291 / 1082
S278
ESTRO 36 2017
_______________________________________________________________________________________________
profiles of the individual beams into account was fitted to
the data. The lateral profiles of the WFF beams were
assumed to be uniform while those of the FFF beams were
approximated using 4th or 6th order polynomials. The
scatter functions of the FFF beams were recalculated
using a uniform lateral profile (the same as the physical
profile of the WFF beams) and are henceforth denoted as
virtual uniform FFF beams (VUFFF). The field sizes of the
VUFFF beams having the same scatter contribution as the
corresponding FFF beams at a given field size were defined
as the EQUSFs.
Results
In total, the data of four different LINACs and 18 different
beam energies were analysed. The average values of
EQUSFs over all investigated LINACs of the conventional 10
x 10 cm² reference fields of 6 MV and 10 MV FFF beams for
C-arm LINACs and machine specific reference fields for
CyberKnife and TomoTherapy were 9.5 cm, 9 cm, 5.3 cm
and 6.5 cm, respectively. The standard deviation of these
averaged EQUSFs was below 0.1 cm. Figure 1 illustrates
the process of finding the EQUSFs of a FFF field. The
scatter functions of a 10 MV WFF, a 10 MV FFF and a 10 MV
VUFFF beam are plotted as a function of field size. In this
example the 10 x 10 cm² reference field of the 10 MV FFF
beam has the same scatter contribution as the 9 x 9 cm²
field of the 10 MV VUFFF beam which was defined as
EQUSFs. The scatter function of the 10 MV WFF was
plotted for comparison purposes.
Conclusion
It has been shown that with the introduction of a VUFFF
beam, EQUSFs can be consistently defined for a variety of
energies and collimations. These EQUSFs serve as basis for
a unified reference dosimetry protocol for all different
types of FFF machines.
References
[1] Fogliata A, Fleckenstein J, Schneider F, et al.
Flattening filter free beams from TrueBeam and Versa HD
units: Evaluation of the parameters for quality assurance.
Med. Phys. 2016;43.
[2] Chang Z, Wu Q, Adamson J, et al. Commissioning and
dosimetric characteristics of TrueBeam system:
Composite data of three TrueBeam machines. Med Phys.
2012;39:6981–7018.
OC-0531 The influence of detector resolution on pre-
treatment quality assurance in SBRT
A. Bruschi
1
, S. Russo
2
, M. Esposito
2
, S. Pini
2
, A. Ghirelli
2
,
G. Zatelli
2
, P. Bastiani
3
1
University od Florence, University of Florence, FIrenze,
Italy
2
Azienda USL Toscana Centro- Sede di Firenze, Medical
Physics Unit, Florence, Italy
3
Azienda USL Toscana Centro- Sede di Firenze,
Radiotherapy Unit, Florence, Italy
Purpose or Objective
2D detector arrays have become the standard device for
verification of VMAT dose distributions. The detector pixel
size is a key parameter to reproduce complex dose
distributions.
Aim of this work is to compare the gamma analysis of three
ionization chamber systems and to test the ability of each
system to detect deliberate errors.
Material and Methods
Measurements performed by PTW Octavius 4D 729
(5x5x5mm
3
ionization chamber, 10 mm spacing), PTW
Octavius 4D 1500 (4.4x4.4x3mm
3
ionization chambers, 7.1
mm spacing) and PTW Octavius 4D 1000 SRS (2.3x 2.3x 0.5
mm
3
liquid filled ionization chamber, 2.5 mm spacing) in
the PTW Octavius 4D phantom were used to validate the
dosimetric accuracy of the VMAT delivery. Firstly, 50
VMAT SBRT treatment plans from a variety of clinical sites
were considered. Secondly, systematic variations in
collimator (2° and 5° rotation) and gantry angle (shift of
2° and 5°) and lack of monitor units were applied to four
clinical treatments (2 lung tumors, 1 spine and 1
abdominal lymph node) in order to establish the detection
sensitivity of the three devices. Measurements were
compared with TPS Elekta Monaco computed doses via
local gamma analysis ( 2%L/ 2 mm, 2%L/ 1 mm and 1% L/1
mm). For the 729 and 1500 detectors, the resolution was
improved by merging two measurements performed with
5 mm couch shift.
The threshold for a success in error detectability was
established, by using the concept of confidence limit (CL),
as suggested by AAPM Task Group 119 [1]:
CL = (100 - D) + σ
where D and σ are respectively the mean dose and the
standard deviation of the distribution of the gamma
passing rate (35 plans for 6 MV and 15 plans for 10 MV)
measured by each dosimeter; the detectability threshold
(DT) has been calculated as:
DT = 100 - CL
Results
The average pass rate with 2%L/2 mm criterion for the 6MV
cases was 86.6 ± 5.2 (no shift) and 90.3 ± 4.3 (merged) for
729, 91.5 ± 3.7 (no shift) and 94.7 ± 2.9 (merged) for 1500
and 98.9 ± 1.1 for 1000 SRS. Box plot relative to 6 MV VMAT
SBRT 2%L/2 mm pass rate is presented in figure 1. Similar
results were achieved for 10 MV plans. The results for the
plans with errors, normalized to the DT for a success in
error detectability, show that only the SRS system can
distinguish the delivery errors, as shown in figure 2 for
6MV. Only a 2%L/1 mm gamma criterion proved to be
sensitive enough to detect errors.
Figure 1 : Box plot for 2%L/2 mm gamma passing rate of 6
MV VMAT SBRT plans for the investigated 2D arrays.