Physica Medica: European Journal of Medical Physics - Volume 32 - Supplement 1 - February 2016

28 / 146

Result: The dosimetric comparison supplied agreements within 2% for the
isocenter doses; gamma % greater than 0.97% for head and neck, 98% for
lung and 99% for pelvic regions and the gamma-mean values within 0.4.
The PTV V95 and mean doses were within 2%, while the mean doses of
principal OARs were within 3%.
Conclusion: The CBCT calibration method is accurate enough to supply
useful information for adaptive radiotherapy strategy.
A.76
IMPROVING THE QUALITY OF PROTONTHERAPY TREATMENT PLANS AND
THEIR VERIFICATION WITH MONTE CARLO METHODS
F. Fracchiolla M. Schwarz a Azienda Provinciale per i Servizi Sanitari
(APSS), Trento, Italy; b Post Graduate School of Medical Physics ‘Università La
Sapienza’, Rome, Italy; c INFN-TIFPA, Trento, Italy
Introduction: In radiotherapy an independent tool for MU verification is
necessary to ensure the best treatment safety. We developed a Monte
Carlo (MC) tool for independent dose calculation and we propose here
further applications to the clinical practice of our protontherapy center.
Methods andmaterials: We used the MC code for three clinical applications:
• Recalculation of a nominal plan from the TPS;
• Recalculation of a treatment session by using log file from the beam
delivery system;
• Simulation of patient specific QA (PQA) procedure (retrospective study
on 27 fields for 10 patients). For each PQA TPS and MC dose distribu-
tions calculated in water equivalent material at different depth were
compared with measurements performed with an array of 1020 ion-
ization chamber detector via γ analysis (3%, 3 mm). In our PQA protocol
the γ passing rate (PR) has decision thresholds:
• PR > 95%: field accepted;
• 90% < PR < 95%: if accepted, a justification must be added to PQA report;
• PR < 90%: field rejected.Measurement and simulation times were com-
pared too.
Results: For a neuroma case we recalculated the nominal plan and the dose
based on the log file. The recalculation based on the log file showed a re-
duction of max dose (1 Gy) in brainstem and in the fifth cranial nerve.
MC has a PR always greater than 95% for every depth, while TPS results
are always in the range of 90–95%. The measurement time of a PQA takes
almost 1.5 hours, while the simulations can be performed in parallel and
take 15 min on average.
Conclusion: Our tool let us to estimate the effects on the dose distribu-
tion due to delivery fluctuations of the machine.
We also proposed a method to drastically reduce PQA verification time. Our
suggestion is to substitute measurements with simulations that showed
a very high accordance in terms of γ PR; only one field per patient should
be measured at one single depth as safety check.
Reference
[1]
A.77
A NEW PHANTOM FOR DAILY QA IN PROTONTHERAPY: A FAST, RELIABLE
AND INEXPENSIVE SOLUTION
F. Fracchiolla N. Bizzocch C. Algranat M. Schwarz a Protontherapy
Department, Azienda Provinciale per i Servizi Sanitari (APSS), Trento, Italy; b Post
Graduate School of Medical Physics, ‘Università La Sapienza’, Roma, Italy;
c INFN-TIFPA, Via Sommarive 14, 38123 Trento, Italy
Introduction: In a radiotherapy center daily QA (DQA) measurements need
to be fast, accurate and sensitive to any variation of the output of the de-
livery machine that may have a clinical impact. In a pencil beam scanning
protontherapy center, spot positioning, spot size, range and dose output
are usually verified every day before treatments. We designed and built a
new reliable, sensitive and inexpensive phantom for DQAs that reduces the
execution times preserving the accuracy of the test.
Materials and methods: The phantom is provided with two couple of
wedges: thanks to these the Bragg peak is sampled at different depths and
the image becomes a transposition on the transverse plane of the depth
dose. Three ‘boxes’ in the center are used to check spot positioning and de-
livered dose. The box thickness helps ‘spreading’ the single spot and to fit
a Gaussian profile on a low resolution (7.6 mm) 2D detector.
We tested whether our new QA solution could detect errors of 1 mm in
spot positioning, 2 mm in range and 10% in spot size. Execution time was
also investigated and compared with the previous workflow.
Results: Our method is able to correctly detect 98% of spots that are ac-
tually in tolerance and 99% of spots out of 1 mm tolerance. All range
variations greater than the threshold (2 mm) were correctly detected. A
month of DQA showed that the repeatability of the sigma is very high (the
standard deviation of each spot is less than 0.7%). The aim of the sigma
analysis is to check the constancy of this parameter: it is not a check of
the sigma in air but in a medium. The sigma in air is measured during weekly
QA.
This phantom let us halve the execution time (20 min on average) com-
pared with the former multi-device procedure.
Conclusions: We designed and built a phantom precise, accurate, inex-
pensive and fast in terms of execution time; thanks to our new procedure,
we can preserve the reliability of DQAs and save time, favoring more slots
for patient treatments.
A.78
PATIENT QUALITY ASSURANCE FOR HD TOMOTHERAPY®: A 3-YEAR
REVIEW OF PRE-TREATMENT IN-PHANTOM DOSIMETRY
C. Fulcheri *, A. Chiappiniello, M. Marcantonini, C. Zucchetti, M. Iacco,
A.C. Dipilato, A. Didona, R. Tarducci. Medical Physics Unit, Santa Maria della
Misericordia Hospital, Perugia, Italy
Introduction: The purpose of this study was to analyze the results of de-
livery quality assurance (DQA) dosimetry performed at our hospital over
the first 3 years of TomoTherapy® use.
Materials and methods: 349 Tomotherapy® DQA results were consid-
ered. Patient DQAs were grouped according to pathology and irradiated site,
which implies taking into account the DQA dependences on the machine
parameters, the dose prescription and the in-axis/off-axis position of ir-
radiated volumes. The DQA measurements were performed with an
ArcCHECK dosimetry phantom in combination with an internal A1SL ion-
ization chamber. Comparisons between planned and measured dose
distributions were performed using a gamma index metric of 3%/3 mm in
relative dose only, with a 10% low dose threshold. For absolute dose com-
parisons, the relative differences (RDs) between measured and calculated
doses were assessed. Statistical analysis was performed with the non-
parametric Friedman test.
Results: All the plans analyzed passed our quality criteria: gamma passing
rate GP > 95% and RD < 3%. A statistically significant difference in DQA results
was found among the different sites: the median values of GP resulted 99.5%
and 99.4% for the breast and thoracic wall (with or without lymph nodes),
respectively, while for all other sites it resulted to being always greater than
99.8%. Likewise, similar results were observed for RD: 1.7% for the breast
and 1.9% for the thoracic wall, while for all other sites it was always <1%.
The authors hypothesize that the lower agreement in dose and dose dis-
tributions recorded for the breast and thoracic wall may be ascribed to the
off-axis positions of the lesions. Further investigations are required in this
issue.
Conclusions: In this review patient DQAs were always found to satisfy our
quality criteria; however, at the same time differences in RD and GP values
among different sites were observed. These results could be used to rede-
sign DQA frequencies and/or tolerances.
e23 Abstracts/Physica Medica 32 (2016) e1–e70