15269914, 0, Downloaded from https://aapm.onlinelibrary.wiley.com/doi/10.1002/acm2.13917 by Cochrane Poland, Wiley Online Library on [25/02/2023].

See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Received: 3 January 2022 Revised: 16 November 2022 Accepted: 5 January 2023

DOI: 10.1002/acm2.13917

R A D I AT I O N O N C O L O G Y P H Y S I C S

Impact of CT scan parameters on deformable image

registration accuracy using deformable thorax phantom

Ryutaro Ikeda1,2 Noriyuki Kadoya1 Yujiro Nakajima3 Shin Ishii2

Takayuki Shibu2 Keiichi Jingu1

1
Department of Radiation Oncology, Tohoku
University Graduate School of Medicine,
Sendai, Japan
2
Department of Radiology, Japanese Red
Cross Ishinomaki Hospital, Ishinomaki, Japan
3
Department of Radiological Science,
Komazawa University, Tokyo, Japan
Correspondence
Noriyuki Kadoya, Department of Radiation
Oncology, Tohoku University Graduate School
of Medicine, Sendai, 980-8574, Japan.
Email: kadoya.n@rad.med.tohoku.ac.jp
Abstract
The purpose of this study was to evaluate the deformable image registra-
tion (DIR) accuracy using various CT scan parameters with deformable thorax
phantom.
Our developed deformable thorax phantom (Dephan, Chiyoda Technol Corp,
Tokyo, Japan) was used. The phantom consists of a base phantom, an inner
phantom, and a motor-derived piston. The base phantom is an acrylic cylin-
der phantom with a diameter of 180 mm, which simulates the chest wall.
The inner phantom consists of deformable, 20 mm thick disk-shaped sponges
with 48 Lucite beads and 48 nylon cross-wires which simulate the vascular
and bronchial bifurcations of the lung. Peak-exhale and peak-inhale images
of the deformable phantom were acquired using a CT scanner (Aquilion LB,
TOSHIBA). To evaluate the impact of CT scan parameters on DIR accuracy,
we used the four tube voltages (80, 100, 120, and 135 kV) and six reconstruc-
tion algorithms (FC11, FC13, FC15, FC41, FC44, and FC52). Intensity-based
DIR was performed between the two images using MIM Maestro (MIM soft-
ware, Cleveland, USA). Fiducial markers (beads and cross-wires) based target
registration error (TRE) was used for quantitative evaluation of DIR.
In case with different tube voltages, the range of average TRE were 4.44–
5.69 mm (reconstruction algorithm: FC13). In case with different reconstruction
algorithms, the range of average TRE were 4.26–4.59 mm (tube voltage:
120 kV). The TRE were differed by up to 3.0 mm (3.96–6.96 mm) depending
on the combination of tube voltage and reconstruction algorithm.
Our result indicated that CT scan parameters had moderate impact of TRE,
especially for reconstruction algorithms for the deformable thorax phantom.

KEYWORDS
deformable image registration, lung, physical phantom, radiotherapy, target registration error

1 INTRODUCTION ventilation to indirectly evaluate lung function,4,5 taking

In recent years, deformable image registration (DIR)
method has been used in the field of radiotherapy. Appli-
cations of DIR enable dose accumulation,1,2 automatic
contour propagation,3 and computed tomography (CT)
into account changes in physique and position changes
during the treatment period by using this technology.
However, there is uncertainty with regard to DIR, and
it has been reported that the accuracy of DIR differs
depending on the software used, and the institution

This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided
the original work is properly cited.
© 2023 The Authors. Journal of Applied Clinical Medical Physics published by Wiley Periodicals, LLC on behalf of The American Association of Physicists in Medicine.

J Appl Clin Med Phys. 2023;e13917. wileyonlinelibrary.com/journal/acm2 1 of 8

https://doi.org/10.1002/acm2.13917

2 of 8
using the technology.6 Therefore, the accuracy needs to
be examined by each institution when using DIR. The
American Association of Medical Physics (AAPM) pub-
lished the TG132 guideline7 summarizing the use of
image registration techniques including DIR, and image
registration quality assurance (QA) methods; thus, we
can say that the examination of the accuracy of DIR
is an important matter. As one method of QA in DIR,
it is recommended to perform QA using a physical
moving phantom that is actually capable of deforma-
tion. Regarding the DIR QA with this phantom, we can
obtain the actual deformation vector field using a land-
mark implemented in the phantom. Various studies have
been conducted with regard to phantoms used in this
technique.8,9
Many deformation algorithms applying image inten-
sity are used in the deformation parameters of DIR.10
In the field of radiomics in which analyses are simi-
larly performed using image intensity information,10–12
it has been reported that the quantity of features
obtained changes according to the scan parameters
used to obtain the CT image used,13 and Kim et al.
reported that when different reconstruction algorithms
are used a predominant difference occurs in 19 features
calculated.14
Therefore, similarly in DIR also, it is conceivable
that the accuracy will be affected by changing the
parameters that affect image intensity information when
performing CT scans. However, the impact of CT scan
parameters on DIR accuracy has not been elucidated.
Yet, tests cannot be performed using images of actual
human bodies, and consequently tests must be per-
formed using phantoms. Therefore, in the present study,
we used a dynamic deformable phantom to test the
extent by which DIR accuracy is affected by tube volt-
age and reconstruction algorithm, which are CT scans
parameters that affect image quality.
2 METHODS
2.1 Deformable thoracic phantom
The dynamic deformable phantom (Dephan, Chiyoda
Technol Corp, Tokyo, Japan) consists of a base phan-
tom, an inner phantom, and a motor-derived piston.
An overview of our phantom is shown in Figure 1.
The base phantom was an acrylic cylinder phantom
180 mm in diameter, which simulates the chest wall.
The inner phantom consists of deformable disk-shaped
sponges. Disk- shaped sponges are made of a soft
material (physical density, 0.22 g/cm3 ) 175 mm in diam-
eter and 20 mm thick. There are multiple holes (20.0
mmφ) in the disk-shaped sponge, and embedding-type
cylinder-shaped sponges, in which the Lucite beads (2.0
mmφ) or nylon cross-wires (1.4 mmφ and 18.0 mm)
were inserted, were incorporated into the hollowed-out
15269914, 0, Downloaded from https://aapm.onlinelibrary.wiley.com/doi/10.1002/acm2.13917 by Cochrane Poland, Wiley Online Library on [25/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
IKEDA ET AL.
spaces in the disk-shaped sponge. Furthermore, the
soft tissue sphere (30.0 mmφ) can be inserted. This
enables us to place many beads and nylon cross-wires
at multiple arbitrary points to simulate various inner
phantom settings. The Lucite beads and nylon cross-
wires (48 nylon wires and 48 Lucite beads simulate
the vascular and bronchial bifurcations of the lung. The
piston, simulating the diaphragm, is fastened to a pro-
grammable motor via a metal rod that compresses and
decompresses the lung according to an arbitrary, pro-
grammable breathing profile. The details are described
in the article by Sugawara et al.15
2.2 Acquisition of peak-exhale and
peak-inhale images using various scan
parameter and reconstruction settings
Peak-exhale and peak-inhale CT scan images of the
phantom were taken using a CT scanner (Aquilion LB,
Canon Medical Systems, Tochigi, Japan) with different
four scanning parameters (i.e.,80,100,120,and 135 kV).
In general, 120 kV is preferred to use for treatment plan-
ning CT scan. To investigate the impact of tube voltage
on DIR accuracy, we additionally used the uncommon
scanning parameter (e.g., 80 kV). The tube current was
set to 150 mA, and with an exposure time of 0.5 s, the
mAs value was 75mAs. The slice thickness was set to
2 mm, and the voxel size was 0.98 mm × 0.98 mm
× 2 mm. The volume CT dose index were 3.3 mGy
(80 kV), 5.9 mGy (100 kV), 9.0 mGy (120 kV), and 11.6
mGy (135 kV), respectively. The peak-exhale image was
acquired at compression phase (the phantom settings
of 20 mm) and the peak-exhale image was acquired at
decompression phase (the phantom settings of 0 mm).
In addition, to evaluate the impact of CT recon-
struction algorithm, we used different six reconstruction
algorithms (FC11, FC13, FC15, FC41, FC44, and FC
52). FC11 is a reconstruction function for abdom-
inal/mediastinum conditions, and FC13 and 15 are
functions of the same series, and the larger the number,
the sharper the reconstruction function. FC41 and 44
are reconstruction functions for the head condition, and
FC44 enables sharper reconstruction than FC41. FC52
is a reconstruction function for lung field conditions and
is a function used for sharp image reconstruction. The
details of the filters are not disclosed due to vendor
business secrets.
2.3 Performing of deformable image
registration
DIR was performed between peak-exhale image and
peak-exhale image. We used the free-form deformation-
based algorithm implemented in MIM ver.6.6 (Cleveland,
OH, USA). The default smoothing factor was set to

IKEDA ET AL.
FIGURE 1 Overview and deformable thorax phantom.
0.5. We set the volume of interest (VOI) within the
inner phantom not including the motor-derived piston
surrounding the phantom.
2.4 Evaluation of DIR accuracy
To assess the accuracy of DIR, we used the target
registration error (TRE). Dice similarity coefficient is a
widely acceptable method for evaluation of DIR accu-
racy. However, this method may not suitable in the case
which the evaluation structure is extremely small (e.g.,
cross-wire). Thus, we used the TRE in this study. We set
96 landmarks (48 points on wire intersections, and 48
points on markers) on both peak-inhale and peak-exhale
images. To calculate TRE using landmarks, we applied
the formula below.
√ (FC41), 3.96–5.56 mm (FC44), and 4.42–6.96 mm
2 2 2
TRE = (x1 − x2) + (y1 − y2) + (z1 − z2) ,
where, x1, y1, and z1 indicate the landmark coordinates
on the fixed image, while x2, y2, and z2 indicate the
landmark coordinates on the deformed image. The TRE
was calculated for all landmarks, and the mean value of
which was calculated.
To correlation between DIR accuracy and CT image
quality, we calculated the standard deviation (SD) of CT
value within cubic ROI which covered one cross-wire at
peak-inhale image.16
15269914, 0, Downloaded from https://aapm.onlinelibrary.wiley.com/doi/10.1002/acm2.13917 by Cochrane Poland, Wiley Online Library on [25/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
3 of 8
In addition, we did the ANOVA statistical analysis to
test the significant difference in TRE among all tube volt-
ages or reconstruction algorithms using JMP software
(version 16.2.0, SAS Institute Inc., USA) (significant:
p < .05).
3 RESULT
Figure 2 shows images obtained with each reconstruc-
tion algorithm and tube voltage. We found that blurring
and sharpness of markers visually change according to
changes in tube voltage and reconstruction algorithm.
Figure 3 presents a summary of the TRE according
to all reconstruction algorithms and tube voltages com-
bined. In case with different tube voltages, the range
of average TRE were 4.59–5.93 mm (FC11), 4.44–
5.69 mm (FC13), 4.26–5.39 mm (FC15), 4.38–5.83 mm
(1) (FC52). In case with different reconstruction algorithms,
the range of average TRE were 4.50–6.55 mm (80 kV),
3.96–4.96 mm(100 kV), 4.26–4.59 mm (120 kV), and
5.39–6.96 mm (FC52). The TRE differed by up to
3.0 mm (3.96–6.96 mm) depending on the combina-
tion of tube voltage and reconstruction algorithm. The
significance differences were observed among tube
voltages (p = 0.002) but they were not observed among
reconstruction algorithms (p = 0.609).
Figure 4 shows the difference images of the fixed
image and deformed image in typical two cases (the
15269914, 0, Downloaded from https://aapm.onlinelibrary.wiley.com/doi/10.1002/acm2.13917 by Cochrane Poland, Wiley Online Library on [25/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
IKEDA ET AL.

Summary of the DIR accuracy for each reconstruction algorithm and tube voltage.
CT image obtained by different reconstruction algorithms and tube voltages.
FIGURE 2

FIGURE 3
4 of 8
 15269914, 0, Downloaded from https://aapm.onlinelibrary.wiley.com/doi/10.1002/acm2.13917 by Cochrane Poland, Wiley Online Library on [25/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
IKEDA ET AL. 5 of 8

F I G U R E 4 Difference images between the fixed image and deformed image for the typical two cases (case with the highest TRE and the
case with the lowest TRE).

FIGURE 5 The relationship between SD and TRE for each reconstruction algorithm.

case with the highest TRE and the case with the low-
est TRE). In the case with the highest TRE, deformed
image showed more deformation compared to the fixed
image. Figures 5 and 6 show the correlation between
SD and TRE when the tube voltage or reconstruction
algorithm was constant. Depending on the tube voltage,
a negative strong correlation between the SD and TRE,
or a positive strong correlation was observed.
4 DISCUSSION
In the present study, we examined how the accu-
racy of DIR is affected by CT scan parameters.
Our result showed that the TRE changed from
4.56 to 4.59 mm due to reconstruction algorithm. In
addition, the TRE changed from 4.44 to 5.69 mm
by changing the tube voltage. This result sug-
gested that DIR accuracy between peak-exhale
and peak-inhale images may be changed due to
combination of tumor voltage and reconstruction
algorithm.
Regarding the reconstruction algorithm, our result
showed that the TRE changed from 4.56 to 4.59 mm
due to the algorithms. Different reconstruction algo-
rithms caused the different amount of blurring for
CT image. The DIR accuracy is reduced in blurred
images. Davis et al. showed the highest HU differences

6 of 8
FIGURE 6 The relationship between SD and TRE for each tube voltage.
(e.g., 56 HU for density to water and 117 HU for
bone-like materials) in CT images when changing recon-
struction algorithm using Toshiba Aquilion LB scanner
which was used in our study.17 Their result is consistent
with our result. In addition, we found the FC52 algo-
rithm caused the largest difference in the TRE. Since
this algorithm is created for lung regions, this has a
higher image resolution. On the other hand, the image
noise was emphasized. This can be explained by the
fact that the images with FC52 had higher SD values
around markers. This emphasized noise may hinder cor-
responding anatomical points between fixed and moving
images during DIR process, resulting in lower DIR
accuracy.
Regarding the tube voltages, the tube voltages
changed the TRE in all reconstruction algorithms (e.g.,
FC13, TRE:4.44–5.69 mm). Overall, the TRE tended
to decrease between 80 and 120 kV and the TRE
increased for 135 kV. The reason for this may be that
increased x-ray intensity, which caused by the increased
tube voltage, reduced the image contrast. Konishi et al.
reported that increasing the tube voltage increased the
x-ray intensity,18 and that increasing the number of x-
ray photons that reach the detector created images with
less noise. However, they reported that high tube volt-
age for CT scans reduces the difference in the linear
attenuation coefficient of the material. Due to this effect,
our result with 135 kV may have the lower DIR accuracy
(e.g., higher TRE). That is, the difference in the linear
attenuation coefficient between the marker and sponge
used in the phantom became smaller, resulting in lower
contract.
15269914, 0, Downloaded from https://aapm.onlinelibrary.wiley.com/doi/10.1002/acm2.13917 by Cochrane Poland, Wiley Online Library on [25/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
IKEDA ET AL.
Our result suggested that the DIR accuracy may be
affected by CT scan parameters. The AAPM TG 132
task report describes the DIR evaluation (e.g., com-
missioning and regularly QA for DIR) method using a
physical phantom, virtual phantom, and actual patient
images.7,19 When the virtual phantom and actual patient
images were used, we can check the impact of the
DIR software configuration and DIR parameter setting
on DIR accuracy, but we cannot check the impact of
CT scanning parameters on DIR accuracy. On the other
hand, physical phantom can do the end-to-end test
for DIR accuracy including the impact of CT scanning
parameters.
To achieve higher DIR accuracy, we may optimize the
suitable reconstruction algorithm and tube voltage. How-
ever, in clinical practice, since we basically determine
the optimized reconstruction algorithms and tube volt-
age based on the clinical impact, it may be difficult to
optimize the reconstruction algorithm and tube voltage
for DIR performance. Samavati et al. reported that the
hybrid algorithm improved the DIR performance (e.g.,
the TRE by up to 1 mm).12 When you cannot optimize
the reconstruction algorithm and tube voltage for DIR
accuracy, the optimization for DIR method can be useful
for improvement of the DIR accuracy.
In addition, Samavati et al. reported that 1.6 mm
of TRE caused the 1 Gy dose error of DIR-based
total accumulated dose for minimum dose.20 Our result
showed that the combination of the reconstruction algo-
rithm and tube voltage caused the large difference
(TRE >3 mm). Based on the result reported by Sama-
vati et al., we estimated that this TRE difference caused

IKEDA ET AL.
uncertainty of 1 Gy or more in the total accumulated
dose.
The present study has a limitation. First, we used
only one CT scanner. We are scheduled to use the
different CT scanner. Next, we used the only one DIR
software program (e.g., MIM Mestro). Different DIR
algorithm implanted in different DIR software (e.g.,
Velocity, and RayStation) may have different result.
Second, the phantom which we used is designed for
thoracic region and reproduces the only high con-
trast tissue. Other tissue type with low contrast (e.g.,
parotid) may have the different impact of CT scan
parameters on DIR accuracy. Further evaluation is
needed to clearly this. Third, we only evaluated the
two main CT scan parameters (i.e., tube voltage and
reconstruction algorithm). In general, the two CT scan
parameters (i.e., tube voltage and reconstruction filter)
are relatively easy to change in acquisition of treatment
planning CT images. Thus, we used these two CT scan
parameters.
There are other CT scan parameters (e.g., pitch and
CT slice thickness). These parameters may have impact
for DIR accuracy. In the near future, we will investigate
the different other imaging parameters. Finally, The TRE
values were relatively large compared to those that can
be found in the literature.6,7 This may due to usage of
simple physical phantom which cannot provide the rich
information about surrounding tissues for DIR. In addi-
tion, the reason for this may be that in both fixed and
moving images, FOV for DIR was mostly occupied by
air, so there was little information to optimally perform
the DIR.
5 CONCLUSION
We evaluated the effect of CT image scan parameters
on DIR accuracy (TRE). Our result indicated that CT
scan parameters had moderate impact of TRE, espe-
cially for reconstruction algorithms for the deformable
thorax phantom which had the high contrast tissue. The
TRE were differed by up to 3.0 mm (3.96–6.96 mm)
depending on the combination of tube voltage and
reconstruction algorithm.
AU T H O R C O N T R I B U T I O N S
RI and NK contributed to the conception and design of
the study. RI and YN wrote the program. RI and NK per-
formed the analysis. RI mainly drafted the manuscript.
NK, YN, SI, TS, and KJ reviewed the manuscript. All
authors read and approved the final manuscript.
AC K N OW L E D G M E N T S
The authors would like to thank the anonymous referees
for their extremely thorough review, which led to signifi-
cant changes that substantially improved the quality of
this paper.
15269914, 0, Downloaded from https://aapm.onlinelibrary.wiley.com/doi/10.1002/acm2.13917 by Cochrane Poland, Wiley Online Library on [25/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
7 of 8
C O N F L I C T O F I N T E R E S T S TAT E M E N T
The authors declare no conflict of interest with regard to
this manuscript.
REFERENCES
1. Abe Y, Kadoya N, Arai K, et al. Effect of DIR uncertainty on
prostate passive-scattering proton therapy dose accumulation.
Phys Med. 2017;39:113-120.
2. Lu W, Olivera GH, Chen Q, et al. Deformable registration
of the planning image (kVCT) and the daily images (MVCT)
for adaptive radiation therapy. Phys Med Biol. 2006;51:4357-
4374.
3. Takagi H, Kadoya N, Kajikawa T, et al. Multi-atlas-based
auto-segmentation for prostatic urethra using novel predic-
tion of deformable image registration accuracy. Med Phys.
2020;47:3023-3031.
4. Ieko Y, Kadoya N, Kanai T, et al. The impact of 4DCT-
ventilation imaging-guided proton therapy on stereotactic body
radiotherapy for lung cancer. Radiol Phys Technol. 2020;13:230-
237.
5. Kadoya N, Nemoto H, Kajikawa T, et al. Evaluation of four-
dimensional cone beam computed tomography ventilation
images acquired with two different linear accelerators at vari-
ous gantry speeds using a deformable lung phantom. Phys Med.
2020;77:75-83.
6. Kadoya N, Nakajima Y, Saito M, et al. Multi-institutional validation
study of commercially available deformable image registration
software for thoracic images. Int J Radiat Oncol Biol Phys.
2016;96:422-431.
7. Brock KK, Mutic S, McNutt TR, Li H, Kessler ML. Use of image
registration and fusion algorithms and techniques in radiotherapy:
report of the AAPM radiation therapy committee task group no.
132. Med Phys. 2017;44:e43-e76.
8. Miyakawa S, Tachibana H, Moriya S, Kurosawa T, Nishio T.
Evaluation of deformation parameters for deformable image
registration-based ventilation imaging using an air-ventilating
non-rigid phantom. Phys Med. 2018;50:20-25.
9. Chang J, Suh TS, Lee DS. Development of a deformable lung
phantom for the evaluation of deformable registration. J Appl Clin
Med Phys. 2010;11:3081.
10. Pukala J, Johnson PB, Shah AP, et al. Benchmarking of
five commercial deformable image registration algorithms for
head and neck patients. J Appl Clin Med Phys. 2016;17:25-
40.
11. Lambin P, Rios-Velazquez E, Leijenaar R, et al. Radiomics:
extracting more information from medical images using advanced
feature analysis. Eur J Cancer. 2012;48:441-446.
12. Samavati N, Velec M, Brock K. A hybrid biomechanical intensity
based deformable image registration of lung 4DCT. Phys Med
Biol. 2015;60:3359-3373.
13. Mackin D, Fave X, Zhang L, et al. Measuring computed tomog-
raphy scanner variability of radiomics features. Invest Radiol.
2015;50:757-765.
14. Kim YJ, Lee HJ, Kim KG, Lee SH. The effect of CT scan parame-
ters on the measurement of CT radiomic features: a lung nodule
phantom study. Comput Math Methods Med. 2019;2019:8790
694.
15. Sugawara Y, Kadoya N, Kotabe K, et al. Development of a
dynamic deformable thorax phantom for the quality manage-
ment of deformable image registration. Phys Med. 2020;77:100-
107.
16. K I. CTmeasure. Japanese society of CT technology. 2012–2014.
17. Davis AT, Palmer AL, Pani S, Nisbet A. Assessment of
the variation in CT scanner performance (image quality and
Hounsfield units) with scan parameters, for image optimisa-
tion in radiotherapy treatment planning. Phys Med. 2018;45:59-
64.
 15269914, 0, Downloaded from https://aapm.onlinelibrary.wiley.com/doi/10.1002/acm2.13917 by Cochrane Poland, Wiley Online Library on [25/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
8 of 8 IKEDA ET AL.

18. Konishi T, Asai Y, Uemura M, Yamada K, Matsumoto M. Vari-

ation in tube voltage dependence of X-ray CT image contrast How to cite this article: Ikeda R, Kadoya N,
attributed to the difference of object materials. Med Imag Inform
Sci. 2008;25:90-95.
Nakajima Y, Ishii S, Shibu T, Jingu K. Impact of
19. McGurk R, Seco J, Riboldi M, Wolfgang J, Segars P, Paganetti CT scan parameters on deformable image
H. Extension of the NCAT phantom for the investigation of intra- registration accuracy using deformable thorax
fraction respiratory motion in IMRT using 4D Monte Carlo. Phys phantom. J Appl Clin Med Phys. 2023;e13917.
Med Biol. 2010;55:1475-1490. https://doi.org/10.1002/acm2.13917
20. Samavati N, Velec M, Brock KK. Effect of deformable registra-
tion uncertainty on lung SBRT dose accumulation. Med Phys.
2016;43:233.