ESTRO 36 Abstract Book

489 / 1082

S474 ESTRO 36 2017
_______________________________________________________________________________________________
which may result in under-dosage of target. We
investigated the applicability of a decision suppo rt system
developed for photon therapy in a proton therapy setting.
Material and Methods
Twenty-three consecutive NSCLC patients stage 1B to IV
treated with adaptive photon therapy were
retrospectively planned using intensity modulated proton
therapy. The adaptive protocol was based on geometrical
measures of target positioning and large anatomical
changes as e.g. atelectasis, as observed on daily CBCT
scans. Two surveillance CT-scans were acquired during the
treatment course. The consequences of anatomical
changes were evaluated by recalculation of the proton
plans on the surveillance scans. The CTV receiving 95% of
the prescribed dose was analyzed. Proton treatment plans
were scaled to prescribed doses of 70, 74 or 78Gy, to
investigate if full CTV coverage at 95% of 66Gy = 62.7Gy
could be maintained by increasing the prescribed dose.
Results
Fourteen (61%) patients needed adaptations when treated
with protons, given that 95% of the CTV must be covered
by 95% of the dose. In comparison, no patients needed
adaptation when treated with photons using this criterion.
Figure 1 shows CTV coverage for all patients. For proton
therapy, the adaptive protocol was found to identify
patients with large target under-dosage (six patients,
group A). Additionally, under-dosage was observed for
another eight patients (group B) with non-rigid changes up
to 15mm in the positioning of the bones. The median
decrease in coverage for all patients was 92.8% [48.1-
100%]. Robust optimization reduces the decrease in target
coverage, but does not eliminate the under-dosage, see
Fig.2.All patients in group B would be treated sufficiently
when prescribing 74Gy with all CTVs receiving 95% of
66Gy. For patients in group A, only two patients would be
treated sufficiently with a 78Gy prescription. A geometric
decision support protocol as the present is thus mandatory
in order to maintain target coverage of the patients in
group A. When increasing the prescribed dose, the
maximum dose to important normal tissue such as the
oesophagus, trachea, bronchi, and heart increases and
may thus be the dose limiting factor.
Conclusion
Large anatomical changes can be corrected for by an
adaptive protocol. Non-rigid positioning erro rs are not
identified by the geometrical criteria used for photo ns
but can be compensated by an increase in the prescribed
dose keeping in mind that this requires additional
attention to organs at risk. Robust optimisation reduces,
but does not eliminate the risk of under-dosage. Daily
imaging and treatment adaptation for a high fraction of
patients is mandatory in proton therapy for loco-regional
lung cancer.
PO-0877 Proton therapy of oesophageal cancer is more
robust against anatomical changes than photons
D.S. Møller1, M. Alber2, T.B. Nyeng1, M. Nordsmark3, L.
Hoffmann1
1Aarhus University Hospital, Department of Medical
Physics, Aarhus C, Denmark
2Heidelberg University Hospital, Department of
Radiation Oncology, Heidelberg, Germany
3Aarhus University Hospital, Department of Oncology,
Aarhus C, Denmark
Purpose or Objective
Anatomical changes such as changes in the mediastinum
and the diaphragm position are seen in oesophageal
cancer patients during the course of radiotherapy. Field
entrance through areas with a high risk of changes is often
unavoidable with intensity modulated photon
radiotherapy (IMRT) if target conformity and reduction of
dose to especially lungs and heart is pursued. Delivery of
proton therapy is highly sensitive to anatomical changes,
but using only one posterior field may avoid high risk
entrances. We investigate the sparing of normal tissue and
the potential gain in robustness towards anatomical
changes using intensity modulated proton therapy (IMPT)
instead of IMRT.
Material and Methods
Twenty-six consecutive patients with medial or lower
oesophageal or gastroesophageal junction(GEJ) cancer
treated with IMRT (5-8 fields) were retrospectively
planned with IMPT using one posterior beam. The
fractionation schedules were either 41.4 Gy/23fx (pre-
operative regime, 22 patients) or 50Gy/27fx (definitive
regime, 4 patients). To ensure dose coverage of the CTV
for photon plans, a PTV (5 mm AP, 5mm LR, 8 mm CC) was
used to account for uncertainties in planning and
delivery. For protons, three different strategies were
pursued. Robust optimization of the CTV (IMPTCR), robust
optimization of the CTV and full coverage of the PTV
(IMPTPR) and no robust optimization, but full coverage of
the PTV (IMPTP). Robust optimization was performed
accounting for 3mm isocenter shifts and 3% density
uncertainty.
IMRT and IMPT plans were compared in terms of dose to
lungs and heart. For all patients, an additional
surveillance CT-scan was obtained at fraction 10 and used
for recalculation of both IMRT and IMPT plans, analysing
the percentage of CTV receiving 95% of the prescribed
dose.
Results
Using IMPT instead of IMRT reduced the lung and heart
dose significantly regardless of the IMPT strategy (p<0.001
using a Wilcoxon signed rank test). The mean lung and
heart doses decreased from sample median = 8.7Gy
[1.6;16.3] and 17.1Gy [1.1;24.1] using IMRT to 2.2 Gy
[0.5;8.5] and 9.1 Gy [0;15.5], using IMPTPR.
Recalculation on the surveillance scans demonstrated that
7/26 (27%) IMRT plans showed CTV coverage < 99%. For