Abstract
Background/Aim: The focus of this report is establishing an irradiation arrangement to realize an ultra-high dose-rate (uHDR; FLASH) of scanned carbon-ion irradiation possible with a compact commonly available medical synchrotron. Materials and Methods: Following adjustments to the operation it became possible to extract ≥1.0×109 carbon ions at 208.3 MeV/u (86 mm in range) per 100 ms. The design takes the utmost care to prevent damage to monitors, particularly in the nozzle, achieved by the uHDR beam not passing through this part of the apparatus. Doses were adjusted by extraction times, using a function generator. After one scan by the carbon-ion beam it became possible to create a field within the extraction time. The Advanced Markus chamber (AMC) and Gafchromic film are then able to measure the absolute dose and field size at a plateau depth, with the operating voltage of the chamber at 400 V at the uHDR for the AMC. Results: The beam scanning utilizing this uHDR irradiation could be confirmed at a dose of 6.5±0.08 Gy (±3% homogeneous) at this volume over at least 16×16 mm2 corresponding to a dose-rate of 92.3 Gy/s (±1.3%). The dose was ca. 0.7, 1.5, 2.9, and 5.4 Gy depending on dose-rate and field size, with the rate of killed cells increasing with the irradiation dose. Conclusion: The compact medical synchrotron achieved FLASH dose-rates of >40 Gy/s at different dose levels and in useful field sizes for research with the apparatus and arrangement developed here.
Recent studies demonstrated that ultra-high dose-rate (uHDR) radiation with electron (1) and proton (2) beams could efficiently inhibit tumor growth at the same level as the currently used conventional dose-rate (typically some cGy/s) while damage to healthy tissues is reduced. This phenomenon is called the FLASH effect. This can potentially increase the therapeutic window between tumor control rate and normal tissue toxicity. Extensive research on this field is ongoing worldwide.
The carbon-ion beam has good biological characteristics where the relative biological effectiveness (RBE) is increasing compared with protons (i.e., up to a factor of 3) because of their higher linear energy transfer (LET) besides physical characteristics such as the Bragg-peak. These characteristics enhance the therapeutic window. The uHDR initiating FLASH effect with a carbon-ion beam could further expand the therapeutic window by increasing the biological effect’s peak-to-entrance ratio (3). Few reports on uHDR with a carbon-ion beam exist. Dokic et al. reported encouraging normal tissue sparing effect with brain organoids of uHDR carbon ion beams with ≈40 Gy/s at 7.4 Gy, at relevant plateau LET level at 12 keV/μm (4). Tinganelli et al. showed the FLASH effect on Chinese hamster ovary cells (CHO-K1) with a carbon-ion beam with 7.4 Gy at 70 Gy/s at 13 keV/μm of dose averaged LET in hypoxia (5) and on C3H/He mice osteosarcoma in the hind limb with 18 Gy at 100 Gy/s at about 15 keV/μm (6). However, no FLASH effect was observed in HFL1 and HSGc-C5 cells with 1-3 Gy at 96-195 Gy/s at 13 or 50 keV/μm under aerobic condition (7). The mechanism of the FLASH effect in carbon-ion beam has been postulated (3), including oxygen depletion, radical recombination, in which a high transient concentration of peroxyl radicals results in less oxidation damage to normal tissue, and intertrack effects, in which intertrack recombination for positively charged particle beams determines radiation chemical driven biological consequences of radiation exposure, but is unknown. A simulation study found that uHDR carbon ions increasingly generate molecular oxygen toward the end of their trajectory at the Bragg peak. Geant4-DNA was modified to simulate the consumption of O2 and the role of multiple ionizations for carbon-ion beams (8). Evaluation of changes in radiation chemical yield of 7-hydroxy-coumarin-3-carboxylic acid implied that the influence of the reaction between water radiolysis species formed by neighborhood ionization tracks could be strongly related to the mechanisms of the FLASH effect (9). Verifying the FLASH effect with carbon ions will also help clarify the mechanism.
Because of the technical limitations of carbon-ion accelerators, conducting FLASH studies is difficult. Contrary to a proton beam generated with a continuous beam from cyclotrons which provide the high beam current, typically in the order of 300 nA (≈2×1012 protons/s), it is more challenging to establish FLASH conditions currently defined by delivering total dose and dose-rate with a carbon-ion beam with a compact-type medical synchrotron that provides significantly lower particle rates than cyclotrons (≈1×1010 carbon ions/s), resulting in smaller irradiation volume. A synchrotron delivers the beam in so-called spills, ion beam bunches of variable length in the synchrotron cycles of 2-32 s. The single spills are extracted with a specific pause of a minimum of 2 s in-between, which means that FLASH irradiation must be performed with one synchrotron filling and with a short extraction phase within 100 ms. Hence, modifications to the synchrotron operation pattern for FLASH conditions are needed (10).
Although there is an immediate need to identify the beam-related requirements for uHDR research with carbon ions (3), few systems available worldwide can conduct uHDR scanned carbon-ion research. Contrary to research institutes, clinical facilities where the treatment machine works for patient treatments must pay attention to damages to the treatment system due to the uHDR. This study demonstrated the development of an irradiation system to enable the uHDR (FLASH) and wider irradiation field scanned carbon-ion irradiation with a commercially available compact-type medical synchrotron in a clinical facility.
Materials and Methods
Accelerator settings. The Heavy Ion Medical Accelerator in Kansai (HIMAK) accelerator system (11) comprises a 10 GHz electron-cyclotron-resonance (ECR) ion source (12), a radio-frequency-quadrupole (RFQ)-linac accelerator of 610 keV/u followed by an alternating-phase-focus interdigit H-mode (APF IH)-type linac of 4.0 MeV/u, feeding the beam into a compact synchrotron of 6.6 Tm with a circumference of approximately 57 m. All elements in this accelerator chain determine the achievable number of carbon ions that can be extracted within one synchrotron cycle. To enable uHDR irradiation, only few changes were necessary to the radio-frequency (RF)-knockout extraction system (13) to achieve higher dose rates (14). The accelerator was optimized to reliably deliver more than 1×109 carbon ions per synchrotron cycle (spill). The extraction RF power pattern during the extraction time was returned to allow higher charges to be deflected from their orbit into the extraction channel. An ultra-fast extraction of the entire charge in the spill within 100 ms (i.e., instead of the usual 10,000-30,000 ms) was achieved by adjusting the amplitude of the extraction RF, typically several dozen times higher power level than routine radiotherapy operation [i.e., normal dose-rate (NDR)] (Figure 1A). This resulted in much higher beam currents (i.e., typically 10 nA instead of 0.1 nA) available for irradiation during one accelerator cycle. We also noted that the contribution of the time needed for scanning the beam over the irradiation field was small (<14 ms, 0.025 ms for moving between spots) compared to the overall time of the irradiation field, which was completely dominated by the extraction process. The beam current was measured with a digital oscilloscope (TBS2000, Tektronix, Beaverton, OR, USA) connected to HyBeat Heavy-ion Therapy System (Hitachi, Ltd., Tokyo, Japan) (15).
Setup for the ultra-high dose-rate (uHDR) experiments. (A) The amplitude of the extraction radio frequency (RF) in uHDR compared to that of normal dose-rate (NDR). (B) Monochromatic beam mode in a treatment room. The carbon-ion beam did not pass through the nozzle. Advanced Markus chamber (AMC) and Gafchromic film were used to measure the absolute dose and the field size, respectively, at a plateau depth. (C) Example of scanning pattern used in this study. The cross in green indicates the starting point of the scan. Orange lines are disposed spots area. The field used in the experiment is depicted in blue lines. The carbon-ion beam was scanned once to create a field within the extraction time. The axes are the definition of the beam’s coordinate.
Transport of the beam through the beamline did not need to be altered. The beam profile of X and Y directions is symmetrical and with near Gaussian distribution, with a size FWHM (full width at half maximum) of approximately 5 mm just upstream of the reference chamber.
Adaptations for beam monitoring and dosimetry. The extracted beam was guided into the Osaka Heavy Ion Therapy Center (OHITC). Unfavorably, the high beam current (>1×1010 ions/s) for the uHDR irradiation mode poses additional issues and complicates the accurate beam monitoring for the control of the raster scanner. The uHDR carbon-ion beam delivery system called a monochromatic beam mode was newly designed to acquire a pristine monochromatic beam by avoiding passage through the nozzle (Figure 1B). The spill length and beam current were set to run at pre-programmed values. The extraction time and beam transport were controlled by the accelerator system.
We performed redundant measurement with radiochromic films (Gafchromic film, EBT3; International Specialty Products, Wayne, NJ, USA), scanned with an EPSON DC-G20000 flatbed scanner (J331B, Epson Seiko Corporation, Nagano, Japan) at 150 dpi) that should prove the consistency of the uHDR dose distributions in lateral dose profile. It can be assumed that the radiochromic films do not show any dose-rate dependency (16). The profiles were extracted using ImageJ software (17) (version 1.53k, National Institutes of Health, Bethesda, MA, USA). The flatness was calculated within ±8 mm defined by (dmax − dmin)/(dmax + dmin).
For a measurement of saturation curve, an Advanced Markus chamber (AMC) (type 34045, PTW-Freiburg GmbH, Freiburg, Germany) connected to electrometer (6517B, Keithley Instruments, Cleveland, OH, USA) and StingRay (IBA Dosimetry GmbH, Schwarzenbruck, Germany) connected to UNIDOSwebline (PTW-Freiburg GmbH, Freiburg, Germany) were used. To test the recombination behavior of the detectors, saturation curves with electrode voltages from 50 to 400 V for AMC and from 25 to 150 V for StingRay were performed in uHDR conditions. The ion recombination factor was extracted for absolute dosimetry. For this measurement, the AMC was set to a dedicated acryl holder (Accelerator Engineering Corporation, Chiba, Japan), and the surface of the holder was aligned to the isocenter with the room laser. The StingRay was aligned on the holder with the room laser and functioned as a second chamber. The reading of the StingRay was divided by the reading of the AMC. The initial recombination correction factor was determined from a Jaffé-plot (18).
For absolute dosimetry, we just used the AMC with the same equipment as the measurement of the saturation curve. The dose-rate depends on the definition of the dose-rate (e.g., dose-averaged dose-rate and sliding window) (19). In this study, the dose-rate was calculated as the total delivered dose divided by the total delivery time, referred to as the mean (averaged) dose-rate (20).
Irradiation plan for the raster scanner. For the raster scanning irradiation, a plan file comprising a CP and SP file with a certain irradiation pattern must be loaded into the control system. The CP file consists of the number of spots in the control point, accelerator energy, range shifter setting ID, spill length, and dose-rate. The SP file basically consists of the position of the beam spots over which the center of the pencil beam is guided over, and the corresponding number of ions represented in the monitor unit, that should be applied to the individual beam positions. The pattern for the uHDR irradiation consists of 7×7 beam spot positions, as shown in Figure 1C, with a grid spacing of 3 mm and a fixed number of 3.125×107 ions per beam position. Disposed spots were introduced not to use a rising region of the beam current for creating the field (Figure 1C). This yields in total 5×108 carbon ions on an area of approximately 2 cm2. The beam was irradiated once at each spot position (i.e., no repainting) to simplify the dose-rate calculation.
Colony formation assay. HSGc-C5 cells were irradiated 0.7, 1.5, 2.9, 5.4, and 6.5 Gy by uHDR. After the irradiation, the HSGc-C5 cells were washed in Phosphate Buffered Saline (PBS) and trypsinized. The cells were seeded to 60-mm diameter dishes, cultured for 2 weeks, fixed with formalin, and stained with 0.25% crystal violet. Colonies comprising >50 cells were scored as survivors, and the survival fraction (SF) was calculated. Survival curves were fitted to a linear-quadratic model (LQM) (21). At least three independent experiments for each condition were performed, and the results are expressed as the average and standard deviation.
Results
Accelerator settings. The tuned accelerator demonstrates a very short beam extraction time of approximately 60 ms for the whole uHDR irradiation, as shown in Figure 2A. Figure 2B shows that the beam intensity is fine structure (ripple), typical for synchrotron extraction. Figure 2C indicates simulated beam current for a spot by spot based on the measured beam intensity in the case of disposed spots with 21 spots. The beam current fluctuated less in the field, colored in the squared yellow hatched region (14.7±4.0 nA).
Measured beam current of the extracted beam as a function of time. (A) uHDR extraction with approximately 60 ms length. The disposed spots area is shown in the orange hatch. The beam in blue can be used for ultra-high dose-rate (uHDR) irradiation. (B) Zoom of (A) from 10 to 20 ms colored in green. (C) Beam current spot by spot. The cross in green indicates the starting point of the scan. Orange dashed lines are disposed spots area. The field used in the experiment is depicted in blue dashed lines. The axes are the beam’s definition.
FLASH dose application. Figure 3 shows the scanning pattern, scanned film image, and lateral dose profiles at the isocenter. The flatness of the profile in uHDR irradiation is indicated at approximately 3%. Disposed spots of 6.5, 5.4, 2.9, 1.5, and 0.7 Gy were 0, 7, 14, 21, 35 spots, respectively, as shown in Figure 3A, D, G, J and M.
Measured lateral dose distribution for uHDR with 208.3 MeV/u carbon ions for 6.5 Gy (A-C), 5.4 Gy (D-F), 2.9 Gy (G-I), 1.5 Gy (J-L), and 0.7 Gy (M-O). The left, middle, and right pictures show the scanning pattern, GafChromic (EBT3) film measurement, and measured lateral dose profiles in x and y directions for each panel. In the left picture, the cross in green indicates the starting point of the scan. Orange lines are disposed spots area. The field used in the experiment is depicted in blue lines. The axes are the beam’s definition. In the middle picture, the dashed lines indicate the section for the profiles in the measured film. In the right picture, lateral dose profiles (x direction in red and y direction in yellow) for uHDR irradiation are shown.
Figure 4A shows the saturation curves at uHDR for the AMC and StingRay. The AMC provided a lower recombination rate showing that the AMC saturation curve reaches the plateau at 400 V. The StingRay is still increasing at 150 V. Jaffé plots obtained in the uHDR carbon-ion beam for both detectors are shown in Figure 4B. The general recombination factor of the AMC and StingRay was 1.005 and 1.019, respectively.
Dosimetry measurements for ultra-high dose-rate (uHDR) irradiations: (A) Saturation measurement varying the voltage of the Advanced Markus chamber (AMC) (blue open points) and StingRay (orange open points) under uHDR conditions. (B) Jaffé plots obtained in the uHDR carbon-ion beam at 208.3 MeV/u for AMC (blue open points) and StingRay (orange open points). (C) Dose of the uHDR irradiations for a series of tests (same plan file for a given dose) during a day. (D) Survival fraction of HSGc-C5 at uHDR (green open points) at the entrance. Fitted curves with the LQM for uHDR (dashed line).
The uHDR irradiation utilizing beam scanning was confirmed with a dose of 0.7±0.07, 1.5±0.04, 2.9±0.07, 5.4±0.06, and 6.5±0.08 Gy for the beam in a volume of at least 16 mm × 16 mm in a square field and a corresponding dose-rate of 86.3±9.3, 92.9±2.3, 93.8±2.1, 89.9±1.1, and 92.3±1.2 Gy/s (Figure 4C). Table I shows the summary of the results.
Dosimetric parameters of the measured dose distribution at the ultra-high dose-rate.
The developed uHDR scanned carbon-ion beam was applied to in vitro experiment. Figure 4D depicts the SF of the HSGc-C5 cells. The cell-killing rate at uHDR condition was increased as the irradiation dose escalated. The SFs by uHDR irradiation were 0.883 in 0.7 Gy, 0.608 in 1.5 Gy, 0.401 in 2.9 Gy, 0.0845 in 5.4 Gy and 0.0258 in 6.5 Gy.
Discussion
This study demonstrated the technical development of an irradiation system to enable the uHDR (≥40 Gy/s). Previous studies were performed in the experiment room (5, 6). In the present study, we showed a wider irradiation field of scanned carbon-ion irradiation in the treatment room of the clinical facility by using a commercially available compact-type medical synchrotron while keeping a sufficient level of dose application accuracy. To our best knowledge, no studies are reporting various dose levels from low to high dose at uHDR with the field size and the dosimetric characteristics, which is essential to performing FLASH research with the scanned carbon-ion beams.
The implementation of uHDR carbon-ion irradiations at OHITC was technically and clinically challenging because the facility was a clinical-oriented hospital. We developed the monochromatic beam mode and extensive adaptions and optimizations for the accelerator parameters comprising the tuning of the synchrotron for the maximum possible intensity of ≥1×109 carbon ions per spill and a fast beam extraction of ≤100 ms. Due to the monochromatic beam mode, the treatment machine did not face any problems after the uHDR experiments.
A time structure to define dose-rate might be one of the important factors for the FLASH effect (20, 22, 23), especially in synchrotrons which generated quasi-continuous beams. In this study, repainting was not used in which the time structure would be simple compared with one with repainting (Figure 2C). The time structure per spill of the synchrotron depends on the accelerated particle energy. The synchrotron might be used for investigating time structure dependence of the FLASH effect.
The AMC worked at 400 V for the uHDR carbon-ion beam. At the same time, the StingRay needs a higher voltage to reach the plateau region of the saturation curve (Figure 4A and B). The applied voltage was not increased to avoid damage to the chambers in this study. For the AMC applied to 400 V, the plateau of the saturation curve can be reached, thus guaranteeing a linear behavior at the beam intensity used for the uHDR experiments.
The compact-type synchrotron limits the number of carbon ions per synchrotron cycle due to the ring size. Thus, the field size in uHDR irradiation with carbon-ion was forced to limit for delivering a certain dose. In this study, 6.5 Gy was the maximum dose with the field size of approximately 20 mm × 20 mm. The beam extraction time can be extended to approximately 100 ms and more doses can be deliverable. One of the factors initiating the FLASH effect is the total dose to be delivered at the moment. The threshold for the FLASH effect for carbon ions is unknown. It might be defined as radiobiological effect weighted dose [i.e., Gy(RBE)] for carbon-ion beam. The irradiation field size contributes to the more accurate experimental results in in vitro assays because the number of evaluated cells increases. Due to the changeable dose level at uHDR, the developed system would be useful for seeking the threshold by utilizing a cell experiment such as a colony formation assay (24).
To the best of our knowledge, currently, the field size is wide enough to conduct in vitro experiment for the uHDR carbon-ion beam. Introducing the disposed spots contributed to the stabilities of the dose delivery (Figure 2C) at different dose levels at uHDR, which were homogeneity of ±3.1% and dose-rate of ±11.3% in the worst case (Figure 3 and Figure 4C, Table I). The best-case marked homogeneity of ±2.0% and dose-rate of ±0.9%. The dose distribution acquired by the GafChromic EBT3 films has a higher spatial resolution (150 dpi), and the data show intrinsic noise due to the films’ grain structure. The dosimetric characteristics were better than those previously reported (5), where the homogeneity and dose-rates were ±5% and ±20%. It can be concluded that the raster scanning irradiation could yield a good lateral dose distribution, showing only very slight asymmetries. The high dose area is large enough for applying the intended dose from 0.7 to 6.5 Gy within a diameter of 16 mm. The developed uHDR scanned carbon-ion beam confirmed the cell-killing effect by showing the SF changes to different doses. The result demonstrated our radiation system succeeded in irradiate with ultra-high dose rate. We are performing detailed in vitro experiments and analyses and will report the results separately.
Many other improvements are needed to further progress FLASH research with the scanned carbon-ion beam. Because the monochromatic beam mode did not use a monitor chamber, it could not measure dose amount to the synchrotron such as the normal dose-rate mode. A uHDR dedicated monitor chamber is needed to control the uHDR beam. It takes longer than 300 ms to change the energy layer in our synchrotron regardless of the change method (i.e., range shifter scanning or active-energy scanning), which is slow for FLASH research. A ridge filter is demanded to spread the Bragg-peak at uHDR, opening for instantaneous volumetric irradiation (abbreviated as IVI). Noteworthily, uHDR irradiation with carbon ions has made very little use of the treatment planning system (TPS) in the modern sense of the term. A TPS considering dose-rate in optimization and evaluation must be developed and used for uHDR irradiation (19, 25).
Conclusion
The compact medical synchrotron achieved FLASH dose-rates of >40 Gy/s at different dose levels and in useful field sizes for research with the apparatus and arrangement developed here. There were also promising beam characteristics of uHDR scanned carbon-ion beams at useful dose levels, and it will be possible to develop carbon-ion FLASH research further. Biological experiments will also be possible with the doses and dose-rates established here with this uHDR irradiation apparatus.
Acknowledgements
JSPS KAKENHI Grant Number 22K07695 supported this work. The authors acknowledge and thank the staff in OHITC for their help with conducting measurements related to the uHDR, the staff in Osaka Heavy Ion Administration Company, and Hitachi, Ltd., for their help in operating the accelerator in this study.
Footnotes
Authors’ Contributions
Conceptualization: MY, SS, KM, NH, TT, MT, MU, TK. Data curation: MY, KM, NH, TT, MT, TN, WM, TK. Formal analysis: MY, KM, NH, TT, MT, TN, WM, TK. Funding acquisition: MY. Investigation: MY, SS, KM, NH, TT, MT, MU, TN, WM, TK. Methodology: MY, SS, KM, NH, TT, MT, MU, TN, WM, TK. Software: MY, NH, TT, MT, TN, WM. Supervision: SS, TN, MK, KO, TK. Validation: MY, NH, TT, MT, MU, TN, WM. Visualization: MY, NH, TT, MT, MU, TN, WM. Writing – Original draft preparation: MY. Writing – Review & editing: MY, SS, KM, NH, TT, MT, MU, TN, WM, TN, MK, KO, TK.
Conflicts of Interest
The Authors, Masumi Umezawa, Takuya Nomura, and Wataru Mukoyoshi are employees of Hitachi, Ltd.
- Received December 26, 2022.
- Revision received January 3, 2023.
- Accepted January 4, 2023.
- Copyright © 2023 International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved.
This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY-NC-ND) 4.0 international license (https://creativecommons.org/licenses/by-nc-nd/4.0).