293 / 1082
S280
ESTRO 36 2017
_______________________________________________________________________________________________
torso contains lungs and ribs and is made of materials
mimicking human tissue Hounsfield Unit. The lungs are
shaped thus to host four water equivalent inserts (WEI)
that simulate lung lesions and a fillable QA sphere (QAS)
to test reconstruction performances of 4D scanners. Films
can be inserted between two WEIs and a specific housing
for a diamond detector has been drilled into one WEI. 6
small tin markers have also been included near the fourth
WEI to test markers guided tracking procedure. The
anterior part of the phantom moves up and down in sync
with lungs movements driven by an Arduino programmable
board hosted in the caudal phantom portion. Elliptical
paths (axes up to 2cm), pre-programmed in the
microcontroller, and patient specific respiratory
movements, programmable by users, can be chosen on
the LCD screen placed on the caudal extremity of the
phantom. Preliminary tests were performed to assess
Adam usability and its performances in terms of HU, WEI
motion repeatability and lungs-to-surface motion
correlation. Finally a VMAT plan, to deliver 18Gy to one
internal WEI, was planned on an average reconstructed
4DCT data set and delivered to ADAM. Dose was prescribed
to 95% of the PTV = ITV (encompassing all motion and
delineated on MIP) + 5 mm margin. The delivered dose
distribution, measured with a Gafchromic EBT3 film, was
overlapped on the WEI to assess moving target dose
coverage.
Results
In Figure 1 ADAM and some internal details, as appear in
CT images, are presented.
CT acquisitions demonstrate realistic human tissues HU
values: -860±37, 77±30, 83±20, 1098±84 respectively for
lung, thorax soft tissue, WEI and bone. The absence of
artifacts of reconstructed QAS and WEIs in all phases,
demonstrate lungs-to-surface motion correlation.
Movement tests show a long (20 days) and short-term
(30min) amplitude repeatability <1 mm along both axes.
In Figure 2 measured dose distribution delivered to a
moving target is shown together with the QAS volume
measured in static CT and in some 4DCT reconstructed
phases.
Conclusion
ADAM demonstrates suitable performances to test
instruments and methods used to treat moving lesions.
OC-0534 Establishment of patient-specific quality
assurance procedure for Dynamic WaveArc delivery
technique
H. Hirashima
1
, N. Nakamura
1
, Y. Miyabe
1
, N. Mukumoto
1
,
M. Uto
1
, K. Nakamura
1
, T. Mizowaki
1
, M. Hiraoka
1
1
Kyoto University Hospital, Department of Radiation
Oncology and Image applied Therapy, Kyoto, Japan
Purpose or Objective
Dynamic WaveArc (DWA) is a novel delivery techniquethat
uses the Vero4DRT system (Mitsubishi Heavy Industries
[MHI], Ltd., Tokyo, Japan, and Brainlab, Feldkirchen,
Germany) for volumetric modulated arc therapy, with
continuous non-coplanar delivery. DWA achieves high-dose
conformality by optimizing the dose rate, gantry speed,
ring speed, and positions of the dynamic multi-leaf
collimator (MLC). The purpose of this study was to
establish the patient-specific quality assurance (QA)
procedure for the DWA.
Material and Methods
Twenty DWA plans, 10 for brain tumors and 10 for prostate
cancer, were created using RayStation version 4.7
(RaySearch Laboratories, Stockholm, Sweden). The
prescribed dose for the brain tumor was set as 52.2 Gy at
1.8 Gy per fraction, and that for the prostate tumor was
set as 76 Gy at 2 Gy per fraction. The patient-specific QA
included the accuracy verification of the measured dose
distribution, machine movement, and reconstructed dose
distribution. First, absolute dose distributions measured
by ArcCHECK (Sun Nuclear Inc, Melbourne, FL) were
assessed by using 3%/3 mm global γ-tests for the area
receiving more than 10% isodose, based on the AAPM
TG119 report. Next, the log files were analyzed to
evaluate the accuracy of the machine movement. The log
files, including the actual gantry angle, ring angle, MLC
position, and monitor unit (MU) were obtained at 50 ms
intervals. Root mean square errors (RMSEs) between the
planned and actual values in the log files were calculated.
Finally, the delivered dose distributions were
reconstructed based on the log files. The in-house
software was used to load the original DICOM-RT plan file,
and searched the actual values at each control point.
Thereafter, the in-house software replaced the planned
values at each control point with the actual values based
on the log file, and then generated a reconstructed
DICOM-RT plan file. The reconstructed plan was imported
into RayStation, and then recalculated on the planning CT.
The originally-planned dose-volumetric indices were
compared with the reconstructed ones.
Results
In the ArcCHECK dosimetry, the means ± standard
deviations (SDs) of the γ-pass rates were 97.0% ± 1.2%
(range, 94.3%–98.7%) and 98.4% ± 1.0% (range, 96.6%–
99.5%) for the brain and prostate tumors, respectively. In
the log file analysis, the RMSEs for the gantry angle, ring