ESTRO 36 Abstract Book

861 / 1082

S845 ESTRO 36 2017
_______________________________________________________________________________________________
Furthermore, a new integrated strategy in biological
planning module was proposed and verified.
Material and Methods
In this study, twenty patients of advanced stage cervical
carcinoma were enrolled. For each patient, dose volume
based optimization (DVBO), biological model based
optimization (BMBO) and integrated strategy based
optimization (ISBO) plans were produced with the same
treatment parameters. Different biological models,
including LKB and Poisson model, were also used for organ
at risk (OAR) respectively. To evaluate the plan quality of
BMBO plans, the dosimetry differences between BMBO
plans and their corresponding DVBO plans were evaluated.
And ISBO plans were compared with DVBO and BMBO plans
respectively to verify the performance of the integrated
strategy proposed in this study.
Results
Compared with DVBO plans, BMBO plans produced slight
inhomogeneity and worse coverage of planning target
volume (PTV) (V95＝96.787, HI=0.098: p<0.01). However,
the tumor control probability (TCP) value were
comparable. And BMBO plans produced lower normal
tissue complication probability (NTCP) of rectum
(NTCP=0.108) and bladder (NTCP=0.144) compared with
corresponding DVBO plans (NTCP=0.186 and 0.178 for
rectum and bladder, respectively) with significant
differences. Better results of V95, D98, CI and HI values
could be found in ISBO plans (V95=98.307, D98=54.181Gy,
CI=0.762, HI=0.087) compared with BMBO plans
(V95=96.787, D98=53.419Gy, CI=0.707, HI=0.098) with
significant differences. Furthermore, ISBO plans produced
lower NTCP values of rectum (NTCP=0.14) and bladder
(NTCP=0.159) than DVBO plans (NTCP=0.186 and 0.178 for
rectum and bladder, respectively; with p<0.01 and p<0.01
for rectum and bladder, respectively).
Conclusion
For cervical carcinoma cases, BMBO plans produced lower
NTCP values of OARs than DVBO plans with a little worse
target coverage and homogeneity. And the integrated
strategy could produce better coverage, conformity and
homogeneity of PTV than BMBO plans, and reduce the
NTCP values of OARs compared with DVBO plans.
EP-1590 Can bolus range shifting improve plan quality
in the IMPT of head and neck cancer?
S. Michiels1, A. Barragán2, K. Souris2, K. Poels3, W.
Crijns3, J. Lee2, E. Sterpin1,2, S. Nuyts1,3, K.
Haustermans1,3, T. Depuydt 1,3
1 KULeuven - University of Leuven, Department of
Oncology, Leuven, Belgium
2 Université catholique de Louvain, Center of Molecular
Imaging- Radiotherapy and Oncology - Institut de
Recherche Experimentale et Clinique, Woluwe-Saint-
Lambert, Belgium
3 University Hospitals Leuven, Department of Radiation
Oncology, Leuven, Belgium
Purpose or Objective
In intensity-modulated proton therapy (IMPT) of head and
neck cancer (HNC), range shifter (RS) air gap is known to
widen the pencil beams reaching the patient. The actual
effect on dose distributions however remained
unassessed. Moreover, emerging technologies such as 3D
printing enable to implement RS as bolus, hereby
removing the air gap and reducing the risk of collision
compared to nozzle-mounted RS. In this study, we assess
the impact of clinically applied air gaps on IMPT plan
quality, the potential of Monte Carlo (MC) dose calculation
plan optimization to mitigate air gap effects, and the
potential advantage of bolus RS compared to currently
used RS solutions.
Material and Methods
Oropharyngeal cancer patients were selected for IMPT
based on potential reduction of normal tissue
complication probability (NTCP) for xerostomia and
dysphagia. Prescription was 54 Gy to the elective neck and
66 Gy to the primary tumor in 30 fractions. Table 1 shows
the considered organs-at-risk (OAR). For the treatment
planning, a 5 mm CTV-to-PTV-margin was used, and gantry
angles 50°, 180° and 310°, with 4 cm of PMMA as RS.
Pencil beams from clinical beam data were placed with 5
mm spot spacing and target margin. Beamlets were
calculated with a 2x2x2 mm³ voxel size using the
MCsquare fast-MC dose engine. Beamlet weight
optimization was performed using the IPOPT non-linear
solver. For each patient, 3 different RS cases were
compared (Figure 1): applied as bolus, mounted on a 25
cm snout accessory or on a 40x30 cm² nozzle. Firstly, the
air gap effect was determined by a MC recalculation of the
optimal bolus plan (=baseline) on the snout and nozzle
case. This represents the scenario in which the air gap is
disregarded during optimization. Secondly, the snout and
nozzle case were optimized to re-establish the initial
clinical goals in the presence of air gap. Finally,
differences in OAR dose and NTCP between all optimized
RS cases were calculated.
Results
Results for 3 patients were compiled in Table 1.
Suboptimal planning by disregarding the air gap yielded
underdosage ranging from 3.4 Gy to 7.7 Gy for PTV66Gy
(D98%) and from 3.8 Gy to 10.1 Gy for PTV54Gy . Case-specific
optimization restored PTV coverage, but increased the
mean dose to most OARs due to the compromised lateral
penumbra of the pencil beams. Compared to the bolus
plan, these OAR dose increments translated into
xerostomia NTCP increases between 3.9% and 7.1% and
dysphagia NTCP increases between 1.5% and 5.0%.
Conclusion
The actual effect of RS air gaps in clinical circumstances
was quantified for IMPT in HNC. Including the air gap in
the plan optimization is essential, but cannot cancel out
the impact of spot degradation on OAR dose. In this
regard, bolus RS can considerably improve plan quality
compared to snout/nozzle mounted RS. NTCP reductions
for xerostomia and dysphagia achieved with bolus were
not negligible compared to the 10% threshold proposed by
the model-based selection of patients for IMPT
(Langendijk et al. Radiother Oncol 2013).