791 / 1082
S775
ESTRO 36 2017
_______________________________________________________________________________________________
head and neck (H&N), brain and sarcoma in tomotherapy;
breasts at Linac and skin tumors in Plesio-Röntgen
therapy. For each patient films were located in 1 to 3
reproducible points (see figure 1 and 2) and measurements
were repeated on average in three consecutive fractions.
EBT3 films were read with a flatbed scanner Epson
10000XL and images were analyzed using the red channel
calibration. In vivo dose evaluations were compared with
measurements performed on Cheese phantom both with
and without thermoplastic mask at Linac and in
Tomotherapy.
Results
A total of 117 film measurements were performed on 21
patients. The absolute value of the mean difference
between measured and TPS-calculated dose and its
standard deviation was 11.3% ± 6.5% for all treatments. A
mean absolute difference of 17.7% for Linac plans, 11.6%
in Tomotherapy and 4.6% in Plesio-Röntgen therapy were
achieved. Both at Linac and in Plesio-Röntgen therapy
there was not a clear trend of overestimation of the TPS
with respect to measurements. Instead in Tomotherapy
there was an underestimation of the TPS (-9.1%) for H&N
and brain treatments (in these case measurements were
performed with thermoplastic mask) and an
overestimation for the sarcoma (9.2%). This trend was
confirmed by the measurements made on the Cheese
phantom in Tomotherapy, where there was an
overestimation of the TPS without mask (28.6% vs -0.7%
with mask). Moreover, an improvement of the agreement
between EBT3 measurements and Pinnacle and
Tomotherapy dose estimation was shown in presence of
mask (28.6% to -0.7% in Tomotherapy and -20.7% to -16.3%
at Linac).
Conclusion
Gafchromic films are suitable detectors for skin dose
measurements in radiotherapy. In vivo surface dose
measurements with EBT3 are a useful tool for quality
assurance in radiotherapy, since the TPS does not give
accurate dose values in the first millimiters of skin.
EP-1469 Flattening filter free beam profile analysis
using two different normalization methods
G. Nicolini
1
, A. Fogliata
2
, E. Vanetti
1
, G. Reggiori
2
, A.
Stravato
2
, P. Mancosu
2
, M. Scorsetti
2
, L. Cozzi
2
1
Radiqa Developments, Medical Physics Team,
Bellinzona, Switzerland
2
Humanitas Research Hospital, Radiotherapy and
Radiosurgery Dept, Milan, Italy
Purpose or Objective
Flattening filter free (FFF) beams present a profile peaked
on the beam central axis (cax), unsuitable for flatness and
symmetry description that usually characterize standard
beam profiles. Definitions of unflatness and slope have
been recently proposed, requiring a preliminar suitable
FFF profile normalization. Two main normalization
processes as far published are: the inflection point IP
(Pönish 2006), and the renormalization factor RF (Fogliata
2012). In both formalisms the FFF dose fall-off at the field
edge is superimposed with the corresponding FF profile.
The present study aims to compare FFF specific profile
parameters using the two normalization procedures.
Material and Methods
Dosimetric data from a Varian TrueBeam with 6 and 10 MV,
FF and FFF modes, have been collected at SSD 100cm and
5 depths. The cax normalization value N was evaluated for
the IP method as N=D
cax
·(D
u
/D
f
), where D
cax
and D
f
are the
doses on cax and at the IP of the penumbra region for the
corresponding FF beam, D
u
is the dose at the IP of the FFF
beam.
The N value for the RF method was evaluated by using the
fit dependent on the field size FS and depth:
N=(a+b·FS+c·depth)/(1+d·FS+e·depth), where the fitting
parameters are taken from published data. The main
profile parameters of FFF photon beams were computed:
field size, penumbra, unflatness, slope, and peak-position
parameters. To systematically investigate the impact of
the N value, they were recomputed with a RF value
modified of + 1,2,3,5,7,10% (perturbed RF).
Results
In terms of cax normalization value, in average, the two
methods show an agreement within the 2%, with a
tendency of a greater N with IP respect RF method for
10MV. In any case, some outliers are present, with a
discrepancy that reaches the 10%; this is expected, since
the IP method suffers of the uncertainty of IP position
determination in the practice. Beam parameters values
derived with the approaches (IP/RF) were computed
showing, e.g., for both energies 1.00+0.00 for unflatness
and, respectively for 6 and 10MV, 0.99+0.05 and 1.02+0.04
for slope. Analysis with perturbed RF values, shows that
with a variation up to 10% of N, the peak position remains
within 0.05mm, the unflattens within 0.5% and 1% for 6MV
and 10MV beams, while the slope has a variation almost of
the same amount of N itself. Field size difference is within
1mm if N variation is within 5%.
Conclusion
The two normalization methods are both suitable for
subsequent FFF profile description. Unflatness parameter
resulted similar when computed using the two different
normalization formalisms with no significant differences.
Slope values are more sensitive to normalization value,
and therefore some outliers were observed due to
uncertainty of IP position in the practice. The RF
procedure, with the published fitting parameters is easier
to use and more robust respect to measurements sampling
and
detector size.
EP-1470 Determination of paramagnetic gel sensitivity
in low energy X-ray beam
Y. Ben Ahmed
1
, J. Coulaud
2
, S. Ken
1
, L. Parent
1
1
Institut Universitaire du Cancer Toulouse, Haute