GEANT4 Simulation of Photoneutron Spectrum from Medical Linear Accelerator
Round 1
Reviewer 1 Report
This paper shows the photoneutron spectrum simulated with GEANT4. Radiation therapy is a method of treating cancer. But excessive radiation can also damage organs. The paper obtains the Percentage Depth Dose of different depths through simulation. Help study the effects of radiation on the human body. This is the meaning of this paper. This paper includes an introduction to the significance of the research, research methods and materials, research results, and a discussion of the results. The logic of the paper is reasonable.
Advantages of paper:
1. A large number of relevant research are cited to improve the credibility of the results of the paper.
2. The formula derivation and analysis of simulation results are very detailed.
3. Mutual verification of simulation and measurement has been achieved.
Disadvantages of paper:
1. The paper is not vivid enough. Geant4 simulations produce geometric images. Images generated by Geant4 can be included in the paper. In addition, the measurements mentioned in the paper do not give physical drawings or sketches of the equipment. This increases the difficulty of understanding the paper.
2. As mentioned in the paper the geometry of the linac head is unknown. For different linacs, the simulation results should be different.
3. The reference on line 111 is missing.
In short, this paper has a certain reference value.
Author Response
We would like to express our gratitude for the time and efforts that you spent reviewing our work. Your comments and questions were extremely helpful in increasing the quality of our research. We have submitted a revised version of the manuscript, and we hope that it meets your expectations.
We would also like to address the points raised in your review. We carefully considered all your comments and made the necessary changes to improve the clarity and rigor of our study.
Comment: “The paper is not vivid enough. Geant4 simulations produce geometric images. Images generated by Geant4 can be included in the paper. In addition, the measurements mentioned in the paper do not give physical drawings or sketches of the equipment. This increases the difficulty of understanding the paper.”
We agree with this remark. The scheme of the simulated model, shown in Fig.1 of our paper is GEANT4 visualization image, just inverted to be more clarity. We added the illustration of the experiment with tantalum plates (fig 2). The SSD measurements were carried out by the standard method according to TRS-398. We added the sentence “The measurement was carried according to TRS-398” to the text in line 110.
Comment: “As mentioned in the paper the geometry of the linac head is unknown. For different linacs, the simulation results should be different.”
It is right. As we mentioned in Discussion, we had requested the real linac’s head (line 258) construction from manufacturers. That could help us to estimate uncertainties included in head construction.
Comment: “The reference on line 111 is missing.”
Thank you for your comment, we fixed the reference.
Reviewer 2 Report
This study reports about the photoneutron spectra produced by photonuclear reactions in the head of medical linear accelerator. The data are useful for considering the exposures, however, there are unclear points and logical problem. They should be clearly described and revised.
1. Introduction
P2 L55, 60, 65
The author mentions that Monte Carlo simulations are accurate. Actually, Monte Carlo methods can calculate the details which measurements cannot performed due to the physical limitations. However, that does not always mean accurate. For example, Fig 1 shows a simplified model and the calculated results might not be accurate due to the simplification.
If insisting really accurate, the evidence should be described. In addition, the meanings of “accuracy” and “precision” should be clearly distinguished and discussed.
2. Materials and Methods
P3 L111
The reference number should be written in [?].
P5 L173
wR has been used as radiation weighting factor defined by the ICRP. Why is k used in this manuscript?
P5 L179
The radiation weighting factors of ICRP Publication 60 are shown in Table 1. However, the ICRP recommended the use of revised factors shown in ICRP Publication 103. It would be better to mention both publications and the differences.
3. Results
P6 Figure 2
The measured and calculated values are in good agreement at depths greater than 1 cm, but differ significantly at depths less than 1 cm. The reason should also be mentioned.
4. Discussion
P8 L233-234
The author says “it is important to calculate the photoneutron spectrum for each linac individually.” Generally speaking, it is right. However, to obtain each linac head configuration data would be practically difficult. If the differences among varied linacs are big, they should be considered. If not, the necessity for each calculation is not so important.
P9 L249
It would be more difficult to model the components of the treatment room than the linac. The uncertainties included in the Monte Carlo simulation itself, in the modelling, in the materials of the components and so on should be mentioned qualitatively and quantitatively.
P9 L272-274
The calculation speed of GEANT4 and MCNP were compared. As the Monte Carlo simulation code, FLUKA and PHITS have also been used as typical codes. Are there any comparisons among them in terms of the calculation time?
5. Conclusions
P10 L331, 337
The author concluded that GEANT 4 simulation can perform high or higher accuracy calculations. However, as Fig 4 shows, the simulation cannot reproduce the measurement data due to the simplified model as the author mentioned in the Discussion. If so, accuracy level of the simulation results would greatly depend on the modelling. The conclusion should be consistent with the Results and Discussion.
Author Response
We would like to express our gratitude for the time and efforts that you spent reviewing our work. Your comments and questions were extremely helpful in increasing the quality of our research. We have submitted a revised version of the manuscript, and we hope that it meets your expectations.
We would also like to address the points raised in your review. We carefully considered all your comments and made the necessary changes to improve the clarity and rigor of our study.
Comment: The author mentions that Monte Carlo simulations are accurate. Actually, Monte Carlo methods can calculate the details which measurements cannot performed due to the physical limitations. However, that does not always mean accurate. For example, Fig 1 shows a simplified model and the calculated results might not be accurate due to the simplification. If insisting really accurate, the evidence should be described. In addition, the meanings of “accuracy” and “precision” should be clearly distinguished and discussed.”
We fully agreed with this remark and after we read it, we understand that our terms were not such clear. We changed the sentence in lines 55-57 as “Monte Carlo (MC) simulations seems to be a more preferable strategy for estimating the effective photoneutron dose in the patient because it allows calculating details which measurements cannot performed due to the physical limitations.”
Comment:“The reference number should be written in [?].”
Thank you for your comment, we fixed the reference.
Comments: “wR has been used as radiation weighting factor defined by the ICRP. Why is k used in this manuscript?” and “The radiation weighting factors of ICRP Publication 60 are shown in Table 1. However, the ICRP recommended the use of revised factors shown in ICRP Publication 103. It would be better to mention both publications and the differences.”
We fully agreed with these remarks. We change “k” to “wR” in our paper and added the values of radiation weight factors in agreement to ICRP 103 to the table 1.
Comment: “The measured and calculated values are in good agreement at depths greater than 1 cm, but differ significantly at depths less than 1 cm. The reason should also be mentioned.”
We fully agreed with this remark. We added in the paper the next sentences “In the depth range from 0 to 10 mm, the PDD value significantly depended on measurement conditions. That’s why the values of this range weren't considered.” in lines 193-194.
Comment: “The author says “it is important to calculate the photoneutron spectrum for each linac individually.” Generally speaking, it is right. However, to obtain each linac head configuration data would be practically difficult. If the differences among varied linacs are big, they should be considered. If not, the necessity for each calculation is not so important.”
It is right. As we mentioned in Discussion, we had requested the real linac’s head (line 258) construction from manufacturers. That could help us to estimate uncertainties included in head construction. We add the sentence “We also planning to simulate different linac heads and estimate the differences between them. If the differences would not be essential, it would be possible to use general model of linac head for estimation of photoneutron spectrum.” In lines 264-266.
Comment: “It would be more difficult to model the components of the treatment room than the linac. The uncertainties included in the Monte Carlo simulation itself, in the modelling, in the materials of the components and so on should be mentioned qualitatively and quantitatively.”
We fully agreed with this remark. We added the sentence “However the simulating materials of the wall’s components, their shapes and positions will add the additional uncertainties in our model” in lines 319-320.
Comment: “The calculation speed of GEANT4 and MCNP were compared. As the Monte Carlo simulation code, FLUKA and PHITS have also been used as typical codes. Are there any comparisons among them in terms of the calculation time?”
We agreed with this remark. We have no information about FLUKA computation time in common. Some works compare calculation time for FLUKA/GEANT4 and FLUKA/MCNP within the same task but not for photoneutron calculations. The commonly used software for similar simulation was MCNP, so we compared GEANT4/MCNP computational time in our paper. We inserted additional information about PHITS in lines 279, 282-283.
Comment: “The author concluded that GEANT 4 simulation can perform high or higher accuracy calculations. However, as Fig 4 shows, the simulation cannot reproduce the measurement data due to the simplified model as the author mentioned in the Discussion. If so, accuracy level of the simulation results would greatly depend on the modelling. The conclusion should be consistent with the Results and Discussion.”
We agree with this remark. The sentence on lines 340-341 was changed from “It is expected that it may potentially provide a higher accuracy of calculations of the effective dose from photoneutrons but the model used needs to be improved and optimized” to “The model used needs to be improved and optimized”.
