117 / 1082
S104
ESTRO 36 2017
_______________________________________________________________________________________________
compared? Can the resources of ART be motivated by a
clinical gain, or have we lost the clinical perspective
during the technological development? These are points
that will be further addressed in this talk.
SP-0208 Development of procedures for safe clinical
application of plan-of-the-day adaptive radiotherapy
S. Quint
1
, J. Penninkhof
1
, J. Schiphof-Godart
1
, W. Hilst-
van der
1
, B. Heijmen
1
, M. Hoogeman
1
1
Erasmus MC Cancer Institute, Radiation Oncology,
Rotterdam, The Netherlands
Complex tissue variations during the course of a
radiotherapy treatment combined with IMRT or VMAT
require adaptive approaches using in-room verification of
position and shape of the target volume for optimal dose
avoidance in organs at risk. In Erasmus MC we have
developed an on-line adaptive approach for cervical and
bladder cancer patients. Sofar more than 200 patients
have been treated with adaptive therapy [Heijkoop S.,
IJROBP 2014, Jun;90(3):673-9)]. For each patient, an
individualized library of treatment plans is pre-treatment
established. Each plan in the library is optimally suited for
a patient anatomy that can potentially occur during
treatment. During each fraction, the plan that best
matches the anatomy of the day is selected, based on an
acquired CBCT.
In this presentation, we will discuss aspects for safe
introduction and application of the novel technology,
including formal risk analyses, multi-disciplinary
involvement, education, and definition of tasks and
responsibilities of technologists, physicists, and
physicians.
Symposium: Robust optimisation in protons and photons
SP-0209 What is the actual robustness of the plans we
deliver in particle therapy and what measures do we
take to obtain it
S. Molinelli
1
, M. Ciocca
1
1
Fondazione CNAO, Medical Physics, Pavia, Italy
A prerequisite of high precision radiotherapy (RT) is high
precision of dose planning and delivery. If all involved
uncertainties are not accounted for, this will result in a
reduced benefit of highly conformal techniques, such as
particle therapy (PT). By definition, a plan is robust when
treatment goals are met despite uncertainties in patient
and beam models and the plan remains acceptable over a
range of likely variation. PTV margins are a well-
established strategy to guarantee target coverage in
photon RT, but showed to be a suboptimal solution in PT.
Deviations in particle range entail significant dose
deformations, related to the single beam path and require
beam specific margin expansions. Uncertainties, and
robustness as a consequence, depend on multiple factors:
plan optimization, dose calculation accuracy,
immobilization systems, image guidance protocols and
delivery techniques. First of all, robust beam selection is
essential to reduce heterogeneities across the beam path
and avoid regions subject to intra and inter-fraction
variations in patient anatomy which could determine
unexpected severe dose errors. Set up errors and inherent
deviations in CT calibration values can be included in plan
evaluation and in the optimization process itself. Several
approaches have been proposed for robust plan
optimization, showing that the cost of robustness is often
a reduction of plan conformality and a consequent
increase of OAR doses. Planned dose recalculation based
on machine log files allows for evaluation of the impact of
dose delivery errors, providing important information on
plan sensitivity to beam energy or position deviations. The
consistency between planned and delivered doses may
substantially deteriorate when approximation errors occur
in the dose calculation algorithm. This influences particle
range and causes improper modeling of the Bragg peak
degradation and beam lateral spread in heterogeneous
media. When comparing TPSs based on different beam
models, substantial dose differences can be found,
particularly if passive beam modulators are used. While
for protons the well-known distal end RBE enhancement
can be easily accounted for with a distal margin extension,
a more complex issue concerns carbon ions RBE-weighted
doses. The RBE dependence on depth, dose, energy,
fractionation and cell type is strictly related to the
biological model adopted in the TPS. Changing the model
or model parameters, impacts on RBE-weighted dose
values corresponding to the same absorbed (and
delivered) dose, with a significant influence on clinical
outcomes. Most clinical TPS in use do not provide any tool
for management of plan robustness. Site specific, manual
and cumbersome approaches are often required, based on
beam geometry constraints and the use of avoidance
structure to force and/or prevent radiation pathways.
Recent commercial systems provide robust evaluation and
optimization tools based on the inclusion of set up errors
and CT-HU variation to account for random and systematic
range uncertainties. Few attempts have been made in the
direction of delivery pattern optimization, in terms of
energy layers rescanning, redistribution and filtration and
spot editing. Simultaneous plan optimization on multiple
CT scans, representing different anatomical conditions
involved in the dose delivery phase (e.g.: 4D CT scans, in
case of gated treatments to mitigate plan sensitivity
against residual organ motion) is, to our knowledge, still
missing. A fast and accurate MC engine should be available
for dosimetric accuracy assessment in challenging clinical
cases, where the calculation algorithm is known to present
significant limitations. For carbon ion therapy, TPSs should
provide dose averaged LET and fragment spectra
distributions, in addition to a flexible selection of
different RBE biological models and model parameters.
Common plan evaluation metrics, setting a threshold
between plan robustness and conformality, are still not
available in clinical routine. Retrospective analysis of
delivered plans could help in the definition of reference
robustness databases in centers with consolidated clinical
results. Experimental systems for in-vivo monitoring of
particles range provide a direct measure of the
uncertainties involved. A new PET scanner able to operate
during the actual treatment of H&N tumors has recently
been tested, based on the measurement of the β+ activity
induced by the interaction of the therapeutic beam with
patient tissues. Optimal PT plans should preserve target
dose conformity, healthy tissue sparing and robustness
towards uncertainties. IGRT protocols to minimize inter-
fraction deviations should be integrated with robust plan
geometry, optimization and evaluation. Even in a robust
dose distribution, due to the sensitivity of particle range
to variations in volume, shape and filling of tissues along
the beam path, the implementation of adaptive protocols
is mandatory for a correct treatment.
SP-0210 Minimax robust optimisation applied to IMPT
for oropharyngeal tumours
S. Van de Water
1
, M. Hoogeman
2
, B. Heijmen
2
2
Erasmus MC Cancer Institute, Department of Radiation
Oncology, Rotterdam, The Netherlands
Robust optimization techniques increasingly receive
attention, especially in the field of particle therapy, as
they are considered more effective and more efficient in
dealing with treatment uncertainties compared with the
use of conventional safety margins. During robust
optimization, treatment uncertainties are explicitly
included in the mathematical optimization, thereby
ensuring adequate irradiation when errors occur during
treatment execution. Different approaches to robust