose measurements. Polymer gel containing the antioxidant tetrakis(hydroxymethyl)phosphonium (THP) was irradiated with 133 MeV monoenergetic protons, and the gel absorbed dose response was evaluated using MRI. The LET distribution for a monoenergetic proton beam was calculated as a function of depth using the Monte Carlo code PETRA. There was a steep increase in the Monte Carlo calculated LET starting at the depth corresponding to the front edge of the Bragg peak. This increase was closely followed by a decrease in the relative detector sensitivity (Srel = Dgel/Ddiode), indicating that the response of the polymer gel detector was dependent on LET. The relative sensitivity was 0.8 at the Bragg peak, and reached its minimum value at the end of the proton range. No significant effects in the detector response were observed for LET < 4.9 keV microm(-1), thus indicating that the behaviour of the polymer gel dosimeter would not be altered for the range of LET values expected in the case of photons or electrons in a clinical range of energies.     

Monte Carlo simulation of energy-deposit clustering for ions of the same LET in liquid water

This work presents a Monte Carlo study of energy depositions due to protons, alpha particles and carbon ions of the same linear-energy-transfer (LET) in liquid water. The corresponding track structures were generated using the Geant4-DNA toolkit, and the energy deposition spatial distributions were analyzed using an adapted version of the DBSCAN clustering algorithm. Combining the Geant4 simulations and the clustering algorithm it was possible to compare the quality of the different radiation types. The ratios of clustered and single energy depositions are shown versus particle LET and frequency-mean lineal energies. The estimated effect of these types of radiation on biological tissues is then discussed by comparing the results obtained for different particles with the same LET.     

Measurement of the restricted linear energy transfer of stray radiation close to the treatment volume of 12 and 18 MeV clinical photon beams

The restricted dose mean linear energy transfer (LET) (L500,D) of the stray radiation field a few centimeters outside the treatment volume has been measured for 12 and 18 MV photons produced by a clinical Therac-20 (AECL) accelerator. The measurements were performed as a function of field size and distance from the edge of the treatment volume, using the method of the high-pressure ionization chamber. Contrary to what was found in a previous investigation for a clinical Co-60 unit and despite the presence of photoneutrons (in the case of 18 MV photons), the L500,D outside the beam does not increase significantly relative to the L500,D of the primary beam.     

Cell killing and mutation induction by heavy ion beams

Carbon beam radiotherapy for cancer patients was initiated in Japan in June 1994. This study attempts to clarify the radiobiological effects of heavy ion beams. In this study, human cancer cell lines (RMG-1, MDA-MB231) and V79 cells were used. The cell killing was determined by colony forming assay, and mutation induction was determined by counting the number of 6-thioguanine resistant colonies (hprt locus mutation assay). The cell lines were irradiated with carbon (20 or 80 keV/microm) or neon beams (80 keV/microm). Carbon ions with a higher LET value (80 keV/microm) had an enhanced cytotoxic effect compared to those with a lower LET value (20 keV/microm). Carbon beams produced a slightly stronger cytotoxic effect than neon beams when irradiated at the same LET level (80 keV/microm), but the difference was not remarkable. The mutant fraction was significantly higher in all cell lines when they were irradiated with heavy ion beams, compared to the results for X-ray irradiation. The mutant fraction increased when the LET of the carbon beams increased. At equivalent LET values, the mutant fraction was lower for neon beams than for carbon beams. Fractionation of carbon beam irradiation had no effect on survival, but reduced the mutant fraction. Neon beams might be more appropriate for heavy ion therapy, especially when higher doses are being used. In addition, the fractionation of heavy ion beam administration might be appropriate for reducing the mutant fraction.     

General characteristics of the use of silicon diode detectors for clinical dosimetry in proton beams

The properties of silicon diode detectors, used for dosimetry in clinical proton beams, were investigated with special regard to the measurement of relative dose distributions in water. Different types of silicon diode detector were studied, and the resulting distributions of detector signal versus depth in the water phantom were compared with the corresponding distributions obtained with a plane-parallel NACP ionization chamber. The measurements were performed in a proton beam with an initial energy of 173 MeV. It is shown that the Hi-p silicon detector gives a signal which is proportional to the ionization density in the silicon crystal in all parts of the Bragg curve, and for all levels of accumulated dose to the detector. This is in contrast to detectors based on n-type silicon, or on low resistivity p-type silicon. After pre-irradiation, these latter detectors show a stopping-power dependent recombination, yielding an increase in the detector signal per unit dose with increasing LET. This effect leads to an over-response in the Bragg peak, which increases gradually with the accumulated detector dose. Using the Hi-p silicon diode detector, the depth ionization distribution was found to be equal to the distribution obtained with the plane-parallel NACP ionization chamber at all pre-irradiation levels, within the experimental accuracy. This implies that the quotient between the ionization in the detector and the absorbed dose to the surrounding water is equal for these detectors.     

The influence of target thickness and backstop material on proton-produced neutron beams for radiotherapy

Results are presented of measurements of skin sparing, penetration and total dose per unit of incident charge for various target thicknesses and filtrations for a neutron beam generated by 42 MeV protons on beryllium. These results are contrasted to predictions outlined in a previous paper. The differences from these predictions are attributed to the contribution of low-energy neutrons produced by the residual proton beam in the thick copper target backstop.     

The role of nonelastic reactions in absorbed dose distributions from therapeutic proton beams in different medium

Many new techniques for delivering radiation therapy are being developed for the treatment of cancer. One of these, proton therapy, is becoming increasingly popular because of the precise way in which protons deliver dose to the tumor volume. In order to achieve this level of precision, extensive treatment planning needs to be carried out to determine the optimum beam energies, energy spread (which determines the width of the spread-out Bragg peak), and angles for each patient's treatment. Due to the level of precision required and advancements in computer technology, there is increasing interest in the use of Monte Carlo calculations for treatment planning in proton therapy. However, in order to achieve optimum simulation times, nonelastic nuclear interactions between protons and the target nucleus within the patient's internal structure are often not accounted for or are simulated using less accurate models such as analytical or ray tracing. These interactions produce high LET particles such as neutrons, alpha particles, and recoil protons, which affect the dose distribution and biological effectiveness of the beam. This situation has prompted an investigation of the importance of nonelastic products on depth dose distributions within various materials including water, A-150 tissue equivalent plastic, ICRP (International Commission on Radiological Protection) muscle, ICRP bone, and ICRP adipose. This investigation was conducted utilizing the GEANT4.5.2 Monte Carlo hadron transport toolkit.     

Evaluation of the mean energy deposit during the impact of charged particles on liquid water

The DNA strand break yield due to the impact of ionizing particles on living beings is closely related to the number of inelastic events per unit absorbed dose produced by these particles. The higher this number, the higher the probability of causing DNA strand breaks per unit absorbed dose. In a previous work, it was found that the total number of events produced by primary particles and the secondary electrons is almost independent of the type and energy of the incident particle (or LET). This finding could be supported by a quasi-constant mean energy deposit by inelastic event (ε). In this work, ε was defined and determined for electrons and the non-negative charge states of hydrogen (H?,?) and helium (He?,?,2?) species impacting on liquid water. Ionization, excitation and charge transfer (up to two-electron transfers) processes have been included in present calculations. We found that, for liquid water, ε is within 13.7 ±?4.1 eV, 14.2 ± 1.7 eV and 13.8 ± 1.4 eV for electrons, hydrogen and helium species, respectively, with impact energies changing over three orders of magnitude. Unlike the mean excitation energy, the mean energy deposit per inelastic event depends not only on the target molecule but also on the projectile features. However, this dependence is relatively weak. This fact supports the quasi-independent number of inelastic events per unit absorbed dose found previously when charged particles impact on matter.     

Dosimetric characterization of silicon and diamond detectors in low-energy proton beams

The dosimetric behaviour of a Scanditronix p-type silicon diode and a PTW natural diamond detector was studied in low-energy proton beams in the 8.3-21.5 MeV range. The properties investigated were linearity, reproducibility, dose rate dependence, energy and linear energy transfer (LET) dependence. The influence of detector thickness on the results of depth dose measurements was also demonstrated. A Markus parallel plate ionization chamber was used for reference dosimetry. Silicon diode and diamond detectors showed linearity at therapeutic dose level, reproducibility better than 1% (1sigma) and sensitivity variation with dose rate and proton energy.     

Commissioning stereotactic radiosurgery beams using both experimental and theoretical methods

The purpose of this investigation is to study the feasibility of using an alternative method to commission stereotactic radiosurgery beams shaped by micro multi-leaf collimators by using Monte Carlo simulations to obtain beam characteristics of small photon beams, such as incident beam particle fluence and energy distributions, scatter ratios, depth-dose curves and dose profiles where measurements are impossible or difficult. Ionization chambers and diode detectors with different sensitive volumes were used in the measurements in a water phantom and the Monte Carlo codes BEAMnrc/DOSXYZnrc were used in the simulation. The Monte Carlo calculated data were benchmarked against measured data for photon beams with energies of 6 MV and 10 MV produced from a Varian Trilogy accelerator. The measured scatter ratios and cross-beam dose profiles for very small fields are shown to be not only dependent on the size of the sensitive volume of the detector used but also on the type of detectors. It is known that the response of some detectors changes at small field sizes. Excellent agreement was seen between scatter ratios measured with a small ion chamber and those calculated from Monte Carlo simulations. The values of scatter ratios, for field sizes from 6 x 6 mm2 to 98 x 98 mm2, range from 0.67 to 1.0 and from 0.59 to 1.0 for 6 and 10 MV, respectively. The Monte Carlo calculations predicted that the incident beam particle fluence is strongly affected by the X-Y-jaw openings, especially for small fields due to the finite size of the radiation source. Our measurement confirmed this prediction. This study demonstrates that Monte Carlo calculations not only provide accurate dose distributions for small fields where measurements are difficult but also provide additional beam characteristics that cannot be obtained from experimental methods. Detailed beam characteristics such as incident photon fluence distribution, energy spectra, including composition of primary and scattered photons, can be independently used in dose calculation models and to improve the accuracy of measurements with detectors with an energy-dependent response. Furthermore, when there are discrepancies between results measured with different detectors, the Monte Carlo calculated values can indicate the most correct result. The data set presented in this study can be used as a reference in commissioning stereotactic radiosurgery beams shaped by a BrainLAB m3 on a Varian 2100EX or 600C accelerator.     

An approach to microdosimetry in a pi- meson beam using nuclear emulsions.

K5 emulsions 10 mum thick were exposed at various depths in a perspex phantom to a 70 MeV pi- meson beam and counts taken of the tracks in emulsion volumes 7x7x10 mum3. Data are presented on the number of track events and also the total number of grains associated with each event in each of four categories spanning the LET range of the secondary particles. The number of heavy tracks (category 4) shows an increased incidence in the region of the stopping pi- mesons (14-5 cm in perspex) while the number of single grains (category 1) decreases with depth. Categories 2 and 3 (grain clusters and light tracks) are approximately constant with depth. An estimate of the grain sensitivity is obtained by taking the proton as representative of the whole range of secondary particles. This procedure gives a value of 5 keV per grain in the pi- peak. The LET of light tracks was therefore in the range 1-10 keV mum-1 in emulsion, scaling to 0-4-4 keV mum-1 in water. Heavy tracks have LET values in excess of 4 keV mum-1 in water.     

Off-axis beam quality change in linear accelerator x-ray beams

The effective energy of the x-ray beam from linear accelerators changes as a function of the position in the beam due to nonuniform filtration by the flattening filter. In this work, the transmittance through a water column was measured in good geometry and the beam quality characterized in units of HVL in water. Measurements were made on a variety of linear accelerators from 4 to 10 MV. The beam energy decreased with increasing distance from the central ray for all accelerators measured.     

A dosimetric evaluation of water equivalent phantoms for kilovoltage x-ray beams

Solid phantoms are widely used in radiation therapy for both relative and reference dosimetry. Two water equivalent phantoms, RMI-457 Solid Water and Plastic Water, were evaluated for use in kilovoltage x-ray dosimetry in the energy range from 75 to 300 kVp. Relative and reference dosimetry measurements were performed in the solid phantoms and compared with water. The results indicate that RMI-457 Solid Water could be used for output factor determination for all energies tested and the measurement of percentage depth doses for the 300 kVp x-ray beam, with data agreeing to within 1%, compared to the same measurements in water. For the same criteria, Plastic Water could only be used for output factor determination of the 300 kVp x-ray beam. The superior agreement of the calculated mass-energy absorption coefficients for Solid Water and water, as compared to Plastic Water and water was consistent with the experimental results. Reference dosimetry is not recommended with the solid phantoms for the energies studied due to the lack of published correction factors. It is recommended that any solid phantom be tested by comparison with water in the same manner before being used for the dosimetry of kilovoltage x-ray beams.