Voxelized Breast Phantoms for Dosimetry in Mammography
R. M. Tucciariello
1,2 a
, P. Barca
1b
, D. Del Sarto
1,2 c
, R. Lamastra
1,2
, G. Mettivier
3,4 d
,
A. Retico
2e
, P. Russo
3,4 f
, A. Sarno
3g
, A. C. Traino
5h
and M. E. Fantacci
1,2 i
1
Department of Physics, University of Pisa, Pisa, Italy
2
INFN, Pisa Section, Pisa, Italy
3
INFN, Napoli Section, Napoli, Italy
4
Università di Napoli Federico II, Dipartimento di Fisica “Ettore Pancini”, Napoli, Italy
5
Unit of Medical Physics, Pisa University Hospital “Azienda Ospedaliero-Universitaria Pisana”, Pisa, Italy
mettivier@na.infn.it, alessandra.retico@pi.infn.it, russo@na.infn.it,
c.traino@ao-pisa.toscana.it, maria.evelina.fantacci@unipi.it
Keywords: Monte Carlo Simulations, Digital Mammography, X-Ray Breast Dosimetry, Voxelized Phantoms,
Heterogeneous Breast Phantoms, GEANT4.
Abstract: X-ray breast imaging techniques are an essential part of breast cancer screening programs and their
improvements lead to gain in performance and accuracy. Radiation dose estimate and control play an
important role in digital mammography and digital breast tomosynthesis investigations, since the risk of
radioinduced cancer to the gland must be contained and dose delivered to the gland must be declared in the
medical report. The actual dosimetric protocols suggest the assessment of radiation dose by means of Monte
Carlo calculation on digital breast phantoms, providing the assumption of the homogeneous mixture of
glandular and adipose tissues within the breast organ, leading to a drastic approximation. In line with the trend
of other research groups, with the aim of improving the Monte Carlo model, in the current work a new
heterogeneous digital breast model is proposed, involving a voxelized approach and disengaging from the
concept of homogeneous phantom. The proposed model is based on new findings in the literature and after a
validation process, the model is adopted to evaluate mean glandular dose discrepancies with the traditional
model which is adopted in clinic for decades.
1 INTRODUCTION
Worldwide, breast cancer is the most commonly
diagnosed cancer in female subjects, accounting
about 2.1 million newly diagnosed breast cancer, with
1 in 4 cancer cases among women, and the leading
cause of cancer death, followed by colorectal and
lung cancer for incidence, and vice versa for mortality
(Yuuhaa et al., 2018). In 2018, among European
women, breast cancer was by far the most frequently
a
https://orcid.org/0000-0001-9600-4177
b
https://orcid.org/0000-0001-9692-0730
c
https://orcid.org/0000-0003-3293-1005
d
https://orcid.org/0000-0001-6606-4304
e
https://orcid.org/0000-0001-5135-4472
f
https://orcid.org/0000-0001-9409-0008
g
https://orcid.org/0000-0002-3034-7166
h
https://orcid.org/0000-0003-3521-6293
i
https://orcid.org/0000-0003-2130-4372
diagnosed neoplasm (522,500 , 28.2% of the total),
followed by colorectal (228,000 , 12.3%), lung
(158,000 , 8.5%) and corpus uteri (122,000 , 6.6%)
cancers (Ferlay et al., 2018).
As suggested by the World Health Organization,
early detection is critical to improve breast cancer
outcomes and survival, made possible with screening
procedures consisting in testing women to identify
cancers before any symptom appears. This may lead
to tumour early detection, allowing greater
154
Tucciariello, R., Barca, P., Sarto, D., Lamastra, R., Mettivier, G., Retico, A., Russo, P., Sarno, A., Traino, A. and Fantacci, M.
Voxelized Breast Phantoms for Dosimetry in Mammography.
DOI: 10.5220/0010322901540161
In Proceedings of the 14th International Joint Conference on Biomedical Engineering Systems and Technologies (BIOSTEC 2021) - Volume 3: BIOINFORMATICS, pages 154-161
ISBN: 978-989-758-490-9
Copyright
c
2021 by SCITEPRESS Science and Technology Publications, Lda. All rights reserved
possibilities for medical treatment and reducing the
mortality rate. Since last decades, the principal
technique adopted for breast cancer screening
programs is the Digital Mammography (DM), an X-
ray imaging technique consisting in acquiring two
digital images, a cranio-caudal and a medio-lateral
view; since 2011 when it was introduced in the
clinical routine, the new technique Digital Breast
Tomosynthesis (DBT) (Sechopoulos, 2013a, 2013b)
supports or might even replace DM, because it allows
to reduce the tissue superimposition effect, making it
easier for radiologists to distinguish normal from
cancerous tissues.
Since breast imaging techniques involve ionizing
radiation, radiation dosimetry must be accurately
assessed the glandular tissue is considered as the
target tissue for radiation damage. Radioinduced
cancer risk estimates are performed by using the
Average Glandular Dose (AGD) metric as defined in
UK and EU dosimetry protocol (Van Engen et al.,
2018) or also referred as Mean Glandular Dose
(MGD). It should be stressed that breast screening
procedures involve low-dose radiation and the
carcinogenic risk is small with a very favourable
trade-off of the potential beneficial effects of
screening with respect to the calculated risk of
induced cancer (Pauwels et al., 2016).
For both DM and DBT, dosimetry is performed
by using Monte Carlo simulations (Wu et al. 1994;
Dance and Sechopoulos 2016; Sarno et al. 2018;
Tucciariello et al. 2020) which involve models of the
anatomy of the breast (digital phantoms) with a
homogeneous mixture of glandular and adipose
tissues, surrounded by a skin envelope. This
methodology does not consider real heterogeneous
glandular distribution within the breast and the
assumption of the homogeneous compound is very
drastic (Sarno et al. 2018). Nevertheless, the advent
of the fully 3D quantitative technique of Breast
Computed Tomography (bCT) (Sarno, Mettivier, and
Russo 2015), helped to better characterize the breast
anatomy. On the basis of the results provided by
Huang and Hernandez (Hernandez et al. 2015; Huang
et al. 2011) which involved real bCT investigations
on patients, in this work a new voxelized phantom
model, which takes into account the heterogeneous
distribution of the gland, has been created to be
adopted for improving the accuracy of the Monte
Carlo (MC) simulations in dosimetry for DM and
DBT.
In view of the publication of the Task Group 282
of the AAPM in collaboration with EFOMP for the
development of a new universal breast dosimetry, the
purpose of this work is to propose a new
heterogeneous breast model which is more
representative of the real breast anatomies with
respect to the adoption of an homogeneous breast
model. Indeed, a simple glandular distribution model
is presented and the influence of this methodology on
dose estimates will be evaluated by comparing MGD
values with those of the homogeneous model. Sarno
et al. (2018) followed an alternative approach in
which a dataset of real investigations is involved for
dosimetry purposes and patient-specific dose
estimates are performed. It has to be said that this
represents a complex and time-consuming approach,
which takes in consideration one-by-one women
breast, while the proposed methodology in this work
adopts the average glandular distribution among
women.
2 MATERIALS AND METHODS
The radiation glandular dose cannot be assessed
experimentally and the use of Monte Carlo
simulations (MC) is required. MGD estimates are
performed by using dedicated conversion coefficients
from the incident Air Kerma (𝐾

, mGy) to MGD
values (mGy). In this work, the formalism provided
by Wu (Wu et al., 1994) and extended by Boone
(Boone, 1999) is adopted to estimate the MGD values
(eq. 1).
𝑀𝐺𝐷
[
𝑚𝐺𝑦
]
=𝐷𝑔𝑁𝐾

(1
)
where DgN is the normalized glandular dose
coefficient (mGy/mGy).
In MC calculations, the rationale is to perform
simulations of the X-ray beam by tracing every
simulated photon over its track, from the radiation
source, towards the breast and to register energy
deposits in the volume (or mass) of interest. The MC
method produces the 𝐷𝑔𝑁 coefficients through the
calculation of the 𝑀𝐺𝐷 (which cannot be estimated
experimentally) and 𝐾

(measurable quantity). The
clinical practice uses 𝐷𝑔𝑁 numbers for converting
the measured 𝐾

at the entrance surface of the breast
to the estimated 𝑀𝐺𝐷 values (Sarno et al. 2019).
2.1 The Proposed Model
Using the GEANT4 simulation toolkit (Agostinelli et
al., 2003), following previous studies (Sarno et al.
2017; 2018; 2019) the setup for simulations has been
replicated and reported in Figure 1. The scoring
volume for dose deposit is showed in pink colour and
the skin envelope is considered as a shielding layer of
1.45 mm thick (Huang et al., 2008) not involved for
Voxelized Breast Phantoms for Dosimetry in Mammography
155
the dose computation. Adipose and glandular tissues
was replicated in the simulation setup by using the
elemental composition (Hammerstein et al., 1979),
reported in Table 1. The methodology adopted in this
work employs a new phantom model which
disengages from the traditional method of estimating
the mean glandular dose; indeed, the well-known G-
factor is adopted for “correcting” the dose deposit in
the homogeneous phantom (Boone 1999, 2002;
Nosratieh et al. 2015; Sarno et al. 2018; Sarno,
Mettivier, and Russo 2017; Tucciariello et al. 2019).
The traditional methodology of estimating MGD
values for homogeneous phantoms has been deeply
investigated in literature and is not the intent of this
paper.
Table 1: Elemental composition and density for glandular
and adipose tissues as implemented in the MC code.
Tissue H C N O P density
(g/cm
3
)
glandular 0.102 0.184 0.032 0.677 0.005 1.04
adipose 0.112 0.619 0.017 0.251 0.001 0.93
skin
0.098 0.178 0.050 0.667 0.007 1.09
Figure 1: a) Schematic drawing of the cranio-caudal view
irradiation geometry; b) scheme of the adopted simulation
geometry.
To move towards a heterogeneous approach, the
volume inside the breast tissue has to be divided in
voxels by using a voxel grid, inside which either
adipose or glandular tissues can be included, with
voxel dimension of 1×1×1 mm
3
. Since cubic voxels
have to be fixed inside a non-cubic volume (semi-
cylindrical cross section), some space uninvolved by
glandular tissue occurs, mainly between the breast
tissue and the skin interface on the rounded side of the
phantom, while on the chest wall side the border of
the voxels grid perfectly overlaps with the border of
the breast tissue. Despite uninvolved space reduces
1
The term glandularity indicates the percentage of
glandular tissue respect to the adipose tissue, sometimes
referred as breast density.
while decreasing voxels dimensions, computational
times and memory required have to be considered and
optimized. For purely investigative purposes, in the
presented model a voxel dimension of 1×1×1 mm
3
is
used, leading to a grid generation time of few seconds
to generate the whole voxel grid and only about 100
MB of RAM are required for the digital phantom
initialization. Nevertheless, uninvolved volume is
occupied by adipose material, which surely surrounds
the breast gland on the interface with the skin layer
(Huang et al. 2011) and this geometrical
approximation can be negligible.
Huang and Hernandez (Hernandez et al. 2015;
Huang et al. 2011) characterized the glandular tissue
within the breast, providing a metric able to reproduce
it. bCT investigations performed over patients let to
explore different breast anatomies (various cup sizes
and glandularities
1
) and a representative model has
been chosen for simulations.
In order to create a representative phantom model,
comparable with the homogeneous one adopted
previously, for which dose estimates are available in
the literature, the semi-cylindrical cross section has
been maintained. The use of a single gaussian
distribution, with no right-left displacement (as
showed by Huang) is considered appropriate in order
to replace a representative phantom for both right and
left women breasts, and an average value of 0.34 for
the FWHM provided by Hernandez has been adopted.
Equation (2) shows the distribution for glandular
voxels within the breast over the directions y and z,
represented by d (see Figure 2), for a given
compressed breast thickness 𝑙, μ
the centers of the
distributions and 𝜎
the standard deviations in both
directions (eq. 3).
𝐺
𝑑,𝜎
=
1
2𝜋𝜎
𝑒



(2)
𝜎
=
𝐹𝑊𝐻𝑀
2.35
∙𝑙
(3)
The choice of the material for each voxel is
performed by using a random approach, following the
distribution criteria adopted for the glandular tissue.
The assignment of glandular rather than adipose
material for each voxel is carried out by using
G
𝑑,σ
. The product G𝑦,σ
∙G
𝑧,σ
provides
the probability that a certain voxel in position (y,z) is
composed by glandular material depending of its
position in the breast volume. The randomness with
BIOINFORMATICS 2021 - 12th International Conference on Bioinformatics Models, Methods and Algorithms
156
which voxels are effectively fulfilled or not with
glandular material is provided by the uniform
probability distribution in the interval [0,1] produced
by the GEANT4 random number generator. The
amount of glandular voxels can be tuned in the MC
code to reproduce different breast densities.
Figure 2: a) perspective, b) top and c) front views of the
proposed voxelized breast phantom. Glandular
distributions evolve in y and z directions with gaussian
distributions, while in x direction with a constant
distribution. The voxel grid concerns the whole breast
tissue volume (white semi-cylinder) and for an easy display
purpose only glandular voxels are showed, while the
remaining part has to be considered filled with adipose
tissue.
2.2 Model Validation
A fundamental aspect of MC calculations is the
validation process. The MC code adopted derives
from a previously validated MC code for
homogeneous breast phantoms. The validation
process consisted in comparisons against literature
(Dance, Young, and Van Engen 2011; Sechopoulos
et al. 2014; AAPM TG 2015) and experimental
measurements by using radiochromic films.
Since in this work the code has been upgraded to
produce voxelized heterogeneous breast phantoms,
any consistent change in the model adopted should be
confirmed by some kind of verification. For
validation purposes of the voxelized methodology,
the phantom grid has been set using the uniform
probability distribution for assigning either glandular
or adipose tissues for each voxel. The rationale is to
compare MGD simulations results for the grid
phantom, with voxels randomly filled replicating a
uniform distribution of gland, with those of the
homogeneous phantom with the same breast density.
Three compression thicknesses of 3, 5 and 7 cm, and
five glandularities of 1%, 14.3%, 25%, 50% and 75%
have been replicated. For each model three spectra
have been involved for investigating different beam
qualities. In this case, MGD estimates do not involve
the usual relation provided by (Boone, 1999;
Nosratieh et al., 2015; A. Sarno, Mettivier, Di Lillo,
Tucciariello, et al., 2018), but is directly obtained by
scoring the energy deposited 𝐸
in all the n glandular
voxels with mass 𝑚

due to the X-ray beam (eq. 4)
𝑀𝐺𝐷 =
𝐸
𝑛∙𝑚


(4)
2.3 Dose Estimates
Glandular dose estimates dependencies using
homogeneous phantoms have already been
investigated by many authors (Dance and
Sechopoulos 2016; Sarno et al. 2018, 2019;
Tucciariello et al. 2020; Tucciariello et al. 2019) and
are not within the intents of this work, but no
evidences have been published in literature about the
dependence of dose estimates with respect to the
breast density obtained with heterogeneous glandular
distribution within the breast. In order to quantify
discrepancies between the two methods, simulations
have been performed with both homogeneous and
heterogeneous models, for three compressed
thicknesses of 3, 5 and 7 cm, exposed respectively to
W/Rh 26 kV, W/Rh 31 kV and W/Ag 34 kV spectra.
For each breast thickness, 12 glandularities have been
replicated.
3 RESULTS
3.1 Voxelized Phantom Validation
The voxelized phantom validation process regarded
the comparison between MGD values obtained with
the homogeneous phantom (MGD
hom
) with those
obtained with the voxelized phantom (MGD
vox
) using
the constant distribution of gland among voxels. This
kind of verification, performed by choosing the same
Voxelized Breast Phantoms for Dosimetry in Mammography
157
breast density in both models, let to investigate the
discrepancy produced by the voxelized model,
regardless of the glandular distribution adopted.
MGD discrepancies are showed in Table 2, where a
maximum percentage difference of 2.3% and an
average value of 1.0% confirm the success of the
voxels grid phantom model. Figure 3 highlights the
goodness of the fit (R
2
0.9995) performed over the
data obtained from simulations.
Since the creation of each heterogeneous digital
breast phantom involves a Monte Carlo approach, the
degree of reproducibility is questionable and one has
to wonders how much a certain model will differ with
the next one, with the same requested glandularity, in
terms of the amount of glandular voxels effectively
created and of mean glandular dose estimation. The
rationale is to create multiple models with same
glandularity and to verify the reproducibility
capabilities of the MC code. With the same
methodology, for each breast thickness of 3, 5 and 7
cm, five glandularities have been requested to be
reproduced by the MC code. In addition, for each
glandularity five models have been created, for a total
of 75 models, each of them irradiated with W/Al
spectra. In this case, between models with same
glandularity required, a maximum standard deviation
of 0.05% of glandularity is reached, which traduces
in very low MGD discrepancies, less than 0.5%.
Table 2: Data comparison between MGD values obtained
with homogeneous and voxelized phantoms for five breast
densities. Percentage differences refer to the ratio
(MGD
hom
- MGD
vox
)/ MGD
hom
× 100%.
MGD
vox
vs MGD
hom
1% 14.3% 25% 50% 75%
3 cm – W/Al 27kV 1.5% 1.4% 1.3% 0.9% 0.7%
3 cm – W/Rh 27kV 1.6% 1.1% 1.1% 0.9% 0.7%
3 cm – Mo/Mo 27kV 1.3% 2.0% 1.7% 1.3% 0.9%
5 cm – W/Al 30kV 0.1% 1.1% 1.1% 0.8% 0.7%
5 cm – W/Rh 30kV 2.3% 1.4% 0.7% 0.7% 0.6%
5 cm – Mo/Mo 30kV 0.9% 1.6% 1.5% 1.0% 0.9%
7 cm – W/Al 33kV -1.0% 1.0% 0.7% 0.8% 0.6%
7 cm – W/Rh 33kV -0.6% 0.9% 0.9% 0.9% 0.6%
7 cm – Mo/Mo 33kV 1.7% 2.0% 1.2% 1.1% 0.8%
Figure 3: Linear fit of MGD
vox
versus MGD
hom
in units of
mGy per incident photon. Concatenated data for 3, 5 and 7
cm compressed breast thicknesses. Origin 9.4 data analysis
software.
3.2 Dose Estimates in Mammography
Once the upgraded MC code which uses the proposed
phantom model has been validated, simulations over
the new model can be performed and new MGD
values could be investigated. As one can easily guess,
the adoption of the new model can lead to new MGD
values considering that the gland is mainly spread in
the central part of the volume. Di Franco and
colleagues (Di Franco et al., 2020) investigated the
glandular dose map within patient-derived digital
phantoms and highlighted a major dose distribution
towards the X-ray beam incident side (see Figure 1),
where however it should be mainly adipose voxels.
This kind of considerations led to treat the results
showed in Figure 4a more than reasonable, where
digital mammography investigations have been
performed for homogeneous and heterogeneous
breast models for various glandularities. Indeed, in
the homogeneous phantom a greater amount of
glandular tissue is in the upper layer of the breast with
the respect to the heterogeneous one, and glandular
dose deposit is higher, while in the heterogeneous
phantom the X-ray beam undergoes an X-ray
attenuation mainly due to the adipose tissues in the
upper layers and dose deposit is not scored by the MC
code. This effect also traduces in more discrepancies
for higher breast thicknesses and lower glandularities,
since glandular voxels are mainly disposed farther
from the upper surface.
BIOINFORMATICS 2021 - 12th International Conference on Bioinformatics Models, Methods and Algorithms
158
Figure 4: Mean glandular dose (mGy/photon) comparison
between the homogeneous breast model and the voxelized
one. Points in graphs refer to different glandularities.
Simulations have been performed by using the Hologic
Selenia Dimensions setup and typical irradiation settings in
DM modality for breast thicknesses of 3, 5 and 7 cm.
Discrepancies between MGD
het
and MGD
hom
are
reported in Table 3. Larger percentage variations
occur for breast densities less than 50% and decrease
with the increase of the breast density. It is a clear
evidence that for higher glandularities voxels marked
with gland tissue expand towards the border of the
phantom, approaching the 100% glandularity while
expanding and get closer to the homogeneous model.
As opposed to MGD coefficients for
homogeneous phantoms, which trend follows a 2
nd
order polynomial fit respect to the breast density
(Sarno et al. 2018), dose estimates for the voxelized
phantom show a 4
th
order polynomial fit dependence
against glandularity, for both low and high
compressed breast thicknesses, where, moreover, for
7 cm thickness a non-monotone trend is presented and
a concave curve is highlighted (Figure 4).
Discrepancies highlighted in Table 3 are in line
with literature (Dance et al. 2005; Hernandez et al.
2015) and these results should be emphasized,
because major variations occur for the most popular
breast densities among female subjects.
Table 3: Percentage variation of dose estimates between
heterogeneous and homogeneous models.
(MGD
hom
- MGD
het
)/ MGD
hom
× 100%.
3 cm 5 cm 7 cm
Model
#
Glandularity
(normalized)
variation
(%)
gland. var.
(%)
gland. var.
(%)
1 0.03 -8.3% 0.02 -18.0% 0.03 -28.0%
2 0.08 -9.2% 0.06 -19.7% 0.06 -31.9%
3 0.16 -11.9% 0.13 -24.1% 0.16 -37.3%
4 0.28 -14.6% 0.18 -26.8% 0.28 -41.1%
5 0.40 -13.8% 0.24 -28.8% 0.35 -39.5%
6 0.49 -12.3% 0.39 -27.4% 0.50 -32.7%
7 0.65 -8.9% 0.56 -21.3% 0.60 -27.0%
8 0.74 -6.7% 0.65 -17.0% 0.69 -20.9%
9 0.80 -5.2% 0.74 -12.6% 0.79 -14.0%
10 0.90 -2.6% 0.84 -8.1% 0.86 -9.4%
11 0.93 -1.8% 0.89 -5.3% 0.94 -3.7%
12 0.97 -0.8% 0.97 -1.5% 0.98 -1.0%
4 CONCLUSIONS
Since the 90’s, X-ray breast dosimetry has been
performed with Monte Carlo calculations by adopting
the assumption of the homogenous compound of the
breast tissue. Studies in literature showed a non-
uniform distribution of the glandular tissue and
quantitative data are available. This work was aimed
to overcome the drastic approximation of the current
digital breast phantom and therefore to improve
Monte Carlo accuracy for dosimetry in DM and DBT.
Based on findings in the literature, involving results
from real bCT scans on patients, a new methodology
for creating digital breast phantoms is proposed, in
order to provide a more representative phantom to
adopt for dose estimates; the proposed model
involves a voxelized phantom with glandular voxels
which follow a gaussian distribution among vertical
and lateral axis. The validation phase has been
conducted by comparing the current dosimetry
methodology with the proposed one, showing good
reliability and reproducibility of the voxelized
method. Finally, MC calculations have been
performed for both homogeneous and heterogeneous
models in typical DM investigations, in order to
quantify the discrepancy between the two phantom
models; wide variations have been confirmed, mostly
for low breast densities, the most common
characteristic among women and a new trend curve
has been found regarding the MGD values versus the
glandularity. The underestimates of MGD values with
the adoption of the voxelized phantom are in line with
literature results.
Voxelized Breast Phantoms for Dosimetry in Mammography
159
Nevertheless, a justification for today's adoption
of the old protocols based on homogeneous phantoms
can be given by confirming a conservative approach
of the actual method related to glandular dose. More
alarming would have been if the comparison would
lead to exactly opposite results. However, it must be
said that the MC approach is aimed to reproduce
experimental configuration with a certain degree of
accuracy, and the better the methodology used, the
greater the reliability. Moreover, MC calculations are
particularly useful for optimizing support equipment
for experimental measurements, like physical breast
phantoms for quality assurance or research activities
in the field of imaging (Ivanov et al. 2018;
Tucciariello et al. 2020; Barca et al. 2019; Lamastra
et al. 2020) and efforts in Monte Carlo calculations
represent one of the right ways to go through.
ACKNOWLEDGEMENTS
The presented work is part of the RADIOMA project
which is partially funded by "Fondazione Pisa",
Technological and Scientific Research Sector, Via
Pietro Toselli 29, Pisa (Italy). The authors would like
to thank Fondazione Pisa for giving the opportunity
to start this study.
REFERENCES
Agostinelli, S., Allison, J., Amako, K., Apostolakis, J.,
Araujo, H., Arce, P., Asai, M., Axen, D., Banerjee, S.,
Barrand, G., Behner, F., … Zschiesche, D. (2003).
GEANT4 - A simulation toolkit. Nuclear Instruments
and Methods in Physics Research, Section A:
Accelerators, Spectrometers, Detectors and Associated
Equipment, 506(3), 250–303.
https://doi.org/10.1016/S0168-9002(03)01368-8
Barca, P., Lamastra, R., Aringhieri, G., Tucciariello, R. M.,
Traino, A., & Fantacci, M. E. (2019). Comprehensive
assessment of image quality in synthetic and digital
mammography : a quantitative comparison.
Australasian Physical & Engineering Sciences in
Medicine, Cd. https://doi.org/10.1007/s13246-019-
00816-8
Boone, J. M. (1999). Glandular breast dose for
monoenergetic and high-energy x-ray beams: Monte
Carlo assessment. Radiology, 213(1), 23–37.
https://doi.org/10.1148/radiology.213.1.r99oc3923
Boone, J. M. (2002). Normalized glandular dose (DgN)
coefficients for arbitrary x-ray spectra in
mammography: Computer-fit values of Monte Carlo
derived data. Medical Physics, 29(5), 869–875.
https://doi.org/10.1118/1.1472499
Dance, D. R., Young, K. C., & Van Engen, R. E. (2011).
Estimation of mean glandular dose for breast
tomosynthesis: Factors for use with the UK, European
and IAEA breast dosimetry protocols. Physics in
Medicine and Biology, 56(2), 453–471.
https://doi.org/10.1088/0031-9155/56/2/011
Dance, David R., Hunt, R. A., Bakic, P. R., Maidment, A.
D. A., Sandborg, M., Ullman, G., & Carlsson, G. A.
(2005). Breast dosimetry using high-resolution voxel
phantoms. Radiation Protection Dosimetry, 114(1–3),
359–363. https://doi.org/10.1093/rpd/nch510
Dance, David R., & Sechopoulos, I. (2016). Dosimetry in
x-ray-based breast imaging. Physics in Medicine and
Biology, 61(19), R271–R304.
https://doi.org/10.1088/0031-9155/61/19/R271
Di Franco, F., Sarno, A., Mettivier, G., Hernandez, A. M.,
Bliznakova, K., Boone, J. M., & Russo, P. (2020).
GEANT4 Monte Carlo simulations for virtual clinical
trials in breast X-ray imaging: Proof of concept.
Physica Medica, 74(November 2019), 133–142.
https://doi.org/10.1016/j.ejmp.2020.05.007
Ferlay, J., Colombet, M., Soerjomataram, I., Dyba, T.,
Randi, G., Bettio, M., Gavin, A., Visser, O., & Bray, F.
(2018). Cancer incidence and mortality patterns in
Europe: Estimates for 40 countries and 25 major
cancers in 2018. In European Journal of Cancer (Vol.
103, pp. 356–387). Elsevier Ltd.
https://doi.org/10.1016/j.ejca.2018.07.005
Hammerstein, G. R., Miller, D. W., White, D. R.,
Masterson, M. E., Woodard, H. Q., & Laughlin, J. S.
(1979). Absorbed radiation dose in mammography.
Radiology, 130(2), 485–491.
Hernandez, A. M., Seibert, J. A., & Boone, J. M. (2015).
Breast dose in mammography is about 30% lower when
realistic heterogeneous glandular distributions are
considered. Medical Physics, 42(11), 6337–6348.
https://doi.org/10.1118/1.4931966
Huang, S. Y., Boone, J. M., Yang, K., Kwan, A. L. C., &
Packard, N. J. (2008). The effect of skin thickness
determined using breast CT on mammographic
dosimetry. Medical Physics, 35(4), 1199–1206.
https://doi.org/10.1118/1.2841938
Huang, S. Y., Boone, J. M., Yang, K., Packard, N. J.,
McKenney, S. E., Prionas, N. D., Lindfors, K. K., &
Yaffe, M. J. (2011). The characterization of breast
anatomical metrics using dedicated breast CT. Medical
Physics, 38(4), 2180–2191.
https://doi.org/10.1118/1.3567147
Ivanov, D., Bliznakova, K., Buliev, I., Popov, P., Mettivier,
G., Russo, P., Di Lillo, F., Sarno, A., Vignero, J.,
Bosmans, H., Bravin, A., & Bliznakov, Z. (2018).
Suitability of low density materials for 3D printing of
physical breast phantoms. Physics in Medicine and
Biology, 63(17), aad315. https://doi.org/10.1088/1361-
6560/aad315
Lamastra, R., Barca, P., Bisogni, M. G., Caramella, D.,
Rosso, V., Tucciariello, R. M., Traino, A. C., &
Fantacci, M. E. (2020). Image quality comparison
between synthetic 2D mammograms obtained with 15°
and 40° X-ray tube angular range: A quantitative
BIOINFORMATICS 2021 - 12th International Conference on Bioinformatics Models, Methods and Algorithms
160
phantom study. BIOIMAGING 2020 - 7th Int. Conf.
Bioimaging, Proceedings; Part 13th Int. Jt. Conf.
Biomed. Eng. Syst. Technol., BIOSTEC 2020, Biostec,
184–191. https://doi.org/10.5220/0009147601840191
Nosratieh, A., Hernandez, A., Shen, S. Z., Yaffe, M. J.,
Seibert, J. A., & Boone, J. M. (2015). Mean glandular
dose coefficients (DgN) for x-ray spectra used in
contemporary breast imaging systems. Physics in
Medicine and Biology, 60(18), 7179–7190.
https://doi.org/10.1088/0031-9155/60/18/7179
Pauwels, E. K. J., Foray, N., & Bourguignon, M. H. (2016).
Breast Cancer Induced by X-Ray Mammography
Screening? A Review Based on Recent Understanding
of Low-Dose Radiobiology. Medical Principles and
Practice, 25(2), 101–109.
https://doi.org/10.1159/000442442
Sarno, A., Mettivier, G., Di Lillo, F., Bliznakova, K.,
Sechopoulos, I., & Russo, P. (2018). Homogeneous vs.
patient specific breast models for Monte Carlo
evaluation of mean glandular dose in mammography.
Physica Medica.
https://doi.org/10.1016/j.ejmp.2018.04.392
Sarno, A., Mettivier, G., Di Lillo, F., Tucciariello, R. M.,
Bliznakova, K., & Russo, P. (2018). Normalized
glandular dose coefficients in mammography, digital
breast tomosynthesis and dedicated breast CT. Physica
Medica, 55, 142–148.
https://doi.org/10.1016/j.ejmp.2018.09.002
Sarno, A., Tucciariello, R. M., Mettivier, G., di Franco, F.,
& Russo, P. (2019). Monte Carlo calculation of
monoenergetic and polyenergetic DgN coefficients for
mean glandular dose estimates in mammography using
a homogeneous breast model. Physics in Medicine and
Biology, 64(12), 125012. https://doi.org/10.1088/1361-
6560/ab253f
Sarno, Antonio, Mettivier, G., & Russo, P. (2015).
Dedicated breast computed tomography: Basic aspects.
Medical Physics, 42(6), 2786–2804.
https://doi.org/10.1118/1.4919441
Sarno, Antonio, Mettivier, G., & Russo, P. (2017). Air
kerma calculation in Monte Carlo simulations for
deriving normalized glandular dose coefficients in
mammography. Physics in Medicine and Biology.
https://doi.org/10.1088/1361-6560/aa7016
Sechopoulos, I. (2013a). A review of breast tomosynthesis.
Part I. The image acquisition process. Medical Physics.
https://doi.org/10.1118/1.4770279
Sechopoulos, I. (2013b). A review of breast tomosynthesis.
Part II. Image reconstruction, processing and analysis,
and advanced applications. Medical Physics.
https://doi.org/10.1118/1.4770281
Sechopoulos, I., Sabol, J. M., Berglund, J., Bolch, W. E.,
Brateman, L., Christodoulou, E., Flynn, M., Geiser, W.,
Goodsitt, M., Kyle Jones, A., Von Tiedemann, M.
(2014). Radiation dosimetry in digital breast
tomosynthesis: Report of AAPM Tomosynthesis
Subcommittee Task Group 223. Medical Physics,
41(9). https://doi.org/10.1118/1.4892600
Sechopoulos I., Ali Elsayed S. M., Badal A., Badano A.,
Boone J. M., Kyprianou I. S., Mainegra-Hing E.,
McNitt-Gray M. F., McMillan K. L., Rogers D. W. O.,
Samei Ehsan, T. A. C. (2015). Monte Carlo Reference
Data Sets for Imaging Research. The Report of AAPM
Task Group 195. In Medical Physics (Vol. 42, Issue
195). https://doi.org/10.1118/1.4928676
Tucciariello, R. M., Barca, P., Lamastra, R., Traino, A., &
Fantacci, M. (2020). Monte Carlo Methods to evaluate
the Mean Glandular Dose in Mammography and Digital
Breast Tomosynthesis. In T. B. Hall (Ed.), Monte Carlo
Methods: History and Applications (pp. 73–110). Nova
Science Publishers, Inc.
Tucciariello, R. M., Barca, P., Caramella, D., Lamastra, R.,
Retico, A., Traino, A., & Fantacci, M. E. (2020). 3D
printing materials for physical breast phantoms: Monte
Carlo assessment and experimental validation.
BIODEVICES 2020 - 13th Int. Conf. Biomed. Electron.
Devices, Proceedings; Part 13th Int. Jt. Conf. Biomed.
Eng. Syst. Technol.,
BIOSTEC 2020, 254–262.
https://doi.org/10.5220/0009162302540262
Tucciariello, R. M., Barca, P., Caramella, D., Lamastra, R.,
Traino, C., & Fantacci, M. E. (2019). Monte carlo
methods for assessment of the mean glandular dose in
mammography: Simulations in homogeneous
phantoms. BIOINFORMATICS 2019 - 10th Int. Conf.
Bioinforma. Model. Methods Algorithms, Proceedings;
Part 12th Int. Jt. Conf. Biomed. Eng. Syst. Technol.,
BIOSTEC 2019, 242–249.
Tucciariello, R. M., Barca, P., Lamastra, R., Traino, A., &
Fantacci, M. E. (2020). Monte Carlo Methods to
Evaluate the Mean Glandular Dose in Mammography
and Digital Breast Tomosynthesis. In T. B. Hall (Ed.),
Monte Carlo Methods (pp. 73–110). Nova Science
Publishers, Inc.
https://novapublishers.com/shop/monte-carlo-
methods-history-and-applications/
Van Engen, R. E., Bosmans, H., Bouwman, R. W., Dance,
D. R., Lazzari, P., Marshall, N., Phelan, N.,
Schopphoven, S., Strudley, C., Thijssen, M. A. O., &
Young, K. C. (2018). Protocol for the Quality Control
of the Physical and Technical Aspects of Digital Breast
Tomosynthesis Systems. Euref, March, 84.
Wu, X., Gingold, E. L., Barnes, G. T., & Tucker, D. M.
(1994). Normalized Average Glandular Dose in
Molybdenum target-Rhodium Filter and Rhodium
target-Rhodium Filter Mammography. Radiology, 193,
83–89.
Yuuhaa, M. I. W. C., S, L. S., Bidang, P., Lingkungan, D.,
Amrullah, H., Law, A. O. F., Rahma, S. S., Mutiara, K.,
Murad, C., Baja, I., Utomo, K. S., Muryani, C.,
Nugraha, S., Sjafei, I., ﺕﺩﺎﻴﺳ ﺪﻴﻌﺳ, ﯽﺒﻴﺒﻴﺒﻴﺒﻴﺒﻴﺒﻴ ﺐﺜﺒﺜﺒﺛ , นคเรศ
งคว., Kota, D. I., & Selatan, T. (2018). World Health
Statistics 2018.
https://doi.org/10.20961/ge.v4i1.19180.
Voxelized Breast Phantoms for Dosimetry in Mammography
161