|Home | About | Journals | Submit | Contact Us | Français|
Proton therapy offers low integral dose and good tumor comformality in many deep-seated tumors. However, secondary particles generated during proton therapy, such as neutrons, are a concern, especially for passive scattering systems. In this type of system, the proton beam interacts with several components of the treatment nozzle that lie along the delivery path and can produce secondary neutrons. Neutron production along the beam's central axis in a double scattering passive system was examined using Monte Carlo simulations. Neutron fluence and energy distribution were determined downstream of the nozzle's major components at different radial distances from the central axis. In addition, the neutron absorbed dose per primary proton around the nozzle was investigated. Neutron fluence was highest immediately downstream of the range modulator wheel (RMW) but decreased as distance from the RMW increased. The nozzle's final collimator and snout also contributed to the production of high-energy neutrons. In fact, for the smallest treatment volume simulated, the neutron absorbed dose per proton at isocenter increased by a factor of 20 due to the snout presence when compared with a nozzle without a snout. The presented results can be used to design more effective local shielding components inside the treatment nozzle as well as to better understand the treatment room shielding requirements.
Protons were first proposed for radiotherapy in 1946 (Refs. 1 and 2). The principal advantage of protons is the physical dose distribution compared to conventional radiotherapy as well as the rapid dose falloff at the end of the proton range and a sharp lateral penumbra. These features lower the integral dose and the dose to organs at risk.3,4
Two principal proton delivery systems are currently used: active scanning systems and passive scattering systems. In passive scattering systems, the proton beam interacts with energy degraders and scattering foils in the treatment nozzle to create a spread-out Bragg peak (SOBP). The energy degraders modulate the penetrability of the beam, and the scattering foils spread the beam laterally.5 Collimators and boluses can also be used to shape the beam further. In contrast, active scanning systems have fewer beam-shaping devices in the treatment nozzle because beam spots are magnetically scanned throughout the treatment volume. The beam's interaction with shaping devices will produce secondary particles such as neutrons. In fact, some studies have presented neutron dose differences of one order of magnitude between active and passive systems, suggesting that passive systems produce much higher neutron doses than active systems.6,7 However, more recent studies have shown that the difference in neutron production between the two systems is smaller than previously reported.8–10
In this study, we provide a progress report on estimates of the neutron production in a double scattering system. This work expands on earlier studies by using actual proton fluence weights for a typical SOBP to better understand neutron production within the treatment nozzle.11 In this system, the range modulation wheel (RMW) is made of a combination of high-Z and low-Z materials, which simultaneously modulate the beam's range and spread the beam laterally. The second scatterer, which further spreads the beam laterally, is also made of low- and high-Z materials.12,13 The overall purpose of this work was to determine the neutron contribution from each component in the nozzle that shapes the proton beam. To accomplish this, we used Monte Carlo simulations to model the passive scattering beam nozzle as well as the interactions of the proton beam that lead to the production of secondary neutrons.
The treatment nozzle's major components were modeled following the design of the double scattering passive system used at the Proton Therapy Center at M. D. Anderson Cancer Center.14,15 The model includes a nonproprietary RMW similar, but not identical, to that of the Hitachi IBA system.14 It also includes a secondary scatterer, collimators, and snout. No beam-monitoring components nor additional range-modulation devices were included. Simulations were performed using the Monte Carlo code MCNPX developed by Los Alamos National Laboratory, which has been extensively used in proton therapy applications.16–22 In the simulations, a proton beam was transported through the nozzle and interacted with the nozzle's components before being completely stopped in a water phantom. As part of the beam interactions with the nozzle components, neutrons were produced and were also transported. For neutrons with energies ranging from thermal to 150 MeV, nuclear interaction probabilities were obtained from the Evaluated Nuclear Data File (ENDF) library; for neutrons with energies above 150 MeV, cross sections were calculated using physics nuclear interaction models (specifically, the Bertini intranuclear cascade model).23,24
In this work, two geometries of the same treatment nozzle were simulated (see Figs. 1 and and2).2). First, a 200-MeV proton beam traveled through the treatment nozzle, interacting with the RMW, second scatterer, and a collimating system before stopping in a water phantom (Fig. 1). The cylindrical phantom's face was at isocenter and had a diameter and thickness of 300 mm. The RMW was made of a lamination of high- and low-Z materials, and the second scatterer was made of low- and high-Z materials. The high-Z material was lead, and the low-Z material was polymethyl methacrylate (PMMA). The characteristics of the RMW and double scatterer used in these simulations have been described elsewhere.11 The collimators were made of copper and, for clarity, are classified as primary, secondary, tertiary, and final collimators. The primary collimator had a 70-mm thickness, the secondary, 120 mm, and the tertiary and final collimators were both 60 mm thick. The snout, which is another component of the collimating system, was located at the treatment nozzle's end and had a variable size aperture made of copper. Further nozzle dimensions are presented in Fig. 2.
The snout's square aperture along with the thickness of the RMW were varied to irradiate three treatment volumes in the water phantom: 50 × 50 × 50 mm3, 100 × 100 × 100 mm3, and 150 × 150 × 150 mm3. By varying the RMW thicknesses, as well as the beam weights for each thickness, a uniform SOBP was obtained. The same secondary scatterer was used for all the treatment volumes simulated. To score the neutron fluence as a function of radial distance from the beam's central axis, several concentric cylinders with diameters ranging from 5 to 200mm and a thickness of 1 mm were used as detecting volumes. These were placed at four locations along the beam's central axis: 200 mm downstream of the RMW junction between the high- and low-Z materials, 100mm downstream of the second scatterer material junction, as well as 100mm upstream and downstream of the snout. The detecting volumes were simulated as vacuum since their only purpose was to score the neutrons crossing that region.
For the second set of simulations, the water phantom and the detecting surfaces along the nozzle's central axis were removed (Fig. 2). Instead, cylindrical surfaces of 125 mm in radius and 1 mm thick were place at different angles and at 3.06 m with respect to the RMW. These surfaces were used to determine the variation in neutron absorbed dose per proton around the nozzle due to the interaction of the beam with the individual beam-shaping components. The simulations were first run for the 200-MeV proton beam interacting only with the RMW; subsequently, the remaining nozzle components were added one at a time.
Figures 3a, 3b, and 3c show the neutron fluence as a function of the radial distance for each one of the four positions studied. Neglecting the neutrons scored at the volume just upstream of the snout, whose behavior will be explained later, the highest neutron fluence was found downstream of the RMW. After the RMW, the neutron fluence diminished as distance from the RMW increased. On average, the neutron fluence was ~30% higher for the largest treatment volume (150×150×150 mm3), for which the largest total amount of high- and low-Z material was used, than for the smallest one at the location immediately downstream of the RMW. Figure 4a shows neutron energy spectra at 20- to 25-mm radial distance from the central axis just downstream of the RMW and the snout for the smallest and largest treatment volumes. The maximum fluence occurred just downstream of the RMW and is the largest for the largest treatment volume, although at higher energies above 175 MeV, it is close to that of the smallest treatment volume. This behavior was seen up to ~100 mm from the beam's central axis; beyond this distance, the neutron fluence and energy decreased sharply (Fig. 4b). On the other hand, after the snout, the fluence of energetic neutrons was slightly higher than before as the radial distance increases (Fig. 4b). The reason for this behavior is that high-energy protons were collimated by the snout, generating high-energy neutrons.
A large neutron fluence was observed just upstream of the snout. This is attributed to backscatter effects. To determine if this behavior is in fact because of backscatter, the copper's density at the snout was decreased by 20, 50, and 70%. This was done only for the smallest treatment volume for which the largest amount of collimating material is present in the beamline. As the density was decreased, the neutron fluence upstream of the snout decreased dramatically (Fig. 5a). On the other hand, the opposite behavior was seen downstream of the snout, suggesting that although the snout was a source of neutrons, it also shielded some of them (Fig. 5b). Also, an increase in fluence was observed after a radial distance of 60 mm for the case where the collimator had its full density. This behavior could be attributed to scattered neutrons from the final collimator and snout. Note that the water phantom is present in all of these simulations, so there will be some contribution to the neutron fluence by this.
Figures 6a and 6b show the neutron absorbed dose per proton around the treatment nozzle for the smallest and largest treatment volume configurations used for the first set of simulations. These simulations did not have the water phantom present because their objective was to calculate the contribution to the neutron fluence only by the nozzle components. The neutron absorbed dose was calculated by multiplying the neutron fluence at each surface by the appropriate kerma coefficients obtained from ICRU Report 63 (Ref. 25). The neutron absorbed dose is due only to the interaction of the proton beam with the nozzle shaping components. It was observed that the neutron dose increases for all angles when the snout is present in the geometry. The dose increases by a factor of 20 at isocenter for the smallest treatment volume and by ~10 for the largest one. A slightly higher factor was found at 5 deg. This was ~30 and 20 for the smallest and the largest fields, respectively. Above 5 deg, the neutron absorbed dose starts to decrease, although it is still higher at all angles than in simulations where the snout was not present. The higher neutron absorbed dose could be in part due to the proximity of the detecting surfaces to the snout. In addition, this is the highest for the smallest treatment volume, because the largest amount of collimating material in the snout is present in the beam path for this configuration. Overall, the detecting cylinders are closer to the snout than to the RMW and secondary scatterer.
Using Monte Carlo simulations, we found that the RMW was the highest contributor to neutron production within the treatment nozzle. This result was expected, because the RMW is the nozzle component that has the largest amount of material in the proton beam path. These results may vary from institution to institution because this depends on the type of material used for the RMW. Specifically, the RMW simulated had simultaneously high- and low-Z material, while for some institutions, the RMW is made only from low-Z material and the high-Z material is placed separately along the nozzle. We also found that the neutron fluence decreased as a function of distance from the RMW. The maximum fluence of high-energy neutrons was found at radial distances closer to the beam's central axis, indicating that neutron production is forward directed. Thus, as distance from the beam's central axis increased, the neutron fluence decreased. The snout and collimator also contributed to the neutron fluence. In fact, in the simulations discussed in this paper, the snout dominates the neutron absorbed dose per proton for all angles near isocenter. Our results indicate that the RMW is the primary source of neutrons inside the treatment nozzle, while the snout dominates the neutron dose at isocenter and those locations closer to it. Therefore, it might be possible to decrease the neutron production in the nozzle as well as the required room shielding by adding the appropriate combination of shielding in the treatment nozzle.
This work was supported in part by the National Science Foundation (NSF) and the University of Wisconsin–funded Grid Laboratory (GLOW) computer cluster (NSF award 0320708). Funding was provided by the National Institute of Health Ruth L. Kirschstein National Research Service Award Individual Predoctoral Fellowship Program (award F31 CA119943-03) and by Northern Illinois University through a subcontract of the U.S. Department of Defense (contract W81XWH-08-1-0205) (W. D. Newhauser).
A. Pérez-Andújar, University of Wisconsin, School of Medicine and Public Health, 750 Highland Avenue, 4111 HSLC, Madison, Wisconsin 53705-2221.
W. D. Newhauser, The University of Texas M. D. Anderson Cancer Center, Department of Radiation Physics, Unit 94, 1515 Holcombe Boulevard, Houston, Texas 77030.
P. M. DeLuca, Jr., University of Wisconsin, School of Medicine and Public Health, 750 Highland Avenue, 4111 HSLC, Madison, Wisconsin 53705-2221.