503 / 1096
S488
ESTRO 36
_______________________________________________________________________________________________
realistic VMs can help to optimize clinical imaging
protocols and image analysis tools.
PO-0888 Response monitoring by 18FDG-PET in locally
advanced NSCLC treated with concurrent
chemoradiotherapy
J.N.A. Van Diessen
1
, M. La Fontaine
2
, M. Van den
Heuvel
3
, W. Vogel
4
, J.S.A. Belderbos
1
, J.J. Sonke
2
1
Netherlands Cancer Institute Antoni van Leeuwenhoek
Hospital, Radiation Oncology, Amsterdam, The
Netherlands
2
Netherlands Cancer Institute Antoni van Leeuwenhoek
Hospital, Academic Physics, Amsterdam, The
Netherlands
3
Netherlands Cancer Institute Antoni van Leeuwenhoek
Hospital, Pulmonology, Amsterdam, The Netherlands
4
Netherlands Cancer Institute Antoni van Leeuwenhoek
Hospital, Nuclear Medicine, Amsterdam, The
Netherlands
Purpose or Objective
The randomized phase 2 Raditux-trial (NTR2230) in locally
advanced non-small cell lung cancer (NSCLC),
investigating the additional benefit of Cetuximab to
concurrent chemoradiotherapy (CCRT) did not show
improved survival but revealed a remarkable 5-year
overall survival (OS) of 37.3% [1]. Patients were staged
with
18
FDG-PET-scans before and 4 weeks after CCRT. The
purpose of this study was to investigate whether PET
metrics have prognostic value in relation to local,
regional, and distant failure.
Material and Methods
In the Raditux-trial, 102 stage IIIA-B NSCLC patients were
included. CCRT consisted of 66 Gy in 24 fractions (using
IMRT) combined with daily low dose Cisplatin. A subgroup
of the patients had a repeat
18
FDG-PET-scan for response
evaluation of the primary tumor and lymph nodes after a
median of 4.2 weeks (range, 1.6-10.1). Twenty patients
underwent additional surgery and were excluded. Ten
patients were excluded due to technical reasons. The pre-
and post-treatment
18
FDG-PET-scans from the remaining
42 patients were anatomically registered with the
planning CT-scan. The following pre-and post-treatment
PET metrics were calculated of the primary tumor (PT) as
well as the combined lymph nodes (LNs): SUV
max
, total
lesion glycolysis (TLG) and metabolic tumor volume (MTV).
The response ratio between the pre- and post-treatment
values was also calculated. These parameters were tested
as prognostic factors using the Kaplan-Meier method and
Cox regression analysis for univariate and multivariate
analyses.
Results
Forty-two patients were evaluated for the prognostic
value of the PET metrics. The median follow-up and OS
was 32 and 33 months, respectively. Median GTV of the PT
and the LNs was 80 cc (range, 2-439) and 27 cc (range, 2-
195). The SUV
max
of both PT and LNs decreased
significantly as well as TLG of the PT and MTV of the LNs
(p≤0.001). The post-treatment and the response ratio of
the SUV
max
of the LNs was correlated significantly with
regional failure (p=0.009; p=0.009) (Table 1). The
response ratio of the SUV
max
of the LNs was also
significantly correlated with OS (p=0.014). No parameters
corresponded with local and distant failure.
Table 1
The P-values and HR of the PET metrics of the primary
tumor (PT) related to local failure and combined lymph
nodes (LNs) related to regional failure of the pre- and
post-treatment
18
FDG-PET-scan as well as the response
ratio.
Conclusion
The post-treatment and response SUV
max
of the LNs were
found to be significant prognostic factors for regional
failure and OS in patients with locally advanced NSCLC
treated with hypofractionated CCRT. These parameters
might be useful in the selection of patients for additional
therapy.
PO-0889 FLT PET kinetic analysis biomarkers of
resistance to radiotherapy for nasal tumours in canines
U. Simoncic
1
, T.J. Bradshaw
2
, L. Kubicek
3
, L.J. Forrest
4
,
R. Jeraj
5
1
Jozef Stefan Institute, F-8, Ljubljana, Slovenia
2
University of Wisconsin, Department of Radiology,
Madison, USA
3
Angell Animal Medical Center, Angell Animal Medical
Center, Boston, USA
4
University of Wisconsin, Department of Surgical
Sciences- School of Veterinary Medicine, Madison, USA
5
University of Wisconsin, Department of Medical Physics,
Madison, USA
Purpose or Objective
Imaging biomarkers of resistance to radiotherapy are
prerequisite for precise treatment. Multiple imaging
biomarkers are typically provided by a separate multi-
tracer or multimodal imaging. This study assessed kinetic
analysis as a means to create multiple imaging biomarkers
of resistance to radiotherapy from a dynamic 3’-
(
18
F)fluoro-3’-deoxy-L-thymidine (FLT) positron emission
tomography (PET) scan.
Material and Methods
Sixteen canine cancer patients with spontaneous nasal
tumours were imaged dynamically with FLT PET before
and during the radiotherapy. Images were analysed for
kinetics on a voxel basis using a two tissue, four rate-
constant compartmental model. Overall parameter values
(mean and median over the region of intrests (ROI)) and
heterogeneity measures (coefficient of variation (COV),
ratio of interquartile range to median (IQR/median)) were
evaluated over the tumour gross target volume for the
transport (K
i
=K
1
k
3
/(k
2
+k
3
)), perfusion/permeability (K
1
)
and vascular fraction (V
b
) parametric images. Response
biomarkers were evaluated as a ratio of mid-therapy to
pre-therapy regional values, (i.e. mean, median, COV,
IQR/median). Alternative, spatial responses were
evaluated as a mean, median, COV or IQR/median taken
on a ratio of mid-therapy to pre-therapy prametric
images. The time to progression after radiotherapy (TTP)
was estimated by assessing the therapy response
according to the RECIST. Kaplan-Meier analysis and
univariate Cox proportional hazards (PH) regression were
used to assess the impact of each imaging biomarker on
the TTP.
Results
Pre- or mid-therapy overall
K
i
parameters were significant
predictors of TTP after the radiotherapy. However, many
imaging biomarkers based on
K
1
and
V
b
parameters had
higher predictive power for the radiation therapy
response. Table shows results of univariate Cox
proportional hazard regression for imaging biomarkers
derived from FLT PET parametric images. Hazard is
significantly increased for higher pre- or mid-therapy
overall
K
i
parameter values, higher or increasing pre- or
mid-therapy overall
K
1
parameter value, lower or