Rudi Labarbe (Belgium)
IBA R&DAuthor Of 7 Presentations
Welcome & Introduction
Q&A
CREATING CLINICALLY COMPETITIVE FLASH PROTON THERAPY TREATMENT PLANS
Abstract
Background and Aims
FLASH proton therapy (FLASH-PT) is beginning the transition into clinical practice. The beam delivery, dose rates, and fractionation schemes for FLASH-PT differ from standard radiotherapy. Hence, no guidelines exist regarding the development of treatment plans for this novel technique. This study aims to determine if FLASH-PT treatment plans can be developed and if the in silico results are comparable to the treatment plans produced for standard radiotherapy.
Methods
FLASH-PT and IMPT treatment plans were created using a novel research version of the MIROpt TPS, developed by Ion Beam Applications SA from the open source version of UCLouvain, for nine patient cases of bone (3), brain (3), and lung metastases (3), previously clinically treated with 3DCRT or VMAT. A FLASH proton dose rate of ≥ 40 Gy/s was included as an optimisation criterion for each patient case. Treatment plans were compared using dose volume histograms (DVHs), boxplots, and the Wilcoxon Rank Sum Test with a 5% significance level and using DVH parameters V100%, V95%, V50%, D99%, D95%, and D2% for target and body structures.
Results
No significant differences were found between the optimised FLASH-PT plans and the clinical 3DCRT/VMAT and optimised IMPT plans.
Conclusions
The FLASH-PT treatment plans created in this study produced in silico results comparable to those of clinically competitive treatment plans. Future work involves the verification of the calculated dose against delivered dose and dose rate, to ensure that the produced treatment plans can be delivered safely and accurately, confirming the feasibility of the clinical implementation of conformal FLASH-PT.
MODEL STUDIES OF THE ROLE OF OXYGEN IN THE FLASH EFFECT
Abstract
Background and Aims
Convincing evidence has been provided that lowering the mean dose rate below 30 Gy/s (Montay-Gruel et al. Radiother Oncol 124: 365-369, 2017) or increasing the oxygen tension through carbogen breathing (Montay-Gruel et al. PNAS 116: 10943-10951, 2019) can reduce (if not eliminate) the sparing effect of FLASH in the mouse brain. Three theoretical models have since been proposed to account for the dependence of FLASH on oxygen, namely, (i) Radiation-induced transient oxygen depletion (TOD), (ii) Cell-specific differences in the ability to detoxify and/or recover from injury caused by reactive oxygen species, (iii) Self-annihilation of radicals by bimolecular recombination.
Methods
These theoretical models were confronted to experimental evidence relating to physical chemistry and the response of biological systems to ultrahigh dose rate irradiation in light of the chemical reactions underlying the radiosensitizing effect of O2 from immediate, early processes to late complications.
Results
The radiosensitizing effect of oxygen has long been known to stem from peroxyl radicals (ROO•). Peroxyl radicals areformed by the addition of O2 in its fundamental, triplet state to short-lived carbon-centered radicals (R•) generated upon H• atom abstraction from aliphatic, unsaturated or conjugated substrates (RH).
Conclusions
Any attempt at deciphering the physical-chemical processes that underpin the FLASH effect must take into consideration the fate of these radicals, and in particular the probability of radical recombination. The radiation dose rate emerges as the most important factor. We propose testable procedures to investigate the mechanisms that underly the FLASH effect in normal cells, and differential tumor vs. normal cells responses.
UHDR PROTON BEAM VS. CONVENTIONAL: HYDROGEN PEROXIDE AS FLASH EFFECT SENSOR
Abstract
Background and Aims
Several models assume that delivering radiation at ultra-high dose rates (UHDR) reduces the yields of reactive oxygen species (ROS). Montay-Gruel et al [1] measured a significant decrease of Hydrogen Peroxide (H2O2) after UHDR irradiation compared to conventional dose-rate irradiation (CONV) with electrons. This study aims to verify this assumption in proton beams. We study the UHDR radiation chemistry by the measurement of H2O2 produced by the water radiolysis. Fricke dosimeter is used as dose control and as ROS sensor as well.
Methods
Using ARRONAX facility, we produced proton beams (68MeV) ranging from low (0.25Gy/s, 100Hz, pulse dose rate=6.3Gy/s) to ultra-high dose rates (7500Gy/s, single pulse). Doses ranging from 5Gy to 80Gy were delivered to 1.5ml Eppendorf tubes filled with water. In this investigation, Fricke dosimetry [2] is used to verify the dose for both irradiation modes. The second step is the determination of H2O2 concentrations after irradiation with the Ghormley triiodide method [3].
Results
In this work, we have brought the evidence of the Flash effect by the H2O2 measurement in the two cases: (i) For CONV irradiations, we determined a H2O2 chemical yield close to the literature at 0.9 10-7 mol.J-1 [4], (ii) For UHDR irradiations, this yield is measured to a significant lower value. Observed Fricke values are the same for both modes.
Conclusions
We have revealed a radio-chemical FLASH effect for the proton beam by the H2O2 measurement as radiolytic species during the irradiation of water. Further studies will be performed to determine the beam time structure effect.
TREATMENT PLANNING AND DELIVERY FOR FLASH PROTON THERAPY WITH BRAGG PEAK
Abstract
Background and Aims
We developed a treatment delivery and planning approach to achieve a high dose rate while meeting the initial goals of an IMPT plan for dose conformity.
Methods
First, the proton fluence at the exit of the nozzle was maximized by transmitting beam at maximum energy through the beam line and using a range shifter to match the maximum proton range to the distal edge of the PTV. Spread-out Bragg peaks (SOBP) were obtained by modulating the range of the protons via a patient-specific ridge filter placed between the patient and the nozzle. To calculate the characteristics of this ridge filter, an IMPT plan was first computed to obtain the list of energies and associated weights that the filter should reproduce from beamlets at maximum energy. The filter was then optimized using an analytical model to allow a comparison in acceptable computational time with the reference dose. Finally, the scanning pattern was optimized with respect to local constraints on the dose rate.
Results
A patient-specific ridge filter is shown in Fig. 1. Corresponding SOBPs were simulated by Monte Carlo with the ridge filter inserted in the CT.
Conclusions
We proposed a treatment approach relying on the use of a patient-specific ridge filter to increase the dose rate without compromising the potential of the Bragg peak. The ridge filter and the spot scanning trajectory were optimized based on a standard IMPT plan by using an analytical algorithm inside MIROpt (http://openmiropt.org/), alongside the MCsquare dose engine (http://www.openmcsquare.org/).
PROTON BEAM FLASH ONLINE MONITORING AT ARRONAX CYCLOTRON
Abstract
Background and Aims
The beam monitoring tools used in conventional irradiation may not be suitable in the case of irradiation using ultra-high dose rate > 40 Gy/sec (FLASH conditions). To monitor the beam, a fast detector coupled with a high dynamic range is needed. A photomultiplier tube measuring the nitrogen fluorescence produced by the beam-air interaction
can be a solution (figure below).
Methods
The ARRONAX facility has been used to deliver proton beams (68 MeV) ranging from low (0.25 Gy/s) to very high (30 kGy/s) dose rates. An in-house prototype consisting of a photomultiplier coupled to an air cavity was used. The signal intensity can be adjusted by changing the solid angle. To verify the measured beam charge, several aluminum targets were irradiated and the produced activities of the radioisotope 24Na were measured after the end of the beam. Several beam time structures with a high number of pulses were tested. In the case of a low doses, radiochromic films (for which we have shown a response independent of the dose rate) were used.
Results
The measured beam charge using the photomultiplier tube was found in good agreement with the produced activity (1.5 %) and with the film responses (< 5 %). A wide range of dose rates was successfully monitored in the linear response region of the photomultiplier using multiple detectors.
Conclusions
We have established a new robust noninvasive method to measure doses in Flash irradiations. The next step is to use the same technology to measure beam geometrical characteristics.