172 / 1096
S159
ESTRO 36
_______________________________________________________________________________________________
simulations were performed for spot or raster scanning
(positioning with/without beam pulse between successive
spots), for constant and varied beam current, and for
cyclotron or synchrotron beam dynamics (e.g. continuous
or pulsed beam). The resulting 4D plans were compared
using dose-volume metrics (D5-D95, V95) for the CTV, as
well as total predicted treatment time (TT).
Results
Independent of the delivery scenario and prescribed dose,
neither gating alone nor rescanning alone could mitigate
motion effects completely, with residual interplay effects
(D
5-95
) of more than 10-20% being observed for GW=5mm
w.o. rescanning (shown by purple error-bars). Moreover,
the D
5-95
of synchrotron based simulations were found to
be >5% higher than for cyclotron scenarios. However, with
re-gating (re-scanning + gating, shown by green and blue
error-bars), motion mitigation performance was found to
be similar effective (close to static reference) for all
scanning dynamics and rescan modes, with the main
difference being only in treatment efficiency. Without any
mitigation, mean TT’s for the 2Gy/12Gy plans were 2x/3x
longer for synchrotron than for cyclotron scenarios. For re-
gating (GW5+LS5), mean TT’s of synchrotron based plans
were on average 2.5x higher when combined with LS and
3.5x higher when combined with VS. Moreover, the
advantage of varying beam current has been
demonstrated by the approximately constant TT as a
function of prescription dose. In addition, for the high
dose scenario, variations caused by differences in
geometry, motion amplitude, field direction and starting
phase, are smaller for varying beam current scenarios in
comparison to corresponding constant scenarios.
Conclusion
In sum, independently of PBS delivery scenario, the
treatment of liver tumours under free-breathing
conditions is not recommended for motions over 10mm,
even when applying large numbers of rescans. However,
re-gating (LS5+GW5) is predicted to be sufficient to
achieve acceptable 4D plan quality for all scenarios, even
though synchrotron based delivery could have a significant
added time cost in comparison to cyclotron based systems.
Symposium: Focus on ART: the clinical difficulties
SP-0307 Multi-parametri c functional PET/MR imaging
for RT individualisation
D. Thorwarth
1
1
University of Tübingen, Sectio n for Biomedical Physics-
Department of Radiation Oncology, Tübingen, Germany
Multi-parametric functional imaging using combined
positron emission tomograph y and magnetic resonance
(PET/MR) imaging may be highly beneficial for the
assessment of tumour stage, size, therapy response and
functional information before and during radiotherapy
(RT) in order to guide not only geometrical but also
biological adaptation during the course of RT treatment.
This talk will discuss a number of issues, which might
compromise the use of multi-parametric functional
imaging in a clinical setting. In order to register functional
PET/MR image data to the planning CT of an individual
patient, tailored patient positioning devices and
dedicated algorithms for PET attenuation correction with
respect to the hardware components of the positioning
aids are required. Furthermore, robust and accurate
deformable registration strategies may be required to
match PET/MR data acquired during RT to baseline
imaging data.
In addition, first results of a multi-parametric PET/MR
imaging study in head-and-neck cancer (HNC) patients will
be discussed. Here, [18F]FDG PET/CT images were
acquired in addition to functional PET/MR data consisting
of [18F]FMISO PET, anatomical T1- and T2-weighted as