528 / 1096
S513
ESTRO 36
_______________________________________________________________________________________________
comparison, a model of the total accumulated dose to the
target was calculated by summing the dose contributions
from each time point. This was accomplished by
deformably registering each post-implant CT scan and
associated dose to the day 0 CT scan, scaling the dose
contribution according to the seed activity at the time of
the scan. A dose evaluation volume (DEV) was defined on
all scans as a 5 mm isotropic expansion of the CTV trimmed
to skin and chest wall muscle. Dosimetric indices for the
CTV (V100) and DEV (V90, V100, and V200) were compared
between each individual postplan and the accumulated
dose using either a paired t-test or a Wilcoxon signed rank
test, whichever carried more power given the distribution
of the data. Residuals were also calculated, defined as the
difference in dosimetric indices for a given time point and
the accumulated dose model. As either a positive or a
negative residual represents a deviation from the model,
the median of the errors (where each error is the absolute
value of the residual) was also calculated for each time
point.
Results
The residuals for the DEV V100 and V200 for all 10 patients
at each time point are shown in Figures 1 and 2,
respectively. A statistically significant difference was
observed between the day 60 scan and the accumulated
dose for the DEV V90, V100, and V200 (paired t-test); no
other significant differences were found. The smallest
median (range) error occurred for the day 15 CT scan (as
demonstrated in Figures 1 and 2); 2.4% (0.2-7.3%) and 4.5%
(0.6-15.9%) for the DEV V100 and V200, respectively.
Conclusion
The results of this study indicate that the day 15 scan is
the most representative of the accumulated dose
delivered to target volumes in PBSI. For a 10-patient
cohort, the median error was found to be at a minimum
for the DEV V100 and V200 for the day 15 time point when
compared to the day 0, 30, and 60 scans.
Poster: Brachytherapy: Prostate
PO-0926 Interstitial HDR prostate brachytherapy:
comparison of pre- and post-implant dose distribution.
S. Novikov
1
, S. Kanaev
1
, N. Ilin
1
, R. Novikov
1
, M.
Girshovich
1
1
Prof. N.N. Petrov Research Institute of Oncology,
Radiation Oncology, St. Petersburg, Russian Federation
Purpose or Objective
Prospective planning of interstitial high dose rate
brachytherapy (HDRBT) for prostate cancer permit high
accuracy of dose delivery to the tumour and\or prostate
with excellent sparing of normal organs. On line correction
of post-implant changes of prostate and normal tissues
volumes is the key factor of precious dose delivery.
The aim of the study was to evaluate possible
uncertainties in dose distribution in cases when
brachytherapy procedure is based only on pre-implant
planning with dose distribution after HDRB with post-
implant correction of dose distributiion.
Material and Methods
in 70 primary patients with prostate cancer we
analyzed
dosimetric plans that were obtained during the first
session of HDRBT. Pretreatment planning was performed
according to standard procedure with calculation of the
following dosimetric parameters: V100, D90 – for prostate,
D2cc – for rectum and D10 – for urethra. According to
standard HDRBT procedure after the end of needle
insertion we performed final US 3D-scanning with post
implant correction of prostate, urethra, bladder and
rectal volumes and subsequent post-implant optimization
of treatment plan.
During the study we also performed fusion of pre-implant
and post-implant images. Fusion was based on needle and
base-plan topography. After that we calculated dose
distribution according to the model when pre-implant plan
was used in patients with post-implant prostate and
normal organs volumes.
Results
Analysis of treatment plans with post-I mplantation
correction of the contours demonstrated h igh precision
and excellent dosimetric parameters: mean V100 - 94.1%
(V100 more than 90% in 97.2% cases), mean D90 – 104.3%
(D90 more than 100% in 95.7% observations). On the
contrary, after fusion of non-corrected plans and post-
implant volumes we mentioned high discrepancies
between preplanned and real dose distribution: V100 was
below 80% in 38.6% observations; D90 was below 80% in