Age-Dependent Changes in Effective Dose in Pediatric Brain CT: Comparisons of Estimation Methods

Round 1

Reviewer 1 Report

Dear colleagues!

Radiation exposure in the practice of a pediatrician is an important knowledge necessary for an informed decision on the choice of diagnostic method.

In terms of content, design and structure, your work is a scientific work and can be accepted for consideration.

I have a few clarifying questions about the content and structure,

1. In the introduction it will be useful to find data on the need for brain CT at different ages and on the prevalence of this diagnostic method

2. When choosing a CT scan for research, did you primarily look for healthy people or was there a focus on a specific pathology? It seems to me correct to indicate these circumstances, if they existed, in the criteria for non-inclusion. You also need to indicate decoding errors, because it was carried out by different people at different times

3. It will be useful to present the brain images (CT) in the results in order of comparison by age. The work lacks visibility

4. It will be useful if you write short recommendations for doctors based on the results of your work

5. Unfortunately, I cannot consider the list of references correct, because There are many links that are more than 10 years old. I recommend updating the data or adding new ones, which will strengthen the “relevance” section

Author Response

Comment 1)

In the introduction it will be useful to find data on the need for brain CT at different ages and on the prevalence of this diagnostic method

Reply 1)

We added the following sentence to the 1st paragraph of the Introduction:

“In the dose survey for pediatric CT in the USA, 52.6% of the analyzed examinations were brain CT excluding the scans of the sinus and maxillofacial area (4.3%, 2.8%, 7.0%, and 38.5% for 0, 1, 2–5, and 6-18 years, respectively) [3].”

Comment 2)

When choosing a CT scan for research, did you primarily look for healthy people or was there a focus on a specific pathology? It seems to me correct to indicate these circumstances, if they existed, in the criteria for non-inclusion. You also need to indicate decoding errors, because it was carried out by different people at different times

Reply 2)

The study patients were examined for various clinical indications, as stated in the subsection “2.1. Subjects.” There were no inclusion criteria associated with the basis of referral.

We added the following sentence regarding the inclusion:

“For patients who underwent CT examinations repeatedly, those performed at an interval longer than 1 year were included in analysis.”

We did not experience decoding errors.

Comment 3)

It will be useful to present the brain images (CT) in the results in order of comparison by age. The work lacks visibility

Reply 3)

We added examples of CT images at 0, 1, 5, and 10 years as new Figure 2.

Comment  4)

It will be useful if you write short recommendations for doctors based on the results of your work

Reply 4)

We added the following sentence to the 2nd last paragraph of the Discussion:

“We recommend the use of the curve method to estimate the ED in relation to age.”

Comment 5)

Unfortunately, I cannot consider the list of references correct, because There are many links that are more than 10 years old. I recommend updating the data or adding new ones, which will strengthen the “relevance” section

Reply 5)

We think that the old references are important in this manuscript. They are a large epidemiologic study, ICRP publication, excellent reviews about automatic exposure control, and a basic paper about organ dose modulation. Original articles about patient head size and effective dose in pediatric brain CT are included because such papers are limited.

We added the following paper as a reference:

“Hauptmann, M.; Byrnes, G.; Cardis, E.; Bernier, M.O.; Blettner, M.; Dabin, J.; Engels, H.; Istad, T.S.; Johansen, C.; Kaijser, M.; et al. Brain cancer after radiation exposure from CT examinations of children and young adults: results from the EPI-CT cohort study. Lancet Oncol. 2023, 24, 45–53. https://doi.org/10.1016/S1470-2045(22)00655-6.”

Reviewer 2 Report

Abstract

In the simple method, the ED was estimated using the ICRP conversion factor for the closest age.

“We also analysed the ED estimated by a radiation dose management system. Although the median DLP at each age increased with age, the median ED estimated by the curve method was highest at 0 years, decreased with age, and then plateaued.” These result sentences are really too vague. Please be more specific in the abstract.

 “The linear method yielded mildly different results. The ED estimated by the simple method or the radiation dose management system showed inconsistent changes with age.” Please give quantitative results.

 Please state the aim of the study clearly in the abstract. Also a conclusion is missing in the abstract.

 Introduction

 “Computed tomography (CT) delivers relatively high radiation doses, and the potentially detrimental effects of CT radiation exposure are a significant concern in modern medicine.” Please add a reference, e.g. to Brenner, An Increasing Source of Radiation Exposure, NEJM 2007.

 Please refer to the new European guidelines on DRL for paediatric CT (IAEA and PiDRL project).

 Please shorten the introduction. E.g. paragraph 4 (lines 53-66) is not relevant and can be omitted. Also paragraph 5 can be largely omitted.

 Materials and Methods

Please specify the inclusion criteria of this study. Now only the exclusion criteria are explicitly given.

 Two 64-detector-row CT scanners with the same specifications were used. Does this mean that 2 identical CT scanners were used (i.e. GE Optima 660)? What was the collimation that was used on this 64-slice scanner?

 Three different methods were used to estimate the ED: a curve method, a linear method, and a simple method. However, also a fourth method is used, the Radimetrics method. Please be consistent and list 4 methods, and also the results of these four methods. What is the rationale for using the equation in line 117 in the curve method? Please provide a reference and theoretical background. The linear method is a simple age interpolation? Please provide a reference for the MC method used in Radimetrics.

Results

 Please add R2 and standard deviations to the fitting parameters a, b and c in line 151.

 Fig 1b. How is the relative conversion factor defined? With respect to the curve method?

 Please add the conversion factors according to the Radimetrics method as shown in fig 7 to fig 1, and omit fig 7.

 Please plot in fig 3 the conversion factors according to the studied 4 methods in one figure with appropriate line colours (e.g. like in fig 1a).

 Fig 4 does not add additional information with respect to fig 2. Please omit.

 From fig 6 it appears as if a different pediatric protocol is used for children up to 1 year. In the age group 0 to 1 year the effective dose seems to rise to a sharp peak at 1 year and then suddenly drops and gradually decreases for older ages. Please comment.

 Discussion

 Again, instead of speaking of 3 methods I think that 4 methods were used in this study.

 Please provide an answer to the research question. Which method is preferred, or does give the most accurate estimation of the effective dose in brain CT of children?

 The discussion section is far too wordy and must be shortened significantly, since a lot of text is simply a repetition of the results. Please provide comparison and discrepancies with other studies. Please compare your results with the DRL for pediatric brain CT.

 “This curve method demonstrated reasonable changes in the ED according to age.” What does reasonable mean? Please quantify. And does this mean that the authors prefer this method above the others. And that this is the proposed method, like stated in the last sentence 333-334?

 The brain is the least sensitive in DLP conversion factor to variation in age as compared to neck, chest and especially abdomen and pelvis (Deak 2010). In this study, however, only CT brain has been included. Please address this limitation in the discussion section.

 References

 At least 5 references are given to own work. Please delete a few.

Moderate editing of English language required

Author Response

Comment 1)

Abstract

In the simple method, the ED was estimated using the ICRP conversion factor for the closest age.

“We also analysed the ED estimated by a radiation dose management system. Although the median DLP at each age increased with age, the median ED estimated by the curve method was highest at 0 years, decreased with age, and then plateaued.” These result sentences are really too vague. Please be more specific in the abstract.

“The linear method yielded mildly different results. The ED estimated by the simple method or the radiation dose management system showed inconsistent changes with age.” Please give quantitative results.

Please state the aim of the study clearly in the abstract. Also a conclusion is missing in the abstract.

Reply 1)

We modified the Abstract to present the aim and conclusion explicitly and to explain the results in more detail. Now, the word count is just below the upper limit.

Comment 2)

Introduction

 “Computed tomography (CT) delivers relatively high radiation doses, and the potentially detrimental effects of CT radiation exposure are a significant concern in modern medicine.” Please add a reference, e.g. to Brenner, An Increasing Source of Radiation Exposure, NEJM 2007.

Reply 2)

We added the reference.

Comment 3)

Introduction

Please refer to the new European guidelines on DRL for paediatric CT (IAEA and PiDRL project).

Reply 3)

We referred to the following guidelines in the original manuscript:

“European Commission. European Guidelines on diagnostic reference levels for paediatric imaging; Radiation Protection 185; Publications Office of the European Union: Luxembourg, 2018.”

Comment 4)

Introduction

Please shorten the introduction. E.g. paragraph 4 (lines 53-66) is not relevant and can be omitted. Also paragraph 5 can be largely omitted.

Reply 4)

We deleted the description about SSDE to shorten the Introduction. Additionally, we deleted two sentences in the 2nd paragraph regarding the DRL. Paragraph 4 explains the effective dose and its calculation from the DLP. Paragraph 5 explains automatic exposure control (AEC). We modulated radiation exposure using AEC, and the behavior of AED determines age-dependent changes in the DLP. Descriptions in paragraphs 4 and 5 are essential in our manuscript.

Comment 5)

Materials and Methods

Please specify the inclusion criteria of this study. Now only the exclusion criteria are explicitly given.

Reply 5)

We added the following sentence regarding the inclusion:

“For patients who underwent CT examinations repeatedly, those performed at an interval longer than 1 year were included in analysis.”

Comment 6)

Two 64-detector-row CT scanners with the same specifications were used. Does this mean that 2 identical CT scanners were used (i.e. GE Optima 660)? What was the collimation that was used on this 64-slice scanner?

Reply 6)

As the reviewer wrote, we used two GE Optima 660 Discovery Edition scanners. There were no differences in specifications between the two scanners.

We described the collimation as "beam width, 10 mm" in the original manuscript, and changed it to "beam collimation, 0.625 mm × 16.”

Comment 7)

Three different methods were used to estimate the ED: a curve method, a linear method, and a simple method. However, also a fourth method is used, the Radimetrics method. Please be consistent and list 4 methods, and also the results of these four methods. What is the rationale for using the equation in line 117 in the curve method? Please provide a reference and theoretical background. The linear method is a simple age interpolation? Please provide a reference for the MC method used in Radimetrics.

Reply 7)

We estimated the ED by three ICRP conversion factor-based methods and one Radimetrics method. In this manuscript, the “four” represents the total number of ED estimation methods (lines 84, 140, and 189 in the original manuscript), and the “three” represents the number of the ICRP conversion factor-based methods (lines 111 and 225 in the original manuscript). To improve readability, we revised the 1st sentence of the last paragraph of “2.3. Estimation of ED” as follows:

“In addition to the three conversion factor-based methods, we analyzed EDs estimated automatically by a radiation dose management system Radimetrics (version 3.4.0; Bayer Medical Care Inc., Indianola, PA, USA) (Radimetrics method).”

The basis of the use of the equation in line 117 was explained in the original manuscript (line 229) as follows:

“The equation form y = a(x + b)^c selected empirically, but not theoretically, achieved successful fitting”

The linear method is mainly based on a simple age interpolation and was explained in the original manuscript (line 122) as follows:

“In the linear method, the conversion factors at 2–4 and 6–9 years (integer years) were calculated by linear interpolation of those at 1 and 5 years and 5 and 10 years, respectively. The conversion factors at 11–14 years were calculated by linear extrapolation of those at 5 and 10 years.”

We added the following reference regarding the ED estimation methods in Radimetrics:

“Bayer Medical Care Inc. CT organ dose calculations in Radimetrics® Enterprise Application by Bayer in Radimetrics operation manual v3.4b. PA, USA, 2021, pp. 219–222.”

Comment 8)

Results

Please add R2 and standard deviations to the fitting parameters a, b, and c in line 151.

Reply 8)

We added the residual standard deviation as an indicator of the quality of fitting to the 1st paragraph of the Results.

Comment 9)

Fig 1b. How is the relative conversion factor defined? With respect to the curve method?

Reply 9)

The definition of the relative conversion factor was described in the original manuscript as follows (line 127):

“The relative conversion factor for the linear or simple method at each age was calculated as the ratio of the conversion factor for the respective method to that for the curve method.”

Comment 10)

Please add the conversion factors according to the Radimetrics method as shown in fig 7 to fig 1, and omit fig 7.

Reply 10)

Radimetrics does not use conversion factors. However, in this study, we demonstrated the ratio of the ED estimated by the Radimetrics to the DLP differed among the 0–0.5-, 0.5–2.5-, 2.5–7.5-, 7.5–12.5-, and 12.5–15-year groups and was almost constant within each age group. This indicates that the Radimetrics method is essentially equivalent to multiplying the DLP by a conversion factor for each age group. This was explained in the 3rd last paragraph of the Discussion.

Comment 11)

Please plot in fig 3 the conversion factors according to the studied 4 methods in one figure with appropriate line colours (e.g. like in fig 1a).

Reply 11)

We modified the figure as indicated by the reviewer.

Comment 12)

Fig 4 does not add additional information with respect to fig 2. Please omit.

Reply 12

Age grouping is usually performed in radiation dose management of pediatric brain CT, and we presented the results of the ED estimation with age grouping in the original Figure 5. The original Figure 4 presents DLP values for each age group and is expected to facilitate comparison between the ED and DLP for each age group.

Comment 13)

From fig 6 it appears as if a different pediatric protocol is used for children up to 1 year. In the age group 0 to 1 year the effective dose seems to rise to a sharp peak at 1 year and then suddenly drops and gradually decreases for older ages. Please comment.

Reply 13)

This issue was discussed in the 7th paragraph of the Discussion. We estimated the ED using the DLP and a fixed conversion factor up to 1 year. The head growth during the first year after birth caused an DLP increase due to the result of the AEC function; therefore, the ED increased with age. The results presented in the original Figure 6 suggest that the curve method should not be applied to the assessment of changes in the ED during the first year after birth.

Comment 14)

Discussion

Again, instead of speaking of 3 methods I think that 4 methods were used in this study.

Reply 14)

In line 225 in the original manuscript, the number "three" represents the number of methods to determine the conversion factor. The Radimetrics method does not determine the conversion factor.

Comment 15)

Discussion

Please provide an answer to the research question. Which method is preferred, or does give the most accurate estimation of the effective dose in brain CT of children?

Reply 15)

We discussed the superiority of the curve method over the simple method (4th paragraph in the Discussion), linear method (5th paragraph in the Discussion), and Radimetrics method (8th paragraph in the Discussion). We stated the usefulness of the curve method (9th paragraph in the Discussion and the Conclusions).

Comment 16)

Discussion

The discussion section is far too wordy and must be shortened significantly, since a lot of text is simply a repetition of the results. Please provide comparison and discrepancies with other studies. Please compare your results with the DRL for pediatric brain CT.

Reply 16)

We presented interpretation of the results in the Discussion.

Age group-based analysis showed that the median DLP values were lower than the DRL values in our country. This was added to the 6th paragraph of the Discussion.

Comment 17)

 “This curve method demonstrated reasonable changes in the ED according to age.” What does reasonable mean? Please quantify. And does this mean that the authors prefer this method above the others. And that this is the proposed method, like stated in the last sentence 333-334?

Reply 17

We discussed complicated, up-and-down changes with age as unreasonable. We replaced “reasonable changes with age” in the 2nd paragraph of the Discussion by “reasonable, consistent changes with age”.

We added the following sentence to the 2nd last paragraph of the Discussion:

“We recommend the use of the curve method to estimate the ED in relation to age.”

In lines 333-334 in the original manuscript, we replaced “the proposed method” by “the proposed curve method.”

Comment 18)

 The brain is the least sensitive in DLP conversion factor to variation in age as compared to neck, chest and especially abdomen and pelvis (Deak 2010). In this study, however, only CT brain has been included. Please address this limitation in the discussion section.

Reply 18

We added the following sentence to the last of the Discussion:

“Additionally, the application of the curve method to ED estimation in pediatric body CT should be examined in the future.”

Comment 19)

 References

At least 5 references are given to own work. Please delete a few.

Reply 19

We deleted one reference. The other four papers were cited to disclose the use of the same data.

Round 2

Reviewer 2 Report

Almost all questions have been adressed properly in this revision. The introduction has been shortened, and the manuscript has been improved. However, I still think the discussion is far too long and wordy.

Author Response

We deleted some descriptions to shorten the Discussion.