ESTRO 36 Abstract Book

78 / 1082

S65 ESTRO 36 2017
_______________________________________________________________________________________________
PV-0137 Validation of fast motion-including dose
reconstruction for proton scanning therapy in the liver
E. Colvill1, J. Petersen1, M. Høyer1, E. Worm1, R.
Hansen1 , P . Poulsen1
1Aarhus University Hospital, Radiation Oncology, Aarhus
C , Denmark
Purpose or Objective
Th e large respiratory motion present during liver
treatments means that the interplay effect could be of
grea t concern for hypofractionated proton scanning
therapy. A method has been developed to use recorded 3D
motion to calculate the dose delivered to moving targets
in scanning proton therapy using a commercial treatment
planning system (TPS). In this study, we validate the
method for the application in liver where tumor motion
and other anatomical changes are large.
Material and Methods
The fast dose reconstruction utilizes an in-house computer
program to manipulate dicom plans exported from the TPS
and incorporates the motion by shifting each proton spot
to its position as seen in tumor’s eye view. Depth motion
and SSD variations are modeled by modifying the beam
energy. The motion-included plans are imported and
recalculated in the TPS in the original CT scan. To validate
the accuracy of the dose reconstruction the 4DCTs of 13
liver patients treated with photon SBRT were used. For
each patient, an IGTV was created from exhale and inhale
GTVs. A plan was created on both inhale and exhale phases
using three-field single-field-optimization with ~6mm spot
spacing and mean IGTV coverage of 56.25Gy in three
fractions. Four dose reconstruction plans were created for
each patient to represent potential interplay effects, two
from each of the original inhale and exhale plans; One plan
with all odd layers shifted to the opposite phase and one
with all even layers shifted. These were imported and
calculated on the original CT phase. Corresponding ground
truth doses were created by splitting the original plans
into sub-plans with odd and even layers, calculating one
sub-plan in the original CT phase and the other in the
opposite phase, and then summing the two dose maps. The
reconstructed plans were compared to the ground truth
plans to assess the accuracy of the fast dose
reconstruction and the original to the ground truth plans
to assess the magnitude of interplay effects. The
comparisons included D5, D95 and the homogeneity index
(HI=(D2-D98)/D50) for the GTV and the root-mean-square
(RMS) dose errors for dose points above 70% and 90% cutoff
levels.
Results
To date validation on two patients (8 plans) has been
completed. Figure1 shows the results for one plan. The
mean and range of errors of the fast dose reconstruction
GTV D5, D95 and RMS errors above 70% dose and above 90%
dose are shown in Table1 along with the interplay effect
results for comparison. The mean and range of the HI for
the original, ground truth and reconstructed plans were
0.07(0.05-0.08), 0.19(0.13-0.28) and 0.19 (0.13-0.28)
respectively.
Conclusion
Validation of a method of fast dose reconstruction to
assess the dosimetric impact of tumor motion on scanning
proton therapy for liver SBRT treatments was performed.
The results show that the dose reconstruction can utilize
a single phase of a 4DCT with dose calculation errors that
are much smaller than the dose distortion created by the
interplay effect itself.
PV-0138 Pencil beam scanning treatments in free-
breathing lung cancer patients – is 5 mm motion a
limit?
A. Jakobi1, R. Perrin2, A. Knopf2,3, C. Richter1,4,5,6
1OncoRay - Center for Radiation Research in Oncology,
High Precision Radiotherapy, Dresden, Germany
2Paul Scherrer Institute, Center for Proton Therapy,
Villigen, Switzerland
3University of Groningen - University Medical Center
Groningen, Department of Radiation Oncology,
Groningen, The Netherlands
4German Cancer Consortium Dresden and German Cancer
Research Center Heidelberg, DKTK, Dresden, Germany
5Faculty of Medicine and University Hospital Carl Gustav
Carus - Technische Universität Dresden, Department of
Radiation Oncology, Dresden, Germany
6Helmholtz-Zentrum Dresden – Rossendorf, Institute of
Radiooncology, Dresden, Germany
Purpose or Objective
To evaluate the dose degradation when treating lung
cancer patie nts with proton pencil beam scanning during
free-breathing. We assess if treatments without
rescanning are feasible in order to avoid prolonged
treatment time, especially for slow scanning facilities.
Material and Methods
For 40 lung cancer patients, 4DCT imaging was used to
generate 4D dynamic dose distributions of 3D treatment
plans with 3 pencil beam scanning fields optimised with
the single field uniform dose technique. Simulations
included the use of random breathing states of the patient
at start of irradiation resulting in multiple possible 4D
dynamic dose distributions per fraction. Complete
treatment was assumed to consist of 33 fractions of
probabilistically chosen single fractions. Treatments were
assumed to be delivered with an IBA universal nozzle
without rescanning (1.5ms between spots, 2s between
energy layers, spot sigma 4mm at highest energy). Tumour
motion amplitude was the maximum displacement in
tumour centre-of-mass assessed by the 4DCT. Evaluation
was done by looking at under- and overdosage in the target
structure. In addition, changes in the dose distribution due
to changes in motion and anatomy during treatment were
analysed using a repeated 4DCT for 4D dynamic dose
calculation in one patient case.