ArticlePDF Available

Figures

No caption available
… 
Content may be subject to copyright.
3rd ESTRO Forum 2015 S817
thickness provided the movement be consistent and periodic.
For non-sinusoidal movements phases far from 0% and 50%
may introduce significant discrepancies. Both systems are
compatible but respiratory training would be necessary to
guarantee consistency for a gating strategy while it would not
for a MIP-based ITV strategy.
This work was financed by grant FIS PI12-00521
EP-1500
3D versus 4D cone beam computed tomography for
lipiodolñguided radiotherapy of hepatocellular carcinomas
M.K.H. Chan1, A.C.C. Chiang1, F.A.S. Lee1, R.W.K. Leung1,
M.Y.P. Wong1, G.M.L. Law1, K.H. Lee1, S.Y. Tung1, V.W.Y.
Lee1
1Tuen Mun Hospital, Clinical Oncology, Hong Kong, Hong Kong
(SAR) China
Purpose/Objective: Evaluate the positional uncertainty of
hepatocellular carcinomas (HCC) guided by lipiodol on
respiratory–correlated (4D) and uncorrelated (3D) cone beam
computed tomography (CBCT).
Materials and Methods: Elekta XVI v4.5 (Elekta, Crawley, UK)
was used to acquire 4D CBCT of 15 HCCs treated by
hypofractionated radiotherapy after single trans–arterial
chemoembolization (TACE) with lipiodol. 1320 x–ray
projections per 4D CBCT scan were sorted into 10 CBCT
dataset (132 projections per CBCT dataset). A 4D registration
workflow was followed to register the reconstructed time–
weighted average CBCT with the planning mid–ventilation CT
by an initial bone registration of the vertebrae and then
tissue registration of the lipiodol. For comparison, projection
data of each 4D CBCT were used to synthesize 3D CBCT
images without motion extraction. Uncertainties of the
treatment setup estimated from the absolute lipiodol
position and the interfractional lipidol drift relative to
vertebrae were analyzed separately from 4D and 3D CBCT
images.
Results: Qualitatively, 3D CBCT showed better lipiodol
contrast than 4D CBCT primarily because of a tenfold
increase of projection data applied to the reconstruction.
Some motion artifacts were observed on the 3D CBCT but not
on 4D CBCT (Fig. 1). Group mean, systematic and random
errors estimated from 4D and 3D CBCT are similar, agree to
within 0.7 mm in the cranio–caudal (CC), and anterior–
posterior (AP) directions, and 0.3 mm in left–right (LR)
direction. Systematic and random errors are largest in the CC
direction, amounting to 4.7 mm and 3.9 mm from 3D CBCT
and 5.5 mm and 4.0 mm from 4D CBCT in terms of the
absolute lipiodol position, and 3.7 mm and 3.0 mm from 3D
CBCT and 4.3 mm and 2.8 mm from 4D CBCT in terms of the
lipiodol baseline drift relative to vertebrae, respectively.
Margin calculated from 3D CBCT and 4D CBCT differed by less
than 1.9 mm, 0.2 mm and 0.1 mm in the CC, AP, and LR
directions in the patient cohort.
Conclusions: 3D and 4D CBCT were found equivalent in
localizing HCCs guided by lipiodol, resulting in similar safety
margin. 4D CBCT offers the advantage of measurements of
the changes of tumor motion for assessing the adequacy of
the planning margin.
EP-1501
nfluence of prostate rotation on dose distribution in the
target volume
E. Dabrowska1, B. Brzozowska2, B. Chaber3, P. Kukolowicz1, J.
Starzynski3, A. Zawadzka1
1Maria Sklodowska-Curie Memorial Cancer Center,
Department of Medical Physics, Warsaw, Poland
2University of Warsaw, Department of Biomedical Physics,
Warsaw, Poland
3Warsaw University of Technology, Faculty of Electrical
Engineering, Warsaw, Poland
Purpose/Objective: The inter-fraction prostate motion can
have a significant impact on dose distribution in the target
volume. However, the concept of the PTV was developed to
minimize the influence of all uncertainties of target position
on dose distribution, the problem of rotations is not entirely
analyzed. The aim of this study was to investigate the
influence of the prostate rotation on delivered dose for three
different techniques.
Materials and Methods: For 10 prostate cancer patients
previously treated in our hospital 3D-CRT, IMRT and SIB-IMRT
treatment plans were prepared. Internal rotations of the
prostate in the range from -27º to +27º relative to the apex in
anterior-posterior direction were introduced. Rotations in
rectum direction was defined as a negative angle. Based on
originally prepared plans, the DVH for rotated prostate was
calculated and compared with the DVH obtained for non-
rotated structure. Changes of Dmin, D99%, Dmax, D1%, Dmean,
standard deviation and V95% as a function of the rotation
angle were analyzed. For each treatment planning technique
the influence of rotation on the TCP was also determined.
S818 3rd ESTRO Forum 2015
Results: For prostate the average value of the D99% were
99.19% ± 0.25%, 79.62% ± 13.23%, 91.42% ± 7.19% of
prescribed dose, for 0º, -27º, +27º rotation angles
respectively. For IMRT and SIB-IMRT techniques decrease of
D99% was reduced for maximum angles by 12% and 26.5%
respectively and results were rotation direction independent.
The rotation had no impact on D1%. A negligible influence of
rotation for 3D-CRT and IMRT on the Dmean was observed. For
3D-CRT and IMRT techniques decreased of Dmean was greater
than 2% only for maximum analyzed angles in 1 and 3 cases
respectively. For SIB-IMRT, the 2% decrease of Dmean was
observed for 5 patients for rotations larger than 21º,
regardless of the direction of rotation. For group of 8
patients with low grade tumors the average value of the TCP
for non-rotated prostate were calculated and equal to 83.4%
± 0.2% and 83.3% ± 0.3% for 3D-CRT and IMRT respectively.
For rotations smaller than 18º the TCP was close to 80% for
all patients. For SIB-IMRT plans the TCP decreased from 88.3
% to 80.0% for 18º rotations.
Conclusions: Our results showed that the change of dose
distribution in the target volume depend on the angle of
rotation and the treatment delivery technique. Only rotations
larger than about 21o influence significantly on the DVH and
the TCP. The 3D-CRT showed the smallest sensitivity to
prostate rotations. SIB-IMRT technique was the most sensitive
to rotations, however the increased prescribed dose to
prostate compensated reduction in the TCP. It remained
comparable to the TCP obtained for 3D-CRT and IMRT
techniques.
EP-1502
Lipiodol versus diaphragm as tumor surrogate in 4D CBCT-
guided radiotherapy of hepatocellular carcinomas
M.K.H. Chan1, A.C.C. Chiang1, R.W.K. Leung1, M.Y.P. Wong1,
F.A.S. Lee1, G.M.L. Law1, K.H. Lee1, S.Y. Tung1, V.W.Y. Lee1
1Tuen Mun Hospital, Department of Clinical Oncology, New
Territories, Hong Kong (SAR) China
Purpose/Objective: To compare lipiodol and diaphragm as
tumor surrogate in hypofractionated radiotherapy of
hepatocellular carcinomas (HCC) using 4D cone beam
computed tomography (CBCT).
Materials and Methods: Treatment verification 4DCBCT scan
were acquired using Elekta XVI v.4.5 (Elekta, Crawley, UK)
for 15 HCC patients who had prior single transarterial
chemoembolization (TACE) with lipiodol. Automatic 4DCBCT
image registration with the planning mid-ventilation images
was performed initially by bone registration of the vertebrae
followed by a 4D registration based on either lipiodol or
diaphragm on the reconstructed time-weighted average
images. Uncertainties of treatment setup and interfractional
tumor baseline drift estimated by lipiodol and diaphragm
were analyzed.
Results: All lipiodolized HCCs were clearly visualized on the
time-weighted average 4D CBCT images. Lipiodol visibility
decreased with increasing tumor size due to limited amount
of lipiodol per TACE. Group means and random errors of the
treatment setup and interfractional baseline shift based on
lipiodol and diaphragm are similar, agree to within 0.5 mm in
left-right (LR) and anterior-posterior (AP), and 0.1 mm in
cranio-caudal (CC) directions, and systematic errors differ by
1.5 mm, 0.7 mm and 0.2 mm in the LR, AP and CC directions,
respectively. Using lipiodol instead of diaphragm as tumor
surrogate in our margin calculation led to 0.6 mm decrease in
the CC direction, 1.3 mm and 3.1 mm increase in AP and LR
directions in the patient cohort, respectively.
Conclusions: Lipiodol offers the potential of target
localization with motion blurring minimized on the 4D CBCT
images. Using diaphragm as tumor surrogate can lead to
misalignment of the tumor. The amount of TACE lipiodol may
need to be adjusted according to the tumor size to improve
its visibility on 4D CBCT.
EP-1503
Rigid and non-rigid registration dose propagation for
brachytherapy treatment combined with radiotherapy
J. Krayenbuehl1, M. Guckenberger1, C. Linsenmeier1, S.
Kloeck1, M. Zamburlini1
1University Hospital Zürich, Department of Radiation
Oncology, Zurich, Switzerland
Purpose/Objective: Cervical cancer is the fourth most
common cancer in women with an estimate of 266’000 death
per year worldwide. It has been reported that by combining
HDR brachytherapy (HDR) with external beam radiotherapy
(EBRT) the local tumor control can be further improved. The
dose distributions for these two treatment modalities are
based on CTs which differ in the patient positioning. The
dose planned with EBRT and HDR has to be added together in
order to assess the dose to the target and to the organ at
risk. Furthermore, the medial field border for EBRT will be
based on the dose distribution from the HDR plan. In order to
propagate the dose from one CT to the other, image
registration is required. In this work we evaluated the benefit
of using a non-rigid algorithm compared to a rigid registration
for dose propagation.
Materials and Methods: Ten patients treated with HDR
(5x5Gy, Manchester method) combined with Boost-EBRT to
the parametrium (3x2Gy) were included in this study. The
EBRT treatment consisted of 2 opposing fields, in which the
medial border was based on the 50-80% isodose line from the
first HDR treatment. The total dose was calculated by
registering the CT in a rigid or a non-rigid way using Velocity
(Varian Medical System, Palo Alto, CA). The total dose
obtained with the rigid and non-rigid registration was then
compared together based on dose distribution as well as
dose-volume histogram parameters.
Results: The position of the 50% HDR isodose line propagated
on the EBRT CT between the rigid and non-rigid registration
was different by up to 1cm in all directions, thus
substantially affecting the choice of field size for the
subsequent EBRT. The change of dose distribution between
both registration modalities did not affect the dose to the
femoral heads. Larger dose difference to the rectum and
bladder was observed. Dose to 0.1cm3, 1cm3, 2cm3 and
5cm3 of the rectum volume could increase by up to 37%, 22%,
18% and 10%, respectively. For the bladder, dose to 0.1cm3,
1cm3, 2cm3 and 5cm3 could change by more than 41%, 21%,
22% and 28%, respectively.
Conclusions: The registration modality used for dose
propagation has to be carefully evaluated when matching CTs
were taken with the patient lying in different positions, as is
the case in HDR and EBRT. Wrong registration could results in
cold spots in the tumor region or hot spots in the organs at
risk.
Article
Full-text available
Purpose: The focus of this work is to improve the available kV image quality for continuous intrafraction monitoring of the prostate during volumetric modulated arc therapy. This is investigated using a novel blade collimation system enabling tube current modulated (TCM) volume-of-interest (VOI) imaging of prostate fiducial markers during radiotherapy, and Monte Carlo simulation of MV scatter. Materials and methods: A four-blade dynamic kV collimator was used to track a VOI containing gold fiducial markers embedded in a dynamic pelvis phantom during gantry rotation. For each fiducial a VOI margin around each marker was set to be 2σ of the population covariance matrix characterizing prostate motion. This was used to conform to a single or several fiducials and compared to a static field. DRRs were used to calculate the kV attenuation for each VOI as a function of angle and used to optimize x-ray tube current during acquisition. Image quality was assessed with regard to contrast-to-noise ratio (CNR), fiducial detectability and imaging dose. Monte Carlo simulations in EGSnrc were used to calculate the imaging dose to the phantom and MV scatter fluence to the imaging panel. Results: Fiducials can be accurately located using a VOI containing a single or several fiducials using a relatively high constant kV output. However, when using a 6x6cm(2) field the dose can be upwards of 1.5 Gy in bone for constant kV output and 3.1 Gy when applying TCM at 1 Hz imaging over the course of 40 fractions. This can be mitigated through tailoring the imaging field to a single or several fiducials, in which the integral dose is reduced by a factor of 15.6 and 3.7, respectively. For a constant MV treatment field size, the scattered fluence reaching the kV panel varies by less than a factor of two for a completely rotation of the gantry. However, the MV scatter spectrum overlaps with the detector response for a deleterious effect, with a peak MV scatter energy of approximately 100 keV. TCM can be used to overcome the variability in image quality throughout the rotation and therefore improve fiducial CNR and detectability during periods of high kV attenuation. Conclusions: The combination of VOI and TCM introduces an advantageous approach in intrafraction monitoring of the prostate during radiotherapy by both reducing and localizing the imaging dose, while improving image quality and fiducial detectability during periods of high kV attenuation. Additionally, the influence of MV scatter has been shown to be most important in low attenuation regions, with a variation by a factor of two. This article is protected by copyright. All rights reserved.
ResearchGate has not been able to resolve any references for this publication.