75 / 146
A.235
IN-HOUSE IMAGEJ PLUG-IN FOR THE QUALITY ASSURANCE OF DYNAMIC
MULTI LEAF COLLIMATOR
C. Zucchetti
*
, M. Iacco, C. Fulcheri, R. Tarducci.
Medical Physics Department,
Santa Maria della Misericordia Hospital, Perugia, Italy
Introduction:
The purpose of this work is to introduce an automated pro-
cedure for quality control for the linac multi leaf collimator (MLC), working
in dynamic mode, using Gafchromic film. An in-house image-processing
tool was developed, capable of performing an automatic analysis of the film
pattern produced by an irradiation test available on the VARIAN system.
Materials and methods:
Quality-control procedures were performed ac-
cording to AAPM report 82. EBT2 films were irradiated with a VARIAN DHX
linac, equipped with a Millennium 120 MLC, and subsequently scanned.
A garden fence and gap moving pattern were used to monitor positions
and leaf speeds, respectively. An automated image processing tool was de-
veloped using ImageJ, a Java-based free software for image analyses. Our
plug-in allows for (1) film orientation, (2) analysis of the pattern created
by the movement of each leaf pair, fitting the profiles along the longitu-
dinal leaf direction with a Lorentzian function and returning the garden
fence line position and (3) analysis of profile homogeneity of gap moving
pattern to check leaf speed constancy. This plug-in has been tested ana-
lyzing the effects introduced by known errors on leaf position or leaf speeds.
Results:
The automatic image processing tool proved to be useful for evalu-
ating leaf position and speed accuracy. The use of this automated procedure
allows for significant reductions in the time required for quality checks and
improves the detection of delivery errors.
Conclusions:
Fully automated image processing leads to excellent repro-
ducibility of the quality controls and significantly reduces the incidence
of human errors. The plug-in has been developed for the analyses of the
dMLC patterns available on the VARIAN system, but it can be easily adapted
to other systems.
http://dx.doi.org/10.1016/j.ejmp.2016.01.239A.236
VMAT PLANNING APPROACH TO AVOID SUPERFICIAL UNDERDOSAGE FOR
ACCELERATED PARTIAL BREAST IRRADIATION
F. Zucconi
*
, P. Mancosu, F. Lobefalo, G. Reggiori, A. Stravato, V. Palumbo,
A. Gaudino, L. Paganini, G. Maggi, F. De Rose, S. Tomatis,
M. Scorsetti.
Department of Radiotherapy and Radiosurgery, Istituto Clinico
Humanitas Cancer Center, Rozzano, Italy
Purpose:
Accelerated partial breast irradiation (APBI) is a RT approach that
treats only the lumpectomy bed rather than the whole breast. To account
for breathing and residual motion, a 10 mm virtual expansion of the breast
is usually applied to the CT series (CT_E) and the optimization is per-
formed on an expanded target. However, the recalculation on the original
CT series (CT_O) could be underdosed close to the surface. In this study, a
strategy to increase the target superficial dose is presented.
Material and methods:
Ten APBI cases were randomly selected from the
internal database (41 patients since 06/14). PTV_O was defined on CT_O,
cropping it of 5 mm to the body. Dose prescription was 30 Gy in 5 frac-
tions. PTV_E was defined on CT_E, expanding PTV_O of 10 mm toward the
surface. PTV_E was divided in: PTV_EI (PTV_E cropped of 7 mm from the
CT_O body), PTV_ES (PTV_E between 5 and 7 mm from CT_O body), PTV_EE
(the remaining from PTV_E). Two plans were optimized on the CT_E: (i)
prescribing the same dose to the three PTVs and (ii) PTV_EI
=
30 Gy,
PTV_ES
=
32 Gy, PTV_EE
=
33 Gy. Final dose calculations were performed on
the CT_O. Plan objectives were: D98% (dose received by 98% of the target)
>
95% and D2%
<
107% for PTV, minimizing the Homogeneity Index
(HI
=
D2%
−
D98%); Plans normalized to PTV_O mean dose were com-
pared in terms of plan objectives findings.
Results:
Opposite behaviors were obtained on the two CT series. On the
CT_E, D98%, D2%, and HI were favorable to (i) (respectively, 94.9% vs 94.5%,
103.7% vs 105.9%, 8.8% vs 11.5%). On the CT_O, D98%, D2%, and HI were fa-
vorable to (ii) (92.3% vs 94.2%, 104.3% vs 104.2%, 12.1% vs 10.1%). In particular,
the superficial volume (i.e. PTV_ES) was the region of highest underdosage
(D98%
=
85.4
±
3.3% for the first approach). Regarding the OAR, minimal
changes were found between the two approaches.
Conclusion:
We demonstrated that a virtual overdosage on the superfi-
cial part of the target is required to fully cover the target.
http://dx.doi.org/10.1016/j.ejmp.2016.01.240e70
Abstracts/Physica Medica 32 (2016) e1–e70