Automatic Exposure Control Attains Radiation Dose Modulation Matched with the Head Size in Pediatric Brain CT
, Hiroyasu Itoh 2, Hiroki Miyatake 2, Hirofumi Hata 2, Ryosuke Sasa 2, Nao Shiibashi 1 and Kohei Mitsui 1
Round 1
Reviewer 1 Report
- The manuscript aims to investigated the relationship between the head size and radiation dose in pediatric 12 brain computed tomography (CT) to evaluate the validity of automatic exposure control (AEC), and the conclusion to use AEC. This conclusion is already approved by many studies and recommended by different radiation protection organization. The authors should provide strong justification and motivation to conduct the study.
- Children younger than 15 years, why you excluded the age between 15-18 yrs old
- What is the effect of scan range?
- What is the effect of head circumference?
Author Response
Comment 1
The manuscript aims to investigated the relationship between the head size and radiation dose in pediatric brain computed tomography (CT) to evaluate the validity of automatic exposure control (AEC), and the conclusion to use AEC. This conclusion is already approved by many studies and recommended by different radiation protection organization. The authors should provide strong justification and motivation to conduct the study.
Reply 1
The most important background of this investigation was that AEC is not necessarily used in pediatric brain CT, as commented in the last paragraph of the Introduction as follows.
“Although AEC is recognized as a valuable tool for CT dose optimization, its application to pediatric brain CT is less common than pediatric body CT, and large surveys demonstrated that about half of the facilities did not use AEC for pediatric brain CT [12,13].”
Additionally, AEC may not provide appropriate dose modulation, as suggested in previous studies, and we believe that the validity of AEC-based dose modulation should be examined in clinical situations.
We added the following description as the first paragraph of the Discussion:
“Although the utility of AEC is widely recognized, it is not necessarily used in pediatric brain CT. An Italian survey of pediatric CT doses published in 2015 demonstrated that AEC was used in 80.5% of abdominal studies, 68.5% of chest studies, and only 50.5% of head studies despite availability on all CT scanners [12]. A French survey published in 2020 revealed that AEC was used in 70% to 100% of body studies and 50 to 63% of brain studies [13]. It should be noted that AEC may modulate radiation dose unreasonably. For example, tube current along the z-axis determined using a given AEC system may vary depending on the direction of the localizer imaging [20–23]. Using a widely used AEC system, much higher tube current was observed in the shoulder region using posteroanterior localizer images than using posteroanterior and lateral images [21,22]. A markedly high dose was observed at the top of the head depending on the patient positioning, which was resolved after modification of the AEC system [24]. Expected, appropriate dose modulation is not always guaranteed, and the validity of AEC should be investigated in clinical situations.”
Comment 2
Children younger than 15 years, why you excluded the age between 15-18 yrs old
Reply 2
In our country, the border between children and adults is usually set at 15 years for clinical medicine. Thus, we included patients younger than 15 years as children.
We recognized that our writing was misleading and made the following revision in the subsection “2.3. Patients” in the revised manuscript:
(before revision)
A total of 980 CT examinations performed in children younger than 15 years were analyzed retrospectively.
(after revision)
A total of 980 CT examinations performed in patients younger than 15 years were analyzed retrospectively. The border of 15 years between children and adults is commonly used for clinical medicine in our country.
Comment 3
What is the effect of scan range?
Reply 3
Prolongation of the scan length increases total radiation dose received by the patient in one imaging series. The DLP is an index of the total radiation dose and is a product of CTDIvol and the scan length.
AEC manipulates tube current, which is proportional to CTDIvol. In this study, we used tube current and CTDIvol as radiation dose indices to evaluate the validity of AEC-based dose modulation.
Comment 4
What is the effect of head circumference?
Reply 4
The X-rays are attenuated along the path through the imaging object. The degree of attenuation and consequently required radiation exposure depend on the path length: the head diameter in brain CT. The water equivalent diameter (WED) is an index of the section diameter calculated considering the different attenuation strengths of different tissues. It better represents the attenuation strength of the section than the effective diameter, a geometric mean of the anteroposterior and lateral diameters. In this study, we used the WED as an index of the head size, hypothesizing that radiation dose should correlate with the WED when appropriate radiation dose modulation is attained.
The head circumference is proportional to the head diameter, and a prolonged head circumference increases the required radiation dose.
Reviewer 2 Report
In this manuscript, the authors retrospectively analyzed pediatric brain CT images to evaluate the relationship between tube current (as determined by the AEC) and subject-related characteristics, focusing on the Water Equivalent Diameter (WED), an index of the head size. The manuscript is well-written, succinct enough and easy-to-follow. Moreover, the results are clearly communicated. My main concern is related to the scope, novelty and expected impact of this work for clinical users. The comments can be found below.
1. Scope and novelty. The authors investigate the relationship between dose modulation by the AEC and head size in pediatric brain CT imaging. However, this is the basic principle of any AEC available in the market and approved for clinical use. Essentially, the authors have demonstrated that the AEC performed as expected/marketed, according to the manufacturer and vendor; i.e., the tube output is reduced in smaller WED. I am very positive towards performing pre-clinical acceptance tests and periodic QC procedures. However, this is not the case. This is a retrospective study employing already acquired pediatric CT images. In my opinion, the authors should more clearly communicate the scope and novelty of this study. What is being investigated? Why? Are the authors trying to identify flaws/limitations in the AEC system? Can/do the authors suggest this methodology as a retrospective QC procedure as a constancy check of the AEC performance? In my opinion, the authors should focus on the latter when defining the novelty and scope of this work.
2. Expected impact. The conclusion of this work is not clearly communicated. What are the “take-home” messages for the clinical users? Simply stating that this specific AEC by the specific manufacturer and using the specific imaging parameters performs as expected/marketed is not enough to justify a publication. This comment is also related to comment #1.
3. The scope of the phantom study is not well-defined, nor explained anywhere in the text. What is the purpose of performing the phantom experiments? Furthermore, the phantom study is nowhere mentioned in the Abstract or the last paragraph of the Introduction.
4. Regarding the phantom study, I suppose the authors have selected a noise index of 4. For the sake of completeness and to demonstrate that the AEC performs well at other noise indices as well, please perform more scans at other noise indices, e.g., 1-5 with a step of 1. Please provide all relevant results in a Table or a Figure.
5. In sections 2 and 3, the phantom study comes after the patient study. This is uncommon. Phantom studies are typically described before the patient studies. I strongly recommend presenting the phantom experiments before the retrospective analysis, in both sections 2 and 3.
6. All WEDs were calculated using the Radimetrics software. No details are given. However, for the sake of completeness, the authors should add a few technical details on how this software calculates the WED. Please also identify and discuss any limitations and uncertainties related to this step.
7. The authors omitted to provide details with respect to the image reconstruction algorithm used to reconstruct the pediatric brain CT images. Was it a filtered back projection algorithm or an iterative reconstruction algorithm? What parameters were selected and why?
Author Response
Comment 1
Scope and novelty. The authors investigate the relationship between dose modulation by the AEC and head size in pediatric brain CT imaging. However, this is the basic principle of any AEC available in the market and approved for clinical use. Essentially, the authors have demonstrated that the AEC performed as expected/marketed, according to the manufacturer and vendor; i.e., the tube output is reduced in smaller WED. I am very positive towards performing pre-clinical acceptance tests and periodic QC procedures. However, this is not the case. This is a retrospective study employing already acquired pediatric CT images. In my opinion, the authors should more clearly communicate the scope and novelty of this study. What is being investigated? Why? Are the authors trying to identify flaws/limitations in the AEC system? Can/do the authors suggest this methodology as a retrospective QC procedure as a constancy check of the AEC performance? In my opinion, the authors should focus on the latter when defining the novelty and scope of this work.
Reply 1
The most important background of this investigation was that AEC is not necessarily used in pediatric brain CT, as commented in the last paragraph of the Introduction as follows.
“Although AEC is recognized as a valuable tool for CT dose optimization, its application to pediatric brain CT is less common than pediatric body CT, and large surveys demonstrated that about half of the facilities did not use AEC for pediatric brain CT [12,13].”
Additionally, AEC may not provide appropriate dose modulation, as suggested in previous studies, and we believe that the validity of AEC-based dose modulation should be examined in clinical situations.
We added the following description as the first paragraph of the Discussion:
“Although the utility of AEC is widely recognized, it is not necessarily used in pediatric brain CT. An Italian survey of pediatric CT doses published in 2015 demonstrated that AEC was used in 80.5% of abdominal studies, 68.5% of chest studies, and only 50.5% of head studies despite availability on all CT scanners [12]. A French survey published in 2020 revealed that AEC was used in 70% to 100% of body studies and 50 to 63% of brain studies [13]. It should be noted that AEC may modulate radiation dose unreasonably. For example, tube current along the z-axis determined using a given AEC system may vary depending on the direction of the localizer imaging [20–23]. Using a widely used AEC system, much higher tube current was observed in the shoulder region using posteroanterior localizer images than using posteroanterior and lateral images [21,22]. A markedly high dose was observed at the top of the head depending on the patient positioning, which was resolved after modification of the AEC system [24]. Expected, appropriate dose modulation is not always guaranteed, and the validity of AEC should be investigated in clinical situations.”
Comment 2
Expected impact. The conclusion of this work is not clearly communicated. What are the “take-home” messages for the clinical users? Simply stating that this specific AEC by the specific manufacturer and using the specific imaging parameters performs as expected/marketed is not enough to justify a publication. This comment is also related to comment #1.
Reply 2
The important message is towards many facilities where AEC is not used in pediatric brain CT. We revised the last part of the Conclusions as follows:
(before revision)
“For optimization of radiation dose, the application of AEC is recommended in pediatric brain CT.”
(after revision)
“This study indicated the validity of AEC-based dose modulation in clinical situations. For optimization of radiation dose, the application of AEC is recommended in pediatric brain CT.”
Comment 3　
The scope of the phantom study is not well-defined, nor explained anywhere in the text. What is the purpose of performing the phantom experiments? Furthermore, the phantom study is nowhere mentioned in the Abstract or the last paragraph of the Introduction.
Reply 3
This study aimed to evaluate the validity of AEC-based dose modulation in pediatric brain CT. Using clinical data, we assessed the appropriateness of dose modulation based on the relationship between the head size and radiation dose on the image-set basis and slice basis. Ideally, it should be investigated whether AEC improves the consistency of image quality among patients and among slice locations; however, such investigation is difficult. Thus, we added the phantom experiments to assess the effect of AEC on image noise along the scan range.
In the original manuscript, the purpose and interpretation of the phantom experiments were described in the 5th paragraph of the Discussion. We made the following revision:
(before revision)
“We evaluated the effect of AEC on image noise, using phantom experiments. ... Although the scope of the phantom experiments in this study was limited, the results support the effectiveness of AEC-based dose modulation in improving the consistency of image quality along the scan range.”
(after revision)
“The primary role of AEC is to aid in keeping image quality constant among patients and among slices. We evaluated the effect of AEC on image noise using phantom experiments. ... Although the phantom experiments in this study did not simulate imaging of different patients, the results support the effectiveness of AEC-based dose modulation in improving the consistency of image quality along the scan range.”
The lack of image quality analysis of patient data was mentioned in the last paragraph of the Discussion.
Moreover, we added the following description to the Abstract:
“Phantom experiments were performed to assess image noise with and without AEC, and indicated that AEC decreased differences in noise between slices of different section sizes.”
Comment 4
Regarding the phantom study, I suppose the authors have selected a noise index of 4. For the sake of completeness and to demonstrate that the AEC performs well at other noise indices as well, please perform more scans at other noise indices, e.g., 1-5 with a step of 1. Please provide all relevant results in a Table or a Figure.
Reply 2-4
As the reviewer commented, we used a noise index of 4 for the original manuscript. We performed phantom experiments using noise indices of 2.5, 3, 3.5, 4, 4.5, and 5, and presented the results in a new table. The noise indices were selected considering the tube current limitation of the CT scanner.
Comment 5
In sections 2 and 3, the phantom study comes after the patient study. This is uncommon. Phantom studies are typically described before the patient studies. I strongly recommend presenting the phantom experiments before the retrospective analysis, in both sections 2 and 3.
Reply 5
We rearranged the manuscript according to the reviewer’s comment.
Comment 6
All WEDs were calculated using the Radimetrics software. No details are given. However, for the sake of completeness, the authors should add a few technical details on how this software calculates the WED. Please also identify and discuss any limitations and uncertainties related to this step.
Reply 6
We added the following description to the last of “2.2 Phantom Experiments” in the revised manuscript:
“For the WED calculation, the area of the imaging object was automatically demarcated on the CT image, and the pixel values in the area were summed together to determine the diameter of a round water disc equivalent to the imaging object in terms of X-ray attenuation. Although truncation may cause errors in body CT, it did not matter in this study regarding brain CT.”
Comment 7
The authors omitted to provide details with respect to the image reconstruction algorithm used to reconstruct the pediatric brain CT images. Was it a filtered back projection algorithm or an iterative reconstruction algorithm? What parameters were selected and why?
Reply 7
We added the following description to the last of “2.1. Imaging Procedures” in the revised manuscript:
“CT images were reconstructed with adaptive statistical iterative reconstruction (ASiR, 60% blending) and a field of view of 250 mm.”
