204 / 1096
S191
ESTRO 36
_______________________________________________________________________________________________
to low dose while the low value of V85 does not raise
particular concern for tissue necrosis.
Conclusion
Considerable differences between TG43 and MC dosimetry
indicate that plan quality of HDR brachytherapy for lip
carcinoma may be compromised.
The ACE algorithm was found to improve dosimetric
accuracy at clinically relevant distances.
TPS dosimetric accuracy close to the source dwell
positions warrants further investigation.
OC-0358 Evaluation of the Advanced Collapsed-cone
Engine dose calculation algorithm for COMS eye
plaques
H. Morrison
1,2
, G. Menon
1,2
, M. Larocque
1,2
, E. Weis
3,4
, R.
Sloboda
1,2
1
University of Alberta, Oncology, Edmonton, Canada
2
Cross Cancer Institute, Medical Physics, Edmonton,
Canada
3
University of Alberta, Ophthalmology, Edmonton,
Canada
4
University of Calgary, Surgery, Calgary, Canada
Purpose or Objective
The current dosimetry protocol for ocular brachytherapy
involves augmenting TG-43 dose calculations with
correction factors or using look-up tables to account for
plaque materials, as the water-based TG-43 calculation
alone overestimates the dose in front of gold eye plaques
by >20%. This work investigates the accuracy with which
the Advanced Collapsed-cone Engine (ACE) algorithm
(Oncentra Brachy (OcB) v4.6.0, Elekta, Sweden) can
account for the ophthalmic applicator materials (gold
backing and Silastic insert) for three different sizes of
COMS eye plaques in a water phantom.
Material and Methods
The 12, 16, and 20 mm COMS eye plaques were introduced
into the applicator library for OcB by creating 3D CAD
models of the plaques and Silastic inserts with virtual
catheter lines along each seed slot. The Nucletron
selectSeed 130.002 I-125 source model for ACE was
created using primary-scatter separated kernel data
(generated by the CLRP (Carleton Laboratory for
Radiotherapy Physics) group) and AAPM consensus TG-43
dosimetry data. Treatment plans were created in OcB for
a single seed in water, a single seed loaded in the central
slot of the 12 and 20 mm COMS plaques (the 16 mm COMS
plaque does not have a central slot), and fully loaded 12,
16, and 20 mm COMS plaques. ACE dose calculations were
performed in high accuracy mode on a high resolution 0.5
mm
3
calculation grid. The resulting dose data was
compared to Monte Carlo (MC) simulated data using
MCNP6, replicating the OcB treatment plans.
Results
The ACE doses for the single seed in water agree with MC
simulations on average within 4.4 ± 2.1% in a 60x60x60
mm
3
cube centered on the seed, with the largest
differences near the end-welds of the seed. Percent
differences between ACE and MC doses along the plaque
central axes (CAX) for all eye plaque plans are shown in
Figure 1. The agreement improves beyond ~3 mm from the
outer scleral surface, and is generally better for the fully
loaded plaques than the single seed plaques, due to more
overlapping dose from each seed washing out ray effects
caused by the ACE calculation. Compared to using the
previous minimum calculation grid size of 1 mm
3
, the
smaller 0.5 mm
3
grid size results in less voxel averaging,
and therefore more accurate doses immediately adjacent
to the plaques, though both agree well with MC in the eye
region (Figure 2).