828 / 1096
S812
ESTRO 36
_______________________________________________________________________________________________
fractionation, even small relative errors can lead to
serious complications to the normal tissue or recurrences
of the tumor. So delivery quality assurance (DQA) in
SRS/SBRT is very critical and poses unique challenges due
to extremely high dose gradients and lack of electronic
equilibrium. For this reason, dose rate independent
dosimeters with precise, high spatial resolution and 3D
capabilities are essential as reported by the Council on
Ionizing Radiation Measurements and Standards (CIRMS).
Material and Methods
The new CrystalBall system (3D Dosimetry, Madison, CT,
USA) is designed for DQA with sub-millimeter spatial
resolution in 3D. The system is composed of a fast laser CT
scanner (OCTOPUS, MGS Research, Inc, Madison, CT) and
reusable tissue-equivalent radiochromic polymer gel
sphere-mounted on a special QA phantom. Gold fiducial
markers are affixed in different locations of the phantom
for image guidance with fiducial tracking for CyberKnife
(CK) robotic SRS/SBRT system (Accuray, Sunnyvale, CA).
The CT images of the CrystalBall gel phantom were
transferred to the CK Multiplan treatment planning
system. A DQA plan was generated by superimposing a
patient plan onto the gel phantom CT data set. The DQA
plan was then sent for CK irradiation. The CrystalBall’s
VOLQA software registers the plan DICOM CT dataset with
the laser CT of the irradiated gel, creates OD/cm to dose
calibration curve and then compares the CrystalBall
irradiation measurements with the Multiplan’s DQA plan.
It generates QA reports that feature overlays of isodoses
in 2D and 3D, profiles, DVHs, voxel statistics, and pass/fail
metrics for dose difference and distance-to-agreement
according to gamma index criteria. In this study, we
performed DQA for four CK patients who received
treatment for brain metastasis, spine metastasis and
trigeminal neuralgia as recommended by AAPM TG-135.
For each patient, the DQA was done three times.
Results
Figures 1 and 2 show the CrystalBall phantom setup with
OD/cm to dose auto-calibration, 2D and 3D overlay of
isodoses for a patient, respectively.
Table 1 shows results of the study for gamma evaluation
passing averages for the DQA of the four patients. For all
patients studied, we found a passing rate of more than 96%
with gamma index criteria of 2 % dose difference and 2
mm distance-to-agreement. For 3 % and 3 mm criteria, the
passing rate is found to be above 99%.
Conclusion
Our DQA results suggest that the newly developed
CrystalBall QA phantom system for robotic radiosurgery
can be ideal tool for 3D dose verification with isotropic
sub-millimeter spatial resolution and film-equivalent
accuracy. This 3D tool can offer unique advantage over
other existing 2D tools and techniques in terms of high-
resolution DQA necessary for radiotherapy with minimal
additional physics resources.
Electronic Poster: Physics track: Radiation protection,
secondary tumour induction and low dose (incl.
imaging)
EP-1514 Planar kV imaging dose reduction study for
Varian iX and TrueBeam linacs
E. Gershkevitsh
1
, D. Zolotuhhin
1
1
North-Estonian Regional Hospital Cancer Center
Radiotherapy, Radiotherapy, Tallinn, Estonia
Purpose or Objective
IGRT has become an indispensable tool in modern
radiotherapy with kV imaging used in many departments
due to superior image quality and lower dose when
compared to MV imaging. Since, the frequency of kV
images continues to increase (intrafractional imaging,
etc.) the reduction of additional dose assumes high
priority. Many departments use manufacturer supplied
protocols for imaging which are not always optimised
between image quality and radiation dose (ALARA).
Material and Methods
Whole body phantom PBU-50 (Kyoto Kagaku ltd., Japan)
for imaging in radiology has been imaged on Varian iX OBI
1.5 and TrueBeam 2.5 accelerators (Varian Medical
Systems, USA). Manufacturer’s default protocols were
adapted by modifying kV and mAs values when imaging
different anatomical regions of the phantom (head,
thorax, abdomen, pelvis, extremities). Images with
different settings were independently reviewed by two
persons and their suitability for IGRT set-up correction
protocols were evaluated. The suitable images with the
lowest mAs were then selected. The entrance surface dose
(ESD) for manufacturer’s default protocols and modified
protocols were measured with RTI Black Piranha (RTI
Group, Sweden) and compared. Image quality was also
measured with kVQC phantom (Standard Imaging, USA) for
different protocols. The modified protocols have been
applied for clinical work.
Results
The default manufacturer’s protocols on TrueBeam linac
yielded 9.4 times lower ESD than on iX linac (range 2.5-
24.8). For most cases it was possible to reduced the ESD
on average by a factor of 3 (range 0.9-8.5) on iX linac by
optimising imaging protocols. Further ESD reduction was
also possible for TrueBeam linac.