258 / 1096
S245
ESTRO 36
_______________________________________________________________________________________________
Results
The median mAs reduction for all patients was 64% (13%
to 85%). Patient corpulence was not correlated to the mAs
reduction achieved (Spearman’s correlation r
s
= 0.465).
Variance analysis, for every direction, shows no significant
difference (p<0.05) between S
0%
/ S
-50%
and S
0%
/ S
-71%
.
Table 1 : For 3 fractions, the variance in matching from
23 therapists and Fisher’s test results
Conclusion
mAs reductions recorded across the 20 patients are highly
variable, due to the subjective assessment of CBCT image
quality, but a median reduction of 64% indicates a great
potential for reducing imaging dose.
For one patient it has been demonstrated that image
quality deterioration has no impact on interobserver
variability.
A 50% mAs reduction for the Pelvis CBCT template is
therefore considered.
PV-0460 Comparison of 3 Image-guided Adaptive
Strategies for Bladder Radiotherapy
V. Kong
1
, A. Taylor
2
, T. Craig
1
, P. Chung
1
, T. Rosewall
1
1
Princess Margaret Cancer Centre, Radiation Medicine
Program, Toronto, Canada
2
Sheffield Hallam University, Faculty of Health & Well-
being, Sheffield, United Kingdom
Purpose or Objective
Due to the significant variation of bladder volume
observed throughout the course of treatment, various
adaptive strategies have been developed to improve the
quality of bladder radiotherapy. The aim of this study was
to use deformable registration and dose accumulation
processes to compare the dosimetric differences of a
population-based PTV approach and three proposed
adaptive strategies: Plan of the Day (POD), Patient-
Specific PTV (PS-PTV) and daily reoptimization (ReOpt).
Material and Methods
Bladder patients (n=10) were included in this
retrospective investigation. Patients were planned and
treated with a full bladder in supine position. Planning CT
and the CBCTs were retrieved and imported into
treatment planning system. After delineating the bladder
and the pelvic lymph node (PLN) on the planning CT, an
expansion of 1.5 cm and 0.5 cm was applied to generate
the population-based Standard PTV
WB
and PTV
PLN
,
respectively. A 7-field IMRT distribution was designed to
deliver a prescription dose of 46Gy in 23 fractions. Each
adaptive strategy was applied according to published
guidelines. After simulating the execution of each
strategy using the daily CBCTs, daily dose was computed
on all CBCTs and then total dose was summed on the
planning CT using the output from the CT-CBCT
deformable image registration. The volume receiving 95%
of prescription dose (V
95
) was compared against the
Standard for each of the adaptive strategies. p < 0.05 was
considered statistically significant.
Results
Mean V
95
(cm
3
) were 1410 (SD: 227), 1212 (SD: 186), 1236
(SD: 199), and 1101 (SD: 180) for Standard, POD, PS-PTV
and ReOpt, respectively. All adaptive strategies
significantly reduced the irradiated volume, with ReOpt
demonstrating the greatest reduction compared to the
Standard (-25%). This was followed by a mean reduction
of 16% with PS-PTV and 12% with POD. The difference in
the magnitude of reduction between ReOpt and the other
2 strategies reached statistical significance (p = 0.0006).
Conclusion
Previous comparisons between bladder adaptive strategies
have been limited due to the inability to account for the
effect of daily motion of the bladder and surrounding
organs. When deformable registration is used to
reconstruct dose in the presence of organ motion, ReOpt
is the best adaptive strategy at reducing the irradiated
volume due to its frequent adaptation based on the daily
geometry of the bladder. However the resource burden
associated with this strategy needs to be quantified to
further assess the feasibility of clinical implementation.
PV-0461 Integrating diagnostic MRI in radical bladder
cancer radiotherapy: Challenges in image registration.
C.L. Eccles
1
, H. McNair
1
, D. McQuaid
2
, K. Warren-Oseni
2
,
V.N. Hansen
2
, A. Sohaib
3
, M.D. Koh
4
, R. Huddart
4
, S.
Hafeez
4
1
The Royal Marsden NHS Foundation Trust and The
Institute of Cancer Research, Radiotherapy, London,
United Kingdom
2
The Royal Marsden NHS Foundation Trust and The
Institute of Cancer Research, Radiotherapy Physics,
London, United Kingdom
3
The Royal Marsden NHS Trust, Radiology, London,
United Kingdom
4
The Royal Marsden NHS Foundation Trust and The
Institute of Cancer Research, Radiotherapy and Imaging,
London, United Kingdom
Purpose or Objective
Radiographer led soft tissue matching has been
fundamental for implementation of adaptive strategies in
bladder cancer. Integrative MRI technology has the
potential to improve tumour and normal soft tissue
visualisation at treatment planning and delivery. This work
investigates the degree of inter and intra observer
variation in image registration between experts, using a
biological target volume (BTV) defined on diffusion
weighted MRI (DW-MRI), in patients muscle invasive
bladder cancer.
Material and Methods
Twenty-two patients with muscle invasive bladder cancer
recruited prospectively to a phase I image guided
radiotherapy protocol (NCT01124682). Prior to
radiotherapy, all patients underwent MRI on a 1.5T
magnet prior to to acquire T1-weighed, and T2-weighted
DW-MRI with b values of 0, 50, 100, 250, 500 and
750s/mm
2
. The BTV was delineated on b 750 s/mm
2
images
and transferred to the treatment planning system
(Pinnacle v9.6, Philips Medical Systems), where DW-MR
images were registered to the corresponding ADC map and
planning CT by three observers (one oncologist and two
radiographers). Registration was guided using the bladder,
and BTV.
Results
Nineteen of the 22 patients accrued to the study had BTVs
visible on DW MRI and were included in this analysis. The
most notable inter-observer variation in image
registration of the BTV occurred in the caudal-cranial (C-
C) direction with a mean difference of 5.4 mm (standard
deviation (sd) 4.7 mm), followed by the anterior-posterior
(A-P) direction (mean 4.5 mm, sd 4.9 mm). The inter-