Commissioning of a clinical pencil beam scanning proton therapy unit for ultrahigh dose rates (FLASH)
K. P. Nesteruk, M. Togno, M. Grossmann, A. J. Lomax, D. C. Weber, J. M. Schippers, S. Safai, D. Meer, S. Psoroulas
CCommissioning of a clinical pencil beam scanning proton therapyunit for ultrahigh dose rates (FLASH)
Konrad P. Nesteruk, ∗ Michele Togno, and Martin Grossmann
Paul Scherrer Institute, Center for Proton Therapy, Villigen, Switzerland
Anthony J. Lomax
Paul Scherrer Institute, Center for Proton Therapy, Villigen, SwitzerlandETH Zurich, Department of Physics, Villigen, Switzerland
Damien C. Weber
Paul Scherrer Institute, Center for Proton Therapy, Villigen, SwitzerlandUniversity Hospital Zurich, Department of Radiation Oncology, SwitzerlandUniversity Hospital Bern, Department of Radiation Oncology, Switzerland
Jacobus M. Schippers
Paul Scherrer Institute, Division Large Research Facilities, Villigen, Switzerland
Sairos Safai, David Meer, and Serena Psoroulas
Paul Scherrer Institute, Center for Proton Therapy, Villigen, Switzerland a r X i v : . [ phy s i c s . m e d - ph ] J a n bstract Purpose:
The purpose of this work was to provide a flexible platform for FLASH research withprotons by adapting a former clinical pencil beam scanning gantry to irradiations with ultrahighdose rates.
Methods:
PSI Gantry 1 treated patients until December 2018. We optimized the beamlineparameters to transport the 250 MeV beam extracted from the PSI COMET accelerator to thetreatment room, maximizing the transmission of beam intensity to the sample. We characterizeda dose monitor on the gantry to ensure good control of the dose, delivered in spot-scanning mode.We characterized the beam for different dose rates and field sizes for transmission irradiations. Weexplored scanning possibilities in order to enable conformal irradiations or transmission irradiationsof large targets (with transverse scanning).
Results:
We achieved a transmission of 86 % from the cyclotron to the treatment room. Wereached a peak dose rate of 9000 Gy/s at 3 mm water equivalent depth, along the central axis of asingle pencil beam. Field sizes of up to 5x5 mm were achieved for single spot FLASH irradiations.Fast transverse scanning allowed to cover a field of 16x1.2 cm . With the use of a nozzle-mountedrange shifter we are able to span depths in water ranging from 19.6 to 37.9 cm. Various dose levelswere delivered with a precision within less than 1 %. Conclusions:
We have realized a protonFLASH irradiation setup able to investigate continuously a wide dose rate spectrum, from 1 to9000 Gy/s in a single spot irradiation as well as in the pencil beam scanning mode. As such, wehave developed a versatile test bench for FLASH research.
Keywords: Ultrahigh dose rates, FLASH, proton therapy, gantry, pencil beam scanning ∗ [email protected] . INTRODUCTION One of the limiting factors in delivering radiation therapy to particularly aggressive orradiation resistant tumors is the limitation on the amount of dose healthy tissues can with-stand. In recent years, several studies have indicated that ultrahigh dose rates might result inreduced toxicities to healthy tissues while keeping the same tumor control as for treatmentswith standard dose rate levels. In 2014, the so called FLASH effect has been demonstratedfor the first time by irradiating lung tumors in mice with 4.5 MeV electrons [1]. A reducedpulmonary fibrosis was observed with electrons for a dose rate of 60 Gy/s when comparedto a conventional dose rate of 0.03 Gy/s. Whole brain irradiation of mice with 4.5 MeVelectrons showed a unique memory sparing for dose rates above 100 Gy/s [2]. Almost noskin toxicity has been observed in mini-pigs irradiated with electrons, as well as in cat pa-tients with locally advanced squamous cell carcinoma of the nasal planum [3]. Eventually, inSwitzerland, the first ever FLASH treatment of a human patient has been performed in 2018at the Lausanne University Hospital (CHUV) on a prototype linac for FLASH radiotherapywith electrons [4].Although the majority of the promising studies of the FLASH effects was performed withelectrons, it is conceptually possible that this FLASH effect could be observed with protonirradiation. Several centers in the world, as well as main vendors in the field, started toinvestigate the possibility of reaching FLASH dose rates with existing treatment units [5–9].Based on the studies with photons and electrons, it has been concluded that the thresholdfor the FLASH effect is at least 40 Gy/s, at least a factor 10 higher than conventional doserates. However, the definition of the dose rate is prone to ambiguities, as the beam has itsmicrostructure and the average dose rate may be substantially different from the maximumdose rate in a short pulse. The definition of the dose rate is also difficult for pencil beamscanning, due to the sequential character of the delivery technique. Since the mechanism ofFLASH has not been fully explained, a potential influence of pauses in dose delivery andfractionation is not understood yet. Therefore, exact conditions for the FLASH effect tooccur remain unknown, which reflects in non-consistent results of studies performed withprotons so far. Thus, flexible FLASH test benches, able to provide different dose rates andirradiation conditions, are required.In this paper, we report on the commissioning of the former clinical unit Gantry 1 at3
IG. 1. Gantry 1 at an angle of -90 degrees seen from the treatment room. Highlighted are spot-scanning axes and motions, the ”isocenter” defined for this paper, and a Faraday cup used forexperiments. the Paul Scherrer Institute for FLASH research with protons. PSI Gantry 1 was the firstgantry in the world to deliver routinely pencil beam scanning (PBS) to cancer patients.The objective of this work was to transform the in-house developed gantry into a versatileplatform for FLASH effect studies. As such, the modified gantry will be able to providevarious dose rates and allow for single spot irradiations as well as pencil beam scanning withultrahigh instantaneous dose rates for biological samples and potentially patients. Dosedelivery in both transmission and conformal (Bragg peak) modes will be possible.
II. MATERIALS AND METHODSA. Pencil beam scanning Gantry 1
Gantry 1 treated patients from 1996 until the end of 2018. It was the first facilityworldwide that used the spot-scanning technique [10, 11]. With its radius of 2 m, it isstill the most compact gantry in the world competing advantageously with low footprintgantries sold by vendors. The compactness was achieved by mounting the patient table4
IG. 2. Schematic diagram of the Gantry 1 beamline. Selected components relevant to this paperare highlighted. off-axis (excentric gantry) as well as by placing a scanning magnet upstream the last dipole.Since the gantry is excentric and the table moves vertically together with the gantry rotation,there is no true isocenter. In this paper, whenever we refer to the isocenter, we have in minda moving isocenter defined as a rotation-invariant distance between the nozzle and the axisof the table at its reference position (Fig. 1).Magnetic scanning is realized in one transverse direction ( U ) by means of a fast scanningmagnet, so-called sweeper, with a dead time of a few ms between spots. Scanning in depth( S ) is achieved with a range shifter in the gantry nozzle, which consists of 39 polystyrene(PS) plates of 4.53 mm thickness and one PS plate of half thickness. They are movedpneumatically in and out of the beam with a dead time of 60 ms per plate. In view of theFLASH applications, this feature gives a unique opportunity to transport an un-degradedproton beam to the isocenter in order to maximize the beam intensity, hence the dose rate,while keeping the possibility of conformal (Bragg peak) irradiations. The slowest scanningdirection ( T ) is performed by moving a treatment table. An overview of the gantry, as seenfrom the treatment room, is presented in Fig. 1.5 . Beam transport for ultrahigh dose rates Proton beam used for treatments is produced by the COMET cyclotron [12]. The energyand intensity of the extracted beam are 250 MeV and up to 1 µ A, respectively. Energymodulation is performed with a degrader system, installed right after the cyclotron. In thecase of Gantry 1 the previously mentioned range shifter was used for an additional finemodulation of the energy pre-selected by the degrader. However, the degrader introducesmassive intensity losses due to the scattering occuring in the degrader material. At lowenergies, the fraction of the beam transported to the treatment room is lower than 1 %. Inorder to reach ultrahigh dose rates, we had to optimize the beamline parameters to transportthe 250 MeV beam produced by the PSI COMET cyclotron to the treatment room withminimum losses. This energy has never been used for treatments, although the gantry wasdesign to accept energies up to 300 MeV. Therefore, a completely new configuration ofbeamline magnet settings, so-called beamline tune, had to be found.The beamline (Fig. 2), including the gantry, is 44 m long and consists of about 40 magnets- dipoles and quadrupoles. Beam transport can be simulated in numerous codes for beamoptics calculations. We used a simulation by means of TRANSPORT [13, 14] to find initialmagnet settings and as a guide for further experimental fine tuning. We considered thefollowing criteria for the desired beam optics: • beam size (2-sigma) always much smaller than apertures of magnets and collimators • point-to-point imaging between the location of the degrader (upstream) and the gantrycoupling point • limited beam tilt at the entrance of the gantry • beam waist in the gantry at the same location in both bending (dispersive) and non-bending planes • limited beam divergence of the beam extracted from the gantry • round beam spot at a certain distance from the gantry • minimized dispersion and its derivative at the isocenter.6he experimental tuning was based on numerous beam profilers and intensity monitorslocated along the beamline. It was realized in two stages. The goal of the first stage was totransport the beam to the gantry coupling point with minimum loses. At the second stage,we optimized the transport through the gantry to achieve the maximum transmission, around beam spot at a defined distance from the gantry nozzle, and a limited beam diver-gence in air. For these optimizations, we used a scintillating foil and a in-room camera foronline tuning, and a scintillator-CCD detector for precise measurements of the beam sizeand its envelope in air. The beam transmission to the gantry coupling point was determinedas a ratio between coupling-point and cyclotron-extracted beam intensities, measured simul-taneously by means of intensity monitors installed along the beamline. The beam currentat the isocenter was calculated based on the total proton charge measured by a Farady cup(see Fig. 1) and measured delivery time. The transmission from the coupling point to theisocenter was defined as a ratio between the calculated Faraday cup current and the currentmeasured at the coupling point. C. Dose monitoring in the FLASH regime
Gantry 1 is equipped with two parallel-plane ionization chambers placed in the nozzle tocontrol the dose to be delivered [11]. The first chamber (monitor 1) is the main monitorwhich controls dose delivery with an ion collection time of 90 µ s, while the second chamber(monitor 2) with a collection time of 350 µ s is a backup element and it was part of thetherapy verification system. Assuming a FLASH irradiation time of the order of 10 ms,monitor 1 guarantees a precision of at least 1 %.Compared to the clinical setting, the instantaneous dose rate we expect in FLASH ex-periments is 100 to 1000 times higher. Ionization chambers at such high dose rates areaffected by significant ion recombination losses and therefore, a thorough characterization ofthese kind of monitors is needed [15]. In order to provide an accurate dose monitoring, wehave performed such a calibration of the nozzle monitor against a Faraday cup (see Fig. 1).We used the same Faraday cup to validate the precision of the dose delivery in terms ofreproducibility. The prediction of the absolute dose with different dose rates was based ona model described in [15]. 7 . Delivery of different dose rates For the definition of dose rate we assume the beam produced by our cyclotron is contin-uous, as the RF frequency of the COMET cyclotron is 72.85 MHz. As such, every 14 ns a0.8 ns long pulse is delivered. Therefore, in a single pulse the dose rate is 17.5 times higherthan under the assumption of continuous beam. However, a single pulse is so short thatthe dose delivered within it is extremely low. Even for a FLASH dose rate of 1000 Gy/s,the dose per pulse is of the order of 10 µ Gy. Also in terms of ion recombination, the dosedelivery can be considered continuous [16].
1. Single spot irradiations
Single spot transmission irradiations are used to deliver homogeneously a dose to smalltargets, such as cell lines. In order to obtain homogeneity in depth we use the flat part ofthe Bragg peak curve on the central axis where the dose, the beam size, and hence the doserate, is typically constant for a few centimeters, as depicted in Fig. 3 and described in [15].For the transverse homogeneity, we define the field size to be the 95 % isodose surface for agiven Gaussian pencil beam.The instantaneous dose rate can be varied by adjusting the beam current and beam size.The former is realized by a vertical deflector in the COMET cyclotron. As such, we canvary the cyclotron current between 0.1 and 800 nA. Up to 20 % higher beam currents aretechnically achievable. However, they are not clinically commissioned and thus were notused in this study. The variation of beam size can be achieved by choosing a differentnumber of range shifter plates in the gantry nozzle. Due to multiple Coulomb scattering thebeam size increases and hence the dose rate decreases in a quadratic manner. This can alsobe used to increase the field size with 95 % homogeneity at the cost of decreased dose rate.In order to characterize the beam and estimate the corresponding dose rate we measuredbeam sizes by means of a CCD-scintillator detector and depth-dose curves in water by meansof a large sensitive area, plane-parallel Bragg peak ionization chamber. The dose rates weredetermined from the measured delivered protons by means of a Faraday cup and recordeddelivery time. The time is recorded by the control system and corrected offline for thereaction time of an upstream kicker magnet which switches the beam off when the requested8
IG. 3. Comparison of Bragg peak curves - integral and on the central axis for a beam energy of230 MeV. The flat part of the curve with the maximum dose rate is highlighted. dose is reached. As such, an accuracy better than 50 µ s is achieved.
2. Pencil beam scanning
We explored scanning possibilities in order to enable conformal irradiations or transmis-sion irradiations of large targets (transverse scanning).Since the upstream degrader is not used anymore, the only possibility to scan in depthis to use the range shifter in the gantry nozzle. We verified the minimum reachable energywith all the range shifter plates inserted by performing a range measurement in water.For scanning in the transverse U direction (the bending direction), we evaluated theperformance of the scanning magnet (sweeper) at an energy of 250 MeV, as this energyhas never been used for treatments. For the main FLASH-optimized tune, a new scanningmagnet calibration, so-called sweeper map, was defined to obtain equal spacing betweenspots. Additionally, we explored the possibility of fast scanning in the T direction (thenon-dispersive direction) with the use of a steering magnet usually utilized for the first spotcorrection (beam centering). This magnet is not as fast as the sweeper. With a dead timebetween consecutive spots of the order of 100 ms it would, however, guarantee much faster9canning in T than the table moving. Eventually, a 2D scanning map has been determined.For all the transverse scanning measurements, a CCD-scintillator detector was used. III. RESULTSA. FLASH-optimized beam transport
We characterized two different beamline tunes to be used for FLASH experiments. Themain FLASH-optimized tune is shown in Fig. 4. The measured beam envelope is comparedwith the one simulated in TRANSPORT with the final magnet settings. Due to the designof the compact gantry, there is no beam profile monitor in the gantry. The observed dis-crepancies are mostly due to the phase space definition. The TRANSPORT simulation isbased on the phase space measurements performed in 2006. We updated the phase space(highlighted in the figure) on the basis of upstream beam profile measurements. However,the beam emittance was assumed to be the same as measured in 2006. The agreementwe reached is satisfactory for our purpose and the simulation in TRANSPORT served as
FIG. 4. Beam envelope corresponding to the main FLASH beamline tune. Negative values of beamsize correspond to the dispersive plane ( U ) and positive values to the non-dispersive plane ( T ).Continuous line - result of simulations with TRANSPORT; red dots - beam profile measurements.
10 useful guide for the beamline tuning, providing an important qualitative description ofour beam transport. The tune presented in Fig. 4 provides the highest transmission to theisocenter and the minimum eccentricy of the beam spot, even when the range shifter is notused. The beam divergence in air was measured to be (0 . ± .
3) mrad and (2 . ± .
3) mradfor the dispersive ( U ) and non-dispersive ( T ) plane, respectively. At a distance of 41 cmfrom the nozzle, 5 cm behind the isocenter, the beam spot was found to be perfectly roundwith a 1 σ size of (3 . ± .
1) mm. The tune reproducibility has been assessed based onseveral measurements of beam profiles in different days. The measured beam sizes along thebeamline were found to be in agreement within 5 % and those in air within 3 %.The transmission to the gantry coupling point was measured to be (86 ±
1) % for cy-clotron currents exceeding 500 nA. A weak dependence on the beam current was observed,as different settings of the vertical deflector slightly affect the initial phase space and hencethe beam transport. The minimum transmission was measured to be (82 ±
1) % for thelowest cyclotron currents, 1 nA and below. The transmission from the coupling point to theisocenter was determined to be 100 % and independent of the beam current.
B. Ultrahigh dose rates
1. Single spot transmission irradiations
With the main FLASH tune we have achieved field sizes ranging from 2.3x2.3 mm fora beam energy of 250 MeV to 5x5 mm for a beam energy of 170 MeV (all range shifterplates inserted). The corresponding maximum peak dose rates are (3600 ± ±
30) Gy/s, respectively. The mentioned peak dose rates correspond to 3.5 cm waterequivalent depth and a cyclotron beam current of 800 nA. For a given field size, we are ableto continuously vary the dose rate from the level of the conventional PBS dose rate (orderof 1 Gy/s) to the maximum dose rate corresponding to the given field size.The absolute maximum we have achieved was (9300 ± σ sizes of the beam spot were measured to be (2 . ± .
05) mm and (1 . ± .
05) mm in11
IG. 5. Characterization of a beam spot in air. Left: Beam spot of 2.3/1.8 mm ( U / T - 1 σ ) in airrecorded with a scintillator-CCD detector. Right: Corresponding dose rate distribution in waterat 3 mm water equivalent depth. the U and T directions, respectively. As such, the maximum field size with this beam spotfor transmission irradiations is 1.1x1.1 mm . The characterization of the beam spot in airfor this maximum dose rate is depicted in Fig. 5.The uncertainty of the dose rate consists of the model uncertainty, dose delivery uncer-tainty and the beam size uncertainty. The beam size uncertainty contributes most to thetotal uncertainty. In our characterization studies, the uncertainty due to the delivery timemeasurement was negligible due to 3 to 4 orders of magnitude longer delivery times thanthe precision of the time measurement.In order to deliver different dose levels with a wide range of dose rates we used thepreviously described monitor 1 calibrated against dose rate independent Faraday cup. Forall cyclotron beam currents up to 5 nA we did not observe any efficiency drop of monitor 1 dueto ion recombination and the ratio between monitor units (MU) of monitor 1 and the totalcharge q measured with the Faraday cup remained constant. Also between 5 and 10 nA theefficiency drop was very subtle, always below 1 %. For cyclotron currents exceeding 10 nA,the monitor efficiency, normalized to the MU /Q ratio corresponding to a cyclotron currentof 1 nA, was observed to drop rapidly with the increasing current, as shown in Fig. 6. Thecorresponding calibration curve had to be divided in two regions - moderate current range(10-300 nA) and high current range (300-800 nA), as the monitor response presents a kinkat about 300 nA and another polynomial had to be fitted to the latter cyclotron currentrange. With this calibration we were able to deliver different doses with different dose rateswith a precision always better than 1 %, as verified with the Faraday cup.12 IG. 6. Efficiency of monitor 1 in the gantry nozzle as a function of cyclotron beam current.
2. Pencil beam scanning mode
The range shifter enables scanning in depth between 19.6 g · cm − and 37.9 g · cm − (R ranges of the proton beam at 80 % distal fall-off). The measured ranges (Fig. 7) correspondto the energies 250 MeV and 170 MeV, respectively. The former is achieved when no rangeshifter (RS) plate is used, the latter corresponds to the use of all the 40 RS plates. FIG. 7. Bragg peak curves measured with a range scanner for two minimum and maximum beamenergies. IG. 8. Transverse pencil beam scanning pattern acquired with a CCD-scintillator detector.
The maximum range of the transverse scanning with the minimum beam spot distortionwas found to be 16 cm ( ± . − . U and T directions,respectively. In Fig. 8 a scanning pattern acquired with a CCD-scintillator detector is shown.The pattern is based on a previously measured 2D scanning map. The spots are separatedby 2 cm in the U direction and by 1.2 cm (maximum range) in the T direction. IV. DISCUSSION
In this paper, we report the commissioning of the former treatment unit Gantry 1 atPSI for FLASH experiments. In order to maximize the dose rate in the treatment room,we optimized the beam transport for 250 MeV beam, and decided to use only the in-roomrange shifter for the energy degradation. We achieved a transmission of 86%, i.e. we couldreach more than 680 nA beam irradiation at isocenter. Higher beam intensity could bepotentially be achieved, as already shown in a proof-of-principle study by Busold et al [17].Moreover, losses have been observed only upstream of the treatment room, and no losshas been observed on the gantry, therefore satisfying the radiation protection requirementsof our facility (where the shielding of the treatment room is thinner than the one of thecyclotron bunker, which can sustain higher beam losses). In order to minimize the losseswe refrained from using collimators which help define better the phase space and make thebeam transport less sensitive to cyclotron phase space variations. This contributes to ourbeam size uncertainty and thus the total uncertainty of the dose rate.Such high beam currents can then be shaped to a dose and dose-rate distribution usingeither multiple- or single-spot deliveries. The latter provides the highest instantaneous doserates available at our facility: we could achieve more than 9000 Gy/s, which matches dose-14ate levels achieved at synchrotron facilities such as the European Synchrotron ResearchFacility (ESRF) - 8000 Gy/s and 16000 Gy/s [18, 19] - and electron machines like the eRT6electron linac at Lausanne University Hospital (CHUV). The gantry control system doesnot limit the length of these single-spot irradiations, therefore we can deliver all possibledoses in both single- and multiple-spots deliveries. The minimum pause between two spotscorresponds to 2 ms, but larger pauses can be introduced by the control system, to helpinvestigating the impact of these variables on irradiation modalities such as clinical spotscanning. This minimum time can be further decreased down to 100 µ s by using the verticaldeflector in the cyclotron and following some modifications of the control system.The dose can be precisely delivered based on the nozzle monitor, which we calibratedagainst the Faraday cup. However, we observed some day-to-day instabilities in the cali-bration. Therefore, we apply a day-specific correction for our calibration by measuring themonitor-1-to-Faraday-cup ratio (MU /Q ) for low and high dose rates, which takes a few min-utes. We observed an increase ratio over time since the original calibration in the MU /Q ratio. We have not understood this effect yet. One of the explanations could be temperatureand pressure changes, for which we do not apply any correction. The maximum deviationswe have observed so far were up to 4 % and can be completely neutralized by the mentionedday-specific correction.Even though the energy can be modified easily inside the treatment room, the presenceof the range shifter at the end of the beamline causes a clear correlation between final energyand dose rate. The beam size increases as function of number of range shifter plates anddistance from the gantry exit, and consequently the dose rate drops. At the minimum energyachievable with our system, 700 Gy/s could be reached at the isocenter. Such performancecould be improved in case a different material for the range shifters is used, or in case thesamples are placed closer to the gantry exit.On the other hand, the relationship between range shifters and beam size allows aneasy way of adapting the beam conditions to the size of the sample to irradiate, withoutthe production of complex single or double-scattering systems. We have tested differentcombinations of number of range shifter plates and distances from the gantry nozzle, simu-lating different sample sizes. This makes our experimental setup particularly interesting forexperiments with cells, as we could adapt flexibly to the requirements of external users.Thanks to its flexibility, PSI Gantry 1 will give a chance to address a few important15uestions concerning the FLASH effect mechanism and to define appropriate dose deliveryconditions. Different studies of the FLASH effect showed different results, from no protectiveeffect of FLASH irradiations to a significant reduction of toxicity to healthy tissues. However,there are several aspects which have to be taken into account while comparing various resultsobtained with different beam parameters. One of the most important is the differencebetween the average and instantaneous dose rate. Depending on the machine type and dosedelivery technique, the two dose rates may be very different. For instance, the Kinetron andOriatron eRT6 electron linacs, used at Orsay and Lausanne University Hospital (CHUV),respectively, allow an instantaneous dose rate in a 1-2 µ s pulse to exceed 10 Gy/s [1, 20–22].However, the average dose rate under typical treatment conditions would be ”only” up to afew kGy/s and would involve pauses between pulses of at least 5 ms. With these machinesthe most promising results have been achieved so far. However, if the instantaneous doserate plays the key role in triggering of the FLASH effect, it could well be that reaching anaverage dose rate of the order of 100 Gy/s or even 1000 Gy/s will not be sufficient unless largeenough instantaneous dose rate is provided. On the other hand, studies in different animaland cells models have shown FLASH effects at such average dose rates. Dose-rate thresholdsobserved are not fully consistent among different studies, though, and this prompted someauthors to suggest that other factors might play a role - not only the instantaneous dose rate,but also the total irradiation time as well as the minimum dose within a (macro-)pulse [23].Different machines are currently able to test at best only a part of the parameter space, ashighlighted in a recent overview of the so-called FLASH-compatible machines [24]; thereforeit is of fundamental importance to develop flexible irradiation facilities, allowing to test inthe same conditions and different animal models different combinations of dose, dose rateand pulse structure.
V. CONCLUSIONS AND OUTLOOK
We have successfully converted the PSI pencil beam scanning Gantry 1 into a test benchfor FLASH experiments. We are able to deliver a wide range of dose rates from 1 to9000 Gy/s, which enables detailed studies of FLASH with protons. Moreover, we can conducttransmission as well as conformal irradiations using only local energy modulation. Field sizeslarger than 5x5 mm can be irradiated by means of fast transverse pencil beam scanning.16lthough the gantry has not been designed to provide scanning in both transverse directions,we have proven that it is possible to include the second direction for fast scanning of smallfields. This feature can be used for pre-clinical studies of FLASH with small animals.Teaming with colleagues from CHUV and Varian Medical Systems, we are currentlycarrying out our first biological experiments at Gantry 1. ACKNOWLEDGMENTS
We thank our collaborators at PSI (M. Schippers, M. Eichin, U. Rechsteiner, B. Rohrer,M. Egloff) and CHUV (M.-C. Vozenin, V. Grilj, C. Bailat). This work is partially fundedby the Swiss National Science Foundation (grant No. 190663).
CONFLICT OF INTEREST STATEMENT
The authors have no conflict to disclose. [1] V. Favaudon, L. Caplier, V. Monceau, F. Pouzoulet, M. Sayarath, C. Fouillade, M.-F. Poupon,I. Brito, P. Hup´e, J. Bourhis, J. Hall, J.-J. Fontaine, and M.-C. Vozenin, Ultrahigh dose-rateFLASH irradiation increases the differential response between normal and tumor tissue inmice, Science Translational Medicine , 245ra93 (2014).[2] P. Montay-Gruel, K. Petersson, M. Jaccard, G. Boivin, J.-F. Germond, B. Petit, R. Doenlen,V. Favaudon, F. Bochud, C. Bailat, J. Bourhis, and M.-C. Vozenin, Irradiation in a flash:Unique sparing of memory in mice after whole brain irradiation with dose rates above 100Gy/s,Radiotherapy and Oncology , 365 – 369 (2017), 15th International Wolfsberg Meeting2017.[3] M.-C. Vozenin, P. De Fornel, K. Petersson, V. Favaudon, M. Jaccard, J.-F. Germond, B. Pe-tit, M. Burki, G. Ferrand, D. Patin, H. Bouchaab, M. Ozsahin, F. Bochud, C. Bailat, P. De-vauchelle, and J. Bourhis, The Advantage of FLASH Radiotherapy Confirmed in Mini-pigand Cat-cancer Patients, , 35–42 (2019).[4] J. Bourhis, W. J. Sozzi, P. G. Jorge, O. Gaide, C. Bailat, F. Duclos, D. Patin, M. Ozsahin,F. Bochud, J.-F. Germond, R. Moeckli, and M.-C. Vozenin, Treatment of a first patient with LASH-radiotherapy, Radiotherapy and Oncology , 18 – 22 (2019), FLASH radiotherapyInternational Workshop.[5] A. Patriarca, C. Fouillade, M. Auger, F. Martin, F. Pouzoulet, C. Nauraye, S. Heinrich,V. Favaudon, S. Meyroneinc, R. Dendale, A. Mazal, P. Poortmans, P. Verrelle, and L. DeMarzi, Experimental Set-up for FLASH Proton Irradiation of Small Animals Using a ClinicalSystem, International Journal of Radiation Oncology*Biology*Physics , 619 – 626 (2018).[6] E. Beyreuther, M. Brand, S. Hans, K. Hidegh´ety, L. Karsch, E. Leßmann, M. Sch¨urer, E. R.Szab´o, and J. Pawelke, Feasibility of proton FLASH effect tested by zebrafish embryo irradi-ation, Radiotherapy and Oncology , 46 – 50 (2019), FLASH radiotherapy InternationalWorkshop.[7] M. Buonanno, V. Grilj, and D. J. Brenner, Biological effects in normal cells exposed to FLASHdose rate protons, Radiotherapy and Oncology , 51 – 55 (2019), FLASH radiotherapyInternational Workshop.[8] A. Darafsheh, Y. Hao, T. Zwart, M. Wagner, D. Catanzano, J. F. Williamson, N. Knutson,B. Sun, S. Mutic, and T. Zhao, Feasibility of proton FLASH irradiation using a synchrocy-clotron for preclinical studies, Medical Physics , 4348–4355 (2020).[9] E. S. Diffenderfer et al., Design, Implementation, and in Vivo Validation of a NovelProton FLASH Radiation Therapy System, International Journal of Radiation Oncol-ogy*Biology*Physics , 440 – 448 (2020).[10] E. Pedroni, R. Bacher, H. Blattmann, T. B¨ohringer, A. Coray, A. Lomax, S. Lin, G. Munkel,S. Scheib, U. Schneider, and A. Tourovsky, The 200-MeV proton therapy project at the PaulScherrer Institute: Conceptual design and practical realization, Medical Physics , 37–53(1995).[11] S. Lin, T. Boehringer, A. Coray, M. Grossmann, and E. Pedroni, More than 10 years experienceof beam monitoring with the Gantry 1 spot scanning proton therapy facility at PSI, MedicalPhysics , 5331–5340 (2009).[12] J. Schippers, R. D¨olling, J. Duppich, G. Goitein, M. Jermann, A. Mezger, E. Pedroni, H. Reist,and V. Vrankovic, The SC cyclotron and beam lines of PSI’s new protontherapy facilityPROSCAN, Nuclear Instruments and Methods in Physics Research Section B: Beam Inter-actions with Materials and Atoms , 773 – 776 (2007), The Application of Accelerators inResearch and Industry.
13] K. L. Brown, D. C. Carey, F. C. Iselin, and F. Rothacker,
TRANSPORT: a computer programfor designing charged-particle beam-transport systems , CERN Yellow Reports: Monographs,CERN, Geneva, 1980, Also publ. as SLAC and FERMILAB.[14] U. Rohrer, Graphic Transport Framework, 2007.[15] S. Safai and et al, Dosimetry and biological experiments with proton beams at ultra-highdose rates (FLASH): experimental setup and first dosimetry results, Physics in Medicine andBiology (2020), in preparation.[16] C. Gom`a, S. Lorentini, D. Meer, and S. Safai, Proton beam monitor chamber calibration,Physics in Medicine and Biology , 4961–4971 (2014).[17] S. Busold and H. R¨ocken, Beam Intensity Modulation Capabilities for Varian’s ProBeam ® Isochronous Cyclotron, Proceedings of the 21st Int. Conf. on Cyclotrons and their ApplicationsCyclotrons2016 , THP05 (2017).[18] R. Serduc, P. V´erant, J.-C. Vial, R. Farion, L. Rocas, C. R´emy, T. Fadlallah, E. Brauer,A. Bravin, J. Laissue, H. Blattmann, and B. van der Sanden, In vivo two-photon microscopystudy of short-term effects of microbeam irradiation on normal mouse brain microvasculature,International journal of radiation oncology, biology, physics , 1519—1527 (2006).[19] E. Br¨auer-Krisch et al., Medical physics aspects of the synchrotron radiation therapies: Mi-crobeam radiation therapy (MRT) and synchrotron stereotactic radiotherapy (SSRT), Physicamedica : PM : an international journal devoted to the applications of physics to medicine andbiology : official journal of the Italian Association of Biomedical Physics (AIFB) , 568—583(2015).[20] P. Lansonneur, V. Favaudon, S. Heinrich, C. Fouillade, P. Verrelle, and L. De Marzi, Simula-tion and experimental validation of a prototype electron beam linear accelerator for preclinicalstudies, Physica Medica , 50 – 57 (2019).[21] M. Jaccard, M. Dur´an, K. Petersson, J. Germond, P. Liger, M.-C. Vozenin, J. Bourhis,F. Bochud, and C. Bailat, High dose-per-pulse electron beam dosimetry: Commissioningof the Oriatron eRT6 prototype linear accelerator for preclinical use, Medical Physics ,863–874 (2018).[22] K. Petersson, M. Jaccard, J.-F. Germond, T. Buchillier, F. Bochud, J. Bourhis, M.-C. Vozenin,and C. Bailat, High dose-per-pulse electron beam dosimetry — A model to correct for the ionrecombination in the Advanced Markus ionization chamber, Medical Physics , 1157–1167 , 11 –17 (2019).[24] N. M. Esplen, M. S. Mendonca, and M. Bazalova-Carter, Physics and biology of ultrahighdose-rate (FLASH) radiotherapy: a topical review, Physics in Medicine and Biology (2020)., 11 –17 (2019).[24] N. M. Esplen, M. S. Mendonca, and M. Bazalova-Carter, Physics and biology of ultrahighdose-rate (FLASH) radiotherapy: a topical review, Physics in Medicine and Biology (2020).