S390 ESTRO 35 2016

______________________________________________________________________________________________________
(MU) were delivered at each angle: The reference dose using transit and through-air EPID images. DC is composite of
without couch attenuation is the average dose at 0⁰,90⁰ and two parts: a deconvolution kernel that converts EPID images
270⁰. Gantry angles ranging from 235⁰ to 223⁰ in 1⁰ to fluence, and a pencil beam algorithm to calculate the
increments are used to measure the edge attenuated dose. dose. It can be used for pre treatment QA verification and for
And gantry angles ranging from 220⁰ to 180⁰ in 10⁰ in vivo dosimetry. In this work we evaluated the suitability of
increments to measure the dose attenuation of the central DC software for in vivo dosimetry of VMAT treatments.
region of the couch.
Skin dose was measured with radiochromic films and FilmQA Material and Methods: DC (v4.10) was used along with Elekta
Pro. Several films were placed between RW3 slabs in Synergy® Linac (6 and 10 MV beams) equipped with a-Si
different depths.The center of RW3 phantom coincides with Electronic Portal Imaging Device (EPID) Iview-GT. Twenty
linac isocenter. VMAT (5 prostate, 5 whole pelvis, 5 lung, 5 head and neck),
First,films were located at the surface, 0.5cm and 1.5cm elaborated by treatment planning system (TPS) Elekta
from the surface, and in the center of the RW3 phantom. Monaco® 5.0 were measured. Through-air (EPID T-A) and
Then 200MU were delivered with an open 10x10 field and transit EPID images were used for three dimensional dose
with zero gantry angle. The irradiated films were removed maps reconstruction in homogeneous phantoms. Octavius 4D
and other films were placed under the phantom, 0.5cm and with 729 2D array was used as reference. Gamma analysis at
1.5cm from the couch and in the center. The opposite beam 3% local dose /3mm DTA was performed. Doses from though-
was delivered, so we measure the effect of the couch to the air measurements were also reconstructed in the planning CT
dose distribution in the buildup region. (T-A in plan TC) and compared with the treatment planning
dose maps.Gamma pass rate of DC dose maps were compared
Results: Table 1A Comparison results between measured and with those of 729 in the Octavius 4D.
calculated relative transmitted dose (T%), with and without
the couch. Table 1B Evaluation of skin dose increment and Results:
comparison results between scanned and calculated
increment of skin dose.

Table1 Gamma pass rate. EPID T-A are computed in the

synthetic Ocatvius 4D phantom,
EPID transit is measured and computed in the homogeneous
Octavius II phantom.

Conclusion: The couch model improves the discrepancy

between measured and computed attenuated dose. If we
take into account the couch in treatment planning
calculations, this average difference decreases from 3.3% to Figure1: Results of a whole pelvis SIB treatment. In the upper
0.4%. The couch increases 4 times the skin dose and the part comparison of profiles for Octavius 4D (A), T-A EPID (B)
couch model provides an accurate calculated dose in the and Transit EPID (C). In yellow the TPS profile. The lower
buildup region. part of the image contains central coronal map of gamma>1
points. In D the comparison of DVH measured by DC in plan
PO-0825 CT (solid) versus TPS (dotted lines) is shown.
Characterization of a commercial EPID 3d software for in Gamma pass rates are shown in the table1.The assessment of
vivo dosimetry. VMAT plans shows a mean of 93.9% points with gamma<1 for
M. Esposito1, P. Bastiani2, A. Bruschi1, A. Ghirelli1, S. Pini1, G. Octavious 4d (3.5% SD), 89.2% and 7.9%SD for EPID T-A, 89%
Zatelli1, S. Russo1 and 8.5% SD for EPID Transit and 89.8% and 6.1%SD for T-A in
1
Azienda Sanitaria Firenze, Fisica Sanitaria, Firenze, Italy plan CT. Transit and through-air EPID acquisitions produced
2
Azienda Sanitaria Firenze, Radioterapia, Firenze, Italy similar gamma pass rates.Through-air EPID images computed
by DC in the plannning CT showed gamma pass rate in
Purpose or Objective: Dosimetry Check (DC) is a commercial agreement with those of Octavious 4d in the prostate, whole
software that allows reconstruction of 3d dose distributions
ESTRO 35 2016 S391
________________________________________________________________________________
pelvis and head and neck, in lung instead, gamma pass rates found deviating according to statistics derived from the
were lower in 4/5 cases. histogram of 3D γ-analysis results. These alerts are mostly
found to be patient-related or attributable to limitations of
Conclusion: DC is a suitable tool for VMAT in vivo dosimetry. our back-projection and dose calculation algorithm. After
The pencil beam algorithm can be inaccurate in the presence inspection, an intervention is considered for only 0.3% of the
of low-density inhomogeneities. treatments. The purpose of this study is to develop a
principal component analysis (PCA) based classification
PO-0826 method to improve the specificity of our EPID dosimetry
Benchmarking computed IDD curves for four proton system. In particular, in contrast to our current classification
treatment planning systems against measured data method, PCA allows for the spatial distribution of γ-values to
J. Alshaikhi1,2, D. D'Souza2, C.G. Ainsley3, I. Rosenberg2, G. be taken into account for deviation detection.
Royle1, R.A. Amos2
1
University College London, Medical Physics & Biomedical Material and Methods: The input for PCA consisted of 3D γ-
Engineering, London, United Kingdom distributions (3%/3mm), one per treatment arc per fraction.
2
University College London Hospitals, Radiotherapy Physics, In total, 2024 3D γ-distributions from 499 H&N VMAT
London, United Kingdom treatment-plans were included. As an initial choice,
3
University of Pennsylvania, Roberts Proton Therapy Center, components describing at least 1% of the variance were
Philadelphia, USA selected. The distribution of variances over the components
was inspected to validate this choice. Using these
Purpose or Objective: Accurate beam modelling is an components, new 3D γ-distributions were created by
essential function of a treatment planning system (TPS) to projecting each input 3D γ-distribution on only these
ensure that plans can be calculated that are deliverable components and then projecting the result to the original
within clinically acceptable tolerances. The purpose of this coordinate system of the 3D γ-distributions. If the selected
work is to evaluate the computed integral depth dose (IDD) components describe the original γ-distribution well, the new
curves of four commercially available proton TPSs, and original γ-distributions will be similar. This similarity was
benchmarked against measured data. The four TPSs quantified by the root mean square (RMS) d of the difference
(EclipseTM, XiO®, Pinnacle3, RayStation®) were between the two γ-distributions; a γ-distribution was marked
commissioned using pencil beam scanning data from the as deviating when d exceeded a threshold. All true positive γ-
University of Pennsylvania (UPenn) facility. distributions (n = 2) in the dataset, as identified by
experienced medical physicists, were used to determine this
Material and Methods: A water cube phantom (40cm3) was threshold for identification of alerts.
created in each TPS for calculation of IDD curves. Calculation
grid size set to 1mm in all TPSs. Individual IDDs for 27 Results: The first 16 components were each found to
nominal energies, ranging from 100 to 226.7MeV, were describe at least 1% of the variance; cumulatively, they
calculated by integrating the calculated depth dose account for 83% of the variance in the dataset. Figure 1
distributions. These were all benchmarked against measured shows the cumulative variance accounted for as a function of
data from UPenn, comparing the clinical range at 80% distal selected components and indicates that the choice for
dose (D80), Bragg peak width between distal and proximal selecting components is reasonable. After finding and
80% (D80-P80), range at 0.5% (R0.5), and distal penumbra applying the appropriate threshold for detecting the
between D80 and R0.5. Gamma-index analysis with pass identified true positives, a drop in alert rate from 49% to of
criteria of 1mm/1% was also used to compare computed and 11% was observed, corresponding to an increase in specificity
measured IDDs. from 0.51 to 0.89.

Results: Mean percentage of IDDs with >95% pass rate for

1mm/1% criteria were 96.7% (SD 4.9) for XiO®, 94.1% (SD 8.9)
for EclipseTM, 95.4% (SD 8.6) for RayStation®, and 49.2 (SD
26.0) for Pinnacle3. Maximum differences between computed
and measured IDD data are shown below. No correlation with
nominal energy was observed.

Conclusion: Characteristics of computed IDDs were compared

to measured data for four commercially available TPSs. All
were within clinically acceptable tolerances, with XiO
showing the closest agreement. Differences observed were
attributed to TPS specific beam modelling. Further
investigation will assess the cumulative impact of these Conclusion: The PCA-based classification method presented
discrepancies on verified clinical treatment plans. in this study enhances the specificity of deviation detection
in 3D in vivo EPID dosimetry of H&N VMAT from 0.51 to 0.89,
PO-0827 compared to our current clinical γ-histogram based method.
Principal component analysis for deviation detection in 3D Before clinical implemention, a rigorous validation is
in vivo EPID dosimetry required.
R.A. Rozendaal1, B. Mijnheer1, I. Olaciregui-Ruiz1, P.
Gonzalez1, J.J. Sonke1, A. Mans1 PO-0828
1
Netherlands Cancer Institute Antoni van Leeuwenhoek Dosimetric assessment of a second generation Multi-Leaf
Hospital, Department of Radiotherapy Physics, Amsterdam, Collimator for robotic radiotherapy
The Netherlands P.H. Mackeprang1, D. Schmidhalter1, D. Henzen1, M.
Malthaner1, D.M. Aebersold1, P. Manser1, M.K. Fix1
Purpose or Objective: One of the clinical issues our institute 1
Division of Medical Radiation Physics and Department of
faces regarding in vivo EPID dosimetry is the number of raised Radiation Oncology Inselspital, Bern University Hospital, and
alerts. For example, alerts are raised for 49% of the University of Bern, Switzerland
treatments in case of head-and-neck (H&N) VMAT
treatments; an alert is raised when dosimetry results are