344 / 1082
S329
ESTRO 36 2017
_______________________________________________________________________________________________
Conclusion
In this small series of consecutive CNS tumor patients,
psychological support during radiotherapy allows an
improvement in distress. Quality of life remained stable
with a significant improvement of emotional well-being.
PO-0636 Impact of DWI-MRI for gross tumor volume
definition in patients with recurrent glioblastoma
O. Oehlke
1
, S. Bott
1
, A.L. Grosu
1
, I. Mader
2
1
Universitatsklinik Freiburg, Klinik für
Strahlenheilkunde, Freiburg, Germany
2
Universitatsklinik Freiburg, Klinik für Neuroradiologie,
Freiburg, Germany
Purpose or Objective
The low specificity of gadolinium-enhancement in T1-
weighted MRI (T1Gd-MRI) in the definition of gross tumor
volume (GTV) in recurrent glioblastoma (GBM) is a
clinically important problem in the context of treatment
planning for re-irradiation. Advanced MR imaging
techniques like diffusion-weighted imaging (DWI) have
been introduced for diagnosis of tumors. Tumors
demonstrate restricted diffusion and thus low apparent
diffusion coefficient (ADC), which inversely correlates
with tumor cellularity. The goal of this study was to
evaluate possible ADC
low
/T1Gd-MRI non-overlap volumes
that may have an impact on definition of the GTV in
patients with recurrent GBM planned for stereotactic re-
irradiation.
Material and Methods
We evaluated 41 patients treated with stereotactic re-
irradiation for recurrent GBM. All patients initially
received standard of care with resection followed by
radiochemotherapy and adjuvant temozolomide.
Progression was diagnosed either by re-resection, biopsy,
or RANO criteria in serial MR images. The T1Gd-MRI-GTV
was defined as contrast enhancement in 3D T1-weighed
sequences. The DWI data were acquired by using a
singleshot spin-echo echo-planar imaging sequence (
b
=0,
1000 s/mm
2
). ADC values are automatically calculated and
displayed as a parametric map. In order to minimize bias,
T1Gd-MRI-GTV and ADC
low
areas were determined by
different persons.
Results
The median and mean volumes of T1Gd-MRI-GTVs were
7.71ml (range 0.95–48.27ml) and 10.67 ± 9.33ml,
respectively. ADC
low
volumes were significantly smaller
(median 1.68ml (range 0.27–7.25ml) and mean 2.43 ±
1.85ml,
p
<0.001). The ADC
low
/ T1Gd-MRI overlap volume
ranged from 0.03-3.25ml (median 0.62ml) with a mean
volume of 0.84 ± 0.85 ml. The corresponding ADC
low
/T1Gd-
MRI-GTV non-overlap volume was 1.46 ± 1.22ml (
p
=0.0053)
with a median of 1.14ml (range 0.09-5.58ml).
Conclusion
The results of this study suggest that GTV delineation
according to contrast enhancement in T1Gd-MRI in
patients with recurrent GBM may lead to target volumes
that do not encompass all biologically active tumor
detectable in advance DWI MR imaging. Retrospective
correlation of ADC
low
areas with the respective isodose
distribution and topography of further progression after
re-irradiation will help to understand the clinical
significance of ADC
low
/non-Gd enhancing areas.
PO-0637 HFSRT to the resection bed for intracranial
metastases: a large retrospective study of 181 patients.
A. Keller
1
, M. Doré
2
, F. Thillays
2
, F. Proust
3
, H. Cebula
3
,
I. Darié
3
, F. Lefebvre
4
, G. Noel
1
, D. Antoni
1
1
Centre Paul Strauss, Radiotherapy, Strasbourg CEDEX,
France
2
Institut de Cancérologie de l'Ouest, Radiotherapy, Saint
Herblain, France
3
CHU Hautepierre, Neurosurgery, Strasbourg, France
4
Faculty of Medecine, Laboratory of Biostatistics,
Strasbourg, France
Purpose or Objective
We retrospectively report the outcomes of a large
multicenter cohort of patients treated with surgery and
hypofractionated stereotactic radiotherapy (HFSRT) to the
resection cavities of brain metastases (BM).
Material and Methods
Between March 2008 to February 2015, a total of 181
patients (189 cavities) with no prior WBRT were treated
by HFSRT to the surgical bed of BM, at the dose of 23.1 Gy
(3 x 7.7 Gy) prescribed to the 70 % isodose line (with 2-
mm margin PTV). The primary end-point was local control.
Secondary endpoints were distance brain control, overall
survival, risks of radionecrosis (RN) and leptomeningeal
disease (LMD).
Results
Of the 189 resected lesions, 44 % were metastatic from a
non-small cell lung cancer (NSCLC) primary tumor and 138
patients (76.2 %) had a single metastasis at the time of
treatment. The median preoperative tumor volume was 12
mL (0,13 - 92,5), and the median PTV 14.1 mL (0,8 - 65,8).
With a median follow up of 15 months (0.6 – 75), the 6-
and 12- month local control rates were 92.6 % [95 % CI:
88.8 - 96.3] and 88.2 % [83.6 - 92.9], respectively. On
multivariate analysis, PTV (HR = 1.03 [1.01 - 1.05], p =
0.005) and GPA (Graded Prognostic Assessment) score (HR
= 2.02 [1.11 - 3.66], p = 0.021) were predictive of local
failure. The 6-, 12-month distance brain control rates
were 70.4 % [63.9 - 76.8] and 60.7 % [53.7 - 67.7],
respectively. Twenty-six patients (14.4 %) developed signs
of LMD, at a median time of 3.8 months (0.13 - 33.6) after
HFSRT. Preoperative tumor volume was predictive of LMD
(HR = 1.02 [1.0 – 1.04], p = 0.029). The median overall
survival (OS) was 17.3 months (0.59 - 65.9). The 6-, 12-
and 24-month OS rates were 78.8 % [73.2 - 84.9], 62.5 %
[55.9 - 69.9] and 39.1 % [32.4 - 47.2], respectively. RPA
Class 3 (HR = 2.05 [1.12 - 3.76], p = 0.02), piecemeal
resection (HR = 1.56 [1.09 – 2.27], p = 0.017) and
increasing number of BM (HR = 1.71 [1.34 - 2.16], p <
0.0001) were independent unfavorable prognostic factors
for OS. Fifty-four patients (29.8 %) were subsequently
treated with salvage WBRT at a median time of 6.5 months
(0.4 - 44.9), and among the patients who developed