ESTRO 36 Abstract Book

136 / 1082

S123 ESTRO 36 2017
_______________________________________________________________________________________________
be the same for the two cohorts, supporting current RBE
practice. Our alternative hypothesis (H1) was that the
radiographic abnormalities would be greater for the
proton cohort, suggesting an end-of-range RBE > 1.1.
Material and Methods
We analyzed follow-up CTs for 10 proton/X-ray patient
pairs matched for age, chemotherapy regimen, disease
laterality & implant status. 5 patients had a smoking
history (4 X-ray, 1 proton), all 20 were prescribed 50.4 Gy
in 28 fractions. For brevity, we write ‘Gy’ throughout, but
for protons ‘Gy’ should be taken as ‘GyRBE assuming a
fixed RBE of 1.1’. Proton TPS doses were recalculated
using TOPAS Monte Carlo simulations. Deformable
registrations enabled us to calculate changes in median
HU value between pre- & post-treatment CTs for dose bins
of 2-30 in 2 Gy increments. For each patient’s final
(modality-blinded) CT, qualitative abnormality grading
was performed by a radiologist.
Results
Quantitative datasets for a matched pair are included in
Fig 1, with the linear regression fits used to calculate our
endpoint: ΔHU/Gy. For all scans, Fig 2 plots this endpoint
as a function of follow-up time: separation between the
proton and X-ray cohorts is clear with proton scans
exhibiting higher ΔHU/Gy values. To assess the effect of
'modality” on the Fig 2 data, we used the lme4 package in
R to perform a linear mixed effects analysis of log
transformed ΔHU/Gy. As fixed effects, we considered
'modality”, 'mean lung dose”, 'change in IV contrast”,
'change in breathhold” plus 'follow-up interval” (without
interaction terms). Subject was added as a random effect.
A p-value of 0.0007 was obtained for a likelihood ratio test
of the full model against the model without modality.
Similar results were obtained for analysis of the non-
smoker sub-population. A significant difference between
the two modalities also arose from our qualitative
radiological scoring (Wilcoxon signed rank test, p=0.018,
median abnormality score=3/9, for protons, 1.5/9 for X-
rays).
Conclusion
Our data indicate that we should reject H0 in favor of H1,
to conclude that the end-of-range proton RBE for lung-
density changes >1.1. Experiments have demonstrated
that, in-vitro, RBE=1.1 underestimates the capacity of
end-of-range protons to kill cells. We studied
asymptomatic radiographic changes rather than cell kill,
but our work nonetheless supports the thesis that end-of-
range variations in proton RBE prove important in-vivo as
well as in-vitro.
OC-0246 Proton minibeam radiation therapy spares
normal rat brain
Y. Prezado1, G. Jouvion2, A. Patriarca3, C. Nauraye3, S.
Heinrich4, J. Bergs1, D. Labiod4, L. Jourdain5, W.
Gonzalez-Infantes1, M. Juchaux1, C. Sebrie5, F.
Pouzoulet4
1CNRS-Imagerie et Modélisation en Neurobiologie et
Cancérologie, New Approaches in Radiotherapy, Orsay,
France
2Institut Pasteur, HUMAN HISTOPATHOLOGY AND ANIMAL
MODELS, PARIS, France
3Institut Curie, Orsay Proton Therapy Center, Orsay,
France
4Institut Curie, Experimental radiotherapy platform,
Orsay, France
5University Paris Sud, Imagerie par Résonance
Magnétique Médicale et Multi-Modalités, Orsay, France
Purpose or Objective
The morbidity of normal tissues continues being the main
limitation in radiotherapy. To overcome it, we recently
proposed a novel concept: proton minibeam radiation
therapy (pMBRT) [1]. It allies the physical advantages of
protons with the normal tissue preservation observed
when irradiated with submillimetric spatially fractionated
beams (minibeam radiation therapy) [2]. We have recently
implemented the technique [3] at a clinical center (Proton
therapy center in Orsay). The main objective of this work
was to confirm the gain in tissue sparing thanks to pMBRT.
Material and Methods
The whole brain of 7 week-old male Fischer 344 rats (n=16)
was irradiated with 100 MeV protons. Half of the animals
received conventional seamless proton irradiation (25 Gy
in one fraction). The other rats were irradiated with
pMBRT (58 Gy peak dose in one fraction). The average dose
deposited in the same target volume was in both cases 25
Gy. The animals were followed up for 7 months. A
magnetic resonance imaging (MRI) follow up (10 days, 3
months and 6 months) at a 7T small animal MRI scanner as
well as histological analysis were performed.
Results
Rats treated with conventional proton irradiation
exhibited severe moist desquamation and permanent
epilation. The MRI and histology analysis showed
important brain damage (extensive blood-brain barrier
breakdown (BBB), hematomas, necrosis, microglial
activation, etc.). See figure 1. In contrast, the pMBRT
group presented no skin damage, a reversible epilation
and no significant brain damage observed by MRI or
histological analysis.