Journal of X-Ray Science and Technology 21 (2013) 507–514 DOI 10.3233/XST-130399 IOS Press

507

Range verification of proton radiotherapy with prompt gamma rays Andy Lau, Yong Chen and Salahuddin Ahmad∗ Department of Radiation Oncology, Peggy and Charles Stephenson Cancer Center, The University of Oklahoma Health Sciences Center, Oklahoma City, OK, USA Received 7 January 2013 Revised 23 August 2013 Accepted 31 August 2013 Abstract. In-vivo range verification systems for incident protons recently utilize positron emission tomography (PET) based on the phenomenon of positron-emitting nuclei (PEN). However, recent investigations are suggesting that the range can be verified also from the prompt gamma (PG) photon emissions generated from proton interactions. In this work we investigate using the Geant4 Monte Carlo toolkit the clinical viability of a theoretical sequential detector system to verify the range of protons in proton radiotherapy by the PG method for simple geometries and beam configurations. The results show a correlation between selected emitted PG rays and the incident protons range and suggest that our detector system is capable of in-vivo range monitoring for energies typical of radiation oncology applications. Future work will include implementing more realistic scenarios and optimizing current detector parameters. Keywords: Prompt gamma, Geant4 simulation, radiation therapy

1. Introduction Proton radiotherapy is a treatment modality that offers unique advantages as compared to photon therapy in radiation oncology applications. These advantages arise due to the unique dose-distribution produced from the interactions of the incident protons with the target material. The dose-distribution produced from a mono-energetic beam of protons is quasi-constant throughout the majority of the proton’s path up until the last few millimeters of their range where the dose sharply increases to the maximum called the Bragg Peak with little to no dose being absorbed past this range. Clinically, these advantages offer better dose conformality of the tumor, minimizes the dose to the intervening tissue, and spares healthy tissue distal to the treatment site. In addition to the dose-distribution, protons deliver a more potent biological effect (∼ 10% more efficient at producing lethal cellular damage) [1]. Since the proposal for using protons for radiation therapy applications in 1946 by Robert Wilson [2], more than 83,000 patients have been treated with protons and a combined total of more than 96,000 patients for both protons and heavier ions as of 2011 [3]. ∗

Corresponding author: Salahuddin Ahmad, Department of Radiation Oncology, Peggy and Charles Stephenson Cancer Center, The University of Oklahoma Health Sciences Center, 800 N.E. 10th St., OKCC L100, Oklahoma City, OK 73104, USA. Tel.: +1 405 271 3016; Fax: +1 405 271 8297; E-mail: [email protected]. c 2013 – IOS Press and the authors. All rights reserved 0895-3996/13/$27.50 

508

A. Lau et al. / Range verification of proton radiotherapy with prompt gamma rays

Another unique advantage of proton radiation therapy, or indeed carbon ion therapy, is the ability to monitor in vivo the range of the incident particles during a treatment fraction. Knowledge about the incident particle’s range in the patient can reveal information about the delivered dose distribution. Comparisons between the delivered dose and the treatment plan dose-distribution can quantify the accuracy of conforming the dose to the target volume as well as the dose delivered to the healthy intervening tissues. This comparison is indeed important as the uncertainty of the incident proton’s range used in the treatment planning calculation accumulates up to 10–15 mm [4]. The range uncertainty, however, arises primarily from the local changes in the patient anatomy and daily setups [5]. Recently, significant interests are shown in the radiotherapy community for two specific methods of in-vivo monitoring of the incident proton’s range in proton radiation therapy. These methods involve measuring the photons created from the positron-emitting nuclei (PEN) and prompt gamma (PG) emitting from excited daughter nuclei produced during inelastic interactions of the incident proton with the target material. The daughter fragments used in the PEN method de-excites via positron emission with half-lives from ∼ 1 min to ∼ 20 min. For example, the interaction of a proton with a 12 C nucleus p(12 C, pn)11 C results in the emissions of a proton with ejected energy less than the incident proton, a neutron, and the daughter fragment 11 C which is a positron-emitter that decays with 99.75% probability to 11 B with a half-life of ∼1218 s and a positron of maximum energy of ∼ 0.96 MeV. The emitted positron then annihilates with an orbital electron to produce two coincident photons that can be detected with a PET scanner. A comparison of the reconstructed PET image with an image generated from simulations gives information about the distal range of the protons in the patient, and thus the spatial deviations of the delivered-dose distribution. In addition to protons, this method has been clinically used for 12 C-ions treatments for nearly all 440 patients treated at GSI [2]. The PEN technique monitors the resulting daughter fragments 11 C and 15 O. The choice of monitoring these daughter fragments [6] is because: a) They have relatively long half-lives suitable for postradiotherapy counting; b) these are the most abundant positron-emitting chemical species in the human body; and, most importantly, c) optimized detector systems for measuring the resulting annihilation photons already exist. Monitoring the range of the incident protons based on the prompt gamma emissions from various daughter nuclei has been very recently investigated and shown to be clinically viable [7]. A prompt gamma emission is an alternative de-excitation mechanism for the unstable target nuclei to de-excite by emitting a photon (γ ) with time scales of ∼10−19–10−9 s [4]. For example, the reaction 12 C(p,pγ)12 C leads to the 12 C target nuclei to de-excite promptly via pure energy with a characteristic gamma ray of ∼4.44 MeV [8–11]. Spectrum studies suggest that the prominent characteristic gamma rays from 12 C and 16 O range from ∼4 to ∼6 MeV and the depth-distribution of these de-excitation gamma rays show a maximum in their production in close proximity to the end of the incident proton’s range. Overall, these gamma rays have better counting statistics (about an order of magnitude higher as compared to the production of PEN daughter fragments), and can be detected on the nanosecond time scale suitable for real-time range monitoring [4,7]. Currently many authors are actively researching both the PEN and PG methods for in vivo monitoring of the incident proton’s range. Although both methods are capable of monitoring the range of the incident protons, the PEN method yields results post treatment fraction; while the PG method can yield results in real time. More in-depth discussions of the history, current developments, as well as future outlook on the PEN technique only can be found in a recent review article [2]. Current research into PG monitoring is primarily focused on developing detector systems that are capable of meeting the demands of a clinical

A. Lau et al. / Range verification of proton radiotherapy with prompt gamma rays

509

Fig. 1. Schematic of the setup used for the simulations. The proton beam enters the phantom from vacuum. The shaded top-left block represents a virtual detector. This virtual detector has dimensions of 5 cm × 5 cm × 1 mm and was replicated 500 times to produce one bank of detectors. We used two banks of detectors for sequential requirement. Schematic is not drawn to scale.

verification system [4,5,8,10,12–15]. These include the ability to detect the high-energy gamma rays with sufficient efficiency and process the information quickly enough to track the range of the incident protons in-real time. In this work we investigate the clinical viability of a detector system based on sequential discrimination for monitoring the range of incident protons based on the detection of prompt gamma rays. The detector system is simulated with the Geant4 Monte Carlo Toolkit [16,17] in order to study features of the emitted prompt gamma radiation as well as correlate the PG signal to the range of the incident protons. 2. Materials and methods This work models a sequential detector system and concentrates on characterizing both the emitted and registered prompt gamma spectra simulated by the Geant4 Monte Carlo toolkit version 9.4.9 p02. According to the Geant4 philosophy the entire simulation environment must be fully described by the user. These include the beam parameters, geometry and material of the target and detectors, the sequential detection logic, and physical interactions. The interaction of the pencil like incident proton beam of energy 110 MeV (without energy or spatial spread) with the target phantom was used in the simulation. The dimension of the phantom was 5 cm × 5 cm × 50 cm and consisted of PMMA. The chemical composition, the density, and the ionization potential of PMMA is C5 H8 O2 , 1.18 g/cm3 , and 68.5 eV, respectively. A virtual detector (volume in vacuum) was constructed with dimensions of 5 cm × 5 cm × 1 mm. This detector was replicated 500 times to form a bank of virtual detectors with dimensions of 5 cm × 5 cm × 50 cm. A lower bank of virtual detectors was placed 1 cm above the target phantom while another bank (upper bank) of virtual detectors was placed 1 cm above the lower bank. Our simulation records various parameters associated with all secondary particles produced in the simulation environment. We define the quantity γ 1 as any photon produced in the phantom that passed

510

A. Lau et al. / Range verification of proton radiotherapy with prompt gamma rays

Fig. 2. The prompt gamma spectrum is shown for 800 million incident protons onto the PMMA phantom. The vertical dashed lines indicate the energy selection window used in the prompt gamma analysis for the photon emitted from the 12 C and 16 O.

through the upper bank of detectors (independent from which lower detector segment the photon traversed). The quantity γ 2 is defined as the photons that sequentially pass through a lower detector and the corresponding top detector segment directly above it. The quantity of γ 2 reflects prompt gammas being filtered by the restrictive sequential requirements reflecting suppressed background (γ 1 > γ 2) that yields localization information about the production point. A schematic of our setup is shown in Fig. 1. Our investigation used the built-in physics list “QGSP_BIC_HP”. The choice for this physics list came from previous experience from our institution [1,6,18,19] as well as an investigation from Bom et al. [14]. The list consisted of electromagnetic as well as nuclear interactions. The electromagnetic interactions were described by the G4EmStandardPhysics and EmExtraPhysics packages. A cut value of 700 μm was used with resulting energies in PMMA for photons, electrons, positrons, and protons of 2.40, 303.52, 297.46, 70 keV respectively. The nuclear interactions were separated into elastic and inelastic categories described by the G4HadronElasticPhysicsHP as well as HadronPhysicsQGSP_BIC_HP and G4IonBinaryCascadePhysics packages, respectively. Decays are also described in the G4DecayPhysics package. 3. Results and discussion 3.1. Emitted prompt gamma spectrum Prompt gamma spectra γ 1 and γ 2 arising out of the interaction of 800 million 110 MeV incident protons onto PMMA are shown in Fig. 2. These spectra display three dominant peaks of higher energy (> 4 MeV), a peak centered around 511 keV due to the annihilation photons, and few other peaks below 4 MeV. The 511 keV photons production location does not necessarily originate from the incident proton path and thus were not used to monitor the range as suggested in a recent study [14]. We have concentrated in this study only on the higher energy peaks originating from the de-excitation of carbon

A. Lau et al. / Range verification of proton radiotherapy with prompt gamma rays

511

Table 1 The incident proton energy (in MeV) or SOBP widths (in mm), the range (in mm) as defined by the depth of the distal 90% dose, the depth (in mm) corresponding to the 50% distal edge of the prompt gamma produced from excited daughter nuclei 12 C and 16 O, and the differences (in mm) between the depths of D90% and the 50% levels are tabulated Energy or SOBP 90 MeV 95 MeV 100 MeV 105 MeV 110 MeV 5 mm SOBP 10 mm SOBP 15 mm SOBP 20 mm SOBP

D90% (mm) 55 60 66 73 78 78 78 78 78

C50% (mm) 57 61 68 74 78 77 78 77 77

O50% (mm) 59 65 70 77 82 82 81 82 82

ΔC (mm) −2 −1 −2 −1 0 1 0 1 1

Δ0 (mm) −4 −5 −4 −4 −4 −4 −3 −4 −4

Fig. 3. Prompt gamma counts per proton as a function of depth for 800 million 110 MeV protons incident onto a PMMA phantom.

(∼ 4.4 MeV) and oxygen (∼ 6 MeV) produced by the interaction of the incident proton on PMMA; and thus are suitable for range monitoring. The amount of γ 2 counts registered is drastically reduced compared to γ 1 (roughly a factor of 10) due to sequential restriction suppressing background and removing large angled scattered gammas. 3.2. In vivo range monitoring using prompt gamma rays Of the three high energy prompt gamma peaks, we have studied only two: a) gammas detected in the energy range 4400 ± 187.2 keV emitted from 12 C; and b) gammas detected in the energy range 6050 ± 257.4 keV emitted from 16 O. The yields of these gammas as a function of proton interaction depths are investigated to determine the range of the incident protons. Figure 3 shows prompt gamma counts per incident proton as a function of interacting depth for 800 million 110 MeV incident protons on PMMA. The normalized dose distribution including the Bragg Peak is also shown in the figure.

512

A. Lau et al. / Range verification of proton radiotherapy with prompt gamma rays

Fig. 4. Comparison between the PEN yields [6] and the PG yields for incident 110 MeV protons onto a PMMA phantom.

Fig. 5. SOBP produced from 110-95 MeV protons incident onto a PMMA phantom. The prompt gamma from the daughter nuclei 12 C and 16 O are also plotted.

The range of the incident protons is determined by the 50% level of the emitted prompt gamma curves. In Table 1, C50% and O50% indicates the depth (in mm) of the prompt gamma yield at the distal 50% level for the photons arising from 12 C and 16 O, respectively. ΔC and ΔO (in mm) represents the differences between the proton range (as defined by D90% ) and C50% and O50% as calculated in this study. This Δc and Δo values will be added to the measured prompt gamma C50% and/or O50% to determine the proton beam range.

A. Lau et al. / Range verification of proton radiotherapy with prompt gamma rays

513

The depths corresponding to the maximum prompt gamma counts are in close proximity to the range of the protons as determined by the Bragg Peak. A comparison between the PEN yields acquired from 4 million primary protons as described in our previous publication [6] and the PG yields (this study) for incident 110 MeV protons onto a PMMA phantom is shown in Fig. 4. We note that the PEN yields have their maximum and distal 50% yield more superficially than the incident particles Bragg Peak. This is clearly different compared to the PG yields that show their maximum in closer proximity to the Bragg Peak. The 50% distal PEN yields also lie shallower to the Bragg peak whereas to the 50% distal PG yields lie deeper with respect to the Bragg Peak. Several Spread-Out Bragg Peaks (SOBPs) have been constructed with the maximum proton energy of 110 MeV incorporating different modulation widths (0.5 cm to 2.0 cm). The D90% of all SOBPs considered in this study aligned well with the D90% of the pristine Bragg peak of 110 MeV proton beam and ΔC, ΔO as defined in Table 1 are independent of the SOBP modulation width. The prompt gamma curve for a typical SOBP is shown in Fig. 5. 4. Conclusions This investigation focused on the simulation of prompt gamma spectrum produced from 110 MeV proton interactions in the target PMMA phantom and registered in a sequential detector system. The findings of this work suggest that the range of the incident protons can be measured in vivo with a suitable detector system utilizing the prompt gamma technique. More realistic simulations by implementing and optimizing various detector materials, detector spacing, timing discrimination (for real time monitoring), etc. are now being planned. References [1]

Y. Chen and S. Ahmad, Empirical model estimation of relative biological effectivness for proton beam therapy, Radiat Prot Dosim 149 (2012), 116–123. [2] D. Schardt, T. Elsässer and D. Schulz-Ertner, Heavy-ion tumor therapy: Physical and radiobiological benefits, Reviews of Modern Physics 82 (2010), 383–425. [3] PTCOG-http://ptcog.web.psi.ch/Archive/Patientstatistics-updateMar2012.pdf. [4] J. Smeets et al., Prompt gamma imaging with a slit camera for real-time range control in proton therapy, Physics in Medicine and Biology 57 (2012), 3371. [5] S.W. Peterson, D. Robertson and J. Polf, Optimizing a three-stage Compton camera for measuring prompt gamma rays emitted during proton radiotherapy, Physics in Medicine and Biology 55 (2010), 6841. [6] A. Lau, Y. Chen and S. Ahmad, Yields of positron and positron emitting nuclei for proton and carbon ion radiation therapy: A simulation study with GEANT4, Journal of X-Ray Science and Technology 20 (2012), 317–329. [7] M. Moteabbed, S. España and H. Paganetti, Monte Carlo patient study on the comparison of prompt gamma and PET imaging for range verification in proton therapy, Physics in Medicine and Biology 56 (2011), 1063. [8] J.C. Polf et al., Measurement and calculation of characteristic prompt gamma ray spectra emitted during proton irradiation, Physics in Medicine and Biology 54 (2009), N519. [9] J.M. Verburg, H.A. Shih and J. Seco, Simulation of prompt gamma-ray emission during proton radiotherapy, Physics in Medicine and Biology 57 (2012), 5459. [10] J.C. Polf et al., Measurement of characteristic prompt gamma rays emitted from oxygen and carbon in tissue-equivalent samples during proton beam irradiation, Physics in Meidicne and Biology 58 (2013), 5821. [11] U. Oelfke et al., Proton dose monitoring with PET: Quantitative studies in Lucite, Physics in Medicine and Biology 41 (1996), 177–196. [12] D. Robertson et al., Material efficiency studies for a Compton camera designed to measure characteristic prompt gamma rays emitted during proton beam radiotherapy, Physics in Medicine and Biology 56 (2011), 3047. [13] B.-H. Kang and J.-W. Kim, Monte Carlo Design Study of a Gamma Detector System to Locate Distal Dose Falloff in Proton Therapy, IEEE Transactions on Nuclear Science 56 (2009), 46–50.

514 [14] [15] [16] [17] [18] [19]

A. Lau et al. / Range verification of proton radiotherapy with prompt gamma rays V. Bom, L. Joulaeizadeh and F. Beekman, Real-time prompt gamma monitoring in spot-scanning proton therapy using imaging through a knife-edge-shaped slit, Physics in Medicine and Biology 57 (2012), 297–308. M. Frandes, V. Maxim and R. Prost, A Tracking Compton-Scattering Imaging System for Hadron Therapy Monitoring, IEEE Trans Nucl Sci 57 (2010), 144–150. S. Agostinelli et al., Geant4, a simulation toolkit, Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment 506 (2003), 250–303. J. Allison et al., Geant4 developments and applications, Nuclear Science, IEEE Transactions on 53 (2006), 270–278. Y. Chen, Nuclear Interactions and Relative Biological Effectiveness in Proton Radiation Therapy: A Simulation Study with GEANT4. Ph.D. dissertation. University of Oklahoma Health Science Center. Oklahoma City, OK, 2010. Y. Chen and S. Ahmad, Evaluation of inelastic hadronic processes for 250 MeV proton interactions in tissue and iron using Geant4, Radiat Prot Dosim 136 (2009), 11–16.

Copyright of Journal of X-Ray Science & Technology is the property of IOS Press and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.

Range verification of proton radiotherapy with prompt gamma rays.

In-vivo range verification systems for incident protons recently utilize positron emission tomography (PET) based on the phenomenon of positron-emitti...
1MB Sizes 0 Downloads 0 Views