795 / 1082
S779
ESTRO 36 2017
_______________________________________________________________________________________________
Results
Relative depth-dependent RBE based on nanodosimetric
quantities are similar to the microdosimetric RBE. For Co-
60 and Ir-192, the RBE increases with depth due to an
increasing contribution of low-energy photons in the
spectra. For the denser ionizing sources, nanodosimetric
RBE values were divided by 1.9. Apart from this factor, the
constant RBE-dependence up to 10 cm for I-125 and the
decrease of RBE for the two EBX sources due to beam
hardening are in good agreement with the
microdosimetric RBE.
Conclusion
RBE based on track structure (nanodoismetric approach)
shows that the average intra-track distance between DNA-
modelling volumes potentially suffering severe damage is
well related to the microdosimetric RBE, based on the
formation of dicentric chromosomes, for several BT-
sources. Apart from a constant normalization factor for
the denser ionizing sources, the depth-dependence is in
excellent agreement. This indicates that the
nanodosimetric photon track characterization performed
in this study is a good descriptor for the radiation quality.
Furthermore, the proposed target volume appears
realistic. Note, that neither the photon fluence nor
biological endpoints were taken into account for this
approach.
EP-1476 Preliminary results of in-vivo dosimetry by
EPID
S. Giancaterino
1
, M. Falco
2
, A. De Nicola
2
, N. Adorante
2
,
M. Di Tommaso
2
, M. Trignani
2
, A. Allajbej
2
, F. Perrotti
2
,
D. Genovesi
2
, F. Greco
3
, M. Grusio
3
, A. Piermattei
3
1
Ospedale Clinicizzato S.S. Annunziata, Radioterapia,
Chieti, Italy
2
University of Chieti SS. Annunziata Hospital,
Department of Radiation Oncology “G. D’Annunzio”-,
Chieti, Italy
3
Università Cattolica del Sacro Cuore, Medical Physics
Institute - Fondazione Policlinico Universitario A.
Gemelli-, Rome, Italy
Purpose or Objective
This study reports in-vivo dose verification (IVD) results
elaborated with SOFTDISO software on 300 cancer patients
treated with 3D-CRT, IMRT and VMAT techniques.
SOFTDISO uses the integral EPID image referred to each
single static or dynamic beam providing a quasi- real-time
test elaboration.
Material and Methods
The selected patients for this study were treated with an
Elekta Synergy Agility LINAC at SS. Annunziata Hospital.
3D-CRT, IMRT and VMAT treatment plans of 300 patients
were randomly selected. IVD tests were processed
with the SOFTDISO software who provides two type of
tests: (i) R ratio between the reconstructed isocenter dose
and the planned one; (ii) transit dosimetry based on γ-
analysis of EPID imaging (P
g
(%) and g
mean
).
Results
We identified class-1 errors, derived from inadequate QCs,
and class-2 errors due to patient morphological changes.
Considering overall (6697) tests, we found out that only
5% of them showed out-of-tolerance mean R values. For
gamma index analysis, in 13% of the overall tests were
found to be out of tolerance. Ignoring class-2 errors, 100%
of patients treated with different radiotherapy techniques
(except 3DCRT breast treatment, for which no class-2
errors were observed) reported mean P
g
(%) values within
tolerance levels. Thus, the percentage of out- of-
tolerance tests decreases from 13% to 7%. However,
considering all the techniques, only 4.4% of mean g
mean
tests resulted out of tolerance. In addition, removing
class-2 errors, this percentage decreases to approximately
3%. Actually the workload of IVD procedures on 9 patients
is 1 hour per day.
Conclusion
IVD performed using SOFTDISO assures: (i) a rapid response
of dose delivery alert with a reduced workload; (ii) a large
number of patients tested daily and (iii) for out- of-
tolerance tests repeating IVD in the subsequent day, the
possibility to verify the efficacy of the adopted
corrections.
EP-1477 Evaluating gamma-index quality assurance
methods for Nasopharynx Volumetric Arc Therapy
(VMAT)
E.M. Pogson
1,2,3
, S. Arumugam
2
, S. Blake
1
, N. Roberts
4
, C.
Hansen
5,6
, M. Currie
7
, M. Carolan
7
, P. Vial
2
, J. Juresic
2
, C.
Ochoa
2
, J. Yakobi
2
, A. Haman
2
, A. Trtovac
2
, L.
Holloway
1,2,3,4,8
, D.I. Thwaites
1
1
University of Sydney, Institute of Medical Physics-
School of Physics- Faculty of Science, Sydney NSW,
Australia
2
South Western Sydney Local Health District, Liverpool
and Macarthur Cancer Therapy Centres, Liverpool,
Australia
3
Ingham Institute, Medical Physics, Liverpool, Australia
4
University of Wollongong, Centre for Medical Radiation
Physics- School of Physics, Wollongong, Australia
5
Odense University Hospital, Laboratory of Radiation
Physics, Odense, Denmark
6
University of Southern Denmark, Faculty of Health
Sciences- University of Southern Denmark- Denmark,
Odense, Denmark
7
Illawarra and Shoalhaven Local Health District,
Illawarra Cancer Care Centre, Wollongong, Australia
8
University of New South Wales, South Western Sydney
Clinical School, Sydney, Australia
Purpose or Objective
Pre-treatment dose verification is often performed on
dose measuring phantoms with some form of gamma
evaluation. However it has been shown that the clinical
relevance of a 3% and 3mm pass rate tolerance is
questionable. The purpose of this study is to simulate
machine errors of clinical significance for nasopharynx
patients and test if these errors can be detected on a
standard commercial phantom. In this study systematic
errors including collimator rotation, gantry rotation, MLC
shifts, and MLC field sizes are investigated.
Material and Methods
Ten retrospective VMAT patients were planned with a
department protocol. Machine errors were deliberately
introduced to all plans. Plans were modified by increments
using Python to create simulated error plans; -5 to 5° for
gantry and collimator angles and -5 to +5mm for MLC shift
and MLC field size, considering each parameter
separately. Simulated error plans (Dose
error
) were
compared to the original non-error plan (Dose
Baseline
)
utilising
equation
(1).
(1)
All error plans doses were then recalculated in Pinnacle
3
.
Plans were reviewed against acceptable tolerance limits.
Plans were above tolerance and considered unacceptable
if PTV D95%, Brainstem D1cc or spinal cord D1cc were
beyond a ±5% deviation in dose. Additionally if either of
the left or right parotid mean doses were beyond ±10%,
this was also considered an unacceptable plan.
The smallest unacceptable error plan for each error type
(including the Gantry (G), Collimator (C), MLC Shift (S),
and MLC Field Size (F) error was delivered on an Elekta
Linac and dose was measured using an ArcCheck. Gamma
analysis was performed in SNCpatient version 6.6 utilising
a global 3%/3 mm (10% threshold with correction off)
gamma pass rate. Before measurement, the Linacs were
tested for MLC, gantry and dose accuracy. Only one