Journal of Radiological Protection

ACCEPTED MANUSCRIPT

Assessment of patient dose and noise level of clinical CT images:

Automated measurements
To cite this article before publication: Choirul Anam et al 2019 J. Radiol. Prot. in press https://doi.org/10.1088/1361-6498/ab23cc

Manuscript version: Accepted Manuscript

Accepted Manuscript is “the version of the article accepted for publication including all changes made as a result of the peer review process,
and which may also include the addition to the article by IOP Publishing of a header, an article ID, a cover sheet and/or an ‘Accepted
Manuscript’ watermark, but excluding any other editing, typesetting or other changes made by IOP Publishing and/or its licensors”

This Accepted Manuscript is © 2019 Society for Radiological Protection. Published on behalf of SRP by IOP Publishing Limited. All
rights reserved..

During the embargo period (the 12 month period from the publication of the Version of Record of this article), the Accepted Manuscript is fully
protected by copyright and cannot be reused or reposted elsewhere.
As the Version of Record of this article is going to be / has been published on a subscription basis, this Accepted Manuscript is available for reuse
under a CC BY-NC-ND 3.0 licence after the 12 month embargo period.

After the embargo period, everyone is permitted to use copy and redistribute this article for non-commercial purposes only, provided that they
adhere to all the terms of the licence https://creativecommons.org/licences/by-nc-nd/3.0

Although reasonable endeavours have been taken to obtain all necessary permissions from third parties to include their copyrighted content
within this article, their full citation and copyright line may not be present in this Accepted Manuscript version. Before using any content from this
article, please refer to the Version of Record on IOPscience once published for full citation and copyright details, as permissions will likely be
required. All third party content is fully copyright protected, unless specifically stated otherwise in the figure caption in the Version of Record.

View the article online for updates and enhancements.

This content was downloaded from IP address 128.151.10.35 on 23/05/2019 at 01:04

Page 1 of 10 AUTHOR SUBMITTED MANUSCRIPT - JRP-101604.R1

1
2
3
4
Assessment of patient dose and noise level of clinical CT images: Automated
5 measurements
6
7 Choirul Anam1,*, Wahyu Setia Budi1, Kusworo Adi1, Heri Sutanto1, Freddy Haryanto2, Mohd Hanafi
8 Ali3, Toshioh Fujibuchi4, and Geoff Dougherty5

pt
9 1
Department of Physics, Faculty of Mathematics and Natural Sciences, Diponegoro University, Jl. Prof. Soedarto
10 SH, Tembalang, Semarang 50275, Central Java, Indonesia.
11 2
Department of Physics, Faculty of Mathematics and Natural Sciences, Bandung Institute of Technology, Ganesha
12 10, Bandung 40132, West Java, Indonesia
13 3
Department of Medical Radiation Sciences, University of Sydney, Sydney, Australia

cri
14 4
Department of Health Sciences, Faculty of Medical Sciences, Kyushu University, 3-1-1 Maidashi, Higashi-ku,
15 Fukuoka 812-8582, Japan
5
16 Department of Applied Physics and Medical Imaging, California State University Channel Islands, Camarillo,
17 CA 93012, USA.
18
19 E-mail: [email protected]
20

us
21 Abstract
22 We investigated comparisons between patient dose and noise in pelvic, abdominal, thoracic and head
23 CT images using an automatic method. 113 patient images (37 pelvis, 34 abdominal, 25 thoracic, and
24 17 head examinations) were retrospectively and automatically examined in this study. Water-equivalent
25 diameter (Dw), size-specific dose estimates (SSDE) and noise were automatically calculated from the
26 center slice for every patient image. The Dw was calculated based on auto-contouring of the patients’
27
28
29
an
edges, and the SSDE was calculated as the product of the volume CT dose index (CTDIvol) extracted
from the Digital Imaging and Communications in Medicine (DICOM) header and the conversion factor
based on the Dw obtained from AAPM 204. The noise was automatically measured as a minimum
30 standard deviation in the map of standard deviations. A square region of interest (ROI) of about 1 cm2
31 was used in the automated noise measurement. The SSDE values for the pelvis, abdomen, thorax, and
32
dM
head were 21.8 ± 7.3 mGy, 22.0 ± 4.5 mGy, 21.5 ± 4.7 mGy, and 65.1 ± 1.7 mGy, respectively. The
33
SSDEs for the pelvis, abdomen, and thorax increased linearly with increasing Dw, and for the head with
34
constant tube current, the SSDE decreased with increasing Dw. The noise in the pelvis, abdomen,
35
36
thorax, and head were 5.9 ± 1.5 HU, 5.2 ± 1.4 HU, 4.9 ± 0.8 HU and 3.9 ± 0.2 HU, respectively. The
37 noise levels for the pelvis, abdomen, and thorax of the patients were relatively constant with Dw because
38 of tube current modulation (TCM). The noise in the head image was also relatively constant because
39 Dw variations in the head are very small. The automated approach provides a convenient and objective
40 tool for dose optimizations.
41
42 Keywords: automated noise calculation, automated size-specific dose estimate (SSDE), water-
43
pte

equivalent diameter (Dw), CT dose optimization, ALARA principle

44
45
46
47
1. Introduction
48 Computed tomography (CT) is a powerful modality in medical imaging [1, 2] and its use continues to
49 increase as a result of ongoing advances in its technology. However, the growing use of CT has
50 increased the potential radiation risk to the population [3], since the dose from CT scans is greater than
51
ce

from other imaging modalities [4]. It was reported that the contribution from CT doses was more than
52 half of the dose from all other radiological modalities [5]. Recently, several publications reported that
53 there was a positive relationship between CT examinations and the subsequent occurrence of cancer [6,
54
7]. Therefore it is crucial to implement CT dose optimization based on the principle of as low as
55
reasonably achievable (ALARA). In the ALARA principle, the intended purpose of the application of
56
radiation in medical examinations is dose optimization, not simply dose reduction [8, 9]. In dose
Ac

57
58 optimization, image quality with a level adequate for diagnosis should be obtained, and based on this
59 image quality level an image is then produced with the lowest possible dose.
60
 AUTHOR SUBMITTED MANUSCRIPT - JRP-101604.R1 Page 2 of 10

1
2
3 For radiation dose optimization, there are two types of variables that should be considered:
4 controllable and uncontrollable variables [10]. Controllable variables are input imaging variables such
5 as tube voltage, tube current, gantry speed, mode of operation, pitch, slice thickness, type of
6
reconstruction filter, modulation parameters, field of view (FOV), etc. Uncontrollable variables are
7
directly related to the patient such as patient size, patient geometry, and attenuation of the scanned body.
8
All these variables affect the dose and the image quality [11]. Thus, the application of the ALARA

pt
9
10 principle requires the quantification of specific quantities, such as radiation dose and image quality, for
11 each patient personally and cannot be sufficiently represented by standardized phantoms.
12 Comparison between radiation doses and image quality has been investigated in previous studies
13 [12, 13], however the studies have several limitations. The dose metric used is based on volume CTDI

cri
14 (CTDIvol) [12, 14], which only characterizes the output dose of CT scanner and does not take into
15 account the size of the patient [13]. As already mentioned, the noise for the same exposure factor is
16
strongly influenced by the size of the patient [15]. For the quantification of radiation doses in CT, a
17
recently developed metric, the size-specific dose estimate (SSDE), has been specifically designed to
18
take into account the individual size of each patient [16-19]. SSDE is a development of the
19
20
quantification of previous doses indices that were based on standardized phantoms, such as the CTDIvol

us
21 [20-23].
22 Another limitation is that the measurement of patient dose and noise is carried out manually [12,
23 13]. This approach is rather subjective and a more objective comparison between patient dose and
24 patient noise needs to be performed with an automated approach. Presently, an automated calculation
25 of patient dose (SSDE) is already available [24, 25]. Christianson et al [24] developed an algorithm for
26
27
28
29
an
automatically calculating SSDE based on an effective diameter (Deff) and more recently Anam et al
[25] developed software for automatic SSDE based on water-equivalent diameter (Dw) [26, 27].
Recently, Christianson et al [28] developed a technique for automatically measuring the noise in
30
abdominal clinical CT images. More recently, Anam et al [29] developed a refined automated technique
31 for calculating noise from patient images that can be applied to all body regions, such as the head, chest,
32 abdomen and pelvis. This opens up the opportunity to optimize patient dose more objectively,
dM
33 efficiently, and conveniently. However, to date, a comparison between the automatic dose and noise
34 measurements for dose optimization purposes has never been conducted. This current study aims to
35 implement the automated patient dose and noise in clinical images as dose optimization for the most
36 common types of CT examinations: pelvis, abdominal, thorax and head.
37
38
39 2. Methods
40
41
2.1. Study design
42 A schematic diagram of the current study is shown in Figure 1. The study was a retrospective study on
43
pte

113 patient images in four CT examinations (pelvis, abdomen, thoracic and head). The number of
44 patients were 37, 34, 25 and 17 for pelvic, abdominal, thoracic and head CT examinations. All images
45 of patient were obtained using a MSCT Toshiba Aquilion-128, with 120 kVp and a slice thickness of 2
46 mm. The pitch of 0.938 for the pelvis, 1.438 for the abdomen and thorax, and 0.688 for the head, and
47
the rotation times were 500 ms for the pelvis, abdomen, and thorax, and 750 ms for the head. The tube
48
current modulation (TCM) technique was implemented in pelvis, abdomen and thorax examinations,
49
50
and a fixed tube current (FTC) technique of 300 mA was used in the head examination. In the head
51 examinations, constant tube current or FTC was usually employed due to the very small head size of
ce

52 the adult patients [30].

53 The Dw, SSDE and noise calculations were automatically calculated using one image of the
54 middle slice (e.g. the orbitals (head), sacrum (pelvis), umbilicus (abdomen), and xiphoid process
55 (thorax)). The SSDEs and noises for each examination were then compared as a function of Dw.
56 Comparisons of SSDE values and noises were also shown using boxplot graphics.
Ac

57
58
59
60
Page 3 of 10 AUTHOR SUBMITTED MANUSCRIPT - JRP-101604.R1

1
2
3
4
5
6
7
8

pt
9
10
11
12
13

cri
14
15
16
17
18
19 Figure 1. Schematic diagram of the current study.
20

us
21
22 2.2. Algorithm of automated Dw and SSDE calculations
23
24 The water-equivalent diameter (Dw) is a characterization of patient size and is the “gold standard”
25 metric for patient size [31]. The Dw does not only surrogate the geometrical size of the patient, but it
26 also the attenuation characteristics [32, 33]. Automated Dw calculation was adopted from the previous
27
28
29
an
study [25]. The steps in the automated Dw calculation are shown in Figure 2. The calculation was started
with automated patients contouring. Dw was calculated using equation:
30 1 𝐴𝑅𝑂𝐼
31
𝐷𝑤 = 2√(1000 ̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅
𝐻𝑈(𝑥, 𝑦)𝑅𝑂𝐼 + 1) 𝜋
(1)
32
dM
33 The region of interest (ROI) was a result of automated patient contouring. AROI was the area of
34
the patient and HU(x, y)ROI was the average HU value within the patient. Anam et al [25] reported that
35
36
the percentage differences between the automated and the manual Dw calculations were less than 0.5%.
37
38
39
40
41
42
43
pte

44
45
46
47
48
49
50 Figure 2. Steps of automated Dw calculation. (a) Image of patient, (b) Result of auto-contouring of the
51
ce

image, (c) Water-equivalent diameter (Dw) based on the result of auto-contouring.

52
53 A metric for characterizing patient dose estimates is the size-specific dose estimate (SSDE) [33].
54 The SSDE values were calculated automatically based on the results of automated Dw calculations [25].
55 The equations to calculate the SSDE based on head and body PMMA phantoms were obtained from
56
four different research groups organized by the American Association of Physicists in Medicine
Ac

57
58 (AAPM) in 2011 [34]. The four groups used different methods, i.e. physical measurements using an
59 anthropomorphic phantom, physical measurements using cylindrical PMMA phantoms, Monte Carlo
60 measurements on voxelized phantoms, and Monte Carlo measurements on simple cylindrical phantoms
 AUTHOR SUBMITTED MANUSCRIPT - JRP-101604.R1 Page 4 of 10

1
2
3 [34]. All data were then combined to produce the equations for calculating SSDEs taking into account
4 the patient size. The results from the four different studies were highly correlated with a R2 of 0.942
5
6
[34]. The equation for SSDE for the pelvis, abdominal and thoracic examinations was:
7 32
𝑆𝑆𝐷𝐸 = 𝐶𝑇𝐷𝐼𝑣𝑜𝑙 × 4.378094 × 𝑒 −0.04331124 × 𝐷𝑤 (2)
8

pt
9 The equation for SSDE for the head examinations was:
10 16
11 𝑆𝑆𝐷𝐸 = 𝐶𝑇𝐷𝐼𝑣𝑜𝑙 × 1.874799 × 𝑒 −0.03871313 × 𝐷𝑤 (3)
12 32
𝐶𝑇𝐷𝐼𝑣𝑜𝑙 16
is the metric of CT output dose measured using 32 cm PMMA phantom, and 𝐶𝑇𝐷𝐼𝑣𝑜𝑙 is
13 32 16
measured using16 cm PMMA phantom. The values of 𝐶𝑇𝐷𝐼𝑣𝑜𝑙 and 𝐶𝑇𝐷𝐼𝑣𝑜𝑙 values were taken from

cri
14
15 the Digital Imaging and Communications in Medicine (DICOM) header.
16
17
18 2.3. Algorithm of automated noise calculation
19
20 Noise is generally determined by calculating the standard deviation within a homogeneous area in the

us
21 region of interest (ROI) [35, 36]. Automated noise calculations require this be done automatically.
22 However, the determination of a ROI in a patient is not trivial. Recently, Christianson et al [29]
23 introduced an automated noise calculation algorithm without ROI detection. However, it was only able
24 to be applied in abdominal images. Anam et al [29] developed a refinement algorithm that is applied to
25 all body regions. Automated noise calculation was measured in the most homogeneous area, i.e. it was
26 identified as the smallest standard deviation (SD) in the standard deviation map. The deviation map was
27
28
29
an
calculated using a sliding window with kernel size of about 1 cm2. Figure 3 shows the original images
of patients (first row) and its corresponding noise maps and locations of the smallest SD in the SD map
to decide the noise of the image (second row).
30
31
32
dM
33
34
35
36
37
38
39
40
41
42
43
pte

44
45
46
47
48
49
50 Figure 3. Original images (first row) and their corresponding standard deviation maps (second row).
51
ce

The blue boxes indicate the location of the smallest standard deviation which characterizes the noise of
52
the image. (a) pelvis, (b) abdomen, (c) thorax, and (d) head.
53
54
55
56
Ac

57
58
59 3. Results
60
Page 5 of 10 AUTHOR SUBMITTED MANUSCRIPT - JRP-101604.R1

1
2
3 The SSDE levels for the pelvis, abdomen, thorax and head of the patients as a function of Dw are shown
4 in Figure 4. The SSDE values increase linearly with increasing Dw in the pelvis, abdomen, and thorax
5 images. The R2 are 0.74, 0.54 and 0.82 for pelvis, abdomen, and thorax, respectively. A different trend
6
is observed in the head images, the SSDE exponentially decreases with increasing Dw. The noise for
7
the pelvis, abdomen, thorax and head of the patients as a function of Dw are superimposed in Figure 5.
8
The noise in the pelvis, abdomen, thorax, and head images are rather constant (with R2 around 0.01) for

pt
9
10 varying Dw.
11
12
13

cri
14
15
16
17
18
19
20

us
21
22
23
24
25
26
27
28
29
an
30
31
32
dM
33
34
35
36
37
38
39
40
41
42
43
pte

44
45
46
47
48
49
50
51
ce

52
53
54
55
56
Ac

57
58
59
60
 AUTHOR SUBMITTED MANUSCRIPT - JRP-101604.R1 Page 6 of 10

1
2
3
4
5
6
7
8

pt
9
10
11
12
13

cri
14
15
16
17
18
19
20

us
21
22
23
24
25
26
27
28
29
an
30
31
32
dM
33
34
35
36
37
38
39
40
41
42
43
pte

44
45
46
47
48
49
50
51
ce

52
53
54
55
56
Ac

57 Figure 4. The result of automated noise and SSDE calculations as a function of Dw for different body
58 parts. (a) pelvis, (b) abdomen, (c) thorax, and (d) head.
59
60
Page 7 of 10 AUTHOR SUBMITTED MANUSCRIPT - JRP-101604.R1

1
2
3 Box plots of SSDE and noise are shown in Figure 5. The SSDE values in the head image (65.1 ±
4 1.7 mGy) are significantly higher than for the pelvis (21.8 ± 7.3 mGy), abdomen (22.0 ± 4.5 mGy) and
5 thorax (21.5 ± 4.7 mGy). The noise in the pelvis, abdomen and thorax images are similar with mean
6
values of 5.9 ± 1.5 HU, 5.2 ± 1.4 HU, and 4.9 ± 0.8 HU respectively, while in the head images the noise
7
is significantly lower at 3.9 ± 0.2 HU.
8

pt
9
10
11
12
13

cri
14
15
16
17
18
19
20

us
21
22
23
24
25
26
27
28
29
an
30
31
32
dM
33
34
35
36
37
38 Figure 5. Box plots for different body parts. (a) SSDE, and (b) Noise.
39
40
41 4. Discussion
42
43 The automated noise calculation [28, 29] in conjunction with an automated size-specific dose estimate
pte

44 (SSDE) [24, 25] can be used as a tool to optimize dose in accordance with the ALARA principle. An
45 automated method more objectively and effectively determines the noise and SSDE than the manual
46 method. The determination of noise is strongly influenced by the selection of the ROI by the observer,
47 who may sometimes determines noise in an area that is not the most homogeneous.
48
It is found that the noise in the pelvis, abdomen and thorax images are relatively constant for
49
50 varying Dw because of TCM [37-39]. Even though it did not use TCM, the noise in the head image is
51 also relatively constant, due to very small Dw variations in the head, of about 2 cm. This small variation
ce

52 of Dw leads to only a slight impact on the noise on the head.

53 SSDE is affected by both values of CTDIvol and size-correction factor. SSDE decreases with
54 size in the head due to constant mA (or fixed tube current, FTC). The constant mA leads to a constant
55 CTDIvol. This constant CTDIvol then results in a decreasing of SSDE, because of the decrease in size-
56 correction factor. Conversely, the other examinations use tube current modulation (TCM) for dose
Ac

57
reduction in small patients, where the mA increases with the size. This increase of mA leads to an
58
increase in CTDIvol. If the size increases, the CTDIvol increases but the size-correction factor
59
60
decreases. The combination of the increase of CTDIvol and a decrease of size correction factor can lead
 AUTHOR SUBMITTED MANUSCRIPT - JRP-101604.R1 Page 8 of 10

1
2
3 SSDE to remain constant, or increase, or decrease depending on which one is more dominant. This
4 dominance depends on the magnitude of TCM. In this study, the increase in CTDIvol was more
5 dominant than the decrease in size-correction factor, consequently SSDE increased with the size. In
6
previous studies the reported difference in results depended on the magnitude of TCM. One study
7
reported that the TCM technique caused a relatively constant SSDE [40], although other study reported
8
that the TCM technique caused an increase in SSDE [41].

pt
9
10 The current study found that the SSDE of the head was in the order of 65 mGy, a much higher
11 dose level than other body parts examined which had SSDE values of about 20 mGy. Our results are
12 comparable with a survey of output dose values of about 60 mGy for head CT examinations and of
13 about 20 mGy for body examination in the U.S. in the period 1999 to 2001 [42]. In this study, we found

cri
14 that the noise in head images was about 3.9 HU and it was lower than that in pelvis, abdominal and
15 thorax images (about 5-6 HU). This finding indicated that the protocol for head CT examination
16
produces too high a dose and too low noise, the dose could be further reduced while maintaining
17
diagnostic image quality. Previously, Tipnis et al [43] reported that a decrease in output dose from 75
18
mGy to 60 mGy in head CT examinations produced no clear degradations in depicting neuroanatomy.
19
20
Based on this finding, the noise in the head image could be further increased up to about 5 HU which

us
21 may be obtained with a dose lower than 60 mGy.
22 Comparison of doses on each examination for optimization purposes could also be conducted. In
23 the pelvis and abdominal examination, there are several noise outliers, i.e. the noise is much greater
24 than the average noise value for the certain examination. This phenomenon indicates that in one
25 examination further optimization needs to be carried out. The ranges of noise values in the pelvis,
26
27
28
29
an
abdominal and thoracic examinations are relatively large at around 4 HU, indicating that TCM
optimization needs to be increased so that the dose and noise ranges can be reduced to around 2 HU.
This task is challenging because the pelvis, abdomen, and thorax have relatively large variations in size,
30
geometry and composition. A noise range of less than 2 HU has been achieved on head examinations
31 due to the relatively small variation in head size from one patient to another, which is in the range of 2
32 cm.
dM
33 In the current study, however, the calculated noise level has not been compared with visual
34 observations by expert radiologists [43]. In a follow-up study, we will compare quantitative noise levels
35 from automated methods with qualitative radiological observations determined by a radiologist. Image
36 quality is not only characterized by noise. For a more comprehensive optimization we need to evaluate
37
an automatic approach taking into account other measures of image quality, such as spatial resolution
38
[44], artifacts and so on.
39
40
41
42 5. Conclusions
43
pte

Implementation of automated noise calculation, along with automated size-specific estimation, may
44 provide a convenient tool for optimizing patient doses according to the ALARA principle. We found
45 that the noise in head images was lower than that in pelvic, abdominal and thoracic images, while the
46 dose in the head was about three times that in the other regions. This finding revealed that the protocol
47 for head CT seems to be characterized by higher dose and lower noise, both of which could be further
48
optimized while maintaining diagnostic image quality.
49
50
51
Acknowledgments
ce

52 This work was funded by the Riset Publikasi Internasional Bereputasi Tinggi (RPIBT) 2019,
53 Universitas Diponegoro. The authors give special thanks to Sanggam Ramantisan for providing
54
images of the patients.
55
56
References
Ac

57
58 [1] Voress M 2007 The increasing use of CT and its risks Radiol. Technol. 79 186-190
59
60
Page 9 of 10 AUTHOR SUBMITTED MANUSCRIPT - JRP-101604.R1

1
2
3 [2] Mettler Jr FA, Bhargavan M, Faulkner K, et al 2009 Radiologic and nuclear medicine studies in the United
4 States and worldwide: frequency, radiation dose, and comparison with other radiation sources–1950-
5 2007 Radiology 253 520–531
6 [3] Brenner DJ and Hall EJ 2007 Computed tomography—an increasing source of radiation exposure N. Engl.
7 J. Med. 357 2277-2284
8 [4] Kalender WA 2014 Dose in x-ray computed tomography Phys. Med. Biol. 59 R129–R150

pt
9 [5] NCRP 2009 Ionizing radiation exposure of the population of the United States. NCRP Report 160,
10 Bethesda MD
11 [6] Pearce MS, Salotti JA, Little MP, et al 2012 Radiation exposure from CT scans in childhood and
12 subsequent risk of leukaemia and brain tumours: a retrospective cohort study Lancet 380 499–505
13 [7] Miglioretti DL, Johnson E, Williams A, et al 2013 Pediatric computed tomography and associated radiation

cri
14 exposure and estimated cancer risk. JAMA Pediatr. 167 700–707
15 [8] Kalra MK, Maher MM, Toth TL, et al 2004 Strategies for CT radiation dose optimization Radiology 230
16 619–628
17 [9] Shah NB and Platt SL 2018 ALARA: is there a cause for alarm? Reducing radiation risks from computed
tomography scanning in children Curr. Opin. Pediatr. 20 243–247
18
[10] Larson DB 2014 Optimizing CT radiation dose based on patient size and image quality: the size-specific
19
dose estimate method Pediatr. Radiol. 44 S501–S505
20

us
[11] Karmazyn B, Liang Y, Klahr P, and Jennings S G 2013 Effect of tube voltage on CT noise levels in
21 different phantom sizes Am. J. Roentgenol. 200 1001–1005
22 [12] Spampinato MV, Stalcup S, Matheus MG, Byington K, Tyler M, Bickley S and Tipnis S 2018 Radiation
23 dose and image quality in pediatric head CT Rad. Prot. Dosim. 182 310–316
24 [13] Khoramian D, Sistani S and Firouzjah RA 2019 Assessment and comparison of radiation dose and image
25 quality in multi-detector CT scanners in non-contrast head and neck examinations Pol. J. Radiol. 84
26
27
28
29
[14]
e61-e67
an
Anam C, Fujibuchi T, Haryanto F, Widita R, Arif I and Dougherty G 2019 An evaluation of computed
tomography dose index measurements using a pencil ionisation chamber and small detectors J. Radiol.
Prot. 39 112–124
30 [15] Larson DB, Wang LL, Podberesky DJ and Goske MJ 2013 System for verifiable CT radiation dose
31 optimization based on image quality. Part I. Optimization model Radiology 269 167-176
32 [16] Boone J, Strauss KJ, Cody DD, et al 2011 Size-specific dose estimates (SSDE) in pediatric and adult body
dM
33 CT examinations. AAPM report no. 204. http://www.aapm.org/pubs/reports/RPT_204.pdf. Accessed 2
34 March 2014
35 [17] Wang J, Christner JA, Duan Y, Leng S, Yu L and McCollough CH 2012 Attenuation-based estimation of
36 patient size for the purpose of size specific dose estimation in CT. Part II. Implementation on abdomen
37 and thorax phantoms using cross sectional CT images and scanned projection radiograph images Med.
38 Phys. 39 6772-6778
39 [18] Boos J, Lanzman RS, Heusch P, et al 2016 Does body mass index outperform body weight as a surrogate
40 parameter in the calculation of size-specific dose estimates in adult body CT? Br. J. Radiol. 89
20150734
41
[19] Anam C, Haryanto F, Widita R, Arif I, Dougherty G and McLean D 2017 The impact of patient table on
42
size-specific dose estimate (SSDE) Australas. Phys. Eng. Sci. Med. 40 153-158
43
pte

[20] Christianson O, Li X, Frush D and Samei E 2012 Automated size-specific CT dose monitoring program:
44 Assessing variability in CT dose Med. Phys. 39 7131-7139
45 [21] Anam C, Haryanto F, Widita R, Arif I, Dougherty G and McLean D 2016 Estimation of eye radiation dose
46 during nasopharyngeal CT examination for an individual patient Information (Japan) 19 3951-3962
47 [22] Kidoh M, Utsunomiya D, Oda S, et al 2017 Breast dose reduction for chest CT by modifying the scanning
48 parameters based on the pre-scan size-specific dose estimate (SSDE) Eur. Radiol. 27 2267-2274
49 [23] Anam C, Fujibuchi T, Toyoda T, Sato N, Haryanto F, Widita R, Arif I and Dougherty G 2018 A simple
50 method for calibrating pixel values of the CT localizer radiograph for calculating water-equivalent
51
ce

diameter and size-specific dose estimate Radiat. Prot. Dosim. 179 158-168
52 [24] Christianson O, Li X, Frush D and Samei E 2012 Automated size-specific CT dose monitoring program:
53 assessing variability in CT dose Med. Phys. 39 7131–7139
54 [25] Anam C, Haryanto F, Widita R, Arif I and Dougherty G 2016 Automated calculation of water-equivalent
55 diameter (DW) based on AAPM task group 220 J. Appl. Clin. Med. Phys. 17 320-333
56 [26] Dodge C T, Tamm EP, Cody DD, Liu X, Jensen CT, Wei W, Kundra V and Rong XJ 2016 Performance
Ac

57 evaluation of iterative reconstruction algorithms for achieving CT radiation dose reduction — a

58 phantom study J. Appl. Clin. Med. Phys. 17 511-531.
59
60
 AUTHOR SUBMITTED MANUSCRIPT - JRP-101604.R1 Page 10 of 10

1
2
3 [27] Hussain FA, Mail N, Shamy AM, Alghamdi S and Saoudi A 2016 A qualitative and quantitative analysis
4 of radiation dose and image quality of computed tomography images using adaptive statistical iterative
5 reconstruction J. Appl. Clin. Med. Phys. 17 419-432
6 [28] Christianson O, Winslow J, Frush DP and Samei E 2015 Automated technique to measure noise in clinical
7 CT examinations Am. J. Roentgenol. 205 W93–W99
8 [29] Anam C, Arif I, Haryanto F, Lestari FP, Widita R, Budi WS, Sutanto H, Fujibuchi T and Dougherty G

pt
9 2019 A novel method of automated noise measurement from the most homogeneous region in CT
10 images Biomed. Phys. Eng. Express. Submitted.
11 [30] Nawfel RD and Young GS 2017 Measured head CT/CTA skin dose and intensive care unit patient
12 cumulative exposure AJNR Am. J. Neuroradiol. 38 455–61
13 [31] Sarmento S, Mendes B and Gouvêa M 2018 Automatic calculation of patient size metrics in computed

cri
14 tomography: What level of computational accuracy do we need? J. Appl. Clin. Med. Phys. 19 218–227
15 [32] Anam C, Haryanto F, Widita R, Arif I and Dougherty G 2017 The size-specific dose estimate (SSDE) for
16 truncated computed tomography images Radiat. Prot. Dosim. 175 313-320
17 [33] Burton CS and Szczykutowicz TP 2018 Evaluation of AAPM Reports 204 and 220: Estimation of effective
diameter, water-equivalent diameter, and ellipticity ratios for chest, abdomen, pelvis, and head CT
18
scans J. Appl. Clin. Med. Phys. 19 228–238
19
[34] American Association of Physicists in Medicine 2010 AAPM report no. 111. The future of CT dosimetry:
20

us
comprehensive methodology for the evaluation of radiation dose in x-ray computed tomography
21 (https://aapm.org/pubs/reports/RPT_111.pdf) (Accessed: 15 December 2013)
22 [35] Primak AN, McCollough CH, Bruesewitz MR, Zhang J and Fletcher JG 2006 Relationship between noise,
23 dose, and pitch in cardiac multi–detector row CT RadioGraphics 26 1785–1794
24 [36] Kamezawa H, Arimura H, Shirieda K, Kameda N and Ohki M 2016 Feasibility of patient dose reduction
25 based on various noise suppression filters for cone-beam computed tomography in an image-guided
26
27
28
29
[37]
an
patient positioning system. Phys. Med. Biol. 61 3609-3636
Favazza CP, Yu L, Leng S, Kofler JM and McCollough CH 2015 Automatic exposure control systems
designed to maintain constant image noise: Effects on computed tomography dose and noise relative
to clinically accepted technique charts J. Comput. Assist. Tomogr. 39 437–442
30 [38] Angel E 2009 Monte Carlo simulations to assess the effects of tube current modulation on breast dose for
31 multidetector CT Phys. Med. Biol. 54 497–512
32 [39] Kalender WA, Wolf H and Suess C 1999 Dose reduction in CT by anatomically adapted tube current
dM
33 modulation: phantom measurements Med. Phys. 26 2248–2253
34 [40] Christner JA, Braun NN, Jacobsen MC, Carter RE, Kofler JM and McCollough CH 2012 Size-specific
35 dose estimates for adult patients at CT of the torso Radiology 265 841–847
36 [41] Schindera ST, Nelson RC, Toth TL, Nguyen GT, Toncheva GI, DeLong DM and Yoshizumi TT 2008
37 Effect of patient size on radiation dose for abdominal MDCT with automatic tube current modulation:
38 phantom study AJR Am. J. Roentgenol. 190 W100–W105
39 [42] Martin CJ and Huda W 2013 Intercomparison of patient CTDI surveys in three countries Radiat. Prot.
40 Dosim. 153 431–440
[43] Tipnis S, Thampy R, Rumboldt Z, Spampinato M, Matheus G and Huda W 2016 Radiation intensity
41
(CTDIvol) and visibility of anatomical structures in head CT examinations J. Appl. Clin. Med. Phys.
42
17 293-300
43
pte

[44] Anam C, Fujibuchi T, Budi WS, Haryanto F and Dougherty G 2018 An algorithm for automated
44 modulation transfer function measurement using an edge of a PMMA phantom: Impact of field of view
45 on spatial resolution of CT images J. Appl. Clin. Med. Phys. 19 244–252
46
47
48
49
50
51
ce

52
53
54
55
56
Ac

57
58
59
60